You may notice some images loading slow across the Open Collections website. Thank you for your patience as we rebuild the cache to make images load faster.

Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Development of an active anti-whiplash automotive seat to reduce whiplash injuries following a rear-end… Mang, Daniel 2019

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2020_may_mang_daniel.pdf [ 32.9MB ]
Metadata
JSON: 24-1.0384553.json
JSON-LD: 24-1.0384553-ld.json
RDF/XML (Pretty): 24-1.0384553-rdf.xml
RDF/JSON: 24-1.0384553-rdf.json
Turtle: 24-1.0384553-turtle.txt
N-Triples: 24-1.0384553-rdf-ntriples.txt
Original Record: 24-1.0384553-source.json
Full Text
24-1.0384553-fulltext.txt
Citation
24-1.0384553.ris

Full Text

©	Daniel	Mang,	2019	Development	of	an	Active	Anti-Whiplash	Automotive	Seat		to	Reduce	Whiplash	Injuries	Following	a	Rear-End	Collision			by			Daniel	Mang			B.Sc.,	Boston	University,	2007	M.Sc.,	University	of	British	Columbia,	2010					A	DISSERTATION	SUBMITTED	IN	PARTIAL	FULFILMENT	OF	THE		REQUIREMENTS	FOR	THE	DEGREE	OF			DOCTOR	OF	PHILOSOPHY		in		THE	FACULTY	OF	GRADUATE	AND	POSTDOCTORAL	STUDIES		(Kinesiology)					THE	UNIVERSITY	OF	BRITISH	COLUMBIA		(Vancouver)					October	2019				ii	The	 following	 individuals	 certify	 that	 they	have	 read,	and	 recommend	 to	 the	Faculty	of	Graduate	and	Postdoctoral	Studies	for	acceptance,	the	dissertation	entitled:		Development	of	an	Active	Anti-Whiplash	Automotive	Seat	to	Reduce	Whiplash	Injuries	Following	a	Rear-End	Collision		submitted	by	 Daniel	Mang	 	in	partial	fulfilment	of	the	requirements	for	the	degree	of	 Doctor	of	Philosophy	in	 Kinesiology		Examining	Committee:	Dr.	Jean-Sébastien	Blouin,	Kinesiology	Co-supervisor	Dr.	Gunter	P.	Siegmund,	Kinesiology	Co-supervisor		Dr.	Douglas	P.	Romilly,	Mechanical	Engineering	Supervisory	Committee	Member	Dr.	Anita	Vasavada,	Chemical	Engineering	and	Bioengineering	(Washington	State	University)	External	Examiner	Dr.	Bonita	Sawatzky,	Orthopaedics	University	Examiner	Dr.	Guy	Faulkner,	Kinesiology	University	Examiner					ii	Abstract		 Whiplash	injuries	remain	the	most	common	injury	associated	with	motor	vehicle	crashes	despite	the	 introduction	 of	 anti-whiplash	 seats.	 The	 overall	 goal	 of	 the	 experiments	 presented	 in	 this	dissertation	was	to	design,	build	and	test	a	novel	Experimental	anti-whiplash	automotive	seat	to	prevent	whiplash	 injuries	 following	 low-speed,	 rear-end	collisions.	The	key	safety	 features	of	 the	Experimental	seat	 included	 the	 dynamic	 control	 of	 seat	 hinge	 rotation	 and	 seatback	 cushion	 deformation.	 These	safety	 features	 were	 deployed	 before	 and	 during	 the	 collision	 with	 the	 aim	 to	 reduce	 occupant	kinematic	and	kinetic	responses	and	to	better	minimize	the	relative	motion	between	the	head	and	the	upper	torso.	Four	experiments	were	conducted	to	better	understand	the	performance	of	current	anti-whiplash	 seats	 during	 low	 to	 moderate	 collision	 severities	 (Experiment	 1)	 and	 to	 evaluate	 the	performance	 of	 the	 Experimental	 seat	 (Experiments	 2–4).	 In	 Experiment	 1,	 the	 performance	 of	 four	existing	 anti-whiplash	 seats	 were	 compared	 in	 their	 abilities	 to	 reduce	 anthropomorphic	 test	 device	(ATD)	responses	during	a	series	of	low	to	moderate	collision	speed	changes	(Δv=2–14	km/h).	Good-rated	seats,	 according	 to	 the	 Research	 Council	 for	 Automobile	 Repairs/International	 Insurance	 Whiplash	Prevention	 Group	 (RCAR/IIWPG),	 attenuated	 only	 four	 peak	 ATD	 responses	 compared	 to	 poor-rated	seats.	 The	 next	 three	 experiments	 tested	 the	 two	 safety	mechanisms	 of	 the	 Experimental	 seat:	 seat	hinge	 rotation	 only	 (Experiment	 2),	 seatback	 cushion	 deformation	 only	 (Experiment	 3)	 and	 the	 co-activation	 of	 both	 safety	 mechanisms	 (Experiment	 4).	 In	 comparison	 to	 a	 Control	 seat,	 actively	controlling	seat	hinge	rotation	decreased	most	ATD	responses	and	neck	injury	criteria	by	23–85%	while	modulating	 seatback	 deformation	 attenuated	most	 occupant	 responses	 and	 all	 neck	 injury	 criteria	 by	15–82%.	 In	Experiment	4,	 the	Experimental	seat	combining	both	safety	mechanisms	was	compared	to	four	existing	anti-whiplash	seats	and	yielded	decreases	in	ATD	responses	of	25–99%	and	in	neck	injury	criteria	 of	 9–73%	 for	 collision	 speeds	 of	 4	 km/h	 or	 greater.	 The	 results	 of	 these	 experiments	demonstrated	that	the	Experimental	anti-whiplash	seat	with	the	dynamic	control	of	seat	hinge	rotation	and	 seatback	 cushion	 deformation	 could	 potentially	 be	 an	 effective	 solution	 to	 reduce	 the	 risk	 of	whiplash	injuries	and	improve	occupant	safety.						iii	Lay	Summary	Whiplash	injuries	are	the	most	common	injury	associated	with	low-speed,	rear-end	collisions.	The	goal	 of	 the	 experiments	 presented	 in	 this	 dissertation	was	 to	 develop	 an	 Experimental	 anti-whiplash	seat	that	controlled	seat	hinge	and	seatback	cushion	properties	to	reduce	the	accelerations	experienced	by	 an	 occupant	 and	 to	 better	 support	 an	 occupant’s	 head	 and	 neck	 during	 a	 whiplash	 perturbation.	Compared	to	other	anti-whiplash	seats,	the	Experimental	seat	reduced	occupant	responses	during	rear-end	collisions	less	than	12	km/h.		The	co-activation	of	the	seat	hinge	rotation	and	the	seatback	cushion	deformation	on	the	Experimental	seat	reduced	the	occupant	responses	from	a	12	km/h	collision	to	the	equivalent	 of	 a	 4	 km/h	 collision	 on	 other	 existing	 anti-whiplash	 seats.	 Thus,	 the	 Experimental	 anti-whiplash	seat	with	the	dynamic	control	of	seat	hinge	rotation	and	seatback	cushion	deformation	could	potentially	be	an	effective	solution	to	further	reduce	the	risk	of	whiplash	injuries	and	improve	occupant	safety	in	vehicles.						 				iv	Preface	The	four	experiments	presented	in	this	dissertation	were	conducted	at	MEA	Forensic	Engineers	&	Scientists	 in	Richmond,	British	Columbia.	The	four	experiments	were	written	up	as	separate	studies	to	assist	 with	 future	 publication	 (Chapters	 2	 through	 5).	 	 Each	 chapter	 contains	 complete	 Introduction,	Methods,	 Results	 and	 Discussion	 sections	 relevant	 to	 each	 experiment.	 	 Since	 all	 experiments	 were	conducted	 using	 a	 50th	 percentile	 male	 BioRID	 II	 anthropomorphic	 test	 device	 (ATD)	 and	 most	experiments	 were	 conducted	 on	 the	 novel	 Experimental	 anti-whiplash	 seat	 (Chapters	 3	 through	 5),	there	was	inevitably	some	repetition	within	the	Methods	section	of	these	experiments,	in	particular	the	Anthropomorphic	Test	Device	and	Instrumentation,	Test	Procedures	and	Data	Analysis	subsections.			A	 version	 of	 Chapter	 2	 has	 been	 provisionally	 accepted	 for	 publication	 as	 Mang,	 D.W.H.,	Siegmund,	G.	P.,	&	Blouin,	J.	S.	A	Comparison	of	Anti-Whiplash	Seats	During	Low/Moderate	Speed,	Rear-End	Collisions.	All	authors	contributed	to	the	concept	and	design	of	the	experiment,	data	interpretation,	and	editing	of	the	manuscript.	I	was	solely	responsible	for	the	data	collection,	data	analysis,	and	writing	of	the	manuscript.		Chapters	3,	4	and	5	are	in	preparation	for	publication	with	the	following	authors:	Mang,	D.W.H.,	Siegmund,	G.	P.,	and	Blouin,	J.	S.	All	authors	contributed	to	the	concept	and	design	of	the	experiment,	data	interpretation,	and	editing	of	the	manuscript.	I	was	solely	responsible	for	the	data	collection,	data	analysis,	and	writing	of	the	manuscript.		 				v	Table	of	Contents	 	ABSTRACT	............................................................................................................................................	ii	LAY	SUMMARY	....................................................................................................................................	iii	PREFACE	..............................................................................................................................................	iv	TABLE	OF	CONTENTS	............................................................................................................................	v	LIST	OF	TABLES	....................................................................................................................................	ix	LIST	OF	FIGURES	.................................................................................................................................	xii	ACRONYMS	AND	ABBREVIATIONS	................................................................................................	xxxviii	ACKNOWLEDGEMENTS	........................................................................................................................	xl	DEDICATION	......................................................................................................................................	xlii	CHAPTER	1.	 INTRODUCTION	...........................................................................................................	1	1.1	 LITERATURE	REVIEW	................................................................................................................	1	1.1.1	 Epidemiology	.............................................................................................................	1	1.1.2	 Biomechanics	of	Whiplash	Injury	...............................................................................	3	1.1.3	 Injury	Mechanisms	of	Whiplash	.................................................................................	4	1.1.4	 Anti-Whiplash	Automotive	Seats	and	Existing	Safety	Features	.................................	7	1.1.5	 Effectiveness	of	Current	Anti-Whiplash	Devices	......................................................	10	1.2	 OVERALL	RESEARCH	OBJECTIVE	AND	GOALS	..............................................................................	11	CHAPTER	2.	 A	COMPARISON	OF	ANTI-WHIPLASH	SEATS	DURING	LOW/MODERATE	SPEED,	REAR-END	COLLISIONS	.................................................................................................................................	14	2.1	 INTRODUCTION	.....................................................................................................................	14	2.2	 METHODS	...........................................................................................................................	17	2.2.1	 Anthropomorphic	Test	Device	and	Instrumentation	................................................	17	2.2.2	 Test	Procedures	........................................................................................................	17	2.2.3	 Data	Analysis	...........................................................................................................	22	2.3	 RESULTS	..............................................................................................................................	27	2.4	 DISCUSSION	.........................................................................................................................	31	2.5	 CONCLUSION	........................................................................................................................	34				vi	CHAPTER	3.	 EFFECTS	OF	SEAT	HINGE	ROTATION	ON	ATD	RESPONSES	...........................................	36	3.1	 INTRODUCTION	.....................................................................................................................	36	3.2	 METHODS	...........................................................................................................................	38	3.2.1	 Experimental	Anti-Whiplash	Seat	............................................................................	38	3.2.2	 Anthropomorphic	Test	Device	and	Instrumentation	................................................	41	3.2.3	 Test	Procedures	........................................................................................................	42	3.2.4	 Data	Analysis	...........................................................................................................	44	3.3	 RESULTS	..............................................................................................................................	50	3.4	 DISCUSSION	.........................................................................................................................	61	3.5	 CONCLUSION	........................................................................................................................	67	CHAPTER	4.	 EFFECTS	OF	SEATBACK	CUSHION	DEFORMATION	ON	ATD	RESPONSES	.......................	69	4.1	 INTRODUCTION	.....................................................................................................................	69	4.2	 METHODS	...........................................................................................................................	71	4.2.1	 Experimental	Anti-Whiplash	Seat	............................................................................	71	4.2.2	 Anthropomorphic	Test	Device	and	Instrumentation	................................................	76	4.2.3	 Test	Procedures	........................................................................................................	78	4.2.4	 Data	Analysis	...........................................................................................................	81	4.3	 RESULTS	..............................................................................................................................	87	4.4	 DISCUSSION	.........................................................................................................................	97	4.5	 CONCLUSION	......................................................................................................................	103	CHAPTER	5.	 COMBINED	EFFECT	OF	SEAT	HINGE	ROTATION	AND	SEATBACK	CUSHION	DEFORMATION	ON	ATD	RESPONSES	................................................................................................	104	5.1	 INTRODUCTION	...................................................................................................................	104	5.2	 METHODS	.........................................................................................................................	106	5.2.1	 Experimental	Anti-Whiplash	Seat	..........................................................................	106	5.2.2	 Anthropomorphic	Test	Device	and	Instrumentation	..............................................	111	5.2.3	 Test	Procedures	......................................................................................................	112	5.2.4	 Data	Analysis	.........................................................................................................	117	5.3	 RESULTS	............................................................................................................................	122				vii	5.4	 DISCUSSION	.......................................................................................................................	130	5.5	 CONCLUSION	......................................................................................................................	136	CHAPTER	6.	 GENERAL	DISCUSSION	AND	CONCLUSION	................................................................	138	6.1	 SUMMARY	OF	THE	RESEARCH	AND	DISCUSSION	........................................................................	138	6.2	 IMPLICATIONS	FOR	WHIPLASH	INJURY	RESEARCH	.....................................................................	145	6.3	 GENERAL	LIMITATIONS	AND	FUTURE	DIRECTIONS	.....................................................................	149	6.4	 CONCLUSION	......................................................................................................................	152	BIBLIOGRAPHY	.................................................................................................................................	153	APPENDIX	A:	EXPERIMENTAL	ANTI-WHIPLASH	SEAT	DESIGN	............................................................	165	PROOF-OF-CONCEPT	SEAT	HINGE	ROTATION	COMPUTATION	STUDY	.......................................................	165	EXPERIMENTAL	ANTI-WHIPLASH	SEAT	...............................................................................................	166	SUMMARY	....................................................................................................................................	169	APPENDIX	B:		CREATING	THE	SEAT	HINGE	ROTATION	PROFILE	..........................................................	170	EXPERIMENT	1:	EFFECTS	OF	INDIVIDUAL	SEAT	HINGE	PARAMETERS	........................................................	172	Peak	Seat	Hinge	Angle	...........................................................................................................	172	Initial	Seat	Hinge	Angular	Velocity	........................................................................................	175	Seat	Hinge	Rotation	Onset	....................................................................................................	177	EXPERIMENT	2:	VERIFICATION	OF	IDEAL	SEAT	HINGE	ROTATION	PROFILE	.................................................	179	EXPERIMENT	3.	EFFECT	OF	PULSE	SHAPE	............................................................................................	181	EXPERIMENT	4:	PRE-PERTURBATION	FORWARD	ROTATION	...................................................................	184	EXPERIMENT	5:	FINAL	SEAT	HINGE	PROFILE	COMBINATION	....................................................................	187	SUMMARY	....................................................................................................................................	190	APPENDIX	C:	CREATING	THE	SEATBACK	CUSHION	DEFORMATION	PROFILE	......................................	194	EXPERIMENT	1:	EFFECTS	OF	INDIVIDUAL	SEATBACK	PARAMETERS	...........................................................	197	Peak	Seatback	Motor	Angle	..................................................................................................	197	Initial	Seatback	Motor	Angular	Velocity	................................................................................	201	Seatback	Cushion	Deformation	Onset	...................................................................................	204	EXPERIMENT	2:	VERIFICATION	OF	IDEAL	SEATBACK	DEFORMATION	PROFILE	.............................................	207	EXPERIMENT	3:	FORWARD	PRE-PERTURBATION	SEATBACK	DEFORMATION	..............................................	210				viii	SUMMARY	....................................................................................................................................	213	APPENDIX	D:		COMBINING	THE	SEAT	HINGE	ROTATION	AND	THE	SEATBACK	CUSHION	DEFORMATION	PROFILES	..........................................................................................................................................	215	EXPERIMENT	1:	VARYING	ONSET	OF	SEAT	HINGE	ROTATION	..................................................................	217	EXPERIMENT	2:	VARYING	ONSET	OF	SEATBACK	DEFORMATION	..............................................................	221	EXPERIMENT	3:	VARYING	BOTH	SEAT	HINGE	ROTATION	AND	SEATBACK	DEFORMATION	ONSETS	..................	224	SUMMARY	....................................................................................................................................	228	APPENDIX	E:	SUPPLEMENTARY	MATERIAL	........................................................................................	229	SUPPLEMENTARY	MATERIAL	FOR	CHAPTER	2	......................................................................................	229	SUPPLEMENTARY	MATERIAL	FOR	CHAPTER	5	......................................................................................	235			 				ix	List	of	Tables	Table	 2.1.	 Mean	 (standard	 deviation)	 of	 head	 restraint	 backset	 and	 height	 determined	 by	Optotrak	prior	to	the	onset	of	each	collision.	......................................................................	22	Table	2.2.	Coefficients	of	Variation	(COV)	for	some	peak	ATD	responses	compared	to	previous	literature.	.............................................................................................................................	26	Table	3.1.	Peak	Responses	for	Control	and	Experimental	(EXP)	seats	for	each	collision	severity.		Underlined	results	highlight	cases	where	the	Experimental	trials	had	larger	responses	than	the	Control	trials.	.................................................................................................................	58	Table	4.1.	Mean	(standard	deviation)	and	Coefficient	of	Variation	(COV:	%)	from	five	repeated	whiplash-like	perturbation	(n	=	5)	on	the	Control	and	Experimental	anti-whiplash	seats	at	collision	severities	of	Dv	=	8	km/h	with	Dt	=	148	ms	and	Dv	=	12	km/h	with	Dt	=	194	ms.	..	91	Table	 4.2.	 Peak	 responses	 for	 Control	 and	 Experimental	 anti-whiplash	 (EXP)	 seats	 for	 each	collision	severity.	..................................................................................................................	92	Table	4.3.	One-way	ANOVA	results	and	normalized	percentage	value	to	determine	the	effects	of	 seat	 geometry	 (i.e.	 backset	 and	 head	 restraint	 height)	 on	 the	 reduction	 of	 ATD	responses	in	Control,	No	Motion	and	Experimental	conditions	during	a	Dv	=	12	km/h	speed	change.	 Five	 repeated	 trials	 of	 the	ATD	 seated	on	 the	Control	 (CTRL:	 unmodified	GMHR	seat,	 dbackset	 =	 94.4	 mm),	 No	 Motion	 (Experimental	 seat	 with	 no	 seatback	 cushion	deformation,	dbackset	=	60.7	mm)	and	Experimental	(EXP:	Experimental	seat	with	seatback	cushion	deformation,	dbackset	=	52.0	mm)	conditions.	Normalized	percentage	values	were				x	determined	by	normalizing	the	differences	 in	ATD	responses	between	CTRL	–	No	Motion	with	the	difference	between	Control	–	EXP	(100%).	See	Table	5.1	for	means	and	standard	deviations	for	each	variable.	Blank	cells	represent	p-values	<	0.0000.	Bolded	values	denote	a	non-significant	difference	between	the	indicated	conditions.	..........................................	93	Table	5.1.	Mean	 (standard	deviation)	of	head	restraint	backset	 (dbackset)	and	height	 (dheight)	as	well	as	resulting	RCAR/IIWPG	static	seat	geometry	rating.	...............................................	116	Table	 5.2.	 Mean	 (standard	 deviation)	 and	 Coefficient	 of	 Variation	 (COV)	 from	 five	 repeated	perturbations	on	the	GMHR	and	Experimental	seats	at	Δv	=	8	km/h,	Δt	=	147	ms	and	Δv	=	12	km/h,	Δt	=	187	ms.	The	underlined	COV	values	indicate	a	COV	rating	of	acceptable	(5%	≤	COV	<	10%),	the	bolded	COV	values	indicate	a	COV	rating	of	poor	(COV	>	10%),	and	all	other	values	are	rated	good	(COV	<	5%).	...........................................................................	129	Table	6.1.	Mean	(standard	deviation)	and	percent	change	(%)	from	Control	GMHR	seat	for	the	Experimental	 seat	with	 seat	 hinge	 rotation	 only	 (Chapter	 3),	 seatback	 deformation	 only	(Chapter	 4)	 and	 the	 combination	 of	 both	 seat	 hinge	 rotation	 and	 seatback	 deformation	(Chapter	5)	during	at	a	Δv	=	12	km/h	collision	speed.	 	Five	 repeated	perturbations	were	collected	 of	 the	 ATD	 seated	 on	 the	GMHR	 and	 each	 test	 condition	 of	 the	 Experimental	seat.	The	largest	beneficial	change	is	shown	in	bold.	........................................................	144						xi	Table	E2.1.	Mean	 (Standard	Deviation)	 and	Coefficient	of	Variation	 (COV)	 from	 five	 repeated	whiplash-like	perturbation	at	a	Dv	=	4,	8	and	12	km/h	with	a	Dt	=	141	ms	on	the	2004	GM	Pontiac	Grand	AM	GMHR	seat.	..........................................................................................	229	Table	 E2.2.	Mean	 (standard	 deviation)	 and	Coefficient	 of	 Variation	 (COV)	 from	 five	 repeated	whiplash-like	 perturbation	 at	 a	Dv	 =	 8	 km/h	with	 a	Dt	 =	 141	ms	 for	 the	 2005	 Saab	 9.3	SAHR,	2005	Volvo	S40	WHIPS	and	2004	Volvo	S60	WHIPS	seats.	......................................	230				 				xii	List	of	Figures	Figure	2.1.	Photographs	of	the	experimental	set-up	with	the	BioRID	II	ATD	on	the	four	test	seats	in	 the	 laboratory	 reference	 frame	 (X,	 Z).	 A.)	 2005	Volvo	 S40	WHIPS,	 B.)	 2004	Volvo	 S60	WHIPS,	C.)	2005	Saab	9.3	SAHR	and	D.)	2004	Pontiac	Grand	Am	GMHR.	............................	19	Figure	2.2.	Exemplar	sled	A.)	velocity	and	B.)	acceleration	pulses	for	increasing	collision	speeds	(Δv	=	2,	4,	6,	8,	10,	12	and	14	km/h	with	a	pulse	duration	(Δt)	of	141	ms).	Collision	onset	occurred	at	time	=	0	ms.	......................................................................................................	20	Figure	 2.3.	 ATD	occupant	 responses	 comparing	 an	 8km/h	 collision	 severity	 from	a	 stationary	position	(grey	line)	and	from	a	6km/h	constant	rearward	speed	(black	line)	on	the	Pontiac	Grand	Am	GMHR	seat.	Hollow	circles	represent	peak	responses	for	a	given	variable.	.......	20	Figure	2.4.	Exemplar	data	comparing	a	RCAR/IIWPG	good	rated	(Volvo	S40	WHIPS;	black	 line)	and	a	RCAR/IIWPG	poor	rated	(Pontiac	Grand	Am	GMHR;	grey	line)	seat	for	the	BioRID	II	ATD.	Each	column	represents	occupant	responses	for	a	good	and	poor	seat	while	exposed	to	 various	 collision	 severities	 (Δv	 =	 2,	 4,	 6,	 8,	 10,	 12	 and	14	 km/h	with	 a	 collision	pulse	duration	 (Δt)	 of	 141	 ms).	 Hollow	 circles	 represent	 the	 onset	 of	 head-to-head-restraint	contact	and	peak	responses	of	each	ATD	response	parameter	 for	each	trial.	Similar	plots	for	all	seats	are	given	in	the	Appendix	E:	E2.1	–	E2.4.	.........................................................	29	Figure	2.5.	Experimental	results	of	peak	kinematic	and	kinetic	responses	for	the	BioRID	II	ATD	seated	on	the	WHIPS,	SAHR	and	GMHR	seats.	For	all	the	graphs,	red	circles	represent	the	Volvo	S40	WHIPS	seat,	blue	triangles	represent	the	Volvo	S60	WHIPS	seat,	green	diamonds				xiii	represent	 the	 SAHR	 seat,	 and	 black	 squares	 represent	 the	 GMHR	 seat.	 Grey	 corridors	represent	99%	predictive	interval	for	ATD	responses	on	the	poor	rated	GMHR	seat.	Y-axis	values	 for	 panel	 I	 have	 been	 inverted	 for	 visual	 purposes	 to	 show	 increasing	 responses	from	bottom	to	top.	.............................................................................................................	30	Figure	3.1.	Photographs	of	 the	experimental	set-up	with	 the	BioRID	 II	ATD	on	A.)	 the	Control	seat	and	B.)	the	novel	anti-whiplash	automotive	seat	and	global	reference	frame	(X,	Z).	The	seat	 hinge	 and	 seatback	motors	 on	 the	 left	 side	 of	 the	 anti-whiplash	 seat	 are	 labelled.	Additional	 seat	hinge	and	seatback	motors	are	 located	on	the	right	side	of	 the	seat	 (not	labelled).		The	seatback	motors	are	not	used	in	this	current	study.	....................................	39	Figure	3.2.	The	programed	input	seat	hinge	rotation	profile	(grey	line)	and	the	observed	output	seat	 hinge	 rotation	 (red	 line)	 of	 the	 anti-whiplash	 seat	 during	 a	 12	 km/h	 perturbation	(black	line,	right	axis).		The	cyan	lines	illustrated	measurements	of	each	seat	hinge	rotation	parameter	 used	 to	 define	 the	 input	 rotation	 profile	 (onset	 delay	 of	 pre-perturbation	forward	 rotation:	 tforward-onset,	 peak	 pre-perturbation	 forward	 rotation	 angle:	 θforward-peak,	peak	 rearward	 rotational	 angle:	 θrearward-peak,	 initial	 rearward	 angular	 velocity:	ωrearward-int,	and	onset	delay	of	rotation:	trearward-onset).	............................................................................	40	Figure	3.3.	Exemplar	sled	A)	velocity	and	B)	acceleration	profiles	for	increasing	collision	speeds	(Δv	=	2,	4,	6,	and	8	km/h	with	a	pulse	duration	(Δt)	of	148	ms,	and	Δv	=	12	km/h	with	a	Δt	of	185	ms;	lightest	to	darkest).	Collision	onset	occurred	at	time	=	0	ms.	............................	49				xiv	Figure	 3.4.	 Exemplar	 data	 comparing	 an	 unmodified	 Control	 seat	 (Pontiac	Grand	Am	GMHR)	and	 the	 Experimental	 seat	with	 dynamic	 seat	 hinge	 rotation	 (θseat-hinge)	 for	 the	 BioRID	 II	ATD.	 Each	 column	 represents	 occupant	 responses	 for	 a	 Control	 and	 Experimental	 seat	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6,	and	8	km/h	with	a	collision	pulse	duration	(Δt)	of	148	ms	and	12	km/h	with	Δt	=	185	ms).	Hollow	circles	represent	the	onset	of	head-to-head-restraint	contact	and	peak	responses	of	each	ATD	response	parameter	for	each	trial.	..............................................................................................................................	54	Figure	3.5.	Experimental	results	of	peak	kinematic	and	kinetic	responses	for	the	BioRID	II	ATD	seated	on	the	Control	and	Experimental	anti-whiplash	seats.	For	all	the	graphs,	grey	circles	represent	 the	Control	 seat	 and	black	 triangles	 represent	 the	 Experimental	 seat	with	 the	seat	hinge	rotation.		The	faded	grey	bars	represent	the	99th	percentile	predictive	corridors	for	 the	 Control	 and	 Experimental	 seats.	 Y-axis	 values	 for	 panels	 5C	 and	 5I	 have	 been	inverted	for	visual	purposes	to	show	increasing	responses	from	bottom	to	top.	...............	55	Figure	 3.6.	Exemplar	 horizontal	 head-to-head-restraint	 distance	 in	 the	 X-axis	measured	 from	the	back	of	the	BioRID	II	ATD	head	to	the	front	face	of	the	head	restraint.		Five	repeated	trials	of	 the	ATD	 seated	on	 the	Control	GMHR	 seat	 (grey)	 and	on	 the	Experimental	 seat	with	dynamic	seat	hinge	rotation	(black)	as	well	as	single	trial	of	the	EXP90	condition	(red)	at	a	collision	severity	of	12	km/h	(Δt	=	185	ms).	 	Collision	onset	occurred	at	time	=	0	ms	and	a	head-to-head	restraint	distance	≤	0	mm	indicated	that	the	head	was	in	contact	with	the	head	restraint.	................................................................................................................	56				xv	Figure	 3.7.	 The	 relative	 contributions	 of	 the	 active	 seatback	 response	 and	 head	 restraint	geometry	 to	 the	 improved	 response	 of	 the	 Experimental	 seat	 compared	 to	 the	 Control	seat.	Thirteen	ATD	response	parameters	are	shown	here	for	the	Control	seat	(set	to	0%),	the	Experimental	seat	(set	to	100%)	and	the	extra	condition	(EXP90)	of	the	Experimental	seat	with	a	head	restraint	geometry	similar	to	the	Control	seat	(intermediate	percentage	values).	Most	of	the	improvements	observed	between	the	Control	and	Experimental	seats	comes	from	adding	the	active	seat	hinge	response	(difference	between	Control	and	EXP90	data	 in	 this	 graph)	 rather	 than	 from	 the	 change	 in	 head	 restraint	 geometry	 (difference	between	EXP90	and	Experimental	seat	data	in	this	graph).	................................................	57	Figure	4.1.	Photograph	of	the	experimental	set-up	with	the	BioRID	II	ATD	on	the	Experimental	anti-whiplash	automotive	seat	and	rigid	metal	braces	to	prevent	seat	hinge	rotation.	The	global	laboratory	reference	frame	is	illustrated	with	positive	X-axis	forward	and	positive	Z-axis	downwards.	The	seat	hinge	and	seatback	motors	on	the	left	side	of	the	Experimental	seat	are	labelled.	Additional	seat	hinge	and	seatback	motors	are	located	on	the	right	side	of	the	seat	(not	labelled).		The	seat	hinge	motors	are	not	used	in	this	current	study.	........	72	Figure	 4.2.	 Computer-aided	 design	 (CAD)	 drawings	 of	 Experimental	 seat	with	 three	 seatback	motors	placed	at	 the	 top,	middle	and	bottom	of	 the	seatback,	and	 the	4	 seatbelt	 straps	that	spanned	across	the	perimeter	frame	to	separately	control	the	seatback	deformation	at	the	upper-torso,	mid-torso	and	lower	pelvis	regions.		The	middle	motor	was	connected	to	the	2	middle	straps.	Panel	A.,	B.	&	C.	show	the	right,	left	and	frontal	views,	respectively.				xvi	The	 distances	 between	 the	 center	 of	 the	 seatbelts	 and	 the	 seat	 hinge	 are	 illustrated	 in	Panel	D.	................................................................................................................................	75	Figure	 4.3.	 Estimated	 horizontal	 seatback	 cushion	 deformation	 using	 displacement	 of	 the	 T1	vertebra	 and	 pelvis	 (dpenetration-T1,	 red	 plot	&	 dpenetration-pelvis,	 blue	 plot)	 in	 response	 to	 the	seatback	 motor	 rotation	 profile	 (θseatback-motor,	 grey	 plot)	 with	 θdeformation-peak	 =	 168	 deg,	ωdeformation-int	=	980	deg/s,	and	tdeformation-onset	=	-200	ms	during	a	12	km/h	collision	pulse	with	a	collision	pulse	duration	of	194	ms	(aX-sled,	black	plot).	......................................................	76	Figure	4.4.	Exemplar	sled	a)	velocity	and	b)	acceleration	pulses	 for	 increasing	collision	speeds	(Δv	=	2,	4,	6,	and	8	km/h	with	a	pulse	duration	(Δt)	of	148	ms,	and	Δv	=	12	km/h	with	a	Δt	of	194	ms;	lightest	to	darkest).	Collision	onset	occurred	at	time	=	0	ms.	............................	80	Figure	4.5.	Data	comparing	the	BioRID	II	ATD	response	seated	on	either	the	Control	seat	(grey	lines)	or	the	Experimental	seat	with	the	seatback	cushion	deformation	profile	(black	lines)	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6,	and	8kmh	with	a	collision	pulse	duration	(Δt)	of	148	ms	and	Δv	=	12	km/h	with	a	Δt	=	194	ms).	Hollow	circles	represent	the	onset	 of	 head-to-head-restraint	 contact	 and	 peak	 responses	 of	 each	 ATD	 response	parameter	for	each	trial.	......................................................................................................	94	Figure	4.6.	BioRID	II	ATD	and	seatback	interaction	data	comparing	the	Control	seat	(grey	lines)	and	 the	 Experimental	 anti-whiplash	 seat	with	 seatback	 deformation	 profile	 (black	 lines)	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6,	and	8kmh	with	a	collision	pulse	duration	(Δt)	of	148	ms	and	Δv	=	12	km/h	with	a	Δt	=	194	ms).	From	top	to	bottom	(+X-			xvii	direction	 represents	 towards	 the	 front	 of	 the	 seat):	 Horizontal	 sled	 acceleration	 (aX-sled),	head-to-head	restraint	backset	distance	(dbackset	=	0	mm	represents	head	contact	with	head	restraint),	 and	 rearward	 penetration	 of	 the	 ATD’s	 T1	 (dpenetration-T1)	 and	 pelvis	 (dpenetration-pelvis)	into	the	seatback.	.........................................................................................................	95	Figure	4.7.	Experimental	results	of	peak	kinematic	and	kinetic	responses	for	the	BioRID	II	ATD	seated	 on	 Control	 and	 Experimental	 seats.	 For	 all	 the	 graphs,	 grey	 circles	 represent	 the	control	 seat	 and	 black	 triangles	 represent	 the	 Experimental	 seat	 with	 the	 seatback	deformation	profiles.		The	shaded	bands	represent	the	99th	percentile	predictive	intervals	for	 the	 Control	 and	 Experimental	 seats.	 Y-axis	 values	 for	 panels	 D,	 E	 and	 L	 have	 been	inverted	for	visual	purposes	to	show	increasing	responses	from	bottom	to	top.	...............	96	Figure	5.1.	Photograph	of	the	experimental	set-up	with	the	BioRID	II	ATD	on	the	Experimental	anti-whiplash	automotive	seat	and	laboratory	reference	frame	(X,	Z).		The	seat	hinge	and	seatback	motors	 on	 the	 left	 side	 of	 the	 anti-whiplash	 seat	 are	 labelled.	 Additional	 seat	hinge	and	seatback	motors	are	located	on	the	right	side	of	the	seat	(not	labelled).		(Figure	5.1	is	a	repeat	of	Figure	3.1B.)	...........................................................................................	107	Figure	 5.2.	 Programmed	 inputs	 and	 resulting	 outputs	 for	 A.)	 seat	 hinge	 rotation	 and	 B.)	seatback	cushion	deformation	profiles	during	a	12	km/h	collision	(∆t	=	187	ms;	black	line).	Grey	lines	represent	input	A.	rotation	(θseat-hinge)	and	B.	deformation	(θseatback-motors)	profiles	to	 seat	 hinge	 and	 seatback	 motors,	 respectively.	 Theses	 programmed	 input	 seat	 hinge	rotation	 and	 seatback	 deformation	 profiles	 were	 used	 across	 all	 collision	 speeds.	 Seat				xviii	hinge	 rotation	 (θseat-hinge-output:	 red	 line)	 as	 well	 as	 rearward	 penetration	 of	 ATD	 T1	(dpenetration-T1:	 green	 line)	 and	 pelvis	 (dpenetration-pelvis:	 blue	 line)	 into	 the	 seatback	 indicate	resulting	outputs	to	their	respective	input	profiles.	..........................................................	110	Figure	5.3.	Photographs	of	the	experimental	set-up	with	the	BioRID	II	ATD	on	the	four	existing	anti-whiplash	seats:	A.)	2005	Volvo	S40	WHIPS,	B.)	2004	Volvo	S60	WHIPS,	C.)	2005	Saab	9.3	SAHR	and	D.)	2004	Pontiac	Grand	Am	GMHR	..............................................................	115	Figure	5.4.	Exemplar	sled	A.)	velocity	and	B.)	acceleration	pulses	for	increasing	collision	speeds	(Δv	of	2,	4,	6,	and	8	km/h	with	a	pulse	duration	(Δt)	of	147	ms,	and	Δv	of	12	km/h	with	a	Δt	of	187	ms;	 lightest	 to	darkest).	Collision	onset	occurred	at	 time	=	0	ms.	 	 This	 figure	 is	 a	repeat	of	Figure	3.3.	...........................................................................................................	116	Figure	 5.5.	 Exemplar	 BioRID	 II	 ATD	 response	 data	 comparing	 the	 Experimental	 anti-whiplash	seat	(black	line),	a	RCAR/IIWPG	good-rated	(Volvo	S40	WHIPS;	red	line)	and	a	RCAR/IIWPG	poor-rated	 (Pontiac	 Grand	 Am	 GMHR;	 blue	 line).	 Each	 column	 represents	 occupant	responses	while	exposed	to	various	collision	severities	(Δv	of	2,	4,	6	and	8	with	a	collision	pulse	 duration	 (Δt)	 of	 147	 ms	 and	 Δv	 of	 12	 km/h	 with	 a	 Δt	 =	 187	 ms).	 Hollow	 circles	represent	 the	 onset	 of	 head-to-head-restraint	 contact	 and	 peak	 responses	 of	 each	 ATD	response	parameter	 for	each	trial.	Similar	plots	 for	all	 seats	are	given	 in	 the	Appendx	E:	Figures	E5.1	–	E5.5.	............................................................................................................	126	Figure	5.6.	Exemplar	BioRID	 II	ATD	 response	data	comparing	 the	seat	hinge	 rotation,	backset	and	occupant	penetration	at	the	T1	and	pelvis	for	the	on	the	Experimental	anti-whiplash				xix	seat	(black	line),	a	RCAR/IIWPG	good	rated	(Volvo	S40	WHIPS;	red	line)	and	a	RCAR/IIWPG	poor	 rated	 (Pontiac	 Grand	 Am	 GMHR;	 blue	 line).	 Each	 column	 represents	 occupant	responses	while	exposed	to	various	collision	severities	(Δv	of	2,	4,	6	and	8	with	a	collision	pulse	 duration	 (Δt)	 of	 147	 ms	 and	 Δv	 of	 12	 km/h	 with	 a	 Δt	 =	 187	 ms).	 The	 grey	 lines	represent	 the	 input	seat	hinge	rotation	and	seatback	motor	deformation	profiles	 for	 the	Experimental	anti-whiplash	seat.	.......................................................................................	127	Figure	5.7.	Experimental	peak	kinematic	and	kinetic	responses	for	the	BioRID	II	ATD	seated	on	the	WHIPS,	SAHR,	GMHR	and	Experimental	seats.	For	all	the	graphs,	red	circles	represent	the	 Volvo	 S40	 WHIPS	 seat,	 blue	 triangles	 represent	 the	 Volvo	 S60	 WHIPS	 seat,	 green	diamonds	represent	the	SAHR	seat,	black	squares	represent	the	GMHR	seat	and	magenta	stars	 represent	 the	 Experimental	 seat.	 The	 faded	black	 and	magenta	bars	 represent	 the	99th	percentile	predictive	intervals	for	the	GMHR	and	Experimental	seats,	respectively.	Y-axis	 values	 for	 panels	 6I,	 6J	 and	 6K	 have	 been	 inverted	 for	 visual	 purposes	 to	 show	increasing	responses	from	bottom	to	top.	.........................................................................	128	Figure	 6.1.	Exemplar	 BioRID	 II	 ATD	 response	 data	 comparing	 the	 Experimental	 anti-whiplash	seat	 (black	 line)	with	 the	co-activation	of	both	 seat	hinge	 rotation	and	seatback	cushion	deformation	during	a	12	km/h	collision	(Δt	=	187	ms)	to	a	RCAR/IIWPG	good-rated	(Volvo	S40	WHIPS;	red	 line)	and	a	RCAR/IIWPG	poor-rated	(Pontiac	Grand	Am	GMHR;	blue	 line)	during	 a	 4	 km/h	 collision	 (Δt	 =	 147	ms).	 Hollow	 circles	 represent	 the	 onset	 of	 head-to-			xx	head-restraint	contact	and	peak	responses	of	each	ATD	response	parameter	for	each	trial.	............................................................................................................................................	148		Figure	A1.	Images	of	the	novel	active	anti-whiplash	automotive	seat.	A.)	preliminary	computer	aided	design	(CAD)	model	and	B.)	working	prototype.	......................................................	168	Figure	A2.	Photograph	of	the	experimental	set-up	with	the	BioRID	II	ATD	on	the	experimental	anti-whiplash	automotive	seat	and	laboratory	reference	frame	(X,	Z).		Motors	mounted	to	the	seat	hinge	control	seatback	hinge	rotations;	whereas,	motors	mounted	to	the	seatback	control	seatback	cushion	deformations.	............................................................................	168		Figure	B1.	Rearward	seat	hinge	rotations	(θseat-hinge)	for	the	Control	(blue	line)	and	Experimental	seat	(red	solid	line)	during	a	12	km/h	collision	severity	perturbation	with	a	collision	pulse	duration	 (Δt)	 of	 208	ms.	 The	 dotted	 red	 line	 illustrates	 the	 input	 θseat-hinge	 profile	 to	 the	Experimental	 seat	 to	 generate	 the	 output	 θseat-hinge	 response	 (solid	 red	 line)	 used	 as	 the	initial	seat	hinge	rotation	profile	for	Part	1.	.......................................................................	171	Figure	B2.	 ATD	 responses	 to	 the	Control	 seat	 (blue	 lines	 and	markers)	 and	 the	 Experimental	seat	 with	 various	 amplitudes	 of	 seat	 hinge	 rotation	 (θseat-hinge;	 grey	 and	 red	 lines	 and	markers)	during	a	12	km/h	collision	severity	(Δt	=	208	ms,	black	lines,	right	axis,	Panels	A	&	B).	The	 red	 lines	and	markers	 indicate	 the	seat	hinge	peak	angle	parameter	 selected	 for	further	testing.	Panels	(A)	and	(B),	respectively,	show	the	programmed	input	and	resulting				xxi	output	 seat	 hinge	 rotations	 for	 the	 Experimental	 seat	 with	 increasing	 θrearward-peak.	 The	results	 of	 the	 post-hoc	 test	 (C)	 determined	 the	mean	 rank	 of	 each	 trial	 as	 well	 as	 any	significant	differences	 (*)	between	 the	selected	seat	hinge	 rotation	parameter	 trial	 (red)	and	the	Control	or	other	Experimental	trials.	 	 	Exemplar	ATD	responses	(D	–	F)	and	their	corresponding	peak	responses	(G-I)	for	head	and	torso	accelerations	in	the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	criteria	(NIC),	respectively.	......................................................	174	Figure	B3.	 ATD	 responses	 to	 the	Control	 seat	 (blue	 lines	 and	markers)	 and	 the	 Experimental	seat	with	various	 initial	velocities	of	seat	hinge	rotation	(θseat-hinge;	grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	208	ms,	black	lines,	right	axis,	Panels	A	&	B).	 The	 red	 lines	 and	markers	 indicate	 the	 seat	 hinge	 initial	 angular	 velocity	 parameter	selected	for	further	testing.	Panels	(A)	and	(B),	respectively,	show	the	programmed	input	and	 resulting	 output	 seat	 hinge	 rotations	 for	 the	 Experimental	 seat	 with	 increasing	ωrearward-int.	The	results	of	the	post-hoc	test	(C)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	differences	(*)	between	the	selected	seat	hinge	rotation	parameter	trial	(red)	and	the	Control	or	other	Experimental	trials.	 	 	Exemplar	ATD	responses	(D	–	F)	and	 their	 corresponding	peak	 responses	 (G-I)	 for	 head	 and	 torso	 accelerations	 in	 the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	criteria	(NIC),	respectively.	..................................	176	Figure	B4.	 ATD	 responses	 to	 the	Control	 seat	 (blue	 lines	 and	markers)	 and	 the	 Experimental	seat	with	various	onsets	of	seat	hinge	rotation	(θseat-hinge;	grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	208	ms,	black	lines,	right	axis,	Panels	A	&	B).	The				xxii	red	lines	and	markers	indicate	the	seat	hinge	rotation	onset	parameter	selected	for	further	testing.	Panels	(A)	and	(B),	respectively,	show	the	programmed	input	and	resulting	output	seat	hinge	rotations	 for	 the	Experimental	 seat	with	 increasing	 trearward-onset.	The	results	of	the	 post-hoc	 test	 (C)	 determined	 the	mean	 rank	 of	 each	 trial	 as	 well	 as	 any	 significant	differences	 (*)	 between	 the	 selected	 seat	 hinge	 rotation	 parameter	 trial	 (red)	 and	 the	Control	 or	 other	 Experimental	 trials.	 	 	 Exemplar	 ATD	 responses	 (D	 –	 F)	 and	 their	corresponding	peak	responses	(G-I)	for	head	and	torso	accelerations	in	the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	criteria	(NIC),	respectively.	......................................................	178	Figure	B5.	ATD	 responses	 to	 the	Control	 seat	 (blue	 lines	 and	markers)	 and	 the	 Experimental	seat	with	 various	 seat	 hinge	 rotation	profiles	 (θseat-hinge;	 grey	 and	 red	 lines	 and	markers)	during	a	12	km/h	collision	severity	(Δt	=	208	ms,	black	lines,	right	axis,	Panels	A	&	B).	The	red	lines	and	markers	indicate	the	seat	hinge	rotation	profile	selected	for	further	testing.	Panels	 (A)	 and	 (B),	 respectively,	 show	 the	 programmed	 input	 and	 resulting	 output	 seat	hinge	rotations	for	all	the	different	Experimental	trials.	The	results	of	the	post-hoc	test	(C)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	differences	(*)	between	the	 selected	 seat	 hinge	 rotation	 parameter	 trial	 (red)	 and	 the	 Control	 or	 other	Experimental	 trials.	 	 	 Exemplar	 ATD	 responses	 (D	 –	 F)	 and	 their	 corresponding	 peak	responses	(G-I)	for	head	and	torso	accelerations	in	the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	criteria	(NIC),	respectively.	.......................................................................................	180				xxiii	Figure	 B6.	 Seat	 hinge	 rotation	 (θseat-hinge)	 input	 signals	 for	 the	 seat	 hinge	 rotation	 profile	determined	previously	 (green	 line,	 Experiments	1	&	2)	 and	 the	new	simplified	 sigmoidal	shape	curve	 (red	 line)	during	a	12	km/h	collision	severity	 (Δt	=	208	ms,	black	 lines,	 right	axis).	...................................................................................................................................	182	Figure	B7.	ATD	responses	 to	 the	Control	seat	 (blue	 lines	and	markers),	 the	Experimental	seat	with	the	seat	hinge	rotation	(θseat-hinge)	profile	determined	previously	in	Experiments	1	&	2	(green	lines	and	markers),	and	the	Experimental	seat	with	various	simplified	sigmoidal	seat	hinge	rotation	profiles	(grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	208	ms,	black	lines,	right	axis,	Panels	A	&	B).	The	red	lines	and	markers	indicate	the	sigmoidal	 seat	 hinge	 rotation	 profile	 selected	 for	 further	 testing.	 Panels	 (A)	 and	 (B),	respectively,	show	the	programmed	input	and	resulting	output	seat	hinge	rotations	for	all	the	different	Experimental	trials.	The	results	of	the	post-hoc	test	(C)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	differences	(*)	between	the	selected	seat	hinge	rotation	parameter	trial	(red)	and	the	Control	or	other	Experimental	trials.			Exemplar	ATD	responses	 (D	 –	 F)	 and	 their	 corresponding	 peak	 responses	 (G-I)	 for	 head	 and	 torso	accelerations	 in	 the	X-direction	 (aX-head	 &	aX-T1)	and	neck	 injury	criteria	 (NIC),	 respectively.	............................................................................................................................................	183	Figure	B8.	ATD	responses	 to	 the	Control	seat	 (blue	 lines	and	markers),	 the	Experimental	seat	with	the	sigmoidal	rearward	rotation	profile	(green	lines	and	markers,	Experiment	3),	and	the	Experimental	seat	with	increasing	angles	of	pre-perturbation	only	forward	seat	hinge				xxiv	rotation	(θseat-hinge;	grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	208	ms,	black	lines,	right	axis,	Panels	A	&	B).	The	red	lines	and	markers	indicate	the	pre-perturbation	seat	hinge	rotation	parameter	selected	for	subsequent	testing.	Panels	(A)	and	(B),	respectively,	show	the	programmed	input	and	resulting	output	seat	hinge	rotations	for	the	 Experimental	 seat	 with	 increasing	 forward	 pre-perturbation	 θseat-hinge	 angles.	 The	results	 of	 the	 post-hoc	 test	 (C)	 determined	 the	mean	 rank	 of	 each	 trial	 as	 well	 as	 any	significant	difference	between	 the	selected	seat	hinge	 rotation	parameter	 trial	 (red)	and	the	 Control	 or	 other	 Experimental	 trials.	 	 	 Exemplar	 ATD	 responses	 (D	 –	 F)	 and	 their	corresponding	peak	responses	(G-I)	for	head	and	torso	accelerations	in	the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	criteria	(NIC),	respectively.	......................................................	186	Figure	B9.	ATD	responses	 to	 the	Control	seat	 (blue	 lines	and	markers),	 the	Experimental	seat	with	 the	 sigmoidal	 rearward	 rotation	 profile	 (θseat-hinge;	 green	 lines	 and	 markers,	Experiment	3),	and	the	Experimental	seat	with	a	combination	of	increasing	amplitudes	of	pre-perturbation	forward	seat	hinge	angle	and	the	rearward	rotation	profile	(grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	208	ms,	black	lines,	right	axis,	Panels	A	&	B).	The	red	lines	and	markers	indicate	the	pre-perturbation	seat	hinge	rotation	parameter	 selected	 for	 subsequent	 testing.	 Panels	 (A)	 and	 (B),	 respectively,	 show	 the	programmed	 input	 and	 resulting	 output	 seat	 hinge	 rotations	 for	 the	 Experimental	 seat	with	increasing	forward	pre-perturbation	θseat-hinge	amplitudes.	The	results	of	the	post-hoc	test	 (C)	 determined	 the	 mean	 rank	 of	 each	 trial	 as	 well	 as	 any	 significant	 difference				xxv	between	 the	 selected	 seat	hinge	 rotation	parameter	 trial	 (red)	and	 the	Control	or	other	Experimental	 trials.	 	 	 Exemplar	 ATD	 responses	 (D	 –	 F)	 and	 their	 corresponding	 peak	responses	(G-I)	for	head	and	torso	accelerations	in	the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	criteria	(NIC),	respectively.	.......................................................................................	189	Figure	B10.	The	programed	input	seat	hinge	rotation	profile	(grey	line)	and	the	observed	output	seat	 hinge	 rotation	 (red	 line)	 of	 the	 anti-whiplash	 seat	 during	 a	 12	 km/h	 perturbation	(black	line,	right	axis).		The	cyan	lines	illustrated	measurements	of	each	seat	hinge	rotation	parameter	 used	 to	 define	 the	 input	 rotation	 profile	 (onset	 delay	 of	 pre-perturbation	forward	 rotation:	 tforward-onset,	 peak	 pre-perturbation	 forward	 angle:	 θforward-peak,	 peak	rearward	angle:	θrearward-peak,	 initial	rearward	angular	velocity:	ωrearward-int,	and	onset	delay	of	rearward	rotation:	trearward-onset).	This	figure	is	a	repeat	of	Figure	3.2.	...........................	192		Figure	C1.	Photograph	of	the	experimental	set-up	with	the	BioRID	II	ATD	on	the	Experimental	anti-whiplash	automotive	seat	and	rigid	metal	braces	to	prevent	seat	hinge	rotation.	The	global	laboratory	reference	frame	is	illustrated	with	positive	X-axis	forward	and	positive	Z-axis	downwards.	The	seat	hinge	and	seatback	motors	on	the	left	side	of	the	Experimental	seat	are	labelled.	Additional	seat	hinge	and	seatback	motors	are	located	on	the	right	side	of	 the	seat	 (not	 labelled).	 	The	seat	hinge	motors	are	not	used	 in	 this	 current	appendix.	This	figure	is	a	repeat	of	Figure	4.1.	...................................................................................	196				xxvi	Figure	 C2.	 Seatback	 motor	 profiles	 and	 ATD	 responses	 to	 the	 Control	 seat	 (blue	 lines	 and	markers),	the	No	Motion	condition	(magenta	lines	and	markers)	and	the	Experimental	seat	with	various	peak	angles	of	seatback	motor	rotations	(θdeformation-peak;	grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	195	ms,	black	lines,	Panel	A,	right	axis).	The	 red	 lines	 and	markers	 indicate	 the	 seatback	motor	 parameter	 (θdeformation-peak	 =	 168	deg)	 selected	 for	 further	 testing.	 Panel	A	 shows	 the	 programmed	 input	 seatback	motor	rotation	for	the	Experimental	seat	with	increasing			θdeformation-peak.	The	results	of	the	post-hoc	test	(B)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	differences	(*)	 between	 the	 selected	 seatback	motor	 parameter	 trial	 (red)	 and	 the	 Control,	 the	No	Motion	 or	 other	 Experimental	 trials.	 Panels	 C	 and	 D	 show	 the	 resulting	 seatback	deformation	 at	 the	 T1	 spinal	 level	 and	 at	 the	 pelvis	 (dpenetration-T1	 and	 dpenetration-pelvis,	respectively).	......................................................................................................................	199	Figure	C3.	Exemplar	ATD	responses	for	the	Control	seat	(blue	lines	and	markers),	the	No	Motion	condition	 (magenta	 lines	 and	 markers),	 and	 the	 Experimental	 seat	 with	 various	 peak	angles	of	the	seatback	motors	(θdeformation-peak;	grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	195	ms).	Calibrated	responses	(A	–	C)	and	their	corresponding	peak	responses	(D	–	F)	for	head	and	torso	accelerations	in	the	X-direction	(aX-head	&	aX-T1)	and	 neck	 injury	 criteria	 (NIC),	 respectively.	 	 The	 red	 lines	 and	 markers	 indicate	 the	seatback	motor	parameter	(θdeformation-peak	=	168	deg)	selected	for	further	testing.	..........	200				xxvii	Figure	 C4.	 Seatback	 motor	 profiles	 and	 ATD	 responses	 to	 the	 Control	 seat	 (blue	 lines	 and	markers),	 the	 No	Motion	 condition	 (magenta	 lines	 and	markers),	 and	 the	 Experimental	seat	with	various	 initial	angular	velocities	of	the	seatback	motors	(ωdeformation-int;	grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	195	ms,	black	lines,	Panel	A,	 right	 axis).	 The	 red	 lines	 and	 markers	 indicate	 the	 initial	 angular	 velocity	 parameter	(ωdeformation-int	 =	 980	 deg/s)	 of	 the	 seatback	motors	 selected	 for	 further	 testing.	 Panel	A	shows	 the	 programmed	 input	 of	 the	 seatback	 motors	 for	 the	 Experimental	 seat	 with	increasing	ωdeformation-int.	The	results	of	the	post-hoc	test	(B)	determined	the	mean	rank	of	each	 trial	as	well	as	any	significant	differences	 (*)	between	 the	selected	seatback	motor	parameter	trial	(red)	and	the	Control	seat,	the	No	Motion	condition	or	other	Experimental	trials.	Panels	C	and	D	show	the	resulting	seatback	deformation	at	the	T1	spinal	level	and	at	the	pelvis	(dpenetration-T1	and	dpenetration-pelvis,	respectively).	...................................................	202	Figure	C5.	Exemplar	ATD	responses	for	the	Control	seat	(blue	lines	and	markers),	the	No	Motion	condition	 (magenta	 lines	 and	 markers),	 and	 the	 Experimental	 seat	 with	 various	 initial	angular	 velocity	 of	 the	 seatback	 motor	 (ωdeformation-int;	 grey	 and	 red	 lines	 and	 markers)	during	a	12	km/h	collision	severity	 (Δt	=	195	ms).	Calibrated	responses	 (A	–	C)	and	their	corresponding	peak	responses	(D	–	F)	 for	head	and	torso	accelerations	 in	the	X-direction	(aX-head	 &	 aX-T1)	 and	 neck	 injury	 criteria	 (NIC),	 respectively.	 	 The	 red	 lines	 and	 markers	indicate	the	initial	seatback	motor	rotation	velocity	parameter	(ωdeformation-int	=	980	deg/s)	selected	for	further	testing.	...............................................................................................	203				xxviii	Figure	 C6.	 Seatback	 motor	 profiles	 and	 ATD	 responses	 to	 the	 Control	 seat	 (blue	 lines	 and	markers),	 the	 No	Motion	 condition	 (magenta	 lines	 and	markers),	 and	 the	 Experimental	seat	with	various	onsets	of	 the	 seatback	motor	 rotation	 (tdeformation-onset;	 grey,	 green,	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	195	ms,	black	lines,	Panel	A,	right	axis).	The	green	lines	and	markers	indicate	the	initial	seatback	motor	onset	used	in	the	 previous	 two	 experiments	 (tdeformation-onset	 =	 -130	ms)	 and	 the	 red	 lines	 and	markers	indicate	 the	 seatback	 motor	 onset	 parameter	 (tdeformation-onset	 =	 -200	 ms)	 selected	 for	further	 testing.	 Panel	 A	 shows	 the	 programmed	 input	 of	 the	 seatback	 motors	 for	 the	Experimental	 seat	 with	 increasing	 tdeformation-onset.	 The	 results	 of	 the	 post-hoc	 test	 (B)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	differences	(*)	between	the	 selected	 seatback	motor	 parameter	 trial	 (red)	 and	 the	 Control	 seat,	 the	No	Motion	condition	 or	 other	 Experimental	 trials.	 Panels	 C	 and	 D	 show	 the	 resulting	 seatback	deformation	 at	 the	 T1	 spinal	 level	 and	 at	 the	 pelvis	 (dpenetration-T1	 and	 dpenetration-pelvis,	respectively).	......................................................................................................................	205	Figure	C7.	Exemplar	ATD	responses	for	the	Control	seat	(blue	lines	and	markers),	the	No	Motion	condition	(magenta	 lines	and	markers),	and	the	Experimental	seat	with	various	onsets	of	the	seatback	motor	rotation	(tdeformation-onset;	grey,	green,	and	red	lines	and	markers)	during	a	 12	 km/h	 collision	 severity	 (Δt	 =	 195	 ms).	 Calibrated	 responses	 (A	 –	 C)	 and	 their	corresponding	peak	responses	(D	–	F)	 for	head	and	torso	accelerations	 in	the	X-direction	(aX-head	 &	 aX-T1)	 and	 neck	 injury	 criteria	 (NIC),	 respectively.	 	 The	 green	 lines	 and	markers				xxix	indicate	the	initial	seatback	motor	onset	used	in	the	previous	two	experiments	(tdeformation-onset	 =	 -130	 ms)	 and	 the	 red	 lines	 and	 markers	 indicate	 the	 seatback	 motor	 onset	parameter	(tdeformation-onset	=	-200	ms)	selected	for	further	testing.	....................................	206	Figure	 C8.	 Seatback	 motor	 profiles	 and	 ATD	 responses	 to	 the	 Control	 seat	 (blue	 lines	 and	markers),	 the	 No	Motion	 condition	 (magenta	 lines	 and	markers),	 and	 the	 Experimental	seat	with	various	combinations	of	θdeformation-peak	and	ωdeformation-int	parameters	at	a	constant	tdeformation-onset	 =	 -200	 ms	 (grey	 and	 red	 lines	 and	 markers)	 during	 a	 12	 km/h	 collision	severity	(Δt	=	195	ms,	black	 lines,	Panel	A,	right	axis).	The	red	lines	and	markers	 indicate	the	 seatback	 motor	 profile	 (θdeformation-peak	 =	 168	 deg,	 ωdeformation-int	 =	 980	 deg/s,	 and	tdeformation-onset	=	-200	ms)	determined	from	Experiment	1	and	selected	for	further	testing.	The	results	of	the	post-hoc	test	(B)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	differences	(*)	between	the	selected	seatback	motor	parameter	trial	(red)	and	the	Control,	the	No	Motion	or	other	Experimental	trials.	Panels	C	and	D	show	the	resulting	seatback	deformation	at	 the	T1	 spinal	 level	 and	at	 the	pelvis	 (dpenetration-T1	 and	dpenetration-pelvis,	respectively).	..............................................................................................................	208	Figure	C9.	Exemplar	ATD	responses	for	the	Control	seat	(blue	lines	and	markers),	the	No	Motion	condition	 (magenta	 lines	 and	 markers),	 and	 the	 Experimental	 seat	 with	 various	combinations	of	θdeformation-peak	and	ωdeformation-int	parameters	at	a	constant	tdeformation-onset	=	-200	ms	(grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	195	ms).	Calibrated	responses	(A	–	C)	and	their	corresponding	peak	responses	(D	–	F)	for	head	and				xxx	torso	 accelerations	 in	 the	 X-direction	 (aX-head	 &	 aX-T1)	 and	 neck	 injury	 criteria	 (NIC),	respectively.		The	red	lines	and	markers	indicate	the	seatback	motor	profile	(θdeformation-peak	=	 168	 deg,	 ωdeformation-int	 =	 980	 deg/s,	 and	 tdeformation-onset	 =	 -200	 ms)	 determined	 from	Experiment	1	and	selected	for	further	testing.	..................................................................	209	Figure	 C10.	 Seatback	 motor	 profiles	 and	 ATD	 responses	 to	 the	 Control	 seat	 (blue	 lines	 and	markers),	 the	 No	Motion	 condition	 (magenta	 lines	 and	markers),	 and	 the	 Experimental	seat	with	either	different	combinations	of	pre-perturbation	forward	seatback	deformation	pulses	parameters	(θdeformation-forward	and	tdeformation-forward,	grey	lines	and	markers)	or	without	a	pre-perturbation	(red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	195	ms,	black	 lines,	 Panel	 A,	 right	 axis).	 The	 red	 lines	 and	markers	 indicate	 the	 seatback	motor	profile	 (θdeformation-peak	 =	 168	deg,	ωdeformation-int	 =	 980	deg/s,	 and	 tdeformation-onset	 =	 -200	ms)	determined	 from	 Experiment	 1	 &	 2	 and	 selected	 for	 further	 testing.	 The	 results	 of	 the	post-hoc	 test	 (B)	 determined	 the	 mean	 rank	 of	 each	 trial	 as	 well	 as	 any	 significant	differences	(*)	between	the	selected	seatback	motor	parameter	trial	(red)	and	the	Control	seat,	 the	 No	 Motion	 condition	 or	 other	 Experimental	 trials.	 Panels	 C	 and	 D	 show	 the	resulting	 seatback	 deformation	 at	 the	 T1	 spinal	 level	 and	 at	 the	 pelvis	 (dpenetration-T1	 and	dpenetration-pelvis,	respectively).	..............................................................................................	211	Figure	 C11.	 Exemplar	 ATD	 responses	 to	 the	 Control	 seat	 (blue	 lines	 and	 markers),	 the	 No	Motion	 condition	 (magenta	 lines	 and	 markers),	 and	 the	 Experimental	 seat	 with	 either	different	 combinations	 of	 pre-perturbation	 forward	 seatback	 deformation	 pulses				xxxi	parameters	(θdeformation-forward	and	tdeformation-forward,	grey	lines	and	markers)	or	without	a	pre-perturbation	 (red	 lines	 and	markers)	 during	 a	 12	 km/h	 collision	 severity	 (Δt	 =	 195	ms).	Calibrated	responses	(A	–	C)	and	their	corresponding	peak	responses	(D	–	F)	for	head	and	torso	 accelerations	 in	 the	 X-direction	 (aX-head	 &	 aX-T1)	 and	 neck	 injury	 criteria	 (NIC),	respectively.		The	red	lines	and	markers	indicate	the	seatback	motor	profile	(θdeformation-peak	=	 168	 deg,	 ωdeformation-int	 =	 980	 deg/s,	 and	 tdeformation-onset	 =	 -200	 ms)	 determined	 from	Experiment	1	&	2	and	selected	for	further	testing.	............................................................	212	Figure	C12.	Horizontal	seatback	cushion	deformation	of	the	T1	vertebra	and	pelvis	(dpenetration-T1,	green	line	&	dpenetration-pelvis,	blue	line)	in	response	to	the	seatback	motor	rotation	(θseatback-motor,	grey	line)	with	θdeformation-peak	=	168	deg,	ωdeformation-int	=	980	deg/s,	and	tdeformation-onset	=	-200	 (deformation	profile	parameters,	cyan	 lines)	during	a	12	km/h	collision	severity	 (aX-sled,	Δt	=	195	ms,	black	line,	right	axis).	...............................................................................	214		Figure	 D1.	 Programmed	 inputs	 and	 resulting	 outputs	 for	 A.)	 seat	 hinge	 rotation	 and	 B.)	seatback	cushion	deformation	profiles	during	a	12	km/h	collision	(∆t	=	187	ms;	black	line).		Grey	 lines	 represent	 input	 rotation	 (θseat-hinge)	 and	deformation	 (θseatback-motors)	 profiles	 to	seat	hinge	and	seatback	motors,	respectively.	Seat	hinge	rotation	(θseat-hinge-output:	red	line)	as	well	as	rearward	penetration	of	ATD	T1	(dpenetration-T1:	green	 line)	and	pelvis	 (dpenetration-pelvis:	 blue	 line)	 into	 the	 seatback	 indicate	 resulting	 outputs	 to	 their	 respective	 input	profiles.	..............................................................................................................................	217				xxxii	Figure	D2.	Seat	hinge	and	seatback	motor	profiles	as	well	as	ATD	responses	to	the	Control	seat	(blue	 lines	and	markers),	 the	No	Motion	 condition	 (magenta	 lines	and	markers)	 and	 the	Experimental	 seat	 with	 various	 seat	 hinge	 motor	 onsets	 at	 a	 constant	 seatback	 motor	onset	(trotation-onset	&	tdeformation-onset	=	-200	ms;	grey	and	red	lines	and	markers)	during	a	12	km/h	 collision	 severity	 (Δt	 =	 187	ms,	 black	 lines,	 right	 axis).	 The	 red	 lines	 and	markers	indicate	the	seat	hinge	onset	 (trotation-onset	=	 -90	ms)	selected	 for	 further	 testing.	Panels	A	and	B	show	the	programmed	input	and	output	seat	hinge	rotation	profiles.	Panel	D	shows	the	programmed	input	seatback	motor	profiles	with	resulting	seatback	deformation	at	the	T1	 spinal	 level	 (dpenetration-T1)	 and	 at	 the	 pelvis	 (dpenetration-pelvis)	 in	 Panels	 E	 and	 F,	respectively.	The	results	of	the	post-hoc	test	(C)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	differences	(*)	between	the	selected	seatback	motor	parameter	trial	(red)	and	the	Control,	No	Motion	or	other	experimental	trials.	........................................	219	Figure	 D3.	 Exemplar	 ATD	 responses	 for	 the	 Control	 seat	 (blue	 lines	 and	 markers),	 the	 No	Motion	 condition	 (magenta	 lines	 and	markers),	 and	 the	 Experimental	 seat	 with	 various	seat	hinge	motor	onsets	and	a	constant	seatback	motor	onset	(trotation-onset	&	tdeformation-onset	=	 -200	ms;	grey	and	red	 lines	and	markers)	during	a	12	km/h	collision	severity	 (Δt	=	187	ms).	Calibrated	responses	(A	–	C)	and	their	corresponding	peak	responses	(D	–	F)	for	head	and	 torso	 accelerations	 in	 the	 X-direction	 (aX-head	 &	 aX-T1)	 and	 neck	 injury	 criteria	 (NIC),	respectively.	 	 The	 red	 lines	 and	markers	 indicate	 the	 seatback	motor	parameter	 (trotation-onset	=	-90	ms)	selected	for	further	testing.	.........................................................................	220				xxxiii	Figure	D4.	Seat	hinge	and	seatback	motor	profiles	as	well	as	ATD	responses	to	the	Control	seat	(blue	 lines	and	markers),	 the	No	Motion	 condition	 (magenta	 lines	and	markers)	 and	 the	Experimental	 seat	 with	 various	 seatback	 motor	 onsets	 at	 a	 constant	 seat	 hinge	 motor	onset	 (tdeformation-onset	&	 trotation-onset	 =	 -90	ms;	grey	and	 red	 lines	and	markers)	during	a	12	km/h	 collision	 severity	 (Δt	 =	 187	ms,	 black	 lines,	 right	 axis).	 The	 red	 lines	 and	markers	indicate	the	seat	hinge	onset	(tdeformation-onset	=	-200	ms)	selected	for	further	testing.	Panels	A	 and	 B	 show	 the	 programmed	 input	 and	 output	 seat	 hinge	 rotation	 profiles.	 Panel	D	shows	the	programmed	input	seatback	motor	profiles	with	resulting	seatback	deformation	at	 the	 T1	 spinal	 level	 (dpenetration-T1)	 and	 at	 the	 pelvis	 (dpenetration-pelvis)	 in	 Panels	 E	 and	 F,	respectively.	The	results	of	the	post-hoc	test	(C)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	differences	(*)	between	the	selected	seatback	motor	parameter	trial	(red)	and	the	Control,	No	Motion	or	other	Experimental	trials.	........................................	222	Figure	 D5.	 Exemplar	 ATD	 responses	 for	 the	 Control	 seat	 (blue	 lines	 and	 markers),	 the	 No	Motion	 condition	 (magenta	 lines	 and	markers),	 and	 the	 Experimental	 seat	 with	 various	seatback	motor	onsets	and	a	constant	seat	hinge	motor	onset	(tdeformation-onset	&	trotation-onset	=	-90	ms;	grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	187	ms).		Calibrated	responses	(A	–	C)	and	their	corresponding	peak	responses	(D	–	F)	for	head	and	torso	 accelerations	 in	 the	 X-direction	 (aX-head	 &	 aX-T1)	 and	 neck	 injury	 criteria	 (NIC),	respectively.		The	red	lines	and	markers	indicate	the	seatback	motor	parameter	(tdeformation-onset	=	-200	ms)	selected	for	further	testing.	.......................................................................	223				xxxiv	Figure	D6.	Seat	hinge	and	seatback	motor	profiles	as	well	as	ATD	responses	to	the	Control	seat	(blue	 lines	and	markers),	 the	No	Motion	 condition	 (magenta	 lines	and	markers)	 and	 the	Experimental	 seat	 with	 shifted	 seat	 hinge	 and	 seatback	 motor	 onsets	 (trotation-onset	 &	tdeformation-onset;	grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	187	ms,	black	lines,	right	axis).	The	red	lines	and	markers	indicate	the	seat	hinge	onset	(trotation-onset	 =	 -90	ms	 and	 tdeformation-onset	 =	 -200	ms)	 selected	 for	 further	 testing.	 Panels	A	and	B	show	the	programmed	 input	and	output	 seat	hinge	 rotation	profiles.	Panel	D	 shows	 the	programmed	input	seatback	motor	profiles	with	resulting	seatback	deformation	at	the	T1	spinal	 level	 (dpenetration-T1)	and	at	the	pelvis	 (dpenetration-pelvis)	 in	Panels	E	and	F,	 respectively.	The	results	of	the	post-hoc	test	(C)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	differences	(*)	between	the	selected	seatback	motor	parameter	trial	(red)	and	the	Control,	No	Motion	or	other	experimental	trials.	........................................................	226	Figure	 D7.	 Exemplar	 ATD	 responses	 for	 the	 Control	 seat	 (blue	 lines	 and	 markers),	 the	 No	Motion	 condition	 (magenta	 lines	 and	 markers),	 and	 the	 Experimental	 seat	 with	 shifted	seat	hinge	and	seatback	motors	onsets	(trotation-onset	&	tdeformation-onset;	grey	and	red	lines	and	markers)	during	a	12	km/h	collision	 severity	 (Δt	=	187	ms).	Calibrated	 responses	 (A	–	C)	and	their	corresponding	peak	responses	(D	–	F)	for	head	and	torso	accelerations	in	the	X-direction	 (aX-head	 &	 aX-T1)	 and	 neck	 injury	 criteria	 (NIC),	 respectively.	 	 The	 red	 lines	 and	markers	indicate	the	seat	hinge	and	seatback	motor	parameters	(trotation-onset	=	-90	ms	and	tdeformation-onset	=	-200	ms)	selected	for	further	testing.	.......................................................	227				xxxv		Figure	E2.1.	Experimental	data	for	the	BioRID	II	ATD	seated	on	a	2005	Volvo	S40	WHIPS	seat.	Each	column	represents	occupant	responses	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6,	8,	10,	12	and	14	km/h	with	a	collision	pulse	duration	(Δt)	of	141	ms).	Hollow	circles	represent	the	onset	of	head-to-head-restraint	contact	and	peak	responses	of	each	ATD	response	parameter	for	each	trial.	.............................................................................	231	Figure	E2.2.	Experimental	data	for	the	BioRID	II	ATD	seated	on	a	2004	Volvo	S60	WHIPS	seat.	Each	column	represents	occupant	responses	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6,	8,	10,	12	and	14	km/h	with	a	collision	pulse	duration	(Δt)	of	141	ms).	Hollow	circles	represent	the	onset	of	head-to-head-restraint	contact	and	peak	responses	of	each	ATD	response	parameter	for	each	trial.	.............................................................................	232	Figure	E2.3.	Experimental	data	for	the	BioRID	II	ATD	seated	on	a	2005	Saab	9.3	SAHR	seat.	Each	column	represents	occupant	responses	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6,	8,	10,	12	and	14	km/h	with	a	collision	pulse	duration	(Δt)	of	141	ms).	Hollow	circles	represent	 the	 onset	 of	 head-to-head-restraint	 contact	 and	 peak	 responses	 of	 each	 ATD	response	parameter	for	each	trial.	.....................................................................................	233	Figure	E2.4.	Experimental	data	for	the	BioRID	II	ATD	seated	on	a	2004	Pontiac	Grand	Am	GMHR	seat.	 Each	 column	 represents	 occupant	 responses	 while	 exposed	 to	 various	 collision	severities	 (Δv	=	2,	4,	6,	8,	10,	12	and	14	km/h	with	a	collision	pulse	duration	 (Δt)	of	141				xxxvi	ms).	 Hollow	 circles	 represent	 the	 onset	 of	 head-to-head-restraint	 contact	 and	 peak	responses	of	each	ATD	response	parameter	for	each	trial.	...............................................	234		Figure	E5.1.	Experimental	data	for	the	BioRID	II	ATD	seated	on	the	Experimental	seat	utilizing	both	 seat	 hinge	 rotation	 and	 seatback	 cushion	 deformation	 safety	 mechanisms.	 Each	column	represents	occupant	responses	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6	and	8	km/h,	Δt	=	147	ms	and	Δv	=	12	km/h,	Δt	=	187	ms).	Hollow	circles	represent	the	 onset	 of	 head-to-head-restraint	 contact	 and	 peak	 responses	 of	 each	 ATD	 response	parameter	for	each	trial.	Whiplash	perturbation	onset	occurred	at	t	=	0	ms.	...................	235	Figure	E5.2.	Experimental	data	for	the	BioRID	II	ATD	seated	on	a	2005	Volvo	S40	WHIPS	seat.	Each	column	represents	occupant	responses	while	exposed	to	various	collision	severities	(Δv	 =	 2,	 4,	 6	 and	 8	 km/h,	 Δt	 =	 147	ms	 and	 Δv	 =	 12	 km/h,	 Δt	 =	 187	ms).	 Hollow	 circles	represent	 the	 onset	 of	 head-to-head-restraint	 contact	 and	 peak	 responses	 of	 each	 ATD	response	parameter	for	each	trial.	Whiplash	perturbation	onset	occurred	at	t	=	0	ms.	...	236	Figure	E5.3.	Experimental	data	for	the	BioRID	II	ATD	seated	on	a	2004	Volvo	S60	WHIPS	seat.	Each	column	represents	occupant	responses	while	exposed	to	various	collision	severities	(Δv	 =	 2,	 4,	 6	 and	 8	 km/h,	 Δt	 =	 147	ms	 and	 Δv	 =	 12	 km/h,	 Δt	 =	 187	ms).	 Hollow	 circles	represent	 the	 onset	 of	 head-to-head-restraint	 contact	 and	 peak	 responses	 of	 each	 ATD	response	parameter	for	each	trial.	Whiplash	perturbation	onset	occurred	at	t	=	0	ms.	...	237				xxxvii	Figure	E5.4.	Experimental	data	for	the	BioRID	II	ATD	seated	on	a	2005	Saab	9.3	SAHR	seat.	Each	column	represents	occupant	responses	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6	and	8	km/h,	Δt	=	147	ms	and	Δv	=	12	km/h,	Δt	=	187	ms).	Hollow	circles	represent	the	 onset	 of	 head-to-head-restraint	 contact	 and	 peak	 responses	 of	 each	 ATD	 response	parameter	for	each	trial.	Whiplash	perturbation	onset	occurred	at	t	=	0	ms.	...................	238	Figure	E5.5.	Experimental	data	for	the	BioRID	II	ATD	seated	on	a	2004	Pontiac	Grand	Am	GMHR	seat.	 Each	 column	 represents	 occupant	 responses	 while	 exposed	 to	 various	 collision	severities	(Δv	=	2,	4,	6	and	8	km/h	with	a	collision	pulse	duration	(Δt)	=	147	ms	and	Δv	=	12	km/h	with	 a	 Δt	 =	 187	ms).	 Hollow	 circles	 represent	 the	 onset	 of	 head-to-head-restraint	contact	 and	 peak	 responses	 of	 each	 ATD	 response	 parameter	 for	 each	 trial.	 Whiplash	perturbation	onset	occurred	at	t	=	0	ms.	...........................................................................	239		 				xxxviii	Acronyms	and	Abbreviations	a	 Significance	level	AIS	 Abbreviated	injury	scale	AOJ	 Atlanto-occiptal	joint	aresultant-T1	 Forward	resultant	acceleration	of	T1	ATD	 Anthropomorphic	test	device	aX-head	 Linear	forward	acceleration	of	the	head	aX-sled	 Linear	forward	acceleration	of	the	whiplash	sled	aX-T1	 Linear	forward	acceleration	of	the	T1	vertebra	BioRID	 Biofidelic	rear	impact	dummy	COV	 Coefficient	of	variation	DAQ	 Data	acquisition	dbackset	 Horizontal	backset	of	the	head	restraint	relative	to	the	head	df		 Degrees	of	freedom	dheight	 Vertical	height	of	the	head	restraint	relative	to	the	head	dpenetration-pelvis	 Occupant	penetration	of	the	pelvis	into	the	seatback	Drebound	 Forward	rebound	of	the	head	relative	to	the	head	restraint	EXP	 Experimental	seat	EXP90	 Experimental	condition	with	dbackset	=	86.2	mm	and	dheight	=	-24.6	mm		FSR	 Force	sensitive	resistor	FX	 Upper	neck	shear	force	FZ	 Upper	neck	axial	force	GM	 General	Motor	GMHR	 General	Motor	High	Retention	seat	ICBC	 Insurance	Corporation	of	British	Columbia	IIHS	 Insurance	Institute	for	Highway	Safety	IIWPG	 International	Insurance	Whiplash	Prevention	Group	IRED	 Infrared	light	emitting	diode	MY	 Upper	neck	flexion/extension	bending	moment	n	 Number	of	trials	NHTSA	 National	Highway	Traffic	Safety	Administration	NIC	 Neck	Injury	Criterion	Nij	 Normalized	Neck	Injury	Criterion				xxxix	Nkm	 Neck	Protection	Criterion	p		 Significance	level	RCAR	 Research	Council	for	Automobile	Repairs	RPM	 Revolutions	per	minute	RX	 Retraction	SAHR	 Saab’s	Active	Head	Restraint	SD	 Standard	deviation	T1	 First	thoracic	vertebrae	tdeformation-forward	 Onset	of	seatback	motor	rotation	to	control	deformation	tdeformation-onset	 Onset	of	rearward	seatback	motor	rotation	to	control	deformation	tforward-onset	 Onset	of	forward	seat	hinge	rotation	trearward-onset	 Onset	of	rearward	seat	hinge	rotation	trotation-onset	 Onset	of	seat	hinge	rotation	WAD	 Whiplash-associate	disorders	WHIPS	 Volvo’s	Whiplash	Injury	Prevention	seat	WIL	 Toyota’s	Whiplash	Injury	Lessening	seat	Δt	 Collision	pulse	duration	Δthead-contact	 Head-to-head-restraint	contact	time	Δv	 Collision	speed	change	θdeformation-peak	 Peak	angle	of	rearward	seatback	motor	rotation	to	control	deformation	θdeformation-peak	 Peak	angle	of	forward	seatback	motor	rotation	to	control	deformation	θforward-peak	 Peak	angle	of	forward	seat	hinge	rotation	θhead	 Head	extension	angle	θrearward-peak	 Peak	angle	of	rearward	seat	hinge	rotation	θseat-hinge	 Seat	hinge	rotation	angle	χ2	 Chi-squared	value	ωdeformation-forward	 Initial	angular	velocity	of	forward	seatback	motor	rotation	to	control	deformation	ωdeformation-int	 Initial	angular	velocity	of	rearward	seatback	motor	rotation	to	control	deformation	ωrearward-int	 Initial	angular	velocity	of	rearward	seat	hinge	rotation	ωY-head	 Angular	velocity	of	the	head	ωY-T1	 Angular	velocity	of	the	T1	vertebra				xl	Acknowledgements	Undertaking	this	Doctor	of	Philosophy	(PhD)	degree	has	been	a	truly	memorable	and	life-changing	 experience	 for	 me	 and	 it	 would	 not	 have	 been	 possible	 without	 the	 support	 and	guidance	I	received	from	many	people		Firstly,	 I	 would	 like	 to	 express	 my	 sincerest	 gratitude	 to	 my	 co-supervisors,	 Drs.	 Jean-Sébastien	 Blouin	 and	 Gunter	 Siegmund,	 for	 their	 continued	 patience,	 guidance	 and	 support	throughout	 the	 duration	 of	 my	 PhD.	 	 I	 have	 been	 extremely	 lucky	 to	 have	 supervisors	 who	cared	so	much	about	my	work	and	were	always	available	 to	answer	my	questions	no	matter	how	 trivial	 they	 might	 be.	 	 Thank	 you	 to	 Jean-Sébastien	 and	 Gunter	 for	 your	 wealth	 of	knowledge	in	the	field	of	whiplash	 injuries	and	constant	guidance	to	help	better	my	technical	writing	 skills.	 	 It	 has	 been	 a	 long	 journey,	 but	 without	 your	 constant	 feedback,	 words	 of	encouragement	and	faith	in	me	as	a	researcher	this	PhD	would	not	have	been	achievable.		Many	 thanks	 to	 Dr.	 Douglas	 Romilly	 for	 being	 a	 part	 of	 my	 PhD	 committee	 and	 for	introducing	 me	 to	 the	 AUT021	 Network	 Center	 for	 Excellence.	 	 Thank	 you	 for	 your	encouragement,	support	and	time	to	discuss	my	research.		To	Mr.	Jeff	Nickel	and	Mr.	Mircea	Oala-Florescu	of	MEA	Forensic	Engineers	and	Scientists,	who	work	endlessly	behind	the	scenes	to	ensure	that	all	the	students	have	what	they	need	no	matter	 how	 crazy	 the	 idea	might	 be.	 	 Thank	 you	 both	 for	 your	willingness	 to	 always	 lend	 a	helping	hand	with	big	smiles.		Your	guidance	and	assistance	in	the	design	and	fabrication	of	the				xli	anti-whiplash	seat	and	seat	support	structures	utilized	throughout	this	dissertation	were	greatly	appreciated.	Completing	this	work	would	have	been	all	the	more	difficult	were	it	not	for	the	support	of	my	friends,	fellow	graduate	students	and	colleagues	in	the	Sensorimotor	Physiology	Laboratory.	I	 am	 indebted	 to	 them	 for	 their	 patience,	moral	 support	 and	 years	of	 friendship.	Also,	many	thanks	 to	 the	 numerous	 undergraduate	 students	 who	 helped	 in	 the	 design,	 fabrication	 and	programming	 of	 the	 anti-whiplash	 automotive	 seat:	 Devin	 Luu,	 Roein	Manafi,	 Alex	 Jew	 and	Geoffery	Hohert.		I	would	like	to	say	a	heartfelt	thank	you	to	my	family	for	providing	me	with	support	and	encouragement	to	pursue	my	dreams.	This	accomplishment	was	only	possible	because	of	the	many	opportunities	you	provided	me.	And	finally,	to	my	loving	wife	Betty,	thank	you	for	your	patience,	emotional	support	and	encouragement	 throughout	 my	 entire	 PhD.	 	 You	 were	 always	 there	 to	 keep	 me	 grounded	during	the	high	points	and	always	found	the	right	words	to	support	me	during	the	low	points.		I	am	eternally	grateful	to	you	for	taking	care	of	Oliver	these	past	months	and	making	it	possible	for	me	to	complete	what	I	started.			  	 				xlii	Dedication	To	Oliver,	never	stop	seeking	knowledge	and	exploring	the	world.			 					1	Chapter	1. Introduction		1.1 Literature	Review	1.1.1 Epidemiology		Whiplash	 injuries	 are	 the	most	 common	 type	 of	 injuries	 associated	with	motor	 vehicle	collisions.	Rear-end	vehicle-to-vehicle	collisions	pose	the	greatest	risk	of	whiplash	injury	(38%)	when	compared	to	other	collision	directions	such	as	frontal	 (15.7%)	and	side	(10.8%)	 impacts	(ICBC,	 2006,	 ICBC,	 2007,	 Jakobsson	 et	 al.,	 2000).	 For	 rear-end	 collisions,	 whiplash	 injury	 is	commonly	 defined	 as	 any	 injury	 to	 a	 person’s	 neck	 resulting	 when	 the	 head	 is	 suddenly	accelerated	backwards	into	extension	and	then	forwards	into	flexion.	Whiplash	injuries	account	for	 approximately	 70%	 of	 all	 injury	 claims	 made	 to	 the	 Insurance	 Corporation	 of	 British	Columbia	 (ICBC)	 (ICBC,	 2000).	 	 In	 2000,	 ICBC	 paid	 out	 over	 $500	 million—50%	 of	 all	 injury	payouts—for	whiplash	 injuries	 and	whiplash-associated	 disorders	 (WAD).	 	 Half	 of	 the	money	was	spent	on	medical	treatments	for	occupants	and	the	other	half	covered	the	lost	wages	and	diminished	 earning	 capacity	 of	 drivers	 and	 occupants	 injured	 (ICBC,	 2000).	 	 In	 2006,	 the	estimated	 cost	 of	 whiplash	 injuries	 and	 WAD	 (excluding	 litigation	 costs)	 increased	 to	approximately	$850	million	(ICBC,	2007).		The	annual	incidence	of	whiplash	injuries	in	the	western	world	ranges	from	28	to	834	per	100,000	inhabitants	(Cassidy	et	al.,	2000,	Holm	et	al.,	2008,	Otremski	et	al.,	1989).		Females	are	1.2	to	3	times	more	likely	to	suffer	from	whiplash	injuries	after	a	rear-end	automotive	collision					2	as	compared	to	males	(Mordaka	and	Gentle,	2003,	Harder	et	al.,	1998,	Versteegen	et	al.,	2000).	Females	between	 the	 ages	 20	 to	 24	present	 the	highest	 incidence	 rate	of	 reported	whiplash	injuries	 with	 965	 cases	 per	 100,000	 people	 annually	 (Quinlan	 et	 al.,	 2004).	 	 The	 greater	susceptibility	 of	 younger	 females	 is	 hypothesized	 to	 be	 related	 to	 the	 sex	 differences	 in	anatomical,	physiological,	behavioral	and	sociological	parameters	(Mordaka	and	Gentle,	2003)		as	well	as	the	influence	of	seat	properties	on	neck	biomechanics	and	occupant	dynamics	(Viano,	2003c).		 The	 type	 of	 injury	 and	 recovery	 times	 following	 whiplash	 collisions	 vary	 greatly	 and	depend	on	 factors	 such	as	 impact	 severity,	 seat	position,	 seat	 stiffness,	and	 initial	posture	of	the	 occupant	 (Suissa	 et	 al.,	 2001,	 Viano,	 2003b).	 The	 most	 common	 symptoms	 following	whiplash	 injuries	 are	 neck	 pain	 (88-100%	 of	 patients)	 and	 headaches	 (54-66%	 of	 patients)	(Todman,	 2007).	 	 Other	 symptoms	 of	whiplash	 injuries	 include	 dizziness,	 auditory	 symptoms	(tinnitus	–	perceived	ringing	noise	in	the	ears),	paresthesias	in	the	upper	extremities,	and	back	pain	(Mordaka	and	Gentle,	2003,	Evans,	1992,	Sterner	and	Gerdle,	2004,	Spitzer	et	al.,	1995).	The	 recovery	 time	 from	whiplash	 injuries	depends	on	 the	 initial	whiplash	 injury	 severity,	 but	26%	of	occupants	recover	within	the	first	week	and	the	median	recovery	time	is	approximately	32	days	(Suissa	et	al.,	2001).		However,	12%	of	occupants	do	not	fully	recover	within	six	months	(Suissa	et	al.,	2001)	and	5	to	8%	of	patients	do	not	return	to	work	within	a	year	following	the	collision	(Buitenhuis	et	al.,	2009,	Evans	et	al.,	2001).		Between	14	and	42%	of	injured	occupants	develop	 chronic	 neck	 pain	 and	 approximately	 10%	 are	 left	 with	 permanent	 severe	 pain	 and					3	disability	 (Barnsley	 et	 al.,	 1994).	Due	 to	 the	persistence	of	 chronic	 symptoms	of	WAD,	 these	injuries	 are	 a	 serious	 economic	 and	 social	 burden	 to	 society.	 	 Therefore,	 it	 is	 important	 to	reduce	the	incidence	rate	and	the	risk	of	whiplash	injuries	during	rear-end	collisions.	1.1.2 Biomechanics	of	Whiplash	Injury	The	biomechanics	of	human	occupants	during	a	low-speed,	rear-end	collision	are	variable	and	 depend	on	 the	magnitude	 and	 shape	 of	 the	 acceleration	 pulse,	 seatback	 properties	 (i.e.	seat	 hinge	 stiffness,	 seatback	 cushion	 compliance	 and	 seatback	 angle)	 and	 an	 occupant’s	anthropometry	and	initial	posture.	To	help	understand	the	whiplash	motions,	the	movements	of	 the	 torso,	 head	 and	 neck	 can	 be	 divided	 into	 two	 phases:	 the	 retraction	 phase	 and	 the	rebound	phases	(Pearson	et	al.,	2004,	Brault	et	al.,	2000,	Vasavada	et	al.,	2007).		The	retraction	phase	is	defined	from	onset	of	head	movement	to	peak	linear	rearward	excursion	of	the	head	relative	to	the	torso;	whereas,	the	rebound	phase	is	defined	from	peak	head	retraction	to	peak	forward	 excursion	 of	 the	 head	 relative	 to	 the	 torso.	 During	 the	 retraction	 phase,	 forces	 are	applied	by	the	seatback	to	the	torso	as	the	seat	is	accelerated	forward	by	the	vehicle-to-vehicle	impact.	 	The	torso	 is	accelerated	forward	by	the	seat	earlier	than	the	head,	which	causes	the	head	to	lag	behind	(Luan	et	al.,	2000,	McConnell	et	al.,	1995)	and	forces	the	cervical	spine	into	a	non-physiological	“S”-shaped	curve,	in	which	the	lower	cervical	vertebrae	are	in	extension	and	the	upper	cervical	vertebrae	are	in	flexion.		This	shearing	motion	between	the	top	and	bottom	of	the	cervical	spine	stretches	portions	of	the	cervical	facet	joint	capsular	ligament	(Luan	et	al.,	2000,	Pearson	et	al.,	2004)	and	other	ligaments	in	the	neck.							4	After	peak	head	retraction,	the	head	 is	accelerated	forward	relative	to	the	torso	due	to	the	internally	generated	forces	of	the	neck	(McConnell	et	al.,	1995)	and	rebound	from	the	head	restraint.	 The	 head	 is	 accelerated	 forward	 and	 overtakes	 the	 torso	 before	 being	 actively	decelerated	 by	 the	 neck	muscles	 and	other	 passive	 structures	 of	 the	 neck	 (McConnell	 et	 al.,	1995).	 	 	 After	 the	 head	 reaches	maximum	 forward	 excursion	 with	 respect	 to	 the	 torso,	 the	occupant’s	head	and	torso	return	back	to	their	pre-perturbation	position.	1.1.3 Injury	Mechanisms	of	Whiplash	The	etiology	of	whiplash	 injury	 remains	uncertain,	but	 there	have	been	many	proposed	injury	 mechanisms.	 Some	 possible	 anatomical	 sites	 of	 injury	 in	 the	 neck	 include	 the	 neck	muscles,	 spinal	 ligaments,	 intervertebral	 discs,	 vertebral	 arteries,	 dorsal	 root	 ganglia	 and	 the	facet	 joints.	 Dorsal	 root	 ganglia	 damage	 and	 cervical	 facet	 joint	 disruption	 are	 two	 possible	whiplash	 injury	mechanisms	 that	 have	 been	most	 thoroughly	 examined	 through	 animal	 and	human	 occupant	 experiments,	 as	well	 as	 clinical	 studies.	 The	 dorsal	 root	 ganglia	 are	 located	along	the	vertebral	column	and	consist	of	a	collection	of	cell	bodies	from	afferent	nerve	fibers	that	 encode	 touch,	 stretch,	 temperature,	 pain,	 etc.,	 from	 the	 skin,	 muscle,	 joint	 and	 other	tissues.	Injury	to	the	cervical	dorsal	root	ganglia,	particularly	in	the	lower	cervical	region,	could	explain	most	of	the	symptoms	that	are	typically	associated	with	neck	injury	sustained	in	rear-end	collisions	(Bostrom	et	al.,	1996,	Svensson	et	al.,	1993,	Svensson	et	al.,	2000).		Svensson	et	al.	 (1993)	 proposed	 that	 the	 sudden	 cervical	 extension-flexion	motion	would	 cause	 pressure	gradients	in	the	spinal	canal	that	apply	injurious	stresses	and	strains	to	the	dorsal	root	ganglia.					5	Fluids	inside	the	spinal	canals	are	virtually	 incompressible	and	fluid	transportation	must	occur	during	the	flexion	and	extension	movements	to	relieve	pressure	within	the	cervical	spinal	canal	(Aldman,	 1986,	 Ortengren	 et	 al.,	 1996,	 Svensson	 et	 al.,	 1993).	 This	 injury	 mechanism	 is	supported	by	histopathological	studies	of	anaesthetized	pigs	exposed	to	swift	extension-flexion	motion	 of	 the	 cervical	 spine	 that	 produced	 nerve	 cell	 damage	within	 the	 dorsal	 root	 ganglia	(Ortengren	et	al.,	1996,	Svensson	et	al.,	1993).		Another	possible	source	of	pain	in	whiplash	injury	is	the	cervical	facet	joint.	Barnsley	et	al.	(1994)	performed	clinical	studies	that	identified	the	cervical	facet	joints	as	the	source	of	neck	pain	 in	 40	 –	 68%	 of	 patients	 with	 chronic	 whiplash	 injuries	 following	 a	 rear-end	 collisions	(Barnsley	 et	 al.,	 1995).	 The	 facet	 joint	 injury	 mechanism	 has	 further	 been	 support	 through	animal	studies	(Macnab,	1971),	cadaver	experiments	(Luan	et	al.,	2000,	Winkelstein	and	Myers,	2000)	and	human	volunteer	studies	(Kaneoka	et	al.,	1999).	There	are	three	possible	causes	of	injury	to	the	cervical	facet	joint:		joint	injury,	synovial	fold	injury	or	joint	capsule	injury.			Joint	 injuries	 can	 occur	 when	 the	 sudden	 flexion	 and	 extension	 of	 the	 neck	 causes	rotations	 of	 the	 cervical	 vertebral	 bodies	 about	 abnormally-located	 instantaneous	 axes	 of	rotation	 (Kaneoka	et	 al.,	 1999).	As	 a	 result,	 the	 articular	processes	of	 the	 facet	 joints	do	not	glide	normally	over	one	another,	but	instead	the	inferior	articular	processes	of	one	vertebra	are	forced	into	the	superior	articular	processes	of	the	subjacent	vertebra	(Bogduk	and	Yoganandan,	2001,	 Kaneoka	 et	 al.,	 1999).	 	 At	 low	 impact	 speeds	 these	 abnormal	 rotations	 will	 cause	 no	major	injury,	but	at	higher	speeds	(>	10	km/h)	the	facet	joint	may	be	injured.						6	Another	possible	source	of	cervical	 facet	 joint	pain	 is	 injury	 to	 the	synovial	 fold	 located	between	 the	 two	 articulating	 facet	 joints	 (Kaneoka	 et	 al.,	 1999).	 Inami	 et	 al.	 (2001)	 used	immunohistochemistry	 and	 immunostaining	 to	 show	 the	 presence	 of	 free	 nerve	 fibers,	including	 nociceptive	 fibers,	 in	 human	 cervical	 synovial	 folds.	 It	 has	 been	 hypothesized	 that	facet	joint	rotation	during	low-speed,	rear-end	collision	could	impinge	on	the	synovial	folds	to	cause	cervical	facet	joint	pain	(Kaneoka	et	al.,	1999).	The	 cervical	 facet	 capsule	 and	 its	 role	 in	 whiplash	 injury	 have	 been	 studied	 in	 both	animals	 (Winkelstein	and	Santos,	 2008)	 and	human	cadaver	neck	 segments	 (Siegmund	et	 al.,	2008b,	Winkelstein	et	al.,	2000).		In	rat	experiments,	allodynia	has	been	produced	by	applying	tension	across	the	joint	capsule	(Winkelstein	and	Santos,	2008).	Cadaveric	studies	of	whiplash	injury	have	also	shown	that	capsular	ligament	strains	during	the	whiplash	motion	are	largest	in	the	 lower	 cervical	 spine	 (C6/C7:	 39.9%,	 C5/C6:	 38.5%,	 C4/C5:	 26.5%,	 C3/C4:	 29.9%,	 C2/C3:	16.7%;	Pearson	et	al.,	 2004).	Due	 to	 the	direct	attachment	of	 the	cervical	multifidus	muscles	onto	the	capsular	ligaments	of	C4/5,	C5/6	and	C6/7	(Anderson	et	al.,	2005),	the	early	activation	of	 the	 multifidus	 muscles	 during	 a	 rear-end	 impact	 could	 further	 increase	 the	 capsular	ligaments	strain	 induced	by	the	intervertebral	kinematics	(Anderson	et	al.,	2005,	Siegmund	et	al.,	 2008a,	Winkelstein	 et	 al.,	 2000).	 Failure	 testing	 of	 the	 joint	 capsules	 has	 quantified	 the	maximum	capsular	strain	at	subcatastrophic	(35	–	64.6%	strain)	and	catastrophic	(94	–	103.6%	strain)	failure	(Siegmund	et	al.,	2001,	Winkelstein	et	al.,	2000).		A	head-turned	posture	has	been	shown	 to	 increase	 capsular	 strain	 significantly,	 by	 as	much	 as	 twice	 the	 strain	 in	 the	 neutral					7	head	position,	on	the	side	towards	which	the	head	is	turned	(Siegmund	et	al.,	2008b).	 	These	observations	 suggest	 that,	 in	 some	 individuals,	 cervical	 facet	 joint	 pain	 may	 be	 caused	 by	subcatastrophic	failure	due	to	excessive	capsular	strains,	which	is	significantly	 increased	if	the	occupant’s	head	is	turned.		1.1.4 Anti-Whiplash	Automotive	Seats	and	Existing	Safety	Features	The	automotive	seat	is	not	only	designed	for	comfort	but	is	also	the	primary	safety	device	for	 occupants	 in	 a	 rear-end	 collision.	 To	 motivate	 manufacturers	 to	 design	 seats	 and	 head	restraints	that	reduce	the	risk	of	whiplash	injury	in	low/moderate-speed	rear-end	collisions,	the	Insurance	Institute	for	Highway	Safety	(IIHS)	publishes	ratings	of	seats	and	head	restraints	using	a	 two-part	 protocol	 developed	by	 the	Research	Council	 for	Automobile	 Repairs/International	Insurance	Whiplash	 Prevention	 Group	 (RCAR/IIWPG)	 (Insurance	 Institute	 of	 Highway	 Safety,	2004,	Insurance	Institute	of	Highway	Safety,	2008a).		First,	a	static	test	evaluates	head	restraint	height	 (distance	between	the	top	of	 the	head	and	the	top	of	 the	head	restraint)	and	backset	(distance	between	the	back	of	the	head	and	the	front	of	the	head	restraint)	relative	to	the	head	of	a	median-sized	male	occupant.	If	the	top	of	the	head	restraint	is	no	more	than	8	cm	below	the	top	of	the	head	and	the	backset	is	no	more	than	9	cm	(i.e.,	rated	acceptable	or	good),	the	seat	undergoes	a	dynamic	test	that	simulates	a	rear-end	collision	with	a	peak	acceleration	of	10	g	and	a	speed	change	of	16	km/h	over	91	ms.	The	dynamic	test	is	graded	using	a	combination	of	peak	 upper	 neck	 forces	 (shear	 and	 axial),	 time-to-head-restraint	 contact,	 and	 peak	 forward	acceleration	of	the	T1	vertebra	of	a	BioRID	II	anthropometric	test	device	(ATD).		Seats	and	head					8	restraints	are	rated	as	good,	acceptable,	marginal	or	poor	depending	on	the	combined	rating	of	both	 the	 static	 and	 dynamic	 test	 results.	 In	 a	 comparison	 between	 RCAR/IIWPG	 ratings	 and	real-world	insurance	claims,	occupant	neck	injury	rates	were	only	11	–	15%	lower	in	seats	rated	good	 compared	 to	 seats	 rated	 poor	 (Farmer	 et	 al.,	 2008,	 Trempel	 et	 al.,	 2016).	 This	meager	benefit	 suggests	 that	 there	 remains	 considerable	 room	 for	 improving	 automotive	 seats	 and	perhaps	the	rating	system	to	better	protect	occupants	from	whiplash	injuries.	Currently,	 anti-whiplash	 seats	 attempt	 to	 mitigate	 the	 risk	 of	 whiplash	 injuries	 during	rear-end	collisions	by	reducing	the	relative	accelerations	and	displacements	between	the	head	and	 torso.	 	 Reducing	 these	 relative	 accelerations	 and	 displacements	 reduces	 the	 forces	 and	strains	 that	 develop	 in	 the	 tissues	 of	 the	 neck.	 The	 magnitude	 of	 these	 accelerations	 and	displacements	 are	modulated	 by	 the	 occupant’s	 interaction	with	 the	 seatback,	 and	 can	 vary	with	seatback	stiffness	(Viano,	2003b,	Viano,	2003d).		Stiffer	seatbacks	are	less	likely	to	deform	and	thus	provide	better	occupant	retention	but	result	in	higher	forces	applied	to	the	occupant.	Yielding	seatbacks	attenuate	forces	applied	to	the	occupants	but	increase	the	likelihood	of	the	seatback	 collapse	 and	 occupant	 ejection	 from	 the	 seat.	 	 In	 the	 late	 1990’s,	 General	Motors	(GM)	introduced	the	“high	retention”	seat	design	that	embodied	the	beneficial	characteristics	of	both	a	yielding	and	a	rigid	seat	design.	The	GM	High	Retention	seat	(GMHR)	utilized	a	stiffer	outer	seatback	 frame	with	a	yielding	seatback	center.	The	two	components	work	together	as	the	yielding	seatback	center	section	absorbs	energy	and	more	gradually	decelerates	the	torso					9	during	the	collision,	while	the	stiffer	seatback	frame	enhances	occupant	retention	and	prevents	large	seatback	deflections.		Volvo’s	Whiplash	 Injury	 Prevention	 seat	 (WHIPS)	 is	 another	 implementation	 of	 an	 anti-whiplash	device	 that	 reduces	occupant	accelerations	during	a	 collision.	 	 The	WHIPS	 seat	also	utilizes	a	 yielding	 seatback	 center,	but	 includes	an	energy-absorbing	 recliner	mechanism	 that	controls	the	motion	of	the	seatback	relative	to	the	seat	pan	(Jakobsson	et	al.,	2000,	Lundell	et	al.,	1998a).		The	recliner	mechanism	initially	translates	the	seatback	horizontally	rearward	with	respect	to	the	seat	pan	to	improve	an	occupant’s	head,	neck,	and	torso	alignment	as	well	as	to	absorb	 some	 of	 the	 energy	 generated	 by	 the	 collision.	 	 The	 mechanism	 then	 rotates	 the	seatback	rearward	to	further	reduce	the	accelerations	and	the	occupant’s	subsequent	forward	rebound	 (Lundell	 et	 al.,	 1998a).	 Since	 the	WHIPS	 system	 is	 passive,	 the	 degree	 of	 rearward	translation,	rearward	rotation,	and	overlap	of	the	two	motions	depends	on	parameters	such	as	occupant	mass,	seating	posture	and	crash	severity.			 Minimizing	the	displacement	of	the	head	relative	to	the	upper	torso	also	contributes	to	reducing	 the	 risk	 of	whiplash	 injury.	 	 Proper	 head	 restraint	 positioning	 is	 critical	 to	 reducing	neck	 extension	 and	 head-to-head-restraint	 responses	 during	 collisions	 to	 mitigate	 whiplash	injury	 (Viano,	2002).	Active	head	restraints	 (e.g.,	Saab’s	Active	Head	Restraint,	or	SAHR)	have	been	developed	to	 improve	head-to-head-restraint	alignment	during	collisions.	The	SAHR	 is	a	passive	device	that	reduces	the	need	for	an	occupant	to	manually	adjust	the	head	restraint	to	the	correct	position.	As	with	the	WHIPS	and	GMHR,	the	SAHR	seat	utilizes	a	stiff	outer	seatback					10	frame	with	a	yielding	seatback	center.	The	main	difference	in	the	SAHR	seat	is	the	addition	of	a	metal	 back	 plate	 located	within	 the	 seatback	 that	 is	 connected	 to	 the	 head	 restraint.	 As	 an	occupant’s	 torso	 penetrates	 the	 seatback,	 the	 torso	 triggers	 the	 back	 plate	 and	 linkage	mechanism	 to	move	 the	 head	 restraint	 upwards	 and	 forwards	 to	 the	 back	 of	 an	 occupant’s	head.	 	 The	SAHR	 is	designed	 to	move	 the	head	 restraint	early	 in	 the	 collision	 to	 support	 the	head/neck	complex	through	most	of	the	whiplash	motion.		Decreasing	the	backset	between	the	head	 and	 head	 restraint	 will	 result	 in	 a	 reduction	 in	 the	 rearward	 translation	 of	 the	 head	relative	to	the	torso.	1.1.5 Effectiveness	of	Current	Anti-Whiplash	Devices	Epidemiological	studies	of	current	anti-whiplash	devices	have	shown	the	benefits	of	both	the	WHIPS	 and	 SAHR	 in	 reducing	whiplash	 injury	 following	 rear-end	 collisions	 (Farmer	 et	 al.,	2003,	 Ivancic,	2011,	Viano	and	Olsen,	2001).	 In	a	2003	study,	Farmer	and	colleagues	analyzed	real-life	 insurance	 claims	 in	 the	United	 States	 occurring	 between	 January	 1st,	 1999	 and	 June	30th,	2001	and	observed	that	WHIPS	and	SAHR	decreased	the	neck	injury	rates	by	49%	and	43%,	respectively	 (Farmer	 et	 al.,	 2003).	 In	 another	 study	 conducted	 by	 Volvo,	 self-reported	 injury	data	from	2521	occupants	in	Volvo	cars	of	model	year	1999	(1858	occupants	in	WHIPS	vs.	663	occupants	 in	 conventional	 seats)	 indicated	 a	 neck	 injury	 risk	 reduction	 between	 21	 and	 47%	depending	on	 impact	 severity	and	symptom	duration	 (Jakobsson	et	al.,	2008).	 Similarly,	 Saab	Automobiles	conducted	a	questionnaire	and	phone-interview	study	of	177	insurance	claims	(92	occupants	in	SAHR	vs.	85	occupants	with	standard	head	restraints)	from	September	28th,	1998					11	through	April	4th,	2000	(Viano	and	Olsen,	2001).	 	This	study	 indicated	a	potential	 for	SAHR	to	decrease	the	number	of	injured	occupants	by	20%	(incident	rate:	SAHR:	41%	vs.	standard	head	restraint:	53%)	and	to	reduce	whiplash	injuries	leading	to	medium-	and	long-term	symptoms	in	occupants	 by	 75%	 (incident	 rate:	 SAHR:	 4%	 vs.	 standard	 head	 restraint:	 18%).	 	 Current	 anti-whiplash	devices	(e.g.	Volvo’s	WHIPS	and	Saab’s	SAHR)	have	not	eliminated	the	risk	of	whiplash	injury,	but	they	have	decreased	the	risk	of	whiplash	injury	during	rear-end	collisions	(Farmer	et	al.,	2003,	Jakobsson	et	al.,	2008,	Viano	and	Olsen,	2001).	Thus,	additional	research	addressing	whiplash	 injury	mechanisms	may	 lead	 to	 the	 design	 of	 novel	 and	 efficient	 injury	 prevention	devices	that	further	reduce	the	risk	of	whiplash	injury	following	a	rear-end	collision.					1.2 Overall	Research	Objective	and	Goals	The	overall	objective	of	the	experiments	presented	in	this	dissertation	was	to	develop	a	novel	 anti-whiplash	 seat	 that	 dynamically	 modified	 the	 seat	 hinge	 and	 seatback	 cushion	properties	 to	 attenuate	 the	 occupant	 response	 during	 low-speed,	 rear-end	 collisions	 (see	Appendix	 A	 for	 details	 on	 the	 anti-whiplash	 seat	 design).	 The	 main	 safety	 features	 of	 the	Experimental	 anti-whiplash	 seat	were	 the	 active	 control	 of	 seat	 hinge	 rotation	 and	 seatback	cushion	deformation	to	decrease	the	accelerations	experienced	by	the	occupant	and	to	reduce	the	relative	motion	between	the	occupant’s	head	and	torso.	This	dissertation	was	divided	into	four	experiments	with	the	goal	of	answering	the	following	questions:	Experiment	1:			 How	do	current	anti-whiplash	seats	perform	during	 low-speed,	 rear-end	collisions?					12	Experiment	2:		 Can	active	control	of	seat	hinge	rotation	attenuate	ATD	responses	during	low-speed,	rear-end	collisions?	Experiment	3:		 Can	 active	 control	 of	 seatback	 cushion	 deformation	 attenuate	 ATD	responses	during	low-speed,	rear-end	collisions?	Experiment	4:		 Can	 co-activation	 of	 the	 seat	 hinge	 rotation	 and	 seatback	 cushion	deformation	further	attenuate	ATD	responses	compared	to	current	anti-whiplash	seats	during	low-speed,	rear-end	collisions?			These	 experiments	 were	 designed	 to	 test	 the	 different	 safety	 features	 of	 our	Experimental	 anti-whiplash	 seat.	 To	 minimize	 potential	 sources	 of	 variability	 in	 the	experimental	 design,	 the	 following	 procedures	 were	 chosen	 despite	 restricting	 the	 potential	generalizability	 of	 the	 results.	 First,	 a	 50th	 percentile	 male	 BioRID	 II	 anthropomorphic	 test	device	 (ATD;	 Humanetics	 Innovative	 Solutions,	 Farmington	 Hills,	MI,	 USA)	 was	 selected	 as	 a	surrogate	to	human	occupants.	Even	though	females	are	more	at	risk	of	developing	symptoms	following	rear-end	collisions	(Mordaka	and	Gentle,	2003,	Harder	et	al.,	1998,	Versteegen	et	al.,	2000),	a	female	ATD	is	currently	in	development	and	only	the	50th	percentile	male	BioRID	II	ATD	was	validated	at	the	time	of	performing	these	experiments.	Second,	only	front	passenger	seats	were	used	 for	all	 experiments.	Passenger	 seats	 share	 similar	designs	and	 structures	 to	driver	seats,	but	are	typically	used	less	in	vehicles	(average	vehicle	occupancy	rate	=	1.67;	McGuckin	and	Fucci,	2018).	Third,	neither	a	steering	wheel	nor	a	seatbelt	were	used	in	the	experiments	to	better	 isolate	the	 influence	of	the	seat	properties	on	the	occupant	responses.	Due	to	the	 low					13	collision	 severities	 used	 in	 our	 experiments,	most	 peak	 occupant	 responses	 are	 expected	 to	occur	 during	 the	 retraction	 phase	 of	 whiplash	 motion	 (Siegmund	 et	 al.,	 2005a)	 when	 the	occupant	is	translating	rearward	into	the	seatback	away	from	the	steering	wheel	and	stationary	seatbelt.	 Hence,	 these	 experimental	 decisions	 allowed	 a	 specific	 investigation	 of	 the	 role	 of	anti-whiplash	seats	on	the	occupant	responses	following	rear-end	collisions	but	require	careful	consideration	prior	to	further	development	of	a	seat	that	better	protects	all	occupants.		 					14	Chapter	2. A	Comparison	of	Anti-Whiplash	Seats	During	Low/Moderate	Speed,	Rear-End	Collisions	2.1 Introduction	Whiplash	 injuries	 are	 the	most	 common	 type	 of	 injuries	 associated	with	motor	 vehicle	crashes,	and	rear-end	collisions	pose	the	greatest	risk	of	whiplash	injury	(Jakobsson	et	al.,	2000,	National	Highway	 Traffic	 Safety	Administration,	 2014).	 	 To	motivate	manufacturers	 to	 design	seats	and	head	restraints	that	reduce	the	risk	of	whiplash	 injury	 in	 low/moderate-speed	rear-end	collisions,	 the	 Insurance	 Institute	 for	Highway	Safety	 (IIHS)	publishes	 ratings	of	seats	and	head	 restraints	using	 a	 two-part	protocol	 developed	by	 the	Research	Council	 for	Automobile	Repairs/International	Insurance	Whiplash	Prevention	Group	(RCAR/IIWPG)	(Insurance	Institute	of	 Highway	 Safety,	 2008b,	 Insurance	 Institute	 of	 Highway	 Safety,	 2008a).	 First,	 a	 static	 test	evaluates	head	restraint	height	(distance	between	the	top	of	the	head	and	the	top	of	the	head	restraint)	 and	 backset	 (distance	 between	 the	 back	 of	 the	 head	 and	 the	 front	 of	 the	 head	restraint)	relative	to	the	head	of	a	median-sized	male	occupant.	If	the	top	of	the	head	restraint	is	no	more	than	8	cm	below	the	top	of	the	head	and	the	backset	is	no	more	than	9	cm,	the	seat	undergoes	a	dynamic	test	that	simulates	a	rear-end	collision	with	a	peak	acceleration	of	10	g	and	a	speed	change	of	16	km/h	over	91	ms.	The	dynamic	test	is	graded	using	a	combination	of	peak	 upper	 neck	 forces	 (shear	 and	 axial),	 head-to-head	 restraint	 contact	 time,	 and	 peak	forward	acceleration	of	the	T1	vertebra	of	a	BioRID	II	anthropometric	test	device	(ATD).		Seats	and	 head	 restraints	 are	 rated	 as	 good,	 acceptable,	 marginal	 or	 poor	 depending	 on	 the					15	combined	 rating	 of	 both	 the	 static	 and	 dynamic	 test	 results.	 In	 a	 comparison	 between	RCAR/IIWPG	ratings	and	real-world	insurance	claims,	occupant	neck	injury	rates	were	only	11	–	15%	lower	in	seats	rated	good	compared	to	seats	rated	poor	(Farmer	et	al.,	2008,	Trempel	et	al.,	 2016).	 This	meager	 benefit	 suggests	 that	 there	 remains	 considerable	 room	 for	 improving	both	 automotive	 seats	 and	 the	 rating	 system	 to	 better	 protect	 occupants	 from	 whiplash	injuries.		 Some	manufacturers	have	developed	seats	specifically	 focused	on	reducing	the	risk	of	whiplash	 injury.	 Current	 anti-whiplash	 seats,	 such	 as	 General	 Motors’	 High	 Retention	 seat	(GMHR),	Volvo’s	Whiplash	Injury	Prevention	seat	(WHIPS)	and	Saab’s	Active	Head	Restraint	seat	(SAHR),	attempt	to	reduce	whiplash	injury	risk	by	reducing	key	occupant	kinematics	such	as	the	relative	motion	between	the	head	and	upper	torso.	According	to	the	IIHS	(Insurance	Institute	of	Highway	 Safety,	 2004),	 the	GMHR	was	 rated	poor,	whereas	both	 the	WHIPS	 and	 SAHR	were	rated	good.	The	GMHR	is	designed	with	a	rigid	perimeter	seat	frame	and	a	compliant	seatback	suspension	 to	 allow	 the	 occupant	 to	 “pocket”	 into	 the	 seatback	 and	 increase	 occupant	retention	(Viano,	2003a,	Viano	and	Parenteau,	2015).	The	WHIPS	and	SAHR	seats	are	equipped	with	dynamic	anti-whiplash	devices	that	rely	on	occupant	loading	of	the	seatback	to	deform	a	recliner	mechanism	that	controls	seatback	translation	and	rotation	 (WHIPS)	 (Jakobsson	et	al.,	2008,	Jakobsson	et	al.,	2000)	or	to	move	the	head	restraint	upward	and	forward	(SAHR)	(Viano	and	Olsen,	2001).	Epidemiological	studies	have	shown	that	these	dynamic	anti-whiplash	seats	reduce	 the	 risk	 of	whiplash	 injury	 by	 20	 to	 75%	when	 compared	 to	 previous	 versions	 of	 the					16	same	 seat	 without	 the	 anti-whiplash	 mechanism	 and	 dependent	 on	 how	 injury	 is	 defined.	Although	most	studies	converging	on	a	reduction	in	the	risk	of	whiplash	injury	of	about	40	to	50%	 (Viano	and	Olsen,	2001,	 Jakobsson	et	al.,	 2008,	Kullgren	and	Krafft,	 2010,	 Farmer	et	 al.,	2003,	Kullgren	et	al.,	2007),	these	injury	reduction	rates	suggest	that	many	people	still	do	not	benefit	from	current	dynamic	anti-whiplash	seats.		 One	possible	reason	for	the	incomplete	effectiveness	of	anti-whiplash	seats	is	that	they	have	been	tested	at	the	RCAR/IIWPG	dynamic	test	pulse	(10	g,	16	km/h,	Δt	=	91	ms).	This	pulse	is	more	severe	than	the	speed	change	reported	to	cause	many	whiplash	injuries,	some	of	which	are	reported	following	collisions	with	speed	changes	as	low	as	6	to	8	km/h	(Bartsch	et	al.,	2008,	Krafft	 et	 al.,	 2005).	 The	 RCAR/IIWPG	 test	 pulse	 represents	 a	 generic	 collision	 pulse	 that	 is	 a	necessary	compromise	for	a	safety	standard,	but	nevertheless	eliminates	a	wide	range	of	actual	collision	pulses	in	terms	of	pulse	shape,	duration	and	severity	(Cappon	et	al.,	2001,	Langwieder	and	Hell,	2002,	Linder	et	al.,	2003,	Linder	et	al.,	2001).	Moreover,	seats	optimized	for	a	single	pulse	 (e.g.	 the	 RCAR/IIWPG	 test	 pulse)	 may	 perform	 sub-optimally	 under	 other	 conditions	inducing	whiplash	injuries.	The	goal	of	this	study	was	to	evaluate	the	performance	of	some	anti-whiplash	seats	across	a	range	of	speed	changes	below	16	km/h	to	gain	a	better	understanding	of	 how	 occupant	 kinematic	 and	 kinetic	 parameters	 vary	 with	 collision	 severity.	 We	hypothesized	that	both	the	WHIPS	and	SAHR	anti-whiplash	seats	(rated	good	by	RCAR/IIWPG)	would	cause	lower	occupant	kinematic	and	kinetic	responses	in	comparison	to	the	GMHR	seat	(rated	poor	RCAR/IIWPG)	across	a	range	of	collision	severities.							17	2.2 Methods	2.2.1 Anthropomorphic	Test	Device	and	Instrumentation		 A	 BioRID	 II	 ATD	 (Humanetics,	 Plymouth,	MI,	 USA)	 was	 instrumented	 to	 measure	 the	kinematic	and	kinetic	responses	of	the	head,	neck	and	torso	during	controlled	laboratory	rear-end	 impacts.	 Linear	accelerations	were	measured	using	 two	uni-axial	 accelerometers	 (7264C;	±500	g,	Endevco,	San	Juan	Capistrano,	CA,	USA)	mounted	at	both	the	head	center	of	mass	and	T1	 vertebra.	 	 Uni-axial	 angular	 velocity	 sensors	 were	 also	 mounted	 to	 the	 head	 (ARS-1500;	±26.2	 rad/s,	 DTS,	 Seal	 Beach,	 CA,	 USA)	 and	 to	 the	 T1	 vertebra	 (ARS-04E;	 ±100	 rad/s,	 ATA	Sensors,	 Albuquerque,	 NM,	 USA)	 to	 measure	 angular	 kinematics	 in	 the	 sagittal	 plane	 (i.e.	flexion	and	extension	of	the	head	and	T1).		A	six-axis	load	cell	(Model	4949a,	Robert	A.	Denton,	Inc.,	 Rochester	 Hills,	MI,	 USA)	 was	 installed	 to	measure	 upper	 neck	 forces	 and	moments.	 A	motion	capture	system	(Optotrak	Certus,	Northern	Digital,	Waterloo,	ON,	Canada)	was	used	to	track	 infrared	 light	 emitting	 diode	 (IRED)	 markers	 affixed	 to	 the	 ATD’s	 head,	 C1	 and	 T1	vertebrae	to	measure	head	and	neck	displacements.	Head	restraint	contact	was	detected	with	a	force	sensitive	resistor	(FSR,	Model	406;	Interlink	Electronics,	Camarillo,	CA,	USA)	attached	to	the	 front	 of	 the	 head	 restraint.	 Horizontal	 sled	 acceleration	 was	 measured	 with	 a	 uni-axial	accelerometer	(2220-100;	±100	g,	Silicon	Design	Inc.,	Issaquah,	WA,	USA).				2.2.2 Test	Procedures		 The	BioRID	 II	ATD	was	clad	 in	 two	 layers	of	 lycra	and	seated	on	a	 feedback-controlled	linear	sled	(Kollmorgen	IC55-100A7,	Waltham,	MA,	USA)	fitted	with	one	of	four	different	front					18	passenger	seats:	A)	2005	Volvo	S40	WHIPS,	B)	2004	Volvo	S60	WHIPS,	C)	2005	Saab	9.3	SAHR,	and	 D)	 2004	 Pontiac	 Grand	 Am	 GMHR	 (Figure	 2.1).	 On	 each	 seat,	 the	 unbelted	 ATD	 was	exposed	to	a	series	of	seven	whiplash-like	perturbations	at	increasing	speed	changes	(Δv	=	2,	4,	6,	 8,	 10,	 12	 and	 14	 km/h)	 with	 a	 pulse	 duration	 (Δt)	 of	 141	 ms	 (Figure	 2.2).	 The	 sled	 was	accelerated	 forward	 from	 a	 stationary	 position	 for	 the	 2,	 4	 and	 6	 km/h	 tests,	 and	 from	 a	constant	rearward	speed	of	6	km/h	for	the	higher	speed	changes	(8,	10,	12	and	14	km/h).	Pilot	tests	comparing	ATD	responses	of	an	8	km/h	collision	starting	from	a	stationary	position	to	an	8	km/h	 collision	 starting	 from	 a	 6	 km/h	 rearward	 velocity	 showed	 between	 0.2%	 and	 10.1%	differences	(mean	=	4.7%)	in	the	peak	ATD	responses	and	neck	injury	criteria	(Figure	2.3).	Four	additional	repeated	trials	were	collected	at	the	8	km/h	collision	speed	change	for	all	seats,	as	well	as	at	 the	4	and	12	km/h	speed	changes	only	on	the	GMHR	seat	 (for	a	 total	of	5	 trials	 in	each	 experimental	 condition)	 to	 assess	 the	 repeatability	 and	 reliability	 of	 BioRID	 II	 ATD	occupant	responses.	A	total	of	52	trials	were	collected:	11	trials	were	collected	for	the	SAHR,	S40	WHIPS	and	S60	WHIPS	seats	and	19	trials	were	collected	for	the	GMHR	seat.			 Seat	geometry	was	measured	before	and	after	each	collision	 to	ensure	no	permanent	deformation	 had	 occurred.	 We	 also	 verified	 that	 no	 damage	 occurred	 to	 the	 deformable	elements	in	the	WHIPS	seat	and	that	the	SAHR	mechanism	had	returned	to	its	original	position	following	each	 test.	Additional	Optotrak	markers	were	placed	on	 the	 seats	and	 linear	 sled	 to	determine	 seatback	 and	 head	 restraint	 position	 as	 well	 as	 to	 create	 a	 global	 experimental							19	Figure	2.1.	Photographs	of	the	experimental	set-up	with	the	BioRID	II	ATD	on	the	four	test	seats	in	the	laboratory	reference	frame	(X,	Z).	A.)	2005	Volvo	S40	WHIPS,	B.)	2004	Volvo	S60	WHIPS,	C.)	2005	Saab	9.3	SAHR	and	D.)	2004	Pontiac	Grand	Am	GMHR.			 		 	+X	+Z	C.	 D.	A.	 B.					20		Figure	2.2.	Exemplar	sled	A.)	velocity	and	B.)	acceleration	pulses	for	increasing	collision	speeds	(Δv	 =	 2,	 4,	 6,	 8,	 10,	 12	 and	 14	 km/h	 with	 a	 pulse	 duration	 (Δt)	 of	 141	 ms).	 Collision	 onset	occurred	at	time	=	0	ms.				Figure	 2.3.	 ATD	occupant	 responses	 comparing	 an	 8km/h	 collision	 severity	 from	a	 stationary	position	 (grey	 line)	 and	 from	 a	 6km/h	 constant	 rearward	 speed	 (black	 line)	 on	 the	 Pontiac	Grand	Am	GMHR	seat.	Hollow	circles	represent	peak	responses	for	a	given	variable.						21	reference	 frame	 (+X	 forward,	 +Y	 right,	 and	 +Z	 down).	 The	 initial	 position	 of	 the	 ATD	 was	adjusted	to	a	pre-defined	posture	that	was	constant	within	each	seat	and	as	similar	as	possible	between	seats.	This	pre-defined	posture	was	confirmed	using	the	3D	positions	measured	by	the	Optotrak	at	 the	beginning	of	all	 trials.	 Initial	 seatback	angles	were	 set	 to	27	deg	 rearward	of	vertical	(Siegmund	et	al.,	2005a)	and	the	height	of	the	head	restraint	was	adjusted	according	to	the	RCAR/IIWPG	seat/head	restraint	evaluation	protocol	(Insurance	Institute	of	Highway	Safety,	2008b).	For	all	seats,	the	top	of	the	head	restraint	was	less	than	6	cm	lower	than	the	top	of	the	ATD	head	 (positive	values	 indicate	 that	 the	 top	of	head	was	higher	 than	 the	 top	of	 the	head	restraint)	 and	 would	 receive	 good	 ratings	 for	 head	 restraint	 height	 (Table	 2.1).	 The	 backset	distance	for	each	head	restraint	was	defined	as	the	average	distance	between	the	back	of	the	ATD	head	and	the	front	face	of	the	head	restraint	over	the	100	ms	immediately	preceding	the	onset	 of	 forward	 sled	 acceleration.	 Backset	 distances	 less	 than	 7	 cm	were	 considered	 good,	between	 7	 and	 9	 cm	 were	 considered	 acceptable,	 between	 9	 and	 11	 cm	 were	 considered	marginal,	 and	greater	 than	11	cm	were	considered	poor	according	 to	 the	RCAR/IIWPG	 rating	(Table	2.1).							22			2.2.3 Data	Analysis		 All	accelerometer,	load	cell	and	angular	velocity	transducer	signals	were	simultaneously	sampled	at	2000	Hz	using	a	National	Instruments	Data	Acquisition	(DAQ)	PXI	system	(PXI-4495	&	 PXI-6289,	 National	 Instruments	 Corporation,	 Austin,	 Texas,	 USA)	 and	 a	 custom-written	LabVIEW	virtual	 instrument	 (National	 Instruments	Corporation,	Austin,	 Texas,	USA).	Optotrak	data	were	acquired	at	200	Hz	per	frame	and	the	onset	of	collection	was	triggered	by	the	DAQ	system	to	synchronize	the	data.	All	data	channels	conformed	to	SAE	J211	(Channel	class	180	for	the	ATD	sensors	and	Channel	Class	60	for	the	sled	accelerometer)	(SAE,	1995).	Subsequent	data	analysis	was	performed	using	Matlab	(R2017A,	Mathworks,	Newton,	MA,	USA).	Table	2.1.	Mean	(standard	deviation)	of	head	restraint	backset	and	height	determined	by	Optotrak	prior	to	the	onset	of	each	collision.		Seat	 n	 Backset	(cm)	 Height	(cm)	Volvo	S40	WHIPS	 11	 4.66	(0.51)	 0.42	(0.52)	Volvo	S60	WHIPS	 11	 3.74	(1.5)	 5.41	(0.52)	Saab	9.3	SAHR	 11	 7.86	(0.77)	 3.38	(0.36)	Pontiac	Grand	Am	GMHR	 19	 9.52	(0.70)	 2.50	(0.29)	Notes:	n	denotes	 the	number	of	 trials	 for	each	seat	used	 for	 the	calculations.	Backset:	distance	between	the	back	of	the	head	to	the	front	face	of	the	head	restraint.	 	Height:	distance	between	the	top	of	 the	head	to	the	top	of	 the	head	restraint	 (positive	values	indicated	the	top	of	the	ATD	head	is	higher	than	the	head	restraint).					23		 To	 compare	 the	 BioRID	 II	 ATD	 responses	 between	 the	 different	 seats	 and	 different	collision	 pulses,	 peak	 kinematic	 and	 kinetic	 responses	 were	 extracted	 for	 each	 trial.	Accelerometer	data	were	reported	in	local	head	and	T1	reference	frames	and	were	corrected	to	remove	 the	 earth’s	 gravitation	 field	 using	 the	 head	 and	 T1	 orientations	 determined	 from	Optotrak	 data.	 Peak	 linear	 forward	 acceleration	 of	 the	 sled	 (aX-sled),	 head	 (aX-head)	 and	 T1	vertebra	 (aX-T1)	 were	 extracted	 directly	 from	 the	 transformed	 accelerometer	 data.	 Peak	rotational	velocities	of	 the	head	 (ωY-head)	and	T1	 (ωY-T1)	 in	 the	sagittal	plane	were	determined	from	 the	 angular	 velocity	 sensors.	 	 Peak	 upper	 neck	 shear	 (FX)	 and	 axial	 (FZ)	 forces	 and	 the	flexion/extension	bending	moment	 (MY)	were	determined	 from	 the	upper	neck	 load	 cell	 and	reported	in	the	ATD	reference	frame	as	the	forces/moments	applied	by	the	neck	to	the	head.	Initial	 head	 angle	was	 defined	 as	 the	 average	 angle	 over	 the	 250	ms	 preceding	 the	 onset	 of	forward	sled	acceleration	and	peak	head	extension	angle	(θhead)	was	defined	as	the	maximum	rearward	rotation	of	the	head	into	extension	relative	to	initial	head	angle.	Peak	retraction	(Rx)	was	 defined	 as	 the	 maximum	 horizontal	 rearward	 displacement	 in	 the	 laboratory	 reference	frame	 between	 the	 atlanto-occipital	 joint	 (AOJ)	 pin	 and	 T1	 vertebrae	 with	 rearward	displacements	 defined	 as	 negative	 values.	 	 Head-to-head	 restraint	 contact	 time	 (Δthead-contact)	was	extracted	from	the	onset	of	the	FSR	data	attached	to	the	head	restraint	with	the	onset	of	forward	 sled	 acceleration	 defined	 as	 t	 =	 0	ms.	Onset	 of	 sled	 acceleration	 and	 head	 restraint	contact	were	 determined	when	 the	 accelerometer	 and	 FSR	 signals,	 respectively,	 reached	 1.5	times	 the	 peak	 background	 noise	 level	 measured	 over	 2	 seconds	 prior	 to	 the	 onset	 of	 the	collision	pulse	and	confirmed	visually.					24		 Three	neck	injury	criteria	(NIC,	Nij,	and	Nkm)	were	computed	from	the	accelerometer	and	load	 cell	 data.	 	 The	 Neck	 Injury	 Criterion	 (NIC)	 was	 calculated	 from	 the	 relative	 horizontal	acceleration	and	velocity	 in	the	global	reference	frame	between	the	head	center	of	mass	and	the	T1	joint	(Equation	2.1)	(Bostrom	et	al.,	1996).	The	Normalized	Neck	Injury	Criterion	(Nij)	was	calculated	 from	the	axial	 load	 (FZ)	and	the	 flexion/extension	bending	moment	 (MY)	measured	from	 the	 upper	 neck	 load	 cell	 (Equation	 2.2;	 critical	 Fint	 and	 Mint	 intercept	 values	 used	 for	normalization:	Fint-tension	=	6806	N,	Fint-compression	=	-6160	N,	Mint-flexion	=	310	N,	and	Mint-extension	=	-135	N;	 Eppinger	 et	 al.,	 1999,	 Eppinger	 et	 al.,	 2000).	 The	Neck	 Protection	Criterion	 (Nkm)	was	calculated	 from	 the	 sagittal	 shear	 force	 (FX)	 and	 the	 flexion/extension	bending	moment	 (MY)	measured	from	the	upper	neck	load	cell	(Equation	2.3;	critical	Fint	and	Mint	intercept	values	used	for	normalization:	Fint-shear-positive	=	845	N,	Fint-shear-negative	=	845	N,	Mint-extension	=	47.5	Nm,	and	Mint-flexion	=	88.1	Nm;	Schmitt	et	al.,	2001,	Schmitt	et	al.,	2002).		Peak	values	of	the	three	neck	injury	criteria	were	then	extracted	for	analysis	and	compared	to	proposed	injury	thresholds	(NIC:	15	m2/s2,	 Eichberger	 et	 al.,	 1998;	 Nij	 and	Nkm	 :	 a	 normalized	 value	 of	 one,	 Schmitt	 et	 al.,	 2001,	Schmitt	 et	 al.,	 2002,	 Eppinger	 et	 al.,	 1999).	 	 These	 injury	 thresholds	 correspond	 to	 different	human	tolerance	levels	for	the	causation	of	whiplash-related	injuries	–	NIC:	long-term	whiplash-associated	disorders	(WAD)	levels	1-3	(Bostrom	et	al.,	2000,	Bostrom	et	al.,	1996),	Nij:	22%	risk	of	 abbreviated	 injury	 scale	 (AIS)	 level	 3	 (i.e.,	 fracture;	 Eppinger	 et	 al.,	 1999,	 Eppinger	 et	 al.,	2000),	and	Nkm:	AIS	 level	1	 injury	causation	(i.e.,	minor	 injury;	Schmitt	et	al.,	2001,	Schmitt	et	al.,	2002).					25	NIC t = 	#$%& ' ∙ ) + (,$%&(')).	 	 	 	 …	 Equation	2.1	N01(t) = 23(4)2567 + 89:;(4)8567 	 	 	 	 	 	 …	 Equation	2.2	N<=(t) = 2>(4)2567 + 89:;(4)8567 	 	 	 	 	 …		 Equation	2.3			The	 repeatability	 of	 the	 BioRID	 II	 ATD	 was	 assessed	 using	 the	 Coefficient	 of	 Variation	(COV),	which	was	calculated	as	the	standard	deviation	divided	by	the	mean	and	expressed	as	a	percent	(Davidsson,	1999,	Moorhouse	et	al.,	2012,	Rhule	et	al.,	2005,	Siegmund	et	al.,	2005b).		The	 National	 Highway	 Traffic	 Safety	 Administration	 (NHTSA)	 defines	 a	 COV	 of	 5%	 or	 less	 as	good,	a	COV	of	10%	or	 less	as	acceptable	and	a	COV	greater	 than	10%	as	poor	 (Rhule	et	al.,	2005).	 	To	quantify	the	repeatability,	we	calculated	the	COV	for	the	peak	occupant	responses	from	five	repeated	trials	at	 three	different	collision	severities	 (Dv	=	4,	8	and	12	km/h)	on	the	GMHR	seat	and	at	the	8	km/h	collision	severity	for	the	SAHR	and	WHIPS	seats.	For	the	GMHR	seat,	most	of	 the	COVs	were	below	5%,	except	 for	 the	shear	and	axial	neck	 forces	 (7.9%	and	5.8%	 respectively),	 both	 of	 which	 were	 lower	 than	 values	 reported	 by	 other	 researchers	(Davidsson,	1999,	Moorhouse	et	al.,	2012)	(Table	2.2).	For	the	SAHR	and	WHIPS	seats,	the	COVs	of	 the	upper	 neck	 loads	 (FX	 and	 FZ)	were	 greater	 than	 10%,	 but	 consistent	with	 earlier	work	(Davidsson,	1999).	Detailed	COV	data	for	all	seats	and	variables	are	presented	in	the	Appendix	E:	Tables	E2.1	and	E2.2.	To	 compare	BioRID	 II	ATD	 responses	between	 seats,	 99th	 percentile	predictive	 intervals	were	created	around	the	peak	responses	using	the	maximum	variability	observed	for	the	GMHR					26	seat	at	4,	8	and	12	km/h	(Horslen	et	al.,	2015).	The	maximum	COV	values	were	multiplied	by	2.58	 to	 estimate	 99%	 predictive	 limits	 and	 applied	 to	 the	 respective	 peak	 responses	 for	 all	collision	 severities.	 ATD	 responses	 observed	on	 the	other	 seats	 that	 fell	 outside	 these	point-wise	99%	predictive	corridors	were	judged	to	be	statistically	different	from	the	responses	of	the	ATD	in	the	GMHR	seat.	Averaged	peak	responses	were	used	for	test	conditions	with	repeated		trials	(GMHR:		Dv	=	4,	8,	and	12	km/h	trials;	SAHR	&	WHIPS:	8	km/h	trials).		Table	2.2.	Coefficients	of	Variation	(COV)	for	some	peak	ATD	responses	compared	to	previous	literature.		Parameter	Davidsson	(1999)		BioRID	I	(n	=	5)	Moorhouse	et	al.	(2012)	BIORID	II	(n	=	3)	Current	study,	GMHR	BioRID	II	(n	=	5)	Mean	(SD)	 COV	(%)	 Mean	 COV	(%)	 Mean	(SD)	 COV	(%)	DV	 17	km/h	 17	km/h	 12	km/h	aX-sled	(m/s2)	 104	(1.1)	 1.0	 -	 -	 28.6	(0.1)	 0.4	aX-head	(m/s2)	 283	(13.0)	 4.6	 96.4	 8.8	 130.4	(1.4)	 1.1	aX-T1	(m/s2)	 116	(4.0)	 3.4	 23.9	 7.5	 51.3	(0.5)	 1.0	FX	(N)	 455	(45.9)	 10.1	 -	 9.0	220.9	(17.5)	 7.9	FZ	(N)	2900	(817.0)	 28.1	 820.3	 6.8	684.7	(39.5)	 5.8	MY	(Nm)	 -	 -	 -	 7.9	 14.0(0.6)	 4.2	NIC	(m2/s2)	 -	 -	 -	 -	 11.0	(0.3)	 3.1	Notes:		The	underlined	COV	values	indicate	a	COV	rating	of	acceptable	(5%	≤	COV	<	10%),	the	bolded	COV	values	indicate	a	COV	rating	of	poor	(COV	>	10%),	and	all	other	values	are	rated	good	(COV	<	5%).	Full	list	of	COV	values	for	all	experimental	parameters	is	given	in	Appendix	E:	Tables	E2.1	and	E2.2.					27	2.3 Results		 Positively	 graded	 responses	 with	 increasing	 speed	 change	 were	 observed	 in	 most	kinematic	 and	 kinetic	 variables	 (Figures	 2.4	 and	 2.5,	 see	 also	 Figures	 E2.1	 –	 E2.4	 in	 the	Appendix	E).	For	all	four	seats,	peak	forward	accelerations	of	the	head	(aX-head)	and	torso	(aX-T1),	peak	 angular	 head	 velocity	 (ωhead),	 and	 the	 three	 neck	 injury	 criteria	 (NIC,	 Nij,	 and	 Nkm)	increased	 with	 collision	 severity	 (Figure	 2.5A-C,	 Figure	 2.5G,	 and	 Figure	 2.5I-L).	 Peak	 head	extension	 remained	 relatively	 constant	 across	 all	 speed	 changes	 for	 the	 four	 seats	 (Figure	2.5H).			 The	 ATD	 kinematic	 and	 kinetic	 responses	 were	 generally	 similar	 for	 the	 three	 good	seats,	 but	 the	 good	 seats	were	only	 clearly	 different	 from	 the	GMHR	 seat	 for	 four	 variables:	peak	upper	neck	forces	(FX	and	FZ)	and	moments	(MY)	as	well	as	peak	rearward	head	retraction	(RX).	 Peak	 upper	 neck	 shear	 forces	 (FX)	 and	 retraction	 (RX)	 increased	 with	 increasing	 speed	change	for	the	GMHR	seat,	but	generally	decreased	(or	remained	constant)	for	the	WHIPS	and	SAHR	seats	(FX:	25	to	79%	and	RX:	30	to	55%	of	the	poor	rated	seat	responses	for	Dv	>	6	km/h)	(Figure	2.5E	&	I).		At	and	above	8	km/h	(Dv	≥	8	km/h),	the	WHIPS	and	SAHR	seats	decreased,	or	eliminated,	the	peak	positive	shear	forces	(+FX)	and	extension	bending	moments	(+MY)	((FX:	-1	to	55%	and	MY:	31	to	70%	of	poor	rated	seat	responses)	but	increased	the	peak	negative	shear	forces	(-FX)	and	flexion	bending	moments	(-MY)	(Figure	2.4,	2.5D	&	2.5E,	see	also	Figures	E2.1	–	E2.4	in	the	Appendix	E).							28		 All	three	neck	injury	criteria	(NIC,	Nij	and	Nkm)	increased	with	speed	change	for	all	four	seats	(Figure	2.5J-L).		Neither	NIC	nor	Nij	exceeded	their	respective	proposed	injury	thresholds,	but	Nkm	exceeded	the	injury	threshold	value	of	1	in	the	14	km/h	performed	on	the	SAHR	seat	(Figure	2.5L).			 					29		Figure	2.4.	Exemplar	data	comparing	a	RCAR/IIWPG	good	rated	(Volvo	S40	WHIPS;	black	line)	and	a	RCAR/IIWPG	poor	rated	(Pontiac	Grand	Am	GMHR;	grey	line)	seat	for	the	BioRID	II	ATD.	Each	column	represents	occupant	responses	for	a	good	and	poor	seat	while	exposed	to	various	collision	severities	 (Δv	=	2,	4,	6,	8,	10,	12	and	14	km/h	with	a	collision	pulse	duration	(Δt)	of	141	ms).	Hollow	circles	 represent	 the	onset	of	head-to-head-restraint	contact	and	peak	 responses	of	each	ATD	response	parameter	 for	each	 trial.	Similar	plots	for	all	seats	are	given	in	the	Appendix	E:	Figures	E2.1	–	E2.4.					30		Figure	2.5.	Experimental	results	of	peak	kinematic	and	kinetic	responses	for	the	BioRID	II	ATD	seated	on	the	WHIPS,	SAHR	and	GMHR	seats.	For	all	the	graphs,	red	circles	represent	the	Volvo	S40	WHIPS	seat,	blue	triangles	represent	the	Volvo	S60	WHIPS	seat,	green	diamonds	represent	the	 SAHR	 seat,	 and	 black	 squares	 represent	 the	 GMHR	 seat.	 Grey	 corridors	 represent	 99%	predictive	 interval	 for	ATD	 responses	on	 the	poor	 rated	GMHR	seat.	 Y-axis	 values	 for	panel	 I	have	been	inverted	for	visual	purposes	to	show	increasing	responses	from	bottom	to	top.		 					31	2.4 Discussion		 The	goal	of	this	study	was	to	evaluate	the	performance	of	anti-whiplash	seats	across	a	range	of	speed	changes.	This	goal	was	motivated	in	part	by	the	marginal	reductions	in	whiplash	injury	risk	between	seats	rated	good	and	poor	by	the	IIHS	using	the	RCAR/IIWPG	test	method	(Insurance	Institute	of	Highway	Safety,	2008b).	Moreover,	we	sought	to	examine	the	changes	in	ATD	kinematic	and	kinetic	responses	for	speed	changes	below	the	16	km/h	level	prescribed	by	the	 RCAR/IIWPG	 test	 to	 gain	 a	 better	 understanding	 of	 why	 anti-whiplash	 seats	 have	 not	generated	a	greater	reduction	in	whiplash	injury	risk.			 Although	 the	 exact	 injury	 mechanisms	 causing	 whiplash	 injury	 remains	 unclear,	 the	ability	of	the	WHIPS	and	SAHR	seats	to	attenuate	the	peak	ATD	upper	neck	loads	and	moments,	and	 head	 retraction	 suggested	 that	 these	 responses	may	 be	 associated	with	 a	 higher	 risk	 of	whiplash	 injuries	 (Viano	 and	 Olsen,	 2001,	 Jakobsson	 et	 al.,	 2008,	 Kullgren	 and	 Krafft,	 2010,	Farmer	et	al.,	2003,	Kullgren	et	al.,	2007).	For	these	variables	(FX,	FZ,	MY	and	RX),	the	three	good	rated	 seats	 generated	 lower	 responses	 with	 increasing	 speed	 change	 in	 comparison	 to	 the	GMHR	seat	where	the	responses	continued	to	increase	with	speed	change.	The	poor	rated	seat	did	not	attenuate	these	four	ATD	responses	suggesting	it	was	unable	to	support	the	head	and	neck	 complex	 and,	 thus,	 increased	 both	 the	 relative	 head/torso	 motion	 and	 resulted	 in	additional	stresses/strains	applied	to	neck	structures.	A	shift	in	peak	shear	forces	(FX)	and	neck	bending	 moment	 (MY)	 from	 positive	 shear	 forces	 (+FX)	 and	 extension	 moments	 (+MY)	 to	negative	shear	forces	(-FX)	and	extension	moments	(-MY)	was	also	observed	in	the	good	rated					32	seats	at	speed	changes	greater	than	6	km/h	(Dv	>	6	km/h).		These	results,	in	combination	with	earlier	 head	 contact	 times	 (Δthead-contact)	 and	 decreased	 retraction	 (RX),	 suggested	 that	 the	WHIPS	 and	 SAHR	 seats	 supported	 the	 head	 and	 neck	 complex	 earlier	 into	 a	 forward	 flexed	position	instead	of	the	rearward	extended	position	observed	in	the	GMHR	seat.	This	adaptation	of	the	forward	flexion	position	during	the	whiplash	motion	may	explain	why	good	rated	seats	performed	better	than	poor	rated	seats	at	reducing	the	neck	injury	rate	by	11%	to	15%	(Farmer	et	 al.,	 2008,	 Trempel	 et	 al.,	 2016).	However,	 the	 ability	 of	 the	WHIPS	 and	 SAHR	 seat	 to	only	attenuate	 four	peak	ATD	 responses	 (FX,	 FZ,	MY	 and	RX)	 in	 comparison	 to	 the	GMHR	 seat	may	explain	why	real-world	epidemiological	studies	have	observed	a	limited	reduction	in	the	risk	of	whiplash	injury	of	about	40	to	50%	(Jakobsson	et	al.,	2008,	Kullgren	and	Krafft,	2010,	Viano	and	Olsen,	2001).	Anti-whiplash	seats	are	designed	to	attenuate	the	kinematic	or	kinetic	responses	thought	 to	 be	 responsible	 for	whiplash	 injury	 but	 because	 the	 injury	mechanisms	underlying	whiplash	remain	unclear,	reducing	occupant	accelerations	(head	and	torso)	and/or	minimizing	the	movement	 of	 the	 head	 relative	 to	 the	 upper	 torso,	 although	 potential	 targets,	may	 not	completely	protect	occupants	against	whiplash	injuries	(Siegmund	et	al.,	2009).	Future	work	is	needed	 to	 identify	 the	 biomechanical	 factors	 leading	 to	 whiplash	 injuries	 to	 help	 the	development	of	anti-whiplash	automotive	seats.			 Across	the	four	seats	and	seven	speed	changes	tested,	NIC	did	not	exceed	the	proposed	threshold	of	15	m2/s2	(Eichberger	et	al.,	1998)	and	Nij	remained	well	below	the	injury	thresholds	of	one	(Schmitt	et	al.,	2001,	Schmitt	et	al.,	2002,	Eppinger	et	al.,	1999).	Nkm	was	the	only	neck					33	injury	 to	 exceed	 the	 proposed	 injury	 threshold	 of	 one,	 which	 corresponds	 to	 an	 AIS	 level	 1	injury	(i.e.	minor	injury)	following	the	14	km/h	speed	change	with	the	SAHR	seat	(Schmitt	et	al.,	2001,	Schmitt	et	al.,	2002).	Nkm	attempts	to	capture	the	combination	of	shear	forces	and	neck	flexion/extension	bending	moment	related	to	the	formation	of	the	S-shape	curve	of	the	spine,	where	the	head	lags	the	torso	in	the	retraction	phase	(Schmitt	et	al.,	2001,	Schmitt	et	al.,	2002).	Peak	Nkm	value	occurred	146	ms	after	the	collision	onset,	while	the	head	was	 in	contact	with	the	head	 restraint.	The	neck	was	 in	 the	posterior	 shear	and	 flexion	quadrant	of	 the	Nkm	 load	space,	 indicating	 that	 the	 neck	 was	 applying	 a	 posteriorly-directed	 shear	 force	 and	 a	 flexor	moment	to	the	head	at	the	AOJ.	During	this	time,	the	ATD	initiated	the	rebound	phase	of	the	whiplash	 collision	 and	 the	 head	was	 lifting	 off	 from	 the	 head	 restraint.	 Given	 the	 real-world	success	of	the	SAHR	seat	at	reducing	the	risk	of	whiplash	injury	(Farmer	et	al.,	2003,	Viano	and	Olsen,	2001),	 the	supra-threshold	Nkm	values	observed	during	 the	 flexion	and	posterior	 shear	quadrant	 suggest	 they	 may	 not	 play	 a	 dominant	 role	 in	 the	 etiology	 of	 whiplash	 injuries.	However,	 a	 better	 understanding	of	 the	biomechanical	 factors	 leading	 to	whiplash	 injuries	 is	required	 to	better	 interpret	 current	neck	 injury	 criteria	and	 related	head/neck	kinematic	and	kinetic	responses	or	to	help	develop	a	criterion	more	specific	and	sensitive	to	whiplash	injuries.						 The	collision	severities	used	in	this	study	were	less	severe	than	the	IIWPG	standard	(16	km/h,	Dt	=	91	ms)	but	provided	a	graded	range	of	collision	velocities	applicable	to	some	real-world	collisions	that	cause	whiplash	 injury	(Krafft	et	al.,	2005,	Kullgren	et	al.,	2007).	Although	the	collision	pulse	duration	(Dt	=	141	ms)	was	longer	than	the	RCAR/IIWPG	pulse	(Δt	=	91	ms)					34	due	to	sled	 limitations	at	 the	highest	speed	change,	 it	was	similar	 to	 the	135	ms	observed	 in	prior	vehicle-to-vehicle	rear-end	crashes	(Brault	et	al.,	2000,	Brault	et	al.,	1998,	Siegmund	et	al.,	2000).	Another	limitation	of	the	present	study	was	that	we	did	not	maintain	proper	automotive	floor	 geometry	 relative	 to	 the	 seat	 base.	 The	 Volvo	WHIPS	 and	 Saab	 SAHR	 seats	 contained	electrical	motors	under	the	seat	pan	to	adjust	fore/aft	position	and	seat	pan	angle.	To	mount	these	seats,	we	elevated	the	seat	base	to	allow	for	clearance	under	the	seat	pan.	However,	the	relative	geometry	between	the	BioRID	II	ATD’s	head,	torso	and	pelvis	were	replicated	between	the	different	seats	to	ensure	that	the	ATD	was	in	a	similar	initial	position	and	confirmed	using	IRED	 marker	 positions	 (Figure	 2.1).	 Despite	 the	 resulting	 differences	 in	 ATD	 lower	 limb	positions,	T1	accelerations	were	similar	across	all	tested	seats	(see	Appendix	E:	Table	E2.1	and	Table	E2.2),	which	suggests	that	lower	limb/foot	position	had	little	influence	on	peak	head	and	torso	kinematic	and	kinetic	responses.		2.5 Conclusion	Compared	 to	 the	 poor-rated	 GMHR	 seat,	 the	 good-rated	 WHIPS	 and	 SAHR	 seats	 only	attenuated	 four	 ATD	 responses	 across	 the	 range	 of	 collision	 severities	 tested.	 Differences	 in	peak	upper	neck	forces,	moments	and	peak	head	rearward	retraction	were	observed	between	seats	 rated	 good	 and	 poor	 at	 speed	 changes	 greater	 than	 6	 km/h.	 	 The	 RCAR/IIWPG	 head	restraint	 evaluation	 protocol	 and	 seat	 rating	 system	 captures	 improvements	 in	 some	kinematic/kinetic	parameters	 related	 to	 the	 risk	of	whiplash	 injury,	but	 the	 similar	 responses	observed	 for	 other	 kinematic/kinetic	 parameters	 between	 good	 and	 poor	 seats	may	 explain					35	why	 even	 good	 seats	 have	 not	 achieved	 consistent	 reductions	 of	 more	 than	 about	 50%	 in	whiplash	 injuries	 (Jakobsson	 et	 al.,	 2008,	 Kullgren	 and	 Krafft,	 2010,	 Viano	 and	 Olsen,	 2001).	Further	 research	 is	 required	 to	 develop	 anti-whiplash	 seat	 devices	 that	 reduce	 occupant	responses	and	injuries	across	a	wider	range	of	collision	severities	and	other	collision	conditions.						 					36	Chapter	3. 	Effects	of	Seat	Hinge	Rotation	on	ATD	Responses		3.1 Introduction	Automotive	 seats	 are	 the	 primary	 safety	 devices	 for	 whiplash	 injury	 prevention	 during	rear-end	collisions.	Modern	automotive	seatbacks	are	designed	to	address	two	primary	safety	functions:	1)	 reduce	the	 forces	applied	by	 the	seat	 to	 the	occupant,	and	2)	prevent	occupant	ejection	 out	 of	 the	 seat	 (Viano,	 2003b,	 Viano,	 2003d,	 Viano,	 2008).	 	 Yielding	 and	 compliant	seatbacks	offer	better	energy-absorbing	properties	decreasing	the	accelerations	applied	to	the	occupant	 over	 time	 and	minimize	 head	movements	 relative	 to	 the	 torso	 (Viano,	 2003d).	 If	 a	seatback	is	too	compliant,	however,	large	rearward	deflection	of	the	seatback	can	occur	when	the	occupant	loads	the	seat	during	a	high-speed,	rear-end	collision.	This	loading	can	lead	to	the	seat	collapsing	and	the	occupant	sliding	rearward	under	the	seatbelt	into	the	rear	of	the	cabin,	possibly	being	ejected	 from	 the	 vehicle.	 	 If,	 on	 the	other	hand,	 the	 seatback	 is	 too	 rigid,	 the	forces	applied	by	the	seat	to	the	occupant	during	lower	speed,	rear-end	collisions	will	be	high.	Thus,	automotive	seat	designs	must	balance	the	structural	stiffness	to	perform	well	during	both	low-	and	high-speed,	rear-end	crashes.	There	 are	 currently	 several	 automotive	 seats	 designed	 to	 reduce	 or	 prevent	 whiplash	injuries.	 These	 seats	 include	 the	 General	 Motor’s	 High	 Retention	 seat	 (GMHR),	 Volvo’s	Whiplash	 Injury	 Prevention	 seat	 (WHIPS),	 Saab’s	 Self-Aligning	 Head	 Restraint	 (SAHR),	 and	Toyota’s	Whiplash	 Injury	Lessening	seat	(WIL)	 (Jakobsson	et	al.,	2000,	Viano,	2008,	Viano	and					37	Olsen,	2001,	Wiklund	and	Larsson,	1997,	Sekizuka,	1998).		Epidemiological	studies	have	shown	that	 the	 WHIPS	 and	 SAHR	 reduce	 whiplash	 injury	 risk	 by	 20%	 to	 50%	 following	 rear-end	collisions	 (Farmer	 et	 al.,	 2003,	 Ivancic,	 2011,	 Viano	 and	Olsen,	 2001,	 Jakobsson	 et	 al.,	 2008,	Kullgren	and	Krafft,	 2010).	A	 comparison	of	 four	 anti-whiplash	 seats	 (see	Chapter	2)	 showed	that	even	 the	seats	 rated	good	by	 the	Research	Council	 for	Automobile	Repairs/International	Insurance	Whiplash	 Prevention	 Group	 (RCAR/IIWPG)	 consistently	 attenuated	 only	 four	 peak	occupant	 responses	 (i.e.,	 peak	 upper	 neck	 shear	 and	 axial	 forces,	 flexion/extension	 bending	moment	 and	head	 retraction).	 	 This	 inability	 to	 attenuate	 all	 of	 the	occupant	 responses	may	explain	why	the	good-rated	anti-whiplash	seats	have	yet	to	reduce	the	risk	of	whiplash	injury	by	more	 than	 50%	 (Farmer	 et	 al.,	 2003,	 Ivancic,	 2011,	 Jakobsson	 et	 al.,	 2008,	 Viano	 and	Olsen,	2001).	 Thus,	 an	 additional	 scope	 remains	 to	 improve	 anti-whiplash	 seats	 to	 address	 those	occupants	who	continue	to	be	injured	with	the	currently	available	anti-whiplash	seats.	We	 developed	 a	 novel	 anti-whiplash	 seat	 that	 actively	 rotates	 the	 seat	 hinge	 during	 a	rear-end	collision	to	change	the	apparent	structural	stiffness	of	the	seat	hinge	with	the	goal	of	reducing	 the	kinematics	and	kinetics	experienced	by	an	occupant	and	 thus	 reduce	 the	 risk	of	whiplash	 injury	 (see	Appendix	A).	 The	 seat	 rotates	 forward	before	 the	crash	 to	decrease	 the	head-to-head-restraint	 distance	 (i.e.,	 backset)	 and	 start	 applying	 a	 forward	 force	 to	 the	occupant	 before	 the	 collision	 occurs	 (see	 Appendix	 B).	 The	 seatback	 then	 actively	 rotates	rearward	as	the	occupant	loads	the	seat	in	order	to	attenuate	the	acceleration	experienced	by	the	occupant.	The	seatback	then	stops	and	does	not	rebound	forward,	further	attenuating	the					38	acceleration	and	speed	change	experienced	by	the	occupant.	Preliminary	testing	to	determine	an	 effective	 seat	 hinge	 rotation	 profile	 was	 conducted	 at	 only	 one	 speed	 change	 (Δv)	 of	 12	km/h	with	a	collision	pulse	duration	(Δt)	of	208	ms.	The	primary	goal	of	the	present	study	was	to	 compare	 the	 kinematic	 and	 kinetic	 responses	 of	 a	 BioRID	 II	 anthropomorphic	 test	 device	(ATD)	seated	on	the	anti-whiplash	seat	with	the	seat	hinge	rotation	profile	(Experimental	seat)	and	 on	 an	 unmodified	 automotive	 seat	 (Control	 seat)	 during	 a	 range	 of	 low-speed,	 rear-end	perturbations	 with	 speed	 changes	 of	 2	 to	 12	 km/h.	 A	 secondary	 goal	 of	 the	 study	 was	 to	quantify	 the	 repeatability	 of	 the	 seat	 and	 ATD	 responses.	 In	 comparison	 to	 an	 unmodified	Control	seat,	we	hypothesized	that	the	Experimental	anti-whiplash	seat	would	reduce	most	of	the	peak	ATD	responses	across	all	tested	collision	severities.		3.2 Methods	3.2.1 Experimental	Anti-Whiplash	Seat	The	Experimental	anti-whiplash	seat	design	consisted	of	a	modified	2004	GMHR	seat	and	two	motors	 to	 control	 the	 seat	 hinge	 (Figure	 3.1).	 	 The	 seat	 pan	 and	 head	 restraint	 of	 the	GMHR	 seat	 remained	 unmodified	 in	 this	 design.	 The	 seatback	 consisted	 of	 a	 rigid	 aluminum	outer	frame	and	an	GMHR	upper	seatback	rigidly	mounted	within	this	outer	frame.	Two	large	rotational	servomotors	 (AKM52K,	Kollmorgen,	Waltham,	MA,	USA)	connected	to	helical	 right-		angle	 gearheads	 (VTR014-035,	 35:1	 gear	 ratio,	 Thomson	 Linear,	 Radford,	 VA,	 USA)	 were	mounted	on	either	side	of	the	outer	frame	at	the	seat	hinge	location	and	were	geared	to	rotate					39		Figure	3.1.	Photographs	of	 the	experimental	set-up	with	 the	BioRID	 II	ATD	on	A.)	 the	Control	seat	and	B.)	the	novel	anti-whiplash	automotive	seat	and	global	reference	frame	(X,	Z).	The	seat	hinge	 and	 seatback	motors	on	 the	 left	 side	of	 the	 anti-whiplash	 seat	 are	 labelled.	Additional	seat	hinge	and	 seatback	motors	are	 located	on	 the	 right	 side	of	 the	 seat	 (not	 labelled).	 	 The	seatback	motors	are	not	used	in	this	current	study.			in	 unison	 (one	 in	 a	 positive	 direction	 and	 the	 other	 in	 a	 negative	 direction)	 to	 control	 the	rotation	of	the	seat	hinge	through	predefined	rotation	profiles	(Figure	3.1).	Through	a	series	of	preliminary	experiments	(see	Appendix	B),	we	explored	different	seat	hinge	rotation	profiles	to	find	a	profile	that	effectively	reduced	ATD	responses	at	a	single	speed	change	(Δv	=	12	km/h).	The	 chosen	 dynamic	 seat	 hinge	 rotation	 profile	 (Figure	 3.2,	 grey	 line)	 was	 programmed	 to	rotate	the	seatback	forward	beginning	90	ms	(tforward-onset)	before	the	collision	onset	to	a	peak	pre-perturbation	angle	(θforward-peak)	of	-5.6	deg	and	to	then	rotate	the	seatback	rearward	to	a	peak	 rearward	 angle	 (θrearward-peak)	 of	 5.7	 deg	 at	 an	 initial	 angular	 velocity	 (ωrearward-int)	 of	 3.8	deg/s	 beginning	 90	 ms	 (trearward-onset)	 after	 collision	 onset.	 	 The	 observed	 output	 seat	 hinge	rotation	(Figure	3.2,	red	line)	resulted	in	a	pre-perturbation	θforward-peak	of	-3.6	deg	occurring	at	a					40	tforward-onset	 of	 -90	 ms	 followed	 by	 rearward	 seat	 hinge	 rotation	 to	 a	 θforward-peak	 of	 3.6	 deg	occurring	at	a	trearward-onset	of	40	ms.	Both	seat	hinge	motors	were	controlled	by	separate	digital	 servo	drives	 (Servostar	600,	Kollmorgen,	 Waltham,	 MA,	 USA)	 connected	 to	 an	 universal	 motion	 interface	 and	 a	 motion	controller	(NI	UMI	7774	&	NI	PXI	7350,	National	Instruments	Corporation,	Austin,	Texas,	USA).		A	 custom	 LabVIEW	 program	 (National	 Instruments	 Corporation,	 Austin,	 Texas,	 USA)	 was	created	 to	 send	commands,	monitor	 the	status	of	and	 record	encoder	data	directly	 from	the	motors.			Figure	3.2.	The	programed	input	seat	hinge	rotation	profile	(grey	line)	and	the	observed	output	seat	 hinge	 rotation	 (red	 line)	 of	 the	 anti-whiplash	 seat	 during	 a	 12	 km/h	perturbation	 (black	line,	right	axis).		The	cyan	lines	illustrated	measurements	of	each	seat	hinge	rotation	parameter	used	 to	 define	 the	 input	 rotation	 profile	 (onset	 delay	 of	 pre-perturbation	 forward	 rotation:	tforward-onset,	peak	pre-perturbation	forward	rotation	angle:	θforward-peak,	peak	rearward	rotational	angle:	 θrearward-peak,	 initial	 rearward	 angular	 velocity:	 ωrearward-int,	 and	 onset	 delay	 of	 rotation:	trearward-onset).					41	3.2.2 Anthropomorphic	Test	Device	and	Instrumentation	A	 BioRID	 II	 ATD	 (Humanetics,	 Plymouth,	MI,	 USA)	was	 instrumented	 to	measure	 head,	torso	(T1),	and	pelvis	kinematics	and	kinetics	(Figure	3.1).	A	six-axis	load	cell	(Forces:	FX,	FY,	FZ;	and	Moments:	MX,	MY,	MZ;	Model	4949a,	Robert	A.	Denton,	Inc.,	Rochester	Hills,	MI,	USA)	was	mounted	 at	 the	 atlanto-occipital	 joint	 (AOJ)	 to	 measure	 upper	 neck	 forces	 and	 moments.		Linear	 forward	 accelerations	 of	 the	 head	 and	 T1	 were	 measured	 using	 two	 uni-axial	accelerometers	 (7264C	 sensors;	 ±500	 g,	 Endevco,	 San	 Juan	Capistrano,	 CA,	USA)	mounted	 at	both	 the	 head	 center	 of	 mass	 and	 to	 the	 T1	 vertebra.	 A	 uni-axial	 angular	 rate	 sensor	 was	mounted	 at	 the	 head	 center	 of	 mass	 (ARS-1500;	 ±26.2	 rad/s,	 DTS,	 Seal	 Beach,	 CA,	 USA)	 to	measure	angular	kinematics	in	the	sagittal	plane	(i.e.,	flexion	and	extension).			A	motion	capture	system	(Optotrak	Certus,	Northern	Digital,	Waterloo,	ON,	Canada)	was	used	to	track	infrared	light	emitting	diode	(IRED)	markers	affixed	to	the	head,	AOJ,	T1	and	pelvis	of	the	ATD	to	measure	displacements.	 	Additional	 IRED	markers	were	mounted	to	the	seat	to	record	displacements	and	rotations	of	the	seatback	and	to	create	a	global	reference	frame	(+X	forward,	+Y	right,	and	+Z	down,	Figure	3.1).	Horizontal	sled	acceleration	was	measured	with	a	uni-axial	 accelerometer	 (2220-100;	±100	 g,	 Silicon	 Design	 Inc.,	 Issaquah,	WA,	 USA)	mounted	directly	to	the	base	of	the	linear	sled	frame.		Head	restraint	contact	was	detected	with	a	force	sensitive	 resistor	 (FSR,	Model	 406,	 Interlink	 Electronics,	 Camarillo,	 CA,	 USA)	 attached	 to	 the	front	of	the	head	restraint.					42	All	 accelerometer,	 load	 cell	 and	 angular	 rate	 transducer	 signals	 were	 simultaneously	sampled	at	10	kHz	using	a	National	Instruments	Data	Acquisition	(DAQ)	PXI	system	(PXI-4495	&	PXI-6289,	National	Instruments	Corporation,	Austin,	Texas,	USA)	and	a	custom-written	LabVIEW	virtual	instrument.	Optotrak	data	were	acquired	at	200	Hz	and	collection	was	triggered	by	the	DAQ	system	to	synchronize	the	data.	Subsequent	data	and	statistical	analyses	were	performed	using	Matlab	(Mathworks,	Natick,	MA,	USA).	All	data	channels	were	digitally	filtered	in	Matlab	and	conformed	to	the	SAE	J211	(Channel	class	180	for	the	ATD	sensors	and	Channel	Class	60	for	the	sled	accelerometer)	(SAE,	1995).		3.2.3 Test	Procedures	The	 BioRID	 II	 ATD	 was	 dressed	 in	 two	 layers	 of	 lycra	 and	 was	 seated	 on	 either	 an	unmodified	GMHR	seat	 (Control	 seat)	or	 the	anti-whiplash	 seat	 (Experimental	 seat)	mounted	on	a	10	m	 long,	 feedback-controlled	 linear	sled	 (Kollmorgen	 IC55-100A7,	Waltham,	MA,	USA;	Figure	3.1).	For	both	seats,	the	initial	seatback	angle	was	set	to	27	deg	rearward	from	vertical	(Siegmund	et	al.,	2005a).	To	ensure	the	repeatability	of	the	experiments,	the	initial	position	of	the	 ATD	 was	 adjusted	 to	 a	 pre-defined	 posture	 that	 was	 constant	 within	 each	 seat	 and	 as	similar	 as	 possible	 between	 seats.	 This	 pre-defined	 posture	 was	 confirmed	 using	 the	 3D	positions	measured	by	Optotrak	at	the	beginning	of	all	 trials.	The	seatbelt	was	removed	from	the	seat	to	prevent	interactions	that	could	affect	head,	neck	and	torso	kinematics.						43	For	 both	 the	 Control	 and	 Experimental	 seats,	 the	 ATD	 was	 exposed	 to	 five	 rear-end	perturbations	of	increasing	speed	changes	(Δv	=	2,	4,	6,	and	8	km/h	with	a	Δt	of	148	ms,	and	12	km/h	with	a	Δt	of	185	ms)	(Figure	3.3).	Due	to	limitations	in	the	force-generating	ability	of	the	linear	motors,	 a	 longer	 pulse	 duration	was	 needed	 for	 the	 12	km/h	 speed	 change.	All	 of	 the	pulses	used	here	have	a	 longer	duration	 than	 the	RCAR/IIWPG	pulse	 (Δt	=	91	ms)	 (Insurance	Institute	 of	 Highway	 Safety,	 2008b).	 The	 sled	 was	 accelerated	 forward	 from	 a	 stationary	position	for	the	2,	4	and	6	km/h	speed	changes	and	from	a	constant	rearward	speed	of	6	km/h	for	 the	 8	 and	 12	 km/h	 speed	 changes.	 A	 comparison	 of	 the	 ATD	 responses	 for	 an	 8	 km/h	collision	 starting	 from	 a	 stationary	 position	 to	 an	 8	 km/h	 collision	 starting	 from	 a	 6km/h	rearward	 velocity	 showed	 less	 than	 10%	 differences	 in	most	 of	 the	 occupant	 responses	 and	neck	 injury	criteria	(mean:	2.50%	±	2.53%;	range:	0.5%	–	9.2%;	see	Chapter	2).	 	To	assess	the	repeatability	of	the	ATD’s	responses,	four	additional	repeated	trials	were	collected	at	the	8	and	12	km/h	collision	speed	changes	for	both	the	Control	and	Experimental	seats	(for	a	total	of	5	trials	for	each	seat	at	both	speed	changes).		The	horizontal	position	 (dbackset)	and	vertical	position	 (dheight)	of	 the	head	 relative	 to	 the	head	restraint	is	known	to	affect	the	occupant	response	and	the	risk	of	whiplash	injury	(Nygren	et	al.,	1985,	Siegmund	et	al.,	1999,	Stemper	et	al.,	2006).	For	the	ATD	postures	used	here,	the	head-to-head-restraint	 geometry	was	 different	 for	 the	 Experimental	 seat	 (dbackset	 =	 57.5	mm	and	dheight=	-2.3	mm	prior	to	any	forward	rotation	of	the	seat	hinge)	and	Control	seat	(dbackset	=	95.9	mm	and	dheight	 =	 -29.9	mm).	This	difference	arose	because	we	wanted	 the	Experimental					44	anti-whiplash	 seat	 to	 be	 rated	 as	 good	 according	 to	 the	 RCAR/IIWPG	 criteria.	 An	 extra	experimental	 condition	 (EXP90),	 where	 the	 Experimental	 seat’s	 backset	 and	 head	 restraint	height	 were	 86.2	 mm	 and	 -24.6	 mm	 respectively,	 was	 added	 to	 determine	 whether	 the	observed	 differences	 in	 ATD	 responses	 between	 the	 Experimental	 and	 Control	 seats	 were	caused	by	these	head	restraint	geometry	differences	rather	than	the	active	seat	hinge	control.	This	extra	condition	was	run	only	once	at	a	12	km/h	speed	change.		A	 total	 of	 27	 trials	 were	 collected,	 13	 trials	 for	 the	 Control	 seat	 and	 13	 trials	 for	 the	Experimental	seat,	and	one	trial	for	the	EXP90	condition.			For	test	conditions	with	five	repeated	trials	 (i.e.,	 the	Dv	=	8	 and	12	 km/h	 trials),	 the	average	of	 the	peak	 responses	 across	 the	 five	trials	was	used	for	analysis.			3.2.4 Data	Analysis		 Peak	 horizontal,	 forward,	 sled	 accelerations	 (aX-sled)	 were	 extracted	 directly	 from	 the	accelerometer	mounted	to	the	sled.		Head	and	torso	accelerometer	data	were	reported	in	local	ATD	head	and	T1	reference	frames	and	were	corrected	to	remove	the	earth’s	gravity	using	the	head	and	T1	orientations	determined	from	the	Optotrak	data.	Peak	linear	forward	accelerations	of	 the	 head	 (aX-head)	 and	 T1	 vertebra	 (aX-T1)	 were	 extracted	 directly	 from	 the	 corrected	accelerometer	 data,	 and	 peak	 rotational	 velocities	 of	 the	 head	 (ωY-head)	 in	 the	 sagittal	 plane	were	extracted	directly	from	the	angular	rate	sensor.	Peak	upper	neck	shear	(FX)	and	axial	(FZ)	forces	and	the	flexion/extension	bending	moment	(MY)	were	determined	from	the	upper	neck	load	cell	and	reported	in	the	ATD	reference	frame	as	the	forces/moments	applied	by	the	neck					45	to	the	head.		Initial	head	angle	was	defined	as	the	average	angle	between	-300	ms	and	-200	ms	preceding	 the	 onset	 of	 aX-sled,	 and	 peak	 head	 extension	 angle	 (θhead)	 was	 defined	 as	 the	maximum	 rotation	of	 the	head	 into	extension	 relative	 to	 the	 initial	 head	angle.	A	 foreperiod	between	 -300	ms	 and	 -200	ms	 before	 the	 onset	 of	 forward	 aX-sled	 was	 used	 to	 define	 initial	values	because	the	pre-perturbation	forward	rotation	of	the	seat	hinge	could	start	as	early	as	170	 ms	 before	 collision	 onset	 (see	 Appendix	 D).	 Peak	 retraction	 (Rx)	 was	 defined	 as	 the	maximum	horizontal	displacement	in	the	global	reference	frame	of	the	AOJ	with	respect	to	the	T1	vertebrae	with	rearward	displacements	defined	as	negative	values.		The	shapes	of	the	back	of	the	ATD’s	head	and	front	of	the	head	restraint	were	digitized	relative	to	existing	Optotrak	IRED	markers.	These	shapes	were	then	used	to	calculate	the	initial	head	restraint	backset	and	vertical	position	for	each	trial.	The	initial	backset	(dbackset)	and	height	(dheight)	of	the	head	restraint	were	defined	as	the	average	head-to-head-restraint	horizontal	and	vertical	distances	between	 -300	ms	and	 -200	ms	before	 the	onset	of	 forward	aX-sled.	Negative	values	indicated	that	the	top	of	the	head	restraint	was	lower	than	the	top	of	the	ATD	head.	The	same	 head	 and	 head	 restraint	 shapes	 were	 used	 to	 calculate	 the	 peak	 forward	 rebound	(drebound),	which	was	defined	as	 the	maximum	forward	head-to-head-restraint	distance.	Time-to-head-restraint	 contact	 (Δthead-contact)	was	extracted	 from	 the	 time	of	 force	onset	 in	 the	FSR	attached	to	the	head	restraint	(onset	of	forward	sled	acceleration	defined	as	t	=	0	ms).	Onsets	of	sled	acceleration	and	head	restraint	contact	were	determined	when	the	accelerometer	and	FSR	 signals,	 respectively,	 reached	 1.5	 times	 the	 peak	 background	 noise	 level	 present					46	between	-300	ms	to	-200	ms	before	the	onset	of	the	collision	perturbation	and	were	confirmed	visually.	The	sled	accelerometer	 (aX-sled)	and	head	contact	FSR	signals	had	high	signal-to-noise	ratios	after	filtering	according	to	SAE	J211	Channel	class	180	and	did	not	require	further	manual	corrections	of	the	onsets.			Three	neck	 injury	 criteria	 (NICmax,	Nij,	 and	Nkm)	were	computed	 from	the	accelerometer	and	 load	 cell	 data.	 	 The	 Neck	 Injury	 Criterion	 (NICmax)	 was	 calculated	 from	 the	 relative	horizontal	acceleration	and	velocity	in	the	global	reference	frame	between	the	head	center	of	mass	and	the	T1	joint	(Equation	3.1;	Bostrom	et	al.,	1996).	The	Normalized	Neck	Injury	Criterion	(Nij)	 was	 calculated	 from	 the	 axial	 load	 (FZ)	 and	 the	 flexion/extension	 bending	moment	 (MY)	measured	from	the	upper	neck	load	cell	(Equation	3.2;	critical	Fint	and	Mint	intercept	values	used	for	normalization:	Fint-tension	=	6806	N,	Fint-compression	=	-6160	N,	Mint-flexion	=	310	N,	and	Mint-extension	=	-135	N;	Eppinger	et	al.,	1999,	Eppinger	et	al.,	2000).	The	Neck	Protection	Criterion	(Nkm)	was	calculated	 from	 the	 sagittal	 shear	 force	 (FX)	 and	 the	 flexion/extension	bending	moment	 (MY)	measured	from	the	upper	neck	load	cell	(Equation	3;	critical	Fint	and	Mint	intercept	values	used	for	normalization:	Fint-shear-positive	=	845	N,	Fint-shear-negative	=	845	N,	Mint-extension	=	47.5	Nm,	and	Mint-flexion	=	88.1	Nm;	Schmitt	et	al.,	2001,	Schmitt	et	al.,	2002).		Peak	values	of	the	three	neck	injury	criteria	were	then	extracted	for	analysis	and	compared	to	proposed	injury	thresholds	(NIC:	15	m2/s2,	 Eichberger	 et	 al.,	 1998;	 Nij	 and	Nkm	 :	 a	 normalized	 value	 of	 one,	 Schmitt	 et	 al.,	 2001,	Schmitt	 et	 al.,	 2002,	 Eppinger	 et	 al.,	 1999).	 	 These	 injury	 thresholds	 correspond	 to	 different	human	tolerance	levels	for	the	causation	of	whiplash-related	injuries	–	NIC:	long-term	whiplash-				47	associated	disorders	(WAD)	levels	1-3	(Bostrom	et	al.,	2000,	Bostrom	et	al.,	1996),	Nij:	22%	risk	of	 abbreviated	 injury	 scale	 (AIS)	 level	 3	 (i.e.,	 fracture;	 Eppinger	 et	 al.,	 1999,	 Eppinger	 et	 al.,	2000),	and	Nkm:	AIS	 level	1	 injury	causation	(i.e.,	minor	 injury;	Schmitt	et	al.,	2001,	Schmitt	et	al.,	2002).	NICmax t = maximumfirst	150	ms #$%& ' ∙ ) + (,$%&(')). 	 …	 Equation	3.1	N01(t) = 23(4)2567 + 89:;(4)8567 	 	 	 	 	 	 …	 Equation	3.2	N<=(t) = 2>(4)2567 + 89:;(4)8567 	 	 	 	 	 …		 Equation	3.3			The	repeatability	of	the	ATD	on	the	Control	and	Experimental	seats	was	assessed	for	the	8	and	12	km/h	speed	changes	using	 the	coefficient	of	variation	 (COV),	which	was	calculated	as	the	standard	deviation	(SD)	divided	by	the	mean	and	expressed	as	a	percentage.	The	National	Highway	Traffic	Safety	Administration	(NHTSA)	defines	a	COV	of	5%	or	less	as	excellent,	a	COV	of	10%	or	less	as	acceptable	and	a	COV	greater	than	10%	as	poor	(Rhule	et	al.,	2005).		If	most	of	the	COV	values	for	a	given	seat	(Control	or	Experimental)	were	rated	as	excellent	or	acceptable,	then	 the	 ATD	 responses	 on	 the	 given	 seat	 were	 considered	 repeatable.	 Because	 COV	 is	dependent	 on	 the	 mean	 and	 SD	 values	 of	 the	 responses,	 the	 expected	 decreases	 in	 mean	responses	(which	form	the	denominator	in	the	COV	calculation)	for	the	Experimental	seat	may	generate	 larger	 COV	 values.	 	 A	 comparison	 of	 the	 SD	 values	 between	 the	 GMHR	 and	 the	Experimental	 seats	 would	 provide	 an	 alternate	 assessment	 of	 variability	 and	 would	 help	 to					48	determine	 whether	 the	 larger	 COV	 values	 for	 the	 Experimental	 seat	 were	 likely	 due	 to	 the	lower	mean	responses	or	due	to	the	greater	inherent	variability	(i.e.,	SD)	in	the	ATD	responses.	The	SD	values	for	all	kinematic	and	kinetic	responses	as	well	as	neck	injury	criteria	observed	at	both	the	8	and	12	km/h	collision	speeds	were	used	 in	this	assessment	of	variability.	The	COV	and	SD	values	were	assumed	to	be	independent	observations	and	were	analyzed	with	a	paired	Student’s	 t-test	 after	 a	 Kolmogorov-Smirnov	 test	 determined	 the	 COV	 and	 SD	 values	 were	normally	 distributed.	 	 All	 tests	 were	 performed	 using	 predefined	 functions	 (kstest	 and	ttest)	in	Matlab	(R2017A,	Mathworks,	Natick,	MA,	USA)	at	a	significance	level	a	=	0.05.					To	compare	ATD	responses	between	the	Control	and	Experimental	seats,	99th	percentile	predictive	 intervals	 were	 created	 around	 the	 peak	 responses	 on	 each	 seats	 using	 their	respective	maximum	variability	(COV	value)	observed	from	repeated	trials	at	either	the	8	or	12	km/h	collision	severity	(Horslen	et	al.,	2015).	The	maximum	COV	value	from	either	the	8	or	12	km/h	 collision	 severity	 was	 selected	 to	 represent	 the	 largest	 observed	 variance	 within	 each	response.	 This	 COV	 value	 was	 multiplied	 by	 2.58	 (z-score	 for	 two-tailed	 99th	 percentile)	 to	estimate	 the	 99th	 percentile	 predictive	 limits.	 Since	 the	 COV	 values	 were	 defined	 as	 the	maximum	 variances	 (i.e.,	 SD	 values)	 divide	 by	 the	 mean	 of	 the	 peak	 ATD	 responses,	 these	predictive	 limits	were	represented	as	a	percentage	and	were	then	applied	to	 their	 respective	peak	responses	to	create	predictive	intervals	at	each	collision	severity	(2	–	12	km/h).		Any	ATD	responses	 where	 the	 99th	 percentile	 predictive	 intervals	 from	 the	 Control	 and	 Experimental	seats	did	not	overlap	were	assumed	to	be	statistically	different.	The	99th	percentile	predictive					49	limits	 were	 selected	 to	 account	 for	 the	 numerous	 comparisons	 between	 Control	 and	Experimental	 seats	 for	 the	 ATD	 responses	 at	 each	 collision	 severity	 and	 to	 provide	 a	conservative	estimate	of	statistical	significance.					Figure	3.3.	Exemplar	sled	A)	velocity	and	B)	acceleration	profiles	for	increasing	collision	speeds	(Δv	=	2,	4,	6,	and	8	km/h	with	a	pulse	duration	(Δt)	of	148	ms,	and	Δv	=	12	km/h	with	a	Δt	of	185	ms;	lightest	to	darkest).	Collision	onset	occurred	at	time	=	0	ms.								50	3.3 Results	Both	 the	 Control	 and	 Experimental	 seats	 generated	 graded	 kinematic	 and	 kinetic	responses	to	increasing	collision	severities	(Figure	3.4	&	3.5,	Table	3.1).	 	In	comparison	to	the	Control	 seat,	 active	 seat	 hinge	 rotation	 in	 the	 Experimental	 seat	 decreased	 the	 ATD’s	kinematics	 and	 kinetics	 (aX-head,	 aX-T1,	ωY-head,	 FX,	 FZ,	MY,	 θhead,	 RX	 and	drebound)	 and	neck	 injury	criteria	 (NICmax,	 Nkm,	 and	Nij)	 for	 speed	 changes	 of	 4	 km/h	 and	 higher	 (Figure	 3.5).	With	 the	exception	 of	 Δthead-contact	 (all	 collision	 severities)	 and	 Nij	 (only	 4km/h	 severity),	 there	 was	 no	overlap	between	the	99th	percentile	predictive	intervals	of	the	ATD	response	in	the	Control	and	Experimental	seats.	Peak	forward	head	(aX-head)	and	torso	(aX-T1)	accelerations	decreased	by	50	–	69%	and	24	–	38%	respectively,	while	peak	angular	velocity	of	the	head	(ωY-head)	decreased	by	52	–	66%	 (Figure	3.5A	–	3.5C).	 	 The	Experimental	 seat	also	decreased	upper	neck	 forces	and	moments	compared	to	 the	Control	 seat	 (FX:	43	–	57%	reduction,	FZ:	54	–	74%	reduction,	and	MY:	 53	 –	 76%	 reduction)	 (Figure	 3.5D	 –	 3.5F).	 The	 Experimental	 seat	 also	 reduced	 head	displacements	 relative	 to	 the	 torso:	 head	 angle	 (θhead)	 decreased	 by	 44	 –	 73%	 and	 peak	retraction	(RX)	decreased	by	43	–	75%	(Figure	3.5H	&	3.5I).		At	 the	2	 km/h	 speed	 change,	most	ATD	 responses	 (aX-T1,	ωY-head,	 FX,	 θhead,	 RX	 and	Δthead-contact)	 and	 neck	 injury	 criteria	 (NICmax	 and	Nkm)	 in	 the	 Experimental	 seat	 decreased	 from	 the	Control	seat	by	between	10	to	65%.		Some	of	the	ATD	responses,	however,	were	larger	in	the	Experimental	seat	compared	to	the	Control	seat:	aX-head	increased	by	24%,	FZ	increased	by	95%,	MY	increased	by	433%,	and	Nij	increased	by	4%	(Table	3.1).	Based	on	the	comparison	between					51	99th	percentile	predictive	 intervals,	a	statistically	significant	difference	was	observed	between	Control	 and	 Experimental	 seats	 only	 for	 FZ,	MY,	 θhead,	 and	 RX	 at	 the	 2	 km/h	 collision	 severity	(Figure	3.5).		None	of	the	neck	injury	criteria	(NICmax,	Nij,	nor	Nkm)	exceeded	their	respective	proposed	injury	thresholds	for	either	the	Control	or	Experimental	seat.	The	Experimental	seat	decreased	the	 magnitude	 of	 all	 three	 neck	 injury	 criteria	 (NICmax:	 44	 –	 85%	 reduction,	 Nij:	 23	 –	 66%	reduction,	and	Nkm:	41	–	59%	reduction)	(Figure	3.5J	–	3.5L).			The	 repeatability	of	 the	BioRID	 II	ATD	 responses	 for	both	 the	Control	and	Experimental	seats	were	assessed	using	COV	values	 from	repeated	 trials	at	both	 the	8	and	12	km/h	speed	changes	(Table	3.2).	The	COV	values	for	the	Control	seat	were	excellent	or	acceptable	at	both	speed	 changes	 for	 all	 parameters	 except	 for	 head	 restraint	 height	 (dheight,	 COV	 =	 22.2%	 and	11.3%	 at	 8	 and	 12	 km/h	 respectively).	 The	 Experimental	 seat	 also	 exhibited	 excellent	 or	acceptable	 repeatability,	 except	 for	 head	 restraint	 height	 (COV	 =	 60.4	 and	 75%	 at	 8	 and	 12	km/h	respectively),	NICmax	(COV	=	19.6%	and	10.7%	at	8	and	12	km/h	respectively)	and	Nij	(COV	=	13.3%	at	12	km/h).	The	COV	values	on	the	Experimental	seat	were	significantly	higher	than	those	on	the	Control	seat	for	the	8	and	12	km/h	speed	changes	(paired	t-test;	8	km/h:	t(15)	=	-3.03,	p	=	0.020,	and	12	km/h:	t(15)	=	-3.20,	p	=	0.008).	The	SD	values	of	the	Experimental	and	GMHR	seats	were	not	significantly	different	(paired	t-test;	8	km/h:	t(15)	=	-0.73,	p	=	0.56,	and	12	km/h:	t(15)	=	0.60,	p	=	0.53).		The	larger	COV	values	on	the	Experimental	seat	with	similar	SD					52	values	 to	 the	GMHR	seat	suggested	that	 the	COV	values	considered	poor	 (>	10%)	were	more	likely	due	to	decreased	mean	responses	than	due	to	increased	variability	in	the	ATD	responses.	Prior	to	the	collision,	the	initial	(preset)	head	restraint	backset	(dbackset)	and	head	restraint	height	(dheight)	were	smaller	on	the	Experimental	seat	compared	to	the	Control	seat	(Figure	3.6	and	Table	3.1).		The	average	initial	backset	of	the	Experimental	seat	(57.5	±	10.5	mm)	was	43%	less	 than	 the	 average	 initial	 backset	 of	 the	 Control	 seat	 (95.9	 ±	 2.8	mm)	 across	 all	 collision	severities.	 The	 pre-perturbation	 forward	 seat	 hinge	 rotation	 of	 the	 Experimental	 seat,	which	began	 90	 ms	 before	 the	 collision,	 moved	 the	 head	 restraint	 closer	 to	 the	 ATD’s	 head	 and	decreased	 the	 backset	 to	 13.9	 ±	 6.5	 mm	 (73%	 decrease	 from	 the	 initial	 Experimental	 seat	backset)	prior	to	the	onset	of	the	collision	(t	=	0	ms)	(Figure	3.6).		Across	all	collision	severities,	the	average	initial	head	restraint	height	(dheight)	for	the	Experimental	seat	was	-2.29	±	3.5	mm,	which	placed	the	top	of	the	head	restraint	closer	to,	but	still	below,	the	top	of	the	head	than	for	the	Control	seat	(dheight	=	-29.8	±	8.6	mm)	(Table	3.1).	Peak	forward	rebound	(drebound)	increased	with	 collision	 severity	 for	 both	 the	 Experimental	 and	Control	 seats	 (Table	 3.1).	 	 The	 forward	rebound	for	the	Experimental	seat	was	on	average	48%	lower	than	for	the	Control	seat	at	the	equivalent	collision	severity	(range:	32	–	66%).			A	comparison	of	the	ATD	responses	in	the	EXP90	condition	(the	active	seat	hinge	rotation	with	 head	 restraint	 geometry	 similar	 to	 that	 of	 the	 Control	 seat)	 to	 the	 Control	 and	Experimental	 seats	 showed	 that	 the	 active	 seat	 hinge	 was	 responsible	 for	 most	 of	 the	 ATD	response	 changes	 observed	 between	 the	 Control	 and	 Experimental	 seats	 (Table	 3.3,	 Figure					53	3.7).	Aside	from	the	time-to-head-restraint	contact,	 the	active	seat	hinge	response	generated	an	average	of	85%	(range:	67	to	110%)	of	the	difference	in	ATD	responses	between	the	Control	and	Experimental	seats.		 					54		Figure	 3.4.	 Exemplar	 data	 comparing	 an	 unmodified	 Control	 seat	 (Pontiac	Grand	Am	GMHR)	and	 the	Experimental	 seat	with	dynamic	 seat	hinge	 rotation	 (θseat-hinge)	 for	 the	BioRID	 II	ATD.	Each	column	represents	occupant	responses	for	a	Control	and	Experimental	seat	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6,	and	8	km/h	with	a	collision	pulse	duration	(Δt)	of	148	ms	and	12	km/h	with	Δt	=	185	ms).	Hollow	circles	represent	the	onset	of	head-to-head-restraint	contact	and	peak	responses	of	each	ATD	response	parameter	for	each	trial.						55				Figure	3.5.	Experimental	results	of	peak	kinematic	and	kinetic	responses	for	the	BioRID	II	ATD	seated	 on	 the	 Control	 and	 Experimental	 anti-whiplash	 seats.	 For	 all	 the	 graphs,	 grey	 circles	represent	 the	Control	 seat	 and	black	 triangles	 represent	 the	Experimental	 seat	with	 the	 seat	hinge	rotation.	 	The	 faded	grey	bars	represent	 the	99th	percentile	predictive	corridors	 for	 the	Control	and	Experimental	seats.	Y-axis	values	for	panels	5C	and	5I	have	been	inverted	for	visual	purposes	to	show	increasing	responses	from	bottom	to	top.			 					56		Figure	 3.6.	Exemplar	 horizontal	 head-to-head-restraint	 distance	 in	 the	 X-axis	measured	 from	the	back	of	the	BioRID	II	ATD	head	to	the	front	face	of	the	head	restraint.		Five	repeated	trials	of	the	ATD	seated	on	the	Control	GMHR	seat	 (grey)	and	on	the	Experimental	seat	with	dynamic	seat	 hinge	 rotation	 (black)	 as	 well	 as	 single	 trial	 of	 the	 EXP90	 condition	 (red)	 at	 a	 collision	severity	of	12	km/h	(Δt	=	185	ms).		Collision	onset	occurred	at	time	=	0	ms	and	a	head-to-head	restraint	distance	≤	0	mm	indicated	that	the	head	was	in	contact	with	the	head	restraint.		 					57		Figure	 3.7.	 The	 relative	 contributions	 of	 the	 active	 seatback	 response	 and	 head	 restraint	geometry	 to	 the	 improved	 response	of	 the	 Experimental	 seat	 compared	 to	 the	Control	 seat.	Thirteen	 ATD	 response	 parameters	 are	 shown	 here	 for	 the	 Control	 seat	 (set	 to	 0%),	 the	Experimental	seat	(set	to	100%)	and	the	extra	condition	(EXP90)	of	the	Experimental	seat	with	a	head	restraint	geometry	similar	to	the	Control	seat	(intermediate	percentage	values).	Most	of	the	 improvements	observed	between	the	Control	and	Experimental	seats	comes	 from	adding	the	active	seat	hinge	response	(difference	between	Control	and	EXP90	data	in	this	graph)	rather	than	from	the	change	in	head	restraint	geometry	(difference	between	EXP90	and	Experimental	seat	data	in	this	graph).						58			Table	3.1.	Peak	Responses	for	Control	and	Experimental	(EXP)	seats	for	each	collision	severity.		Underlined	results	highlight	cases	where	the	Experimental	trials	had	larger	responses	than	the	Control	trials.	Dv	(km/h)	 2	km/h,	Dt	=	148	ms	 4	km/h,	Dt	=	148	ms	 6	km/h,	Dt	=	148	ms	 8	km/h,	Dt	=	148	ms	 12	km/h,	Dt	=	185	ms	Trial	 Control	 EXP	 Control	 EXP	 Control	 EXP	 Control	 EXP	 Control	 EXP	dbackset	(mm)	 90.4	 49.1	 91.0	 45.8	 90.2	 48.7	 94.7	 69.2	 95.1	 51.5	dheight	(mm)	 -31.0	 2.7	 -32.3	 -6.4	 -42.3	 -4.7	 -20.0	 -1.7	 -23.7	 -1.4	aX-sled	(m/s2)	 5.8	 5.7	 11.2	 10.9	 15.9	 15.8	 22.8	 21.7	 31.6	 30.2	Dthead	contact	(ms)	 -	 -	 164.5	 -	 138.8	 149.4	 129.5	 112.8	 112.1	 92.8	aX-head	(m/s2)	 8.7	 10.8	 35	 13.4	 57.1	 17.6	 85.2	 42.4	 120.7	 49.3	aX-T1	(m/s2)	 8.2	 7.1	 20.6	 14.2	 26.4	 191.1	 44.1	 27.2	 47.0	 35.7	FX	(N)	 89.5	 73.4	 166.4	 71.7	 193.3	 91.4	 213.5	 121.5	 264.4	 140.7	FZ	(N)	 34.5	 67.4	 173.5	 72.9	 336.5	 87.6	 414.4	 144.3	 586.1	 267.3	MY	(Nm)	 1.0	 5.5	 6.6	 2.6	 18.5	 4.4	 20.1	 7.5	 25.1	 11.9	ω	Y-head	(deg/s)	 110.6	 84.0	 246.3	 86.2	 294.4	 99.8	 335.3	 162.4	 393.8	 189.4	qhead	(deg)	 8.8	 3.1	 15.3	 4.1	 14.5	 6.8	 15.6	 8.7	 16.4	 8.9	drebound	(mm)	 119.6	 81.3	 175.5	 65.0	 269.0	 88.0	 336.1	 185.8	 413.4	 267.5	RX	(mm)	 -34.1	 -14.4	 -57.7	 -14.7	 -58.1	 -29.5	 -61.3	 -35.1	 -63.8	 -35.8	NICmax	(m2/s2)	 1.3	 0.7	 3.5	 0.5	 5.0	 2.8	 7.0	 2.3	 9.8	 4.0	Nij	 0.0	 0.0	 0.0	 0.0	 0.1	 0.0	 0.1	 0.0	 0.2	 0.1	Nkm	 0.2	 0.2	 0.3	 0.2	 0.4	 0.2	 0.5	 0.2	 0.6	 0.3					59	Table	3.2.			Mean	(standard	deviation)	and	Coefficient	of	Variation	(COV)	from	five	repeated	whiplash-like	perturbation	(n	=	5)	on	the	Control	and	Experimental	seats	at	collision	severities	of	Dv	=	8	km/h	with	Dt	=	148	ms	and	Dv	=	12	km/h	with	Dt	=	185	ms.	Dv	(km/h)	 8	km/h,	D	=	148	ms	 12	km/h,	D	=	185	ms	Trial	 Control	 Experimental	 Control	 Experimental	Parameter	 Mean	(SD)	 COV	(%)	 Mean	(SD)	 COV	(%)	 Mean	(SD)	 COV	(%)	 Mean	(SD)	 COV	(%)	Initial	dbackset	(mm)	 94.7	(2.2)	 5.9	 69.3	(4.1)	 5.9	 95.1	(2.6)	 2.8	 51.5	(4.8)	 9.3	Initial	dheight	(mm)	 -23.8	(5.3)	 22.2	 5.3	(3.2)	 60.4	 -25.0	(2.8)	 11.3	 -1.6	(1.2)	 75.0	aX-sled	(m/s2)	 22.8	(1.1)	 4.7	 21.7	(1.1)	 4.6	 31.6	(0.4)	 1.2	 30.2	(0.4)	 1.4	Dthead	contact	(ms)	 125.9	(1.1)	 0.8	 112.8	(1.1)	 0.8	 112.1	(1.4)	 1.3	 82.8	(6.0)	 7.3	aX-head	(m/s2)	 85.2	(1.0)	 1.1	 42.4	(2.1)	 5.0	 120.7	(1.7)	 1.4	 49.3	(3.1)	 6.3	aX-T1	(m/s2)	 44.1	(0.4)	 1.0	 27.2	(0.8)	 3.1	 47.0	(0.3)	 0.7	 35.7	(1.0)	 2.8	FX	(N)	 213.5	(3.8)	 1.8	 121.5	(9.5)	 7.8	 264.4	(14.1)	 5.4	 140.7	(5.7)	 4.0	FZ	(N)	 414.4	(9.3)	 2.2	 144.3	(4.5)	 3.1	 586.1	(16.2)	 2.8	 267.3	(4.4)	 1.7	MY	(Nm)	 20.1	(1.2)	 6.1	 7.5	(0.4)	 4.8	 25.1	(0.1)	 0.2	 11.9	(0.7)	 5.8	w	Y-head(deg/s)	 335.3	(16.7)	 5.0	 162.4	(10.4)	 6.4	 393.8	(7.4)	 1.9	 189.4	(11.5)	 6.1	qhead	(deg)	 15.6	(0.3)	 2.1	 8.7	(0.7)	 7.9	 16.4	(0.6)	 3.8	 8.9	(0.4)	 4.0	RX	(mm)	 -61.3	(1.1)	 1.7	 -35.1	(2.6)	 7.4	 -63.8	(2.3)	 3.6	 -35.8	(1.3)	 3.6	drebound	(mm)	 335.1	(10.4)	 3.5	 185.8	(6.5)	 3.5	 413.4	(7.7)	 1.9	 267.5	(7.0)	 2.6	NICmax	(m2/s2)	 7.0	(0.3)	 4.2	 2.3	(0.4)	 19.6	 9.8	(0.3)	 2.6	 4.0	(0.4)	 10.7	Nij	 0.12	(0.00)	 2.3	 0.04	(0.00)	 3.1	 0.16	(0.00)	 1.5	 0.09	(0.01)	 13.3	Nkm	 0.48	(0.02)	 3.4	 0.22	(0.01)	 5.0	 0.60	(0.02)	 2.9	 0.28	(0.02)	 6.3	Notes:		The	underlined	COV	values	indicate	a	COV	rating	of	acceptable	(5%	≤	COV	<	10%),	the	bolded	COV	values	indicate	a	COV	rating	of	poor	(COV	>	10%),	and	all	other	values	are	rated	good	(COV	<	5%).							60		 	Table	3.3.	 	 	Peak	ATD	responses	and	normalized	percentages	(Norm	%)	for	the	extra	EXP90	condition	relative	to	the	Control	and	Experimental	(EXP)	seats	at	a	speed	change	of	12	km/h.	Normalize	percentages	show	where	the	EXP90	response	falls	within	the	interval	between	the	Control	 (0%)	 and	 Experimental	 (100%)	 seats.	 The	 normalized	 percentages	 for	 the	 ATD	responses	(bottom	13	parameters	in	this	table)	are	shown	graphically	in	Figure	3.7.		Trial	 Control	 EXP90	 EXP	Parameter	 Peak	 Norm	%	 Peak	 Norm	%	 Peak	 Norm	%	Initial	dbackset	(mm)	 95.1	 0	 86.2	 19.8	 51.5	 100	Initial	dheight	(mm)	 -25.0	 0	 -24.6	 1.7	 -1.6	 100	aX-sled	(m/s2)	 31.6	 0	 -31.0	 40.4	 30.2	 100	Dthead	contact	(ms)	 112.1	 0	 111.7	 1.4	 82.8	 100	aX-head	(m/s2)	 120.7	 0	 63.8	 79.8	 49.3	 100	aX-T1	(m/s2)	 47.0	 0	 39.4	 67.0	 35.7	 100	FX	(N)	 264.4	 0	 158.5	 85.7	 140.7	 100	FZ	(N)	 586.1	 0	 277.2	 96.9	 267.3	 100	MY	(Nm)	 25.1	 0	 12.4	 95.9	 11.9	 100	w	Y-head(deg/s)	 393.8	 0	 244.2	 73.2	 189.4	 100	qhead	(deg)	 16.4	 0	 10.1	 84.4	 8.9	 100	drebound	(mm)	 413.4	 0	 288.6	 96.1	 267.5	 100	RX	(mm)	 -63.8	 0	 -40.5	 83.1	 -35.8	 100	NICmax	(m2/s2)	 9.8	 0	 6.0	 66.1	 4.0	 100	Nij	 0.16	 0	 0.08	 109.6	 0.09	 100	Nkm	 0.60	 0	 0.32	 86.9	 0.28	 100					61	3.4 Discussion	The	results	of	this	study	showed	that	actively	controlling	seat	hinge	rotation	before	and	during	a	collision	pulse	consistently	and	significantly	reduced	the	peak	amplitude	of	the	ATD’s	responses	during	rear-end	impacts	for	speed	changes	between	4	and	12	km/h.		In	comparison	to	the	Control	seat,	the	Experimental	seat	reduced	peak	ATD	kinematics	(linear	accelerations,	angular	velocities,	displacements	and	head	angles)	by	24	–	75%,	peak	ATD	kinetics	(forces	and	moments)	by	43	–	76%,	and	neck	injury	criteria	(NICmax,	Nij,	Nkm)	by	23	–	85%	for	speed	changes	from	4	to	12	km/h.	For	the	2	km/h	speed	change,	only	5	out	of	12	peak	ATD	responses	(FZ,	MY,	θhead,	 RX,	 and	 NICmax)	 were	 significantly	 different,	 with	 peak	 FZ	 and	 MY	 larger	 on	 the	Experimental	 seat	 than	on	 the	Control	 seat.	 	 This	equivocal	outcome	at	 a	 speed	 change	of	2	km/h	was	judged	to	be	inconsequential	because	field	data	have	not	shown	whiplash	injuries	at	this	speed	change	(Krafft,	2002,	Krafft	et	al.,	2005).	For	all	other	speed	changes	tested	here	(4	–	12	 km/h),	 with	 the	 exception	 of	 Δthead-contact	 (all	 collision	 severities)	 and	 Nij	 (only	 4	 km/h	severity),	peak	ATD	responses	 for	 the	Experimental	 seat	were	significantly	 lower	 than	 for	 the	Control	seat.			The	observed	dynamics	of	the	seat	hinge	rotation	(Figure	3.2,	red	line)	consisted	of	two	phases:	a	pre-perturbation	forward	rotation	(θforward-peak	=	-3.6	deg,	tforward-onset	=	-90	ms)	and	a	within-perturbation	rearward	rotation	(θrearward-peak	=	3.7	deg,	trearward-onset	=	40	ms).	During	the	pilot	 tests	 used	 to	 develop	 the	 seat	 hinge	 profile	 used	 here	 (see	 Appendix	 B),	 the	 pre-perturbation	 forward	 rotation	 was	 the	 primary	 contributor	 to	 the	 overall	 reduction	 of	 ATD					62	responses.	 	At	 the	12	km/h	speed	change	used	 for	 the	pilot	 tests,	 the	ATD	responses	 for	 the	Experimental	seat	were	11	to	53%	lower	than	the	Control	seat	when	only	the	pre-perturbation	forward-rotation	profile	was	used.	When	only	the	rearward-rotation	profile	was	used,	the	ATD	responses	for	the	Experimental	seat	varied	from	19%	higher	to	10%	lower	than	for	the	Control	seat.	The	combination	of	the	two	rotation	phases	was	similar	to	the	pre-perturbation	forward-rotation-only	profile	and	generated	reductions	in	the	ATD	responses	between	23	–	48%.		One	possible	 explanation	 for	 the	 larger	 effect	 of	 the	 forward	 rotation	may	 be	 an	 increase	 in	 the	effective	 stiffness	 of	 the	 seat	 hinge	 to	 improve	 its	 ability	 to	 resist	 the	 rearward	 seatback	rotation	created	by	the	ATD’s	inertia	as	the	sled	is	accelerated	forward.	Actively	increasing	the	effective	seat	hinge	stiffness	during	the	pre-perturbation	forward	rotation	followed	by	actively	decreasing	 the	effective	seat	hinge	stiffness	during	 the	within-perturbation	rearward	rotation	appeared	 to	 generate	 three	 effects:	 i)	 a	 prolonged	 collision	 pulse,	 ii)	 longer	 head-to-head-restraint	 interaction	 at	 higher	 speed	 changes,	 and	 iii)	 a	 reduced	 forward	 rebound.	 All	 three	effects	appeared	to	contribute	to	attenuating	the	peak	kinematics	and	kinetics	experienced	by	the	ATD,	particularly	at	higher	speed	changes.		The	Experimental	anti-whiplash	seat	was	designed	to	be	rated	good	(dbackset	<	70	mm	and	dheight	>	 -60	mm)	according	 to	 the	RCAR/IIWPG	seat/head	restraint	criteria.	 	The	effect	of	 the	differences	in	backset	and	head	restraint	height	between	the	poor-rated	Control	seat	and	the	good-rated	Experimental	 seat	were	 isolated	 from	 the	effect	of	 the	active	 seat	hinge	 rotation	profile	 by	 adding	 the	 EXP90	 condition	 in	which	 the	 head	 restraint	 backset	 and	 height	 of	 the					63	Experimental	seat	better	matched	that	present	on	the	Control	seat.	Even	with	an	initial	dbackset	and	dheight	similar	to	the	Control	seat,	the	Experimental	seat	reduced	peak	ATD	kinematics	and	kinetics	 by	 an	 average	 of	 85%	 compared	 to	 the	 Control	 seat.	 These	 data	 suggested	 that	 the	differences	in	initial	dbackset	and	dheight	was	a	relatively	small	confounding	factor	in	the	reduction	of	kinematic	and	kinetic	responses	observed	with	the	Experimental	seat.	Even	though	both	the	Control	seat	and	EXP90	condition	had	similar	initial	head	restraint	geometries,	 the	 pre-perturbation	 forward	 rotation	 of	 the	 seat	 hinge	 in	 the	 EXP90	 condition	brought	the	head	restraint	closer	to	the	back	of	the	head	by	the	time	the	collision	pulse	began	(Figure	 3.6).	 Computational	 simulations	 of	 rear-end	 collisions	 have	 shown	 that	 peak	 NICmax	values	decrease	by	approximately	5	m2/s2	for	every	30	mm	decrease	in	backset	and	injury	risk	decreases	by	10%	for	every		25	mm	decrease	in	backset	(Eriksson,	2005).	The	pre-perturbation	forward	 seat	 hinge	 rotation	 decreased	 the	 head-to-head-restraint	 distance	 in	 the	 EXP90	condition	 from	 86.2	 mm	 to	 36.3	 mm	 at	 collision	 onset	 (t	 =	 0	 ms)	 (Figure	 3.6).	 Thus,	 the	observed	49.9	mm	decrease	in	the	EXP90	seat’s	head-to-head-restraint	distance	prior	to	impact	may	have	 reduced	peak	NICmax	by	approximately	8	m2/s2	and	 the	 risk	of	whiplash	 injuries	by	20%.	In	comparison	to	the	Control	seat,	the	total	difference	in	head-to-head-restraint	distance	at	 the	 onset	 of	 the	 collision	 pulse	 for	 the	 EXP90	 condition	 was	 58.8	 mm	 (95.1	 mm	 for	 the	Control	seat	and	36.3	mm	for	 the	EXP90	condition	at	 the	onset	of	 the	collision	pulse).	NICmax	was	reduced	by	3.8	m2/s2	(Control:	NICmax	=	9.8	m2/s2	and	EXP90:	NICmax	=	6.0	m2/s2),	which	was	less	than	half	of	the	reduction	predicted	(Eriksson,	2005)	for	a	58.8	mm	reduction	in	head-to-				64	head-restraint	distance.	This	discrepancy	in	the	predicted	vs	observed	reductions	may	be	due	to	the	smaller	NICmax	values	resulting	from	the	lower	collision	severities	used	in	the	present	study	(∆v	=	2	–	12	km/h,	aX-sled	=	5.7	–	31.6	m/s2)	compared	to	the	computational	study	(∆v	=	12	–	33	km/h,	aX-sled	=	76.5	–	212.9	m/s2).		Regardless,	the	pre-perturbation	forward	seat	hinge	rotation	reduces	 the	 head-to-head-restraint	 distance	 prior	 to	 the	 collision	 onset	 to	 decrease	 ATD	responses	and	can,	potentially,	reduce	the	risk	of	whiplash	injuries	by	20%.			In	 comparison	 to	 current	 anti-whiplash	 seats,	 the	 motion	 of	 the	 Experimental	 seat	combines	features	from	both	the	SAHR	and	WHIPS	seats.	The	SAHR	anti-whiplash	system	uses	the	occupant’s	penetration	into	the	seatback	cushion	to	activate	a	mechanism	that	moves	the	head	 restraint	 upwards	 and	 forwards	 towards	 the	 back	 of	 the	 occupant’s	 head	 (Viano	 and	Olsen,	2001).	 In	 contrast,	 the	WHIPS	 system	uses	a	deformable	element	 in	 the	 seat	hinge	 to	control	rearward	translation	and	rotation	of	the	seatback	(Jakobsson	et	al.,	2008,	Jakobsson	et	al.,	 2000).	 The	 pre-perturbation	 forward	 rotation	 of	 the	 Experimental	 seat	 moves	 the	 head	restraint	closer	to	the	head,	similar	to	the	SAHR	mechanism.		The	subsequent	rearward	rotation	of	the	seatback	reduces	the	effective	rearward	stiffness	of	the	Experimental	seat	and	reduces	forward	 rebound,	 which	 is	 not	 unlike	 the	 net	 effect	 of	 the	 yielding	 element	 in	 the	 WHIPS	system.	In	contrast	to	these	two	prior	designs,	however,	the	Experimental	seat	is	active	rather	than	reactive	and	begins	responding	to	the	crash	before	the	crash	actually	occurs.	These	active	(and	 therefore	programmable)	 features	potentially	allow	 the	 seat	hinge	behavior	 to	adapt	 to					65	occupant	 characteristics	 and	 predictive	 estimates	 of	 the	 collision	 severity—two	 areas	 that	require	further	research.	The	net	effect	that	the	predictive	anti-whiplash	seat	would	have	on	the	real-world	risk	of	whiplash	injury	remains	unknown.	However,	the	SAHR	and	WHIPS	seats,	which	were	previously	shown	to	attenuate	only	four	ATD	responses	in	comparison	to	the	Control	seat	used	here	(see	Chapter	2),	have	been	shown	to	reduce	the	risk	of	whiplash	injuries	by	20	–	75%	(Farmer	et	al.,	2003,	 Ivancic,	 2011,	 Jakobsson	 et	 al.,	 2008,	 Viano	 and	 Olsen,	 2001).	 The	 results	 from	 the	present	 study	 show	 reductions	 in	 all	 measured	 ATD	 kinematic	 and	 kinetic	 responses	 and	suggests	that	the	Experimental	seat	could	further	reduce	the	risk	of	whiplash	injuries.	A	direct	comparison	 the	 magnitude	 of	 these	 kinematic	 and	 kinetic	 reductions	 between	 the	 anti-whiplash	seat	and	SAHR	and	WHIPS	seats	will	be	performed	in	a	subsequent	study	(see	Chapter	5).		The	Experimental	and	Control	seats	generated	ATD	responses	that	had	mostly	excellent	(COV	≤	5%)	and	acceptable	 (5%	<	COV	≤	10%)	 repeatability	at	8	and	12	km/h	speed	changes	(the	 two	 speed	changes	at	which	 the	 repeatability	of	 the	 seat	and	ATD	was	evaluated).	Only	two	ATD	 response	 parameters	 (NICmax	 and	Nij)	 had	 COV	 values	 that	were	 rated	 poor	 (COV	 >	10%).	 Despite	 these	 similarities,	 post-hoc	 comparisons	 of	 the	 COV	 values	 showed	 that	 they	were	signficantly	higher	for	the	Experimental	seat	than	for	the	Control	seat	(paired	t-tests,	p	=	0.020	and	p	=	0.008	for	the	8	and	12	km/h	speed	changes,	respectively).	A	similar	comparison	of	 the	 SD	 values,	 however,	 showed	 no	 signficant	 differences	 between	 the	 Experimental	 and					66	Control	 seats	 (paired	 t-tests,	 p	 =	 0.56	 and	 p	 =	 0.53	 	 for	 the	 8	 and	 12	 km/h	 speed	 changes,	respectively).	Based	on	this	follow-up	analysis,	the	larger	COV	values	for	the	Experimental	seat	were	likely	due	to	lower	mean	values	(which	form	the	denominator	in	the	COV	calculation)	than	due	to	greater	inherent	variability	in	the	ATD	responses.	Thus	we	believe	that	the	Control	and	Experimental	seats	have	sufficient	repeatability	to	justify	the	comparisons	made	in	this	study.	The	 speed	 changes	used	 in	 this	 study	were	 lower	 than	 the	16	 km/h	 speed	 change	 and	longer	 than	 the	 91	 ms	 collision	 pulse	 duration	 used	 to	 rate	 seat	 performance	 by	 the	RCAR/IIWPG	 (Insurance	 Institute	 of	 Highway	 Safety,	 2008b,	 Insurance	 Institute	 of	 Highway	Safety,	 2008a).	Nevertheless,	 the	pulses	used	here	provide	a	 graded	 range	of	 speed	 changes	that	represent	some	real-world	collisions	that	cause	whiplash	injury	(Bartsch	et	al.,	2008,	Krafft	et	al.,	2005).	The	weight	of	the	Experimental	seat,	especially	the	seat	hinge	motors	and	drives,	limited	 the	peak	accelerations	 that	was	achievable	with	 the	 test	 sled.	 These	 large	 seat	hinge	motors	were	selected	to	allow	for	testing	a	wide	range	of	seat	hinge	rotation	parameters	and	to	explore	different	 seat	 hinge	 rotation	profiles.	 	 A	 collision	pulse	 duration	of	Dt	 =	 148	ms	was	selected	for	collision	speeds	up	to	8	km/h	because	it	was	achievable	and	not	dissimilar	to	the	135	ms	duration	observed	in	prior	vehicle-to-vehicle	rear-end	crashes	(Brault	et	al.,	2000,	Brault	et	al.,	1998,	Siegmund	et	al.,	2000,	Siegmund	et	al.,	1997).		However,	a	collision	pulse	duration	of	Dt	=	185	ms	was	needed	to	reach	a	12	km/h	collision	speed.	This	 longer	duration	is	 longer	than	 those	 of	 many	modern	 vehicles	 (Linder	 et	 al.,	 2003,	 Linder	 et	 al.,	 2001,	 Stigson	 et	 al.,	2006),	but	shorter	than	those	observed	in	some	older	vehicles	with	bumper	isolators	(Siegmund					67	et	 al.,	 1994).	 	 Nevertheless,	 further	 work	 is	 needed	 to	 explore	 the	 ATD	 responses	 on	 the	Experimental	anti-whiplash	seat	at	higher	speed	changes	and	shorter	collisions	pulses.		The	BioRID	ATD	used	for	this	study	represented	a	50th	percentile	male,	and	further	work	is	needed	to	evaluate	the	Experimental	seat	for	male	and	female	occupants	of	different	heights	and	weights.	Although	the	BioRID	ATD	was	designed	to	mimic	the	motion	of	human	occupants,	including	 their	 active	muscle	 response,	 further	work	 is	 needed	 to	 evaluate	whether	 the	 pre-perturbation	forward	rotation	of	the	seat	hinge	alters	the	neck	muscle	response	in	humans	and	whether	the	BioRID	ATD	remains	a	valid	surrogate	for	human	occupants	under	these	pre-crash	conditions.	 Further	 work	 is	 also	 needed	 on	 how	 to	 affordably	 and	 efficiently	 implement	 an	active	seat	that	behaves	like	the	large,	heavy	and	expensive	prototype	seat	used	for	this	study.	Despite	these	 limitations,	 the	results	of	this	study	provided	a	better	understanding	of	how	to	actively	control	seat	hinge	rotation	to	minimize	ATD	responses	during	a	rear-end	collision.		3.5 Conclusion	In	 this	 study,	we	 investigated	whether	 a	 novel	 anti-whiplash	 seat	 that	 actively	 controls	seat	 hinge	 rotation	 could	 reduce	 peak	ATD	 responses	 during	 a	 series	 of	 low-speed,	 rear-end	collisions	 of	 varying	 severity	 (Δv	 =	 2,	 4,	 6,	 8	 and	 12	 km/h).	 The	 results	 showed	 that	 in	comparison	to	a	Control	seat	modifying	the	seat	hinge	rotation	could	reduce	all	ATD	responses	by	 24%	 -	 76%	 (aX-head,	 aX-T1,	 ωY-head,	 FX,	 FZ,	MY,	 θhead,	 RX,	 and	 drebound)	 and	 neck	 injury	 criteria	(NICmax,	 Nkm,	 and	 Nij)	 by	 23	 –	 85%	 for	 speed	 changes	 greater	 than	 4	 km/h.	 Based	 on	 these					68	results,	active	 rotation	of	 the	 seat	hinge	 is	a	 repeatable	method	of	decreasing	head-to-head-restraint	distance,	reducing	peak	ATD	responses,	and	thus	could	potentially	reduce	the	risk	of	whiplash	injury	following	low-speed,	rear-end	collisions.		 					69	Chapter	4. Effects	 of	 Seatback	 Cushion	 Deformation	 on	 ATD	Responses	4.1 Introduction	The	 design	 of	 the	 automotive	 seat	 is	 a	 key	 parameter	 in	 preventing	 whiplash	 injuries	following	 rear-end,	 vehicle-to-vehicle	 collisions	 (Hofinger	 et	 al.,	 1999,	 Jakobsson	 et	 al.,	 2008,	Stemper	 et	 al.,	 2006,	 Viano,	 2003d,	 Viano,	 2008,	 Svensson	 et	 al.,	 1996).	 Seatback	 cushion	properties	 are	 not	 only	 designed	 for	 comfort,	 but	 more	 importantly	 determine	 the	 energy	absorption	and	occupant	retention	during	a	collision	(Viano,	2003a,	Viano	and	Parenteau,	2015,	Hofinger	 et	 al.,	 1999).	Most	 current	 automotive	 seatbacks	 are	 comprised	 of	 an	 outer	 frame	with	a	compliant	spring	center	that	is	wrapped	in	cushioning	foam	and	a	seat	cover.	Hofinger	et	al.	 (1999)	 tested	 different	 types	 of	 foam	 (soft	 and	 hard)	 on	 a	 custom-designed	 rigid	 seat	 to	determine	 the	 influence	 of	 seatback	 cushioning	 on	 anthropomorphic	 test	 device	 (ATD)	responses	at	speed	changes	(∆v)	of	9.6	and	14.2	km/h.	The	seatback	was	divided	into	upper	and	lower	halves	to	study	the	effects	of	different	soft	and	hard	foam	combinations.	The	softer	foam	across	 the	 entire	 seatback	 delayed	 the	 onset	 of	 torso	 acceleration,	 increased	 occupant	penetration	 into	 the	 seatback	 and	 increased	 peak	 linear	 torso	 and	 pelvis	 acceleration	magnitudes.	 In	 contrast,	 the	 harder	 foam	minimized	 occupant	 penetration	 into	 the	 seatback	and	distributed	the	pressure	more	evenly	over	the	ATD’s	back	to	decrease	peak	linear	torso	and	pelvis	acceleration	magnitudes.	Hofinger	et	al.	(1999)	suggested	that	the	combination	of	stiffer	foam	at	the	pelvis	and	softer	foam	at	the	torso	best	reduced	occupant	acceleration	by	initiating					70	earlier	torso	and	pelvis	rotation	and	decreasing	the	backset,	whereas	the	combination	of	stiffer	foam	 at	 the	 torso	 and	 softer	 foam	 at	 the	 pelvis	 increased	 head	 accelerations.	 Thus,	 careful	dynamic	 modifications	 of	 seatback	 cushion	 properties	 during	 the	 collision	 may	 represent	 a	viable	option	to	reduce	ATD	responses	and,	consequently,	the	risk	of	whiplash	injuries.		Modification	of	the	seatback	cushion,	however,	may	also	alter	pre-impact	seat	geometry	such	as	 the	horizontal	distance	between	 the	back	of	 the	ATD	head	and	 the	 front	 face	of	 the	head	restraint	 (backset;	dbackset)	as	well	as	 the	vertical	distance	been	the	top	of	 the	head	and	the	 top	 of	 the	 head	 restraint	 (head	 restraint	 height;	 dheight).	 Increasing	 backset	 and	 head	restraint	 height	 has	 been	 shown	 to	 increase	 the	 risk	 of	 whiplash	 injuries	 following	 rear-end	collisions	 (Nygren	 et	 al.,	 1985,	 Siegmund	 et	 al.,	 1999,	 Stemper	 et	 al.,	 2006,	 Eriksson,	 2005).	Specifically,	the	risk	of	neck	injury	was	found	to	be	lowest	when	the	backset	was	zero	and	the	top	 of	 the	 head	 restraint	 was	 level	 with	 the	 top	 of	 the	 occupant’s	 head	 (Eriksson,	 2005).		Eriksson	also	found	that	changes	to	backset	had	a	larger	effect	on	the	risk	of	neck	injury	than	head	restraint	height:	for	every	25	mm	decrease	in	backset,	the	risk	of	neck	injury	decreased	by	10%.	 Thus,	 any	 difference	 in	 pre-impact	 seat	 geometry	 (backset	 and	 head	 restraint	 height)	caused	 by	 the	 dynamic	 modification	 in	 seatback	 cushion	 properties	 may	 influence	 and	potentially	confound	the	reduction	peak	ATD	responses	during	rear-end	collisions.		We	developed	an	Experimental	anti-whiplash	seat	to	control	seatback	deformation	using	motor-driven	 seatbelt	 straps	 that	 translate	 a	 suspended	 seatback	 support	 frame	 with	 fixed	foam	 properties	 (Appendix	 A).	 The	 seatback	 deformation	 profile	 is	 induced	 by	 releasing	 the					71	motor-driven	 seatbelt	 straps	 prior	 to	 the	 collision,	 increasing	 the	 ATD	 penetration	 into	 the	seatback	during	the	collision.	Preliminary	experiments	to	determine	the	seatback	deformation	profile	allowing	the	ATD	to	pocket	into	the	seatback	were	only	conducted	at	one	collision	speed	(Δv:	 12	 km/h,	 Δt:	 195	 ms;	 see	 Appendix	 C).	 The	 primary	 goal	 of	 the	 present	 study	 was	 to	compare	 the	 kinematic	 and	 kinetic	 responses	 of	 a	 BioRID	 II	 ATD	 seated	 on	 the	 Experimental	anti-whiplash	seat	with	the	dynamic	modulation	of	the	seatback	cushion	(Experimental	seat)	to	the	ATD	responses	seated	on	an	unmodified	automotive	seat	(Control	seat)	during	a	range	of	low-speed,	 rear-end	perturbations	 (speed	changes	of	2	 to	12	km/h).	A	 secondary	goal	of	 the	study	was	 to	quantify	 the	 repeatability	of	 the	 seat	 and	ATD	 responses.	 In	 comparison	 to	 the	Control	seat,	we	hypothesized	that	the	Experimental	anti-whiplash	seat	would	reduce	most	of	the	peak	ATD	responses	across	all	tested	collision	severities.	4.2 Methods	4.2.1 Experimental	Anti-Whiplash	Seat	The	Experimental	anti-whiplash	seat	design	consisted	of	a	modified	General	Motor’s	High	Retention	 seat	 (GMHR)	 and	 three	 motors	 mounted	 to	 a	 modified	 seat	 frame	 to	 control	seatback	cushion	deformation	(Figure	4.1).		The	seat	pan	and	head	restraint	of	the	GMHR	seat	remained	 unmodified	 in	 this	 design.	 The	 seatback	 was	 replaced	 by	 a	 rigid	 aluminum	 outer	frame	with	four	straps	made	of	47	mm	seatbelt	webbing	spanning	between	the	sides	of	the						72		Figure	4.1.	Photograph	of	the	experimental	set-up	with	the	BioRID	II	ATD	on	the	Experimental	anti-whiplash	automotive	seat	and	rigid	metal	braces	to	prevent	seat	hinge	rotation.	The	global	laboratory	 reference	 frame	 is	 illustrated	 with	 positive	 X-axis	 forward	 and	 positive	 Z-axis	downwards.	The	seat	hinge	and	seatback	motors	on	the	left	side	of	the	Experimental	seat	are	labelled.	Additional	 seat	hinge	and	 seatback	motors	are	 located	on	 the	 right	 side	of	 the	 seat	(not	labelled).		The	seat	hinge	motors	are	not	used	in	this	current	study.		outer	frame	at	the	upper,	middle	and	lower	back	locations	(upper:	1	strap,	middle:	2	straps	and	lower:	 1	 strap).	 The	 seatback	 cushion	 from	 the	GMHR	 seatback	was	 suspended	on	 the	 front	surface	of	 the	 four	straps.	Seatback	cushion	deformations	were	controlled	by	modulating	the	amount	of	webbing	using	three	rotational	servomotors	 (AKM24D,	Kollmorgen,	Waltham,	MA,	USA)	connected	to	helical	right-angle	gearheads	(VTR006-008,	8:1	gear	ratio,	Thomson	Linear,	Radford,	VA,	USA)	mounted	 staggered	 to	both	 sides	of	 the	 rigid	 seatback.	 	 The	 three	motors					73	were	 placed	 at	 the	 top,	middle	 and	 bottom	 of	 the	 seatback	 and	 attached	 to	 the	 4	 seatbelt	straps	 (with	 the	 middle	 motor	 connected	 to	 the	 2	 middle	 straps)	 that	 spanned	 across	 the	perimeter	frame	to	separately	control	the	seatback	deformation	at	the	upper-torso,	mid-torso	and	lower	pelvis	regions	(Figure	4.2).	Tightening	the	webbing	would	pull	the	suspended	GMHR	seatback	cushion	towards	the	front	of	the	seat	(+X	direction)	and	increase	the	stiffness	of	the	seatback;	 whereas	 loosening	 the	 webbing	 would	 allow	 the	 seatback	 cushion	 to	 translate	towards	the	rear	of	the	seat	(-X	direction)	and	allow	for	deeper	occupant	penetration	into	the	seatback	during	rear-end	collisions.	Slack	was	introduced	into	all	the	webbing	straps	to	set	the	initial	seatback	angle	to	27	deg	rearward	from	vertical	(Siegmund	et	al.,	2005a).	To	isolate	the	effects	of	seatback	cushion	deformation,	seat	hinge	rotation	was	prevented	by	 installing	rigid	metal	braces	between	the	seat	outer	frame	and	the	sled.					Through	 a	 series	 of	 iterative	 experiments	 (see	 Appendix	 C),	 different	 seatback	 motor	parameters	were	 varied	 and	 tested	 to	 create	 a	 seatback	motor	 rotation	profile	 that	 reduced	ATD	responses	during	a	12	km/h	speed	change	(Figure	4.3).	The	seatback	rotation	profile	was	characterized	by	an	input	rotation	to	the	two	lower	motors	with	a	peak	angle	(θdeformation-peak)	of	168	deg	and	an	initial	angular	velocity	of	(ωdeformation-int)	980	deg/s.	The	rotational	 input	to	the	two	lower	motors	introduced	slack	in	the	webbing	and	ultimately	linear	rearward	translation	of	the	 ATD	 with	 respect	 to	 the	 seatback.	 The	 upper-torso	 motor	 was	 locked	 in	 a	 stationary	position,	effectively	creating	a	hinge	point	at	 the	upper	back	and	pocketing	of	 the	pelvis	 into	the	seat	during	the	collisions.	The	onset	of	the	seatback	deformation	profile	occurred	200	ms					74	prior	 to	 the	 collision	 (tdeformation-onset	 =	 -200	 ms).	 A	 motion	 capture	 system	 (Optotrak	 Certus,	Northern	Digital,	Waterloo,	ON,	Canada)	was	used	to	estimate	the	deformation	of	the	seatback	cushion	 during	 low-speed,	 rear-end	 collisions.	 	 Maximum	 rearward	 penetration	 of	 the	 T1	vertebra	 (dpenetration-T1)	 and	 pelvis	 (dpenetration-pelvis)	were	 used	 to	 describe	 the	 seatback	 cushion	deformation	and	were	defined	as	the	horizontal	displacement	 (along	the	global	X-axis,	Figure	4.1)	between	the	 infrared	 light	emitting	diode(IRED)	markers	mounted	to	the	seatback	 frame	and	the	T1	vertebra	and	pelvis,	respectively	(Figure	4.3).	The	initial	position	of	the	ATD	at	rest	was	set	as	 zero	 (dpenetration-T1	=	0	mm	and	dpenetration-pelvis	=	0	mm)	and	rearward	displacements	into	 the	 seatback	were	defined	as	negative	 values.	 For	 a	12	 km/h	 collision,	 the	programmed	seatback	rotation	profile	allowed	maximum	occupant	penetration	into	the	seatback	of	84	mm	at	T1	and	97	mm	at	the	pelvis	(Figure	4.3,	see	Appendix	C).		All	three	seatback	motors	were	controlled	independently	by	separate	digital	servo	drives	(Servostar	600,	Kollmorgen,	Waltham,	MA,	USA)	connected	to	a	universal	motion	interface	and	a	motion	controller	 (NI	UMI	7774	and	NI	PXI	7350,	National	 Instruments	Corporation,	Austin,	Texas,	 USA).	 	 A	 custom	 LabVIEW	 program	 (National	 Instruments	 Corporation,	 Austin,	 Texas,	USA)	was	created	to	send	commands	to,	monitor	the	status	of	and	record	encoder	data	from	the	three	motors.						75		Figure	4.2.	Computer-aided	design	(CAD)	drawings	of	Experimental	seat	with	three	seatback	motors	placed	at	the	top,	middle	 and	 bottom	of	 the	 seatback,	 and	 the	 4	 seatbelt	 straps	 that	 spanned	 across	 the	 perimeter	 frame	 to	separately	 control	 the	 seatback	deformation	at	 the	upper-torso,	mid-torso	and	 lower	pelvis	 regions.	 	The	middle	motor	was	connected	to	the	2	middle	straps.	Panel	A.,	B.	&	C.	show	the	right,	left	and	frontal	views,	respectively.	The	distances	between	the	center	of	the	seatbelts	and	the	seat	hinge	are	illustrated	in	Panel	D.									76		Figure	 4.3.	 Estimated	 horizontal	 seatback	 cushion	 deformation	 using	 displacement	 of	 the	 T1	vertebra	 and	 pelvis	 (dpenetration-T1,	 red	 plot	 &	 dpenetration-pelvis,	 blue	 plot)	 in	 response	 to	 the	seatback	 motor	 rotation	 profile	 (θseatback-motor,	 grey	 plot)	 with	 θdeformation-peak	 =	 168	 deg,	ωdeformation-int	=	980	deg/s,	and	tdeformation-onset	=	-200	ms	during	a	12	km/h	collision	pulse	with	a	collision	pulse	duration	of	194	ms	(aX-sled,	black	plot).		4.2.2 Anthropomorphic	Test	Device	and	Instrumentation	A	 BioRID	 II	 ATD	 (Humanetics,	 Plymouth,	MI,	 USA)	was	 instrumented	 to	measure	 head,	chest	(T1),	and	pelvis	kinematics	and	kinetics	(Figure	4.1).	A	six-axis	load	cell	(Forces:	FX,	FY,	FZ;	and	Moments:	MX,	MY,	MZ;	Model	4949a,	Robert	A.	Denton,	Inc.,	Rochester	Hills,	MI,	USA)	was	mounted	 at	 the	 atlanto-occipital	 joint	 (AOJ)	 to	 measure	 upper	 neck	 forces	 and	 moments.		Linear	 forward	 accelerations	 of	 the	 head	 and	 T1	 were	 measured	 using	 two	 uni-axial					77	accelerometers	 (7264C	 sensors;	 ±500	 g,	 Endevco,	 San	 Juan	Capistrano,	 CA,	USA)	mounted	 at	both	 the	 head	 center	 of	mass	 and	 the	 T1	 vertebra.	 A	 uni-axial	 angular	 rate	 sensor	was	 also	mounted	 at	 the	 head	 center	 of	 mass	 (ARS-1500;	 ±26.2	 rad/s,	 DTS,	 Seal	 Beach,	 CA,	 USA)	 to	measure	angular	kinematics	about	the	sagittal	plane	(i.e.	flexion	and	extension	of	the	neck).			A	motion	capture	system	(Optotrak	Certus,	Northern	Digital,	Waterloo,	ON,	Canada)	was	used	to	track	IRED	markers	affixed	to	the	head,	AOJ,	T1,	and	pelvis	of	the	ATD	to	measure	head,	T1	and	pelvis	displacements.		Additional	IRED	markers	were	mounted	to	the	Experimental	anti-whiplash	 seat	 to	 record	 displacements	 and	 rotations	 of	 the	 seatback	 as	 well	 as	 to	 create	 a	global	 experimental	 reference	 frame	 (+X	 forward,	 +Y	 right,	 and	 +Z	 down).	 	 Horizontal	 sled	acceleration	was	measured	with	 a	 uni-axial	 accelerometer	 (2220-100;	±100	 g,	 Silicon	 Design	Inc.,	Issaquah,	WA,	USA)	mounted	directly	to	the	base	of	the	linear	sled	frame.		Head	restraint	contact	 was	 detected	 with	 a	 force	 sensitive	 resistor	 (FSR;	 Model	 406,	 Interlink	 Electronics,	Camarillo,	CA,	USA)	attached	to	the	front	of	the	head	restraint.		 All	 accelerometer,	 load	 cell	 and	 angular	 rate	 sensor	 signals	 were	 simultaneously	sampled	at	10	kHz	using	a	National	Instruments	Data	Acquisition	(DAQ)	PXI	system	(PXI-4495	&	PXI-6289,	National	Instruments	Corporation,	Austin,	Texas,	USA)	and	a	custom-written	LabVIEW	virtual	instrument	(National	Instruments	Corporation,	Austin,	Texas,	USA).	Optotrak	data	were	acquired	at	200	Hz	and	collection	was	 triggered	by	 the	DAQ	system	 to	 synchronize	 the	data.	Subsequent	data	 and	 statistical	 analyses	were	performed	using	Matlab	 (R2017A,	Mathworks,	Natick,	 MA,	 USA).	 All	 data	 were	 digitally	 filtered	 in	Matlab	 and	 conformed	 to	 the	 SAE	 J211					78	(Channel	class	180	for	the	ATD	sensors	and	Channel	Class	60	for	the	sled	accelerometer;	SAE,	1995).	4.2.3 Test	Procedures	The	 ATD	 was	 dressed	 in	 two	 layers	 of	 lycra	 and	 was	 seated	 on	 either	 the	 Control	 or	Experimental	seats	mounted	on	a	10	m	long,	feedback-controlled	linear	sled	(Kollmorgen	IC55-100A7,	 Waltham,	 MA,	 USA),	 as	 illustrated	 in	 Figure	 4.1.	 To	 ensure	 the	 repeatability	 of	 the	experiments,	the	initial	position	of	the	ATD	was	measured	by	Optotrak	and	adjusted	to	a	pre-defined	 posture	 that	 was	 constant	 within	 each	 seat	 and	 as	 similar	 between	 Control	 and	Experimental	seats.	The	seatbelt	was	removed	from	the	seats	to	prevent	interactions	that	could	affect	ATD	head,	neck	and	torso	kinematics.	Initial	pre-impact	seatback	angle	was	set	to	27	deg	rearward	from	vertical	(Siegmund	et	al.,	2005a)	and	the	vertical	placement	of	the	head	restraint	was	adjusted	to	the	midpoint	between	the	lowest	and	highest	vertical	position	according	to	the	Research	 Council	 for	 Automobile	 Repairs/International	 Insurance	Whiplash	 Prevention	Group	(RCAR/IIWPG)	 seat/head	 restraint	 evaluation	protocol	 (Insurance	 Institute	of	Highway	Safety,	2008b).			The	kinematic	and	kinetic	responses	of	the	ATD	seated	on	the	Experimental	seat	with	the	seatback	 deformation	 profile	were	 first	 compared	 to	 the	ATD	 responses	 seat	 on	 the	 Control	seat.	 For	 each	 seat,	 the	 ATD	 was	 exposed	 to	 a	 series	 of	 five	 rear-end	 whiplash-like	perturbations	at	increasing	speed	changes	(Δv	=	2,	4,	6,	and	8	km/h	with	a	Δt	of	148	ms	and	12					79	km/h	with	a	Δt	of	194	ms)	(Figure	4.4).	Due	to	limitations	in	the	force-generating	ability	of	the	linear	motors,	 a	 longer	pulse	duration	was	needed	 for	 the	12	 km/h	 speed	 change.	All	 of	 the	pulses	had	a	 longer	duration	 than	 the	RCAR/IIWPG	pulse	 (Δt	=	91	ms)	 (Insurance	 Institute	of	Highway	Safety,	2008b).	The	sled	was	accelerated	forward	from	a	stationary	position	for	the	2,	4	 and	 6	 km/h	 speed	 changes	 and	 from	 a	 constant	 rearward	 speed	 of	 6	 km/h	 for	 the	 higher	speed	 changes	 (8	 and	 12	 km/h).	 Preliminary	 results	 comparing	 occupant	 responses	 of	 an	 8	km/h	collision	starting	from	a	stationary	position	to	an	8	km/h	collision	starting	from	a	6	km/h	rearward	velocity	showed	less	than	10%	differences	(mean:	2.50%	±	2.53%;	range:	0.5%	–	9.2%)	in	 most	 of	 the	 occupant	 responses	 and	 neck	 injury	 criterions	 between	 the	 two	 collision	perturbations	(see	Chapter	2).		To	assess	the	repeatability	of	the	ATD	responses,	four	additional	repeated	 trials	 were	 collected	 at	 the	 8	 and	 12	 km/h	 collision	 speed	 changes	 for	 both	 the	Control	and	Experimental	seats	(for	a	total	of	5	trials	for	each	seat	at	these	two	speed	changes).		The	Experimental	anti-whiplash	seat	was	designed	to	receive	a	good-rating	(dbackset	<	70	mm	and	dheight	>	 -60	mm)	whereas	 the	Control	GMHR	seat	 received	a	poor-rating	due	 to	 the	backset	being	greater	than	90	mm	(Insurance	Institute	of	Highway	Safety,	2004).	Consequently,	the	static	pre-impact	position	of	the	head	restraint	on	the	Control	seat	(dbackset	=	94.4	±	2.6	mm,	dheight	 =	 -24.2	 ±	 2.6	mm)	was	 different	 from	 the	 Experimental	 seat	 (dbackset	 =	 52.0	 ±	 7.3	mm,	dheight	 =	2.5	±	7.1	mm).	 	 	 To	determine	whether	differences	 in	pre-impact	backset	potentially	confounded	 the	 results,	 the	 kinematic	 and	 kinetic	 responses	 of	 the	 ATD	 observed	 on	 the	Control	and	Experimental	seats	were	compared	to	a	No	Motion	condition	on	the	Experimental					80	Seat.	For	the	No	Motion	condition,	the	seatback	motors	did	not	unspool	and	remained	at	their	pre-impact	position	during	the	12	km/h	collision.	The	pre-impact	backset	and	height	geometry	of	 the	 head	 restraint	 (dbackset	 =	 60.2	 ±	 5.8	 mm,	 dheight	 =	 -0.23	 ±	 9.6	 mm)	 of	 the	 No	Motion	condition	matched	more	closely	to	the	Control	seat	than	on	the	Experimental	seat.	Five	trials	of	the	No	Motion	condition	were	collected	at	the	12	km/h	collision	speed	change	(Δt	=	194	ms).		A	 total	 of	 31	 trials	 were	 collected,	 13	 trials	 for	 the	 Control	 seat,	 5	 for	 the	 No	Motion	condition	 and	 13	 for	 the	 Experimental	 seat.	 	 Averaged	 peak	 responses	 were	 used	 for	 test	conditions	with	repeated	trials	in	subsequent	analysis	(Control	and	Experimental	seats:		Dv	=	8	and	12	km/h	trials,	and	No	Motion	condition:	Dv	=	12	km/h	trials).					Figure	4.4.	Exemplar	sled	a)	velocity	and	b)	acceleration	pulses	 for	 increasing	collision	speeds	(Δv	=	2,	4,	6,	and	8	km/h	with	a	pulse	duration	(Δt)	of	148	ms,	and	Δv	=	12	km/h	with	a	Δt	of	194	ms;	lightest	to	darkest).	Collision	onset	occurred	at	time	=	0	ms.					81	4.2.4 Data	Analysis		 Head	 and	 torso	 accelerometer	 data	 were	 reported	 in	 local	 head	 and	 T1	 reference	frames	and	were	corrected	 to	 remove	 the	earth’s	gravity	using	 the	head	and	T1	orientations	determined	 from	Optotrak	data.	Peak	 linear	 forward	acceleration	of	 the	head	 (aX-head)	and	T1	vertebra	 (aX-T1)	 were	 extracted	 directly	 from	 the	 transformed	 accelerometer	 data.	 Peak	horizontal	sled	acceleration	(aX-sled)	was	extracted	directly	from	the	accelerometer	mounted	to	the	sled	and	peak	rotational	velocity	of	the	head	(ωY-Head)	in	the	sagittal	plane	was	determined	from	 the	 angular	 rate	 sensor.	 	 Peak	 upper	 neck	 shear	 (FX)	 and	 axial	 (FZ)	 forces	 and	 the	flexion/extension	bending	moment	 (MY)	were	determined	 from	 the	upper	neck	 load	 cell	 and	reported	in	the	ATD	reference	frame	as	the	forces/moments	applied	by	the	neck	to	the	head.		Initial	head	angle	was	defined	as	 the	average	angle	between	 -300	ms	and	 -200	ms	preceding	the	onset	of	sled	acceleration	(aX-sled)	and	peak	head	extension	angle	(θhead)	was	defined	as	the	maximum	 rotation	 of	 the	 head	 into	 extension	 relative	 to	 initial	 head	 angle.	 A	 foreperiod	between	 -300	ms	 and	 -200	ms	 immediately	 before	 the	 onset	 of	 forward	 aX-sled	 was	 used	 to	define	 initial	 values	 to	 account	 for	 any	 possible	 interactions	 of	 the	 early	 onset	 of	 seatback	cushion	deformation	 (tpenetration-onset	 =	 -200	ms)	on	ATD	kinematic	 and	kinetic	 responses.	Peak	retraction	 (Rx)	 was	 defined	 as	 the	 maximum	 horizontal	 displacement	 in	 the	 laboratory	reference	 frame	 of	 the	 AOJ	 with	 respect	 to	 the	 T1	 vertebra	 with	 rearward	 displacements	defined	as	negative	values.	The	back	and	top	of	the	ATD	head	as	well	as	front	face	and	top	of	the	 head	 restraint	 were	 digitized	 relative	 to	 existing	 Optotrak	 IRED	markers.	 These	 digitized					82	points	were	used	to	determine,	 in	the	 laboratory	reference	frame,	the	head-to-head-restraint	distance	(dbackset)	and	the	height	of	the	head	restraint	(dheight;	negative	values	indicate	that	the	top	of	the	head	restraint	was	lower	than	the	top	of	the	ATD	head).	Initial	pre-impact	dbackset	and	dheight	 were	 defined	 as	 the	 average	 head-to-head-restraint	 horizontal	 and	 vertical	 distances	between	 -300	 ms	 and	 -200	 ms	 before	 the	 onset	 of	 forward	 aX-sled.	 Peak	 forward	 rebound	distance	 (drebound)	 was	 defined	 as	 the	 maximum	 forward	 head-to-head-restraint	 distance.	Maximum	 rearward	 penetration	 of	 the	 torso	 and	 pelvis	 into	 the	 seatback	 (dpenetration-T1	 and	dpenetration-pelvis)	 were	 defined	 as	 the	 maximum	 horizontal	 displacement	 in	 the	 laboratory	reference	 frame	between	 the	seatback	 frame	and	T1	vertebra	or	pelvis	markers,	 respectively	(rearward	 displacements	 defined	 as	 negative	 values).	 Time-to-head-restraint	 contact	 (Δthead-contact)	was	extracted	from	the	time	of	force	onset	in	the	FSR	attached	to	the	head	restraint	with	respect	 to	 the	onset	of	 forward	 sled	acceleration	 (t	 =	 0	ms).	Onsets	of	 sled	acceleration	and	head	restraint	contact	were	determined	when	the	accelerometer	and	FSR	signals	reached	1.5	times	 the	peak	background	noise	 level	between	 -300	ms	 to	 -200	ms	before	 the	onset	of	 the	collision	and	were	confirmed	visually.		Three	neck	 injury	 criteria	 (NICmax,	Nij,	 and	Nkm)	were	computed	 from	the	accelerometer	and	 load	 cell	 data.	 	 The	 Neck	 Injury	 Criterion	 (NICmax)	 was	 calculated	 from	 the	 relative	horizontal	 acceleration	 (corrected	 for	 gravity)	 and	 velocity	 in	 the	 global	 reference	 frame	between	 the	head	 center	 of	mass	 and	 the	 T1	 joint	 (Equation	 4.1;	 Bostrom	et	 al.,	 1996).	 The	Normalized	 Neck	 Injury	 Criterion	 (Nij)	 was	 calculated	 from	 the	 axial	 load	 (FZ)	 and	 the					83	flexion/extension	bending	moment	(MY)	measured	from	the	upper	neck	load	cell	(Equation	4.2;	critical	Fint	and	Mint	intercept	values	used	for	normalization:	Fint-tension	=	6806	N,	Fint-compression	=	-6160	N,	Mint-flexion	=	310	N,	and	Mint-extension	=	-135	N;	Eppinger	et	al.,	1999,	Eppinger	et	al.,	2000).	The	Neck	Protection	Criterion	 (Nkm)	was	 calculated	 from	 the	 sagittal	 shear	 force	 (FX)	 and	 the	flexion/extension	bending	moment	(MY)	measured	from	the	upper	neck	load	cell	(Equation	4.3;	critical	Fint	and	Mint	intercept	values	used	for	normalization:	Fint-shear-anterior	=	845	N,	Fint-shear-posterior	=	845	N,	Mint-extension	=	47.5	Nm,	and	Mint-flexion	=	88.1	Nm;	Schmitt	et	al.,	2001,	Schmitt	et	al.,	2002).	 	 Peak	 values	 of	 the	 three	 neck	 injury	 criteria	 were	 then	 extracted	 for	 analysis	 and	compared	to	proposed	injury	thresholds	(NIC:	15	m2/s2,	Eichberger	et	al.,	1998;	Nij	and	Nkm	:	a	normalized	value	of	one,	Schmitt	et	al.,	2001,	Schmitt	et	al.,	2002,	Eppinger	et	al.,	1999).		These	injury	thresholds	correspond	to	different	human	tolerance	levels	for	the	causation	of	whiplash-related	injuries	–	NIC:	long-term	whiplash-associated	disorders	(WAD)	levels	1-3	(Bostrom	et	al.,	2000,	Bostrom	et	al.,	1996),	Nij:	22%	risk	of	abbreviated	injury	scale	(AIS)	level	3	(i.e.,	fracture;	Eppinger	et	al.,	1999,	Eppinger	et	al.,	2000),	and	Nkm:	AIS	 level	1	 injury	causation	 (i.e.,	minor	injury;	Schmitt	et	al.,	2001,	Schmitt	et	al.,	2002).	NICmax t = maximumfirst	150	ms #$%& ' ∙ ) + (,$%&(')). 	 …	 Equation	4.1	N01(t) = 23(4)2567 + 89:;(4)8567 	 	 	 	 	 	 …	 Equation	4.2	N<=(t) = 2>(4)2567 + 89:;(4)8567 	 	 	 	 	 …		 Equation	4.3							84	The	 repeatability	 of	 the	 BioRID	 II	 ATD	 on	 the	 Control	 and	 Experimental	 seats	 were	assessed	 using	 the	 Coefficient	 of	 Variation	 (COV),	 which	 was	 calculated	 as	 the	 standard	deviation	 (SD)	 divided	 by	 the	 mean	 and	 expressed	 as	 a	 percentage.	 The	 National	 Highway	Traffic	Safety	Administration	(NHTSA)	defines	a	COV	of	5%	or	less	as	excellent,	a	COV	of	10%	or	less	as	acceptable	and	a	COV	greater	 than	10%	as	poor	 (Rhule	et	al.,	2005).	 	 To	quantify	 the	repeatability	of	 the	ATD,	 the	COV	for	all	of	 the	peak	responses	were	calculated	 from	the	 five	repeated	trials	at	the	8	and	12	km/h	collision	severities	on	the	Control	and	Experimental	seats.	Because	 COV	 is	 dependent	 on	 the	 mean	 and	 SD	 values	 of	 the	 responses,	 the	 expected	decreases	 in	 mean	 responses	 (which	 form	 the	 denominator	 in	 the	 COV	 calculation)	 for	 the	Experimental	seat	may	generate	larger	COV	values.		A	comparison	of	the	SD	values	between	the	GMHR	 and	 the	 Experimental	 seats	 would	 provide	 an	 alternate	 assessment	 of	 variability	 and	would	help	to	determine	whether	the	larger	COV	values	for	the	Experimental	seat	were	likely	due	 to	 the	 lower	 mean	 responses	 or	 due	 to	 the	 greater	 inherent	 variability	 (i.e.,	 standard	deviation)	in	the	ATD	responses.	The	SD	values	for	all	kinematic	and	kinetic	responses	as	well	as	neck	criteria	observed	at	both	the	8	and	12	km/h	collision	speeds	were	used	in	this	assessment	of	 variability.	 The	 COV	 and	 SD	 values	 were	 assumed	 to	 be	 independent	 observations.	 A	Kolmogorov-Smirnov	 test	 determined	 that	 the	 COV	 values	 of	 the	 Control	 and	 Experimental	seats	at	 the	8	km/h	collision	 severity	were	normally	distributed	and	a	paired	Student’s	 t-test	was	performed.		For	all	other	COV	(12	km/h)	and	SD	(8	and	12	km/h)	values	of	the	Control	and	Experimental	 seats,	 the	 Kolmogorov-Smirnov	 test	 determined	 the	 data	 were	 not	 normally					85	distributed	 and	Wilcoxon	 Signed-Rank	 tests	were	 performed.	 All	 tests	were	 performed	 using	predefined	 functions	 (kstest,	 ttest,	 and	 signrank)	 in	 Matlab	 (R2017A,	 Mathworks,	Natick,	MA,	USA)	at	a	significance	level	a	=	0.05.					First,	to	compare	the	ATD	responses	between	the	Control	and	Experimental	seats	across	the	different	speed	changes,	99th	percentile	predictive	intervals	were	created	around	the	peak	responses	for	each	seat	using	their	respective	maximum	variability	(COV	value)	observed	from	repeated	trials	at	either	the	8	or	12	km/h	collision	severity	(Horslen	et	al.,	2015).	The	maximum	COV	 values	 from	 either	 the	 8	 or	 12	 km/h	 collision	 severity	 were	 selected	 to	 represent	 the	largest	observed	variance	within	each	response	and	were	multiplied	by	2.58	(z-score	for	two-tailed	99th	percentile)	 to	estimate	99th	percentile	predictive	 limits.	Since	the	COV	values	were	defined	as	the	maximum	variances	(i.e.	standard	deviation)	divide	by	the	mean	of	the	peak	ATD	responses,	these	predictive	limits	were	represented	as	a	percentage	and	were	then	applied	to	their	respective	peak	responses	to	create	predictive	intervals	at	each	collision	severity	(2	–	12	km/h).		Any	ATD	responses	where	the	99th	percentile	predictive	intervals	from	the	Control	and	Experimental	 seats	 did	 not	 overlap	 were	 assumed	 to	 be	 statistically	 different.	 	 The	 99th	percentile	 predictive	 limits	 were	 used	 in	 this	 study	 to	 account	 for	 numerous	 comparisons	between	Control	and	Experimental	seats	for	each	ATD	responses	at	each	collision	severity	and	to	provide	a	conservative	estimate	of	statistical	significance.	Because	backset	and	head	restraint	height	varied	between	the	Control	and	Experimental	seats,	 a	 No	 Motion	 condition	 was	 also	 collected	 on	 the	 Experimental	 seat	 to	 isolate	 the					86	contribution	of	seatback	deformation.	The	effects	of	the	Control	seat,	No	Motion	condition	(no	seatback	 deformation),	 and	 Experimental	 seat	 (with	 seatback	 deformation)	 on	 the	 ATD	responses	during	a	12	km/h	collision	severity	were	compared	by	conducting	one-way	analysis	of	 variance	 (ANOVA)	 tests	 for	 each	 peak	 ATD	 responses.	 	 Post-hoc	 comparisons	 were	performed	using	the	Tukey’s	Honest	Significant	Difference	(HSD)	test.	All	statistical	tests	were	performed	 using	 predefined	 functions	 (anova1	 and	 multcompare)	 in	 Matlab	 at	 a	significance	level	a	=	0.05.		To	determine	the	additional	contribution	of	seatback	deformation	from	those	associated	with	pre-impact	 seat	geometry	 (and	other	 seat-related	 factors	 such	as	the	 rigid	 brace	 preventing	 seat	 hinge	 rotation),	 the	 differences	 in	 each	 peak	 kinematic	 and	kinetic	ATD	responses	between	the	Control	seat	and	No	Motion	condition	were	divided	by	the	differences	observed	between	the	Control	and	Experimental	seats	and	expressed	this	ratio	as	a	percentage.		A	score	of	0%	indicates	that	the	No	Motion	yielded	ATD	responses	identical	to	the	Control	 seat	 and	 a	 100%	 score	 that	 the	 No	 Motion	 yielded	 ATD	 responses	 identical	Experimental	 seat.	Thus,	a	 lower	percentage	score	 revealed	 the	 relative	contribution	 (100%	 -	percentage	score)	of	seatback	cushion	deformation	compared	to	the	pre-impact	seat	geometry	(and	other	seat-related	factors)	on	the	overall	reduction	of	ATD	responses.								87	4.3 Results	The	repeatability	of	the	ATD	responses	for	both	the	Control	and	Experimental	seats	were	assessed	using	COV	and	SD	values	from	the	repeated	trials	at	both	the	8	and	12	km/h	collision	speed	changes	 (Table	4.1).	With	 the	exception	of	 initial	head	restraint	height	 (dheight,	8	km/h:	20.4%	and	12	km/h:	11.6%),	the	COV	values	for	the	Control	seat	were	acceptable	or	better	(less	than	 10%)	 for	 both	 collision	 severities.	 	 The	 Experimental	 seat	 on	 the	 anti-whiplash	 seat,	however,	exhibited	poor	peak	COV	values	(greater	than	10%)	in	all	the	neck	injury	criteria	(NIC:	45.9%;	Nij:	22.2%	and	Nkm:	23.3%)	and	in	most	ATD	responses	(dbackset:	14.1%,	dheight:	280.3%,	FX:	26.0%,	FZ:	16.9%,	MY:	25.2%,	wY-head:	15.1%,	qhead:	32.4%,	drebound:	20.6%,	RX:	20.6%	and	Dthead-contact:	19.9%).	The	COV	values	on	the	Experimental	seat	were	significantly	higher	than	those	on	the	Control	seat	for	the	8	and	12	km/h	speed	changes	(paired	t-test;	8	km/h:	t(17)	=	5.01,	p	=	0.001,	and	Wilcoxon	Signed-Rank	Test;	 	12	km/h:	Z	=	3.72,	p	=	0.0002).	The	SD	values	of	 the	Experimental	 and	 Control	 seats	 were	 not	 significantly	 different	 for	 the	 8	 km/h	 (Wilcoxon	Signed-Rank	Test;	Z	=	1.02,	p	=	0.31),	but	were	significantly	different	for	the	12	km/h	(Wilcoxon	Signed-Rank	Test;	Z	=	2.94,	p	=	0.0033).		The	similarities	of	the	SD	values	between	the	Control	and	 the	 Experimental	 seats	 at	 the	 8	 km/h	 suggested	 that	 the	 larger	 COV	 values	 for	 the	Experimental	 seat	 were	more	 likely	 due	 to	 the	 lower	mean	 responses	 than	 to	 the	 inherent	variability	in	the	ATD	responses.		Graded	kinematic	and	kinetic	 responses	 to	 increasing	collision	severities	were	observed	for	both	Control	and	Experimental	seats	(Figure	4.5,	4.6	&	4.7,	Table	4.2).	In	comparison	to	the					88	Control	seat,	the	Experimental	seat	reduced	most	ATD	kinematic	and	kinetic	responses	(aX-head,	ωY-head,	FX,	Fz,	θhead,	RX	and	Δthead-contact)	as	well	as	neck	injury	criteria	(NIC,	Nkm,	and	Nij)	by	15	–	82%	of	peak	Control	responses	at	collision	speeds	of	4	km/h	or	higher.	With	the	exception	of	aX-sled	 and	 aX-T1	 (all	 collision	 severities),	MY	 (2	 –	 6	 km/h	 collision	 speeds),	 and	 Nij	 (2	 &	 4	 km/h	severities),	 there	 were	 no	 overlap	 between	 the	 99th	 percentile	 predictive	 intervals	 of	 the	Control	and	Experimental	seat	responses	and	were	considered	significantly	different	between	the	 two	seats.	 	At	 the	2	km/h	collision	severity,	most	ATD	responses	 (ωY-head,	FX,	θhead,	RX	and	Δthead-contact)	and	neck	injury	criteria	(NIC,	Nij,	and	Nkm)	decreased	from	Control	seat	between	17	to	 58%.	 	However,	 the	 addition	of	 a	 dynamic	 seatback	 cushion	deformation	 increased	 aX-head	(+26%),	FZ	 (+163%),	and	MY	 (+281%)	ATD	responses	 in	comparison	 to	 the	Control	 seat	 (Table	4.2)	In	comparison	to	the	Control	seat,	peak	head	linear	acceleration	(aX-head)	decreased	by	38	–	51%,	while	peak	angular	velocity	of	the	head	(ωY-head)	decreased	by	59	–	81%	(Figure	4.7B	&	4.7F).		Upper	neck	shear	and	axial	forces	and	the	flexion/extension	bending	moment	(MY)	also	decreased	 from	 Control	 trials	 (FX:	 56	 –	 67%,	 FZ:	 36	 –	 72%,	 and	MY:	 29	 –	 77%,	 respectively)	(Figure	 4.7G	–	4.7I).	 No	 head-to-head-restraint	 contact	 (Dthead-contact)	 occurred	 at	 the	 2	 km/h	collision	 severity,	 but	 at	 collision	 severities	 greater	 than	 2	 km/h	 the	 Experimental	 seat	decreased	 Dthead-contact	 by	 50	 to	 82	 ms	 representing	 a	 decrease	 of	 45	 to	 53%	 (Figure	 4.7J).	Kinematic	 responses	 of	 the	 head	 relative	 to	 the	 torso	were	 decreased	 as	measured	 by	 head	angle	(θhead:	56	–	79%)	and	peak	retraction	(RX:	37	–	71%)	(Figure	4.7K	&	4.7L).						89	None	of	the	neck	injury	criteria	(NICmax,	Nij,	nor	Nkm)	exceeded	their	respective	proposed	injury	thresholds	during	the	experiments.	The	Experimental	seat	also	attenuated	all	three	neck	injury	criteria	for	whiplash	injuries	–	Neck	Injury	Criterion	(NICmax:	58	–	69%),	Normalized	Neck	Injury	Criterion	 (Nij:	15	–	81%),	and	Neck	Protection	Criterion	 (Nkm:	41	–	64%)	 (Figure	4.7M	–	4.7O).	Maximum	rearward	penetration	of	the	ATD	torso	and	pelvis	(dpenetration-T1	&	dpenetration-pelvis,	respectively)	also	exhibited	a	graded	response	to	increasing	collision	speed	changes	(Table	4.2).		The	early	onset	of	seatback	cushion	deformation	(tpenetration-onset	=	-200	ms)	prior	to	the	onset	of	the	whiplash	 perturbation	 (t	 =	 0	ms)	 decreased	 the	 backset	 (dbackset)	 and	 increased	 the	 peak	rearward	 occupant	 penetration	 (dpenetration-T1	 &	 dpenetration-pelvis)	 into	 the	 seatback	 from	 the	Control	 seat	 (Figure	 4.6).	 The	 addition	 of	 the	 pre-perturbation	 forward	 seatback	 cushion	deformation	at	-200	ms	before	the	collision	allowed	deeper	‘pocketing’	 into	the	seatback	and	decreased	the	backset	to	13.9	±	6.5	mm	(74%	decrease	from	resting	Experimental	seat	dbackset)	prior	to	the	onset	of	the	whiplash	collision	(Figure	4.6).	Across	the	different	collision	severities,	peak	 forward	rebound	(drebound)	values	 in	 the	Experimental	seat	were	51	–	75%	 lower	than	 in	the	 Control	 seat,	 while	 the	 penetration	 of	 the	 ATD	 into	 the	 seatback	 increased	 at	 the	 T1	(dpenetration-T1:	18	–	26%)	and	pelvis	(dpenetration-pelvis:	75	–	1100%)	(Figure	4.7D	&	4.7E).			The	averaged	pre-impact	head	restraint	geometry	of	the	Control	seat	(dbackset	=	94.3	±	2.2	mm,	dheight	=	-23.8	±	4.0	mm)	were	larger	than	the	Experimental	seat	(dbackset	=	53.8	±	6.0	mm	and	dheight	=	6.3	±	6.5	mm)	and	the	No	Motion	condition	(dbackset	=	60.7	±	6.0	mm	and	dheight	=	-				90	8.7	±	4.4	mm).	The	one-way	ANOVAs	(Table	4.3;	F(2,12)	=	8.43	–	553.10,	p	≤	0.0052)	and	post-hoc	comparisons	revealed	that	the	ATD	responses	observed	on	the	Control	seat	were	different	than	those	observed	in	the	No	Motion	condition	and	all	but	one	response	(aX-T1,	p	=	0.0685)	on	the	Experimental	 seat.	 	 Between	 the	 No	 Motion	 condition	 and	 the	 Experimental	 seat	 most	responses	were	different	with	the	exception	for	five	of	the	eighteen	responses	(dbackset,	aX-sled,	aX-T1,	 MY,	 and	 Nkm:	 p	 =	 0.0765	 –	 0.6646).	 	 These	 results	 suggest	 that	 seatback	 deformation	provides	 a	 unique	 benefit	 over	 pre-impact	 head	 restraint	 geometry.	 The	 percentage	 scores	computed	 to	 determine	 the	 additional	 contribution	 of	 seatback	 deformation	 from	 those	associated	 with	 pre-impact	 seat	 geometry	 revealed	 that	 across	 all	 ATD	 responses,	 the	 No	Motion	 condition	 reduced	 ATD	 responses	 by	 66.7%	 from	 the	 Control	 seat	 (normalized	 %	 =	14.1%	-	163.7%).	This	analysis	suggests	that	approximately	33.3%	of	the	total	reduction	in	ATD	responses	 observed	 between	 the	 Control	 and	 the	 Experimental	 seats	 are	 due	 to	 the	 added	seatback	cushion	deformation.						91		Table	4.1.	Mean	 (standard	deviation)	 and	Coefficient	of	Variation	 (COV:	%)	 from	 five	 repeated	whiplash-like	perturbation	 (n	=	5)	on	 the	Control	and	Experimental	anti-whiplash	seats	at	collision	severities	of	Dv	=	8	km/h	with	Dt	=	148	ms	and	Dv	=	12	km/h	with	Dt	=	194	ms.	Dv	(km/h)	 8	km/h,	D	=	148	ms	 12	km/h,	D	=	194	ms	Trial	 Control	 Experimental	 Control	 No	Motion	 Experimental	Parameter	 Mean	(SD)	 COV	 Mean	(SD)	 COV	 Mean	(SD)	 COV	 Mean	(SD)	 COV	 Mean	(SD)	 COV	dbackset	(mm)	 94.2	(1.9)	 2.0	 55.7	(4.4)	 7.9	 94.4	(2.6)	 2.8	 60.7	(6.0)	 10.0	 52.0	(7.3)	 14.1	dheight	(mm)	 -23.4	(4.8)	 20.4	 -10.0	(3.0)	 30.2	 -24.2	(2.8)	 11.6	 -8.7	(4.4)	 50.8	 2.5	(7.1)	 280.3	aX-sled	(m/s2)	 22.8	(1.1)	 4.7	 22.2	(0.3)	 1.5	 31.6	(0.4)	 1.2	 30.3	(0.4)	 1.3	 30.7	(0.5)	 1.7	dpenetration-T1	(mm)	 -58.1	(2.2)	 3.8	 -68.4	(1.8)	 2.6	 -67.8	(1.0)	 1.5	 -74.3	(2.7)	 3.7	 -78.5	(1.7)	 2.2	dpenetration-pelvis	(mm)	 -38.9	(1.1)	 2.9	 -83.8	(2.7)	 3.2	 -55.8	(2.2)	 2.2	 -62.0	(3.4)	 5.5	 -99.6	(3.8)	 3.8	Dthead-contact	(ms)	 127.2	(1.1)	 0.9	 60.1	(5.8)	 9.6	 112.4	(2.6)	 2.3	 87.0	(4.9)	 5.6	 51.4	(10.2)	 19.9	aX-head	(m/s2)	 85.2	(1.0)	 1.1	 41.5	(1.9)	 4.6	 120.7	(1.7)	 1.4	 75.4	(2.2)	 3.0	 60.7	(4.6)	 7.5	aX-T1	(m/s2)	 45.0	(0.7)	 1.6	 27.2	(0.8)	 3.1	 48.5	(0.4)	 0.9	 46.1	(1.2)	 2.5	 47.1	(1.0)	 2.1	FX	(N)	 213.5	(3.8)	 1.8	 71.6	(10.9)	 7.8	 264.4	(14.1)	 5.4	 172.0	(15.0)	 8.7	 87.8	(22.8)	 26.0	FZ	(N)	 414.4	(9.3)	 2.2	 118.2	(9.8)	 8.3	 586.1	(16.2)	 2.8	 307.3	(37.9)	 12.4	 134.7	(21.6)	 16.0	MY	(Nm)	 20.1	(1.2)	 6.0	 4.6	(1.2)	 25.2	 25.1	(0.1)	 0.2	 4.6	(1.5)	 33.0	 5.9	(1.2)	 19.7	w	Y-head	(deg/s)	 335.3	(16.7)	 5.0	 62.2	(9.4)	 15.1	 393.8	(7.4)	 1.9	 230.5	(19.5)	 8.5	 147.8	(11.9)	 8.1	qhead	(deg)	 15.7	(0.4)	 2.3	 3.8	(1.0)	 26.1	 16.5	(0.6)	 3.9	 12.1	(1.1)	 9.3	 5.3	(1.7)	 32.4	drebound	(mm)	 336.1	(10.4)	 3.1	 83.6	(5.9)	 7.1	 413.5	(7.7)	 1.9	 320.9	(15.2)	 4.7	 126.0	(17.1)	 13.6	RX	(mm)	 -61.8	(1.1)	 1.8	 -26.8	(3.7)	 13.7	 -64.2	(2.4)	 3.7	 -54.0	(4.4)	 8.1	 -32.6	(6.7)	 20.6	NICmax	(m2/s2)	 6.9	(0.2)	 3.1	 1.9	(0.4)	 22.7	 9.7	(0.1)	 1.4	 7.8	(0.5)	 5.9	 2.0	(0.9)	 45.9	Nij	 0.12	(0.003)	 2.3	 0.02	(0.003)	 12.9	 0.16	(0.002)	 1.5	 0.06	(0.011)	 19.1	 0.027	(0.006)	 22.2	Nkm	 0.48	(0.02)	 3.7	 0.17	(0.03)	 17.3	 0.60	(0.02)	 2.9	 0.26	(0.03)	 13.0	 0.23	(0.05)	 23.3	Notes:	 	The	underlined	COV	values	indicate	a	COV	rating	of	acceptable	(5%	≤	COV	<	10%),	the	bolded	COV	values	indicate	a	COV	rating	of	poor	(COV	>	10%),	and	all	other	values	are	rated	good	(COV	<	5%).					92			 	Table	4.2.	Peak	responses	for	Control	and	Experimental	anti-whiplash	(EXP)	seats	for	each	collision	severity.			Dv	(km/h)	 2km/h,	Dt	=	148ms	 4km/h,	Dt	=	148ms	 6km/h,	Dt	=	148ms	 8km/h,	Dt	=	148ms	 12km/h,	Dt	=	194ms	Condition	 Control	 EXP	 Control	 EXP	 Control	 EXP	 Control	 EXP	 Control	 EXP	dbackset	(mm)	 90.5	 58.1	 91.1	 50.2	 90.2	 54.6	 94.2	 55.7	 94.4	 52.0	dheight	(mm)	 -31.0	 8.8	 -32.3	 -3.6	 -42.3	 -3.1	 -23.4	 -10.0	 -24.2	 2.5	aX-sled	(m/s2)	 5.8	 5.6	 11.2	 11.3	 15.9	 16.0	 22.8	 22.2	 31.6	 30.7	dpenetration-T1	(mm)	 -18.4	 -43.9	 -37.0	 -46.5	 -41.7	 -51.4	 -58.1	 -68.4	 -67.8	 -78.5	dpenetration-pelvis	(mm)	 -0.05	 -27.6	 -4.2	 -50.9	 -14.0	 -62.3	 -38.9	 -83.8	 -55.8	 -99.6	Dthead-contact	(ms)	 0.0	 0.0	 164.5	 81.2	 138.8	 72.0	 127.2	 60.1	 112.4	 61.5	aX-head	(m/s2)	 8.7	 12.7	 35	 21.6	 57.1	 34.4	 85.2	 41.5	 120.7	 66.0	aX-T1	(m/s2)	 8.2	 9.4	 20.6	 17.6	 26.4	 27.8	 45.0	 27.2	 48.5	 47.8	FX	(N)	 89.5	 53.8	 166.4	 55.3	 193.3	 85.7	 213.5	 71.6	 264.4	 115.5	FZ	(N)	 34.5	 53.9	 173.5	 111.6	 336.5	 137.4	 414.4	 118.2	 586.1	 166.7	MY	(Nm)	 1.0	 4.0	 6.6	 8.8	 18.5	 13.2	 20.1	 4.6	 25.1	 6.1	w	Y-head	(deg/s)	 110.6	 42.4	 246.3	 50.6	 294.4	 60.6	 335.3	 62.2	 393.8	 163.6	qhead	(deg)	 8.8	 4.0	 15.3	 3.4	 14.5	 3.1	 15.7	 3.8	 16.5	 7.3	drebound	(mm)	 119.6	 58.4	 175.5	 75.0	 260.0	 98.0	 336.1	 83.6	 413.5	 126.0	RX	(mm)	 -34.1	 -18.7	 -57.7	 -16.8	 -58.1	 -24.0	 -61.8	 -26.8	 -64.2	 -40.2	NICmax	(m2/s2)	 1.3	 0.6	 3.5	 2.3	 5.0	 2.1	 6.9	 1.9	 9.7	 3.0	Nij	 0.0	 0.0	 0.0	 0.0	 0.1	 0.0	 0.12	 0.02	 0.16	 0.03	Nkm	 0.2	 0.2	 0.3	 0.2	 0.4	 0.2	 0.48	 0.17	 0.60	 0.30					93	Table	4.3.	One-way	ANOVA	results	and	normalized	percentage	value	to	determine	the	effects	of	seat	geometry	(i.e.	backset	and	head	restraint	height)	on	the	reduction	of	ATD	responses	in	Control,	No	Motion	and	Experimental	conditions	during	a	Dv	=	12	km/h	speed	change.	Five	 repeated	 trials	of	 the	ATD	seated	on	 the	Control	 (CTRL:	unmodified	GMHR	seat,	dbackset	=	94.4	mm),	 No	 Motion	 (Experimental	 seat	 with	 no	 seatback	 cushion	 deformation,	 dbackset	 =	 60.7	 mm)	 and	 Experimental	 (EXP:	Experimental	 seat	 with	 seatback	 cushion	 deformation,	 dbackset	 =	 52.0	mm)	 conditions.	 Normalized	 percentage	 values	 were	determined	by	normalizing	the	differences	in	ATD	responses	between	CTRL	–	No	Motion	with	the	difference	between	Control	–	 EXP	 (100%).	 See	Table	4.1	 for	means	 and	 standard	deviations	 for	 each	 variable.	 Blank	 cells	 represent	p-values	 <	 0.0000.	Bolded	values	denote	a	non-significant	difference	between	the	indicated	conditions.		 ANOVA	 Post	Hoc	Tukey’s	HSD	 Normalized	%	Parameter	 F(2,12)	 p-value	 CTRL	–	No	Motion		 CTRL	-	EXP	 No	Motion	-	EXP	 No	Motion	dbackset	(mm)	 77.42	 	 	 	 0.0765	 79.4	dheight	(mm)	 34.66	 	 0.0012	 	 0.0115	 57.9	aX-sled	(m/s2)	 12.93	 0.0010	 0.0012	 0.0055	 0.6646	 122.4		dpenetration-T1	(mm)	 37.08	 	 0.0006	 	 0.0142	 60.5	dpenetration-pelvis	(mm)	 306.81	 	 0.0189	 	 	 14.1	Dthead-contact	(ms)	 108.09	 	 0.0001	 	 	 41.6	aX-head	(m/s2)	 510.57	 	 	 	 	 75.5	aX-T1	(m/s2)	 8.43	 0.0052	 0.0041	 0.0685	 0.2890	 163.7	FX	(N)	 123.75	 	 	 	 	 52.3	FZ	(N)	 358.98	 	 	 	 	 61.8	MY	(Nm)	 549.87	 	 	 	 0.1810	 106.9	w	Y-head(deg/s)	 407.46	 	 	 	 	 66.4	qhead	(deg)	 103.13	 	 0.0004	 	 	 38.7	drebound	(mm)	 553.10	 	 	 	 	 32.2	RX	(mm)	 56.30	 	 0.0153	 	 	 32.1	NICmax	(m2/s2)	 228.00	 	 0.0009	 	 	 24.1	Nij	 462.34	 	 	 	 0.0001	 78.6	Nkm	 145.87	 	 	 	 0.5462	 93.0					94		Figure	4.5.	Data	comparing	the	BioRID	II	ATD	response	seated	on	either	the	Control	seat	(grey	lines)	or	the	Experimental	seat	with	the	seatback	cushion	deformation	profile	(black	lines)	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6,	and	8kmh	with	a	collision	pulse	duration	(Δt)	of	148	ms	and	Δv	=	12	km/h	with	a	Δt	=	194	ms).	Hollow	circles	represent	the	onset	of	head-to-head-restraint	contact	and	peak	responses	of	each	ATD	response	parameter	for	each	trial.								95		Figure	4.6.	BioRID	II	ATD	and	seatback	interaction	data	comparing	the	Control	seat	(grey	lines)	and	the	Experimental	anti-whiplash	seat	with	seatback	deformation	profile	 (black	 lines)	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6,	and	8kmh	with	a	collision	pulse	duration	(Δt)	of	148	ms	and	Δv	=	12	km/h	with	a	Δt	=	194	ms).	From	top	to	bottom	(+X-direction	represents	towards	 the	 front	 of	 the	 seat):	 Horizontal	 sled	 acceleration	 (aX-sled),	 head-to-head	 restraint	backset	distance	 (dbackset	 =	0	mm	represents	head	 contact	with	head	 restraint),	 and	 rearward	penetration	of	the	ATD’s	T1	(dpenetration-T1)	and	pelvis	(dpenetration-pelvis)	into	the	seatback.							96		Figure	4.7.	Experimental	results	of	peak	kinematic	and	kinetic	responses	for	the	BioRID	II	ATD	seated	on	Control	and	Experimental	seats.	For	all	the	graphs,	grey	circles	represent	the	control	seat	 and	 black	 triangles	 represent	 the	 Experimental	 seat	 with	 the	 seatback	 deformation	profiles.	 	 The	 shaded	bands	 represent	 the	 99th	 percentile	 predictive	 intervals	 for	 the	 Control	and	 Experimental	 seats.	 Y-axis	 values	 for	 panels	 D,	 E	 and	 L	 have	 been	 inverted	 for	 visual	purposes	to	show	increasing	responses	from	bottom	to	top.		 					97	4.4 Discussion	The	goal	of	this	study	was	to	compare	the	performance	of	the	Experimental	anti-whiplash	seat	that	dynamically	modified	seatback	cushion	deformation	to	a	standard	unmodified	GMHR	Control	seat	during	a	series	of	 low-speed,	rear-end	collisions	(Δv	=	2,	4,	6,	8	and	12	km/h).	 In	comparison	 to	 the	 Control	 seat,	 the	 Experimental	 seat	 significantly	 reduced	 most	 ATD	responses	 (aX-head,	 ωY-head,	 FX,	 FZ,	 My,	 θhead,	 RX	 and	 Δthead-contact)	 by	 37	 –	 82%	 and	 neck	 injury	criteria	 (NICmax,	Nkm,	 and	Nij)	 by	15	–	81%	at	 collision	 speeds	4	 km/h	or	higher.	At	 the	 lower	collision	severities	(Δv	<	4	km/h),	the	Experimental	seat	was	less	effective	at	reducing	most	ATD	responses	 (increased	 aX-head	 and	 FZ	 responses	 but	 decreased	 all	 other	ATD	 responses	 by	 17	 –	58%).	 Linear	 acceleration	 of	 T1	 was	 not	 significantly	 different	 between	 Control	 and	Experimental	 seats	 across	 all	 collision	 severities.	 The	 results	 from	 this	 study	 suggest	 that	 the	dynamic	deformation	of	the	seatback	cushion	to	reduce	the	apparent	stiffness	of	the	seatback	may	be	a	novel	safety	mechanism	to	reducing	the	risk	of	whiplash	injuries	following	low-speed,	rear-end	collisions.					In	comparison	to	the	Control	seat,	the	Experimental	seat	with	the	seatback	deformation	mechanism	increased	both	the	maximum	dpenetration-T1	by	-16.4	±	6.9	mm	and	dpenetration-pelvis	by	-52.7	±	18.4	mm	to	allow	deeper	penetration,	or	“pocketing”,	of	the	ATD	into	the	seatback.	Prior	to	the	onset	of	the	collision	(i.e.	at	t	=	0	ms),	the	seatback	deformation	further	decreased	the	pre-impact	 backset	 by	 74%	 to	 13.9	 ±	 6.5	 mm.	 	 During	 an	 impact,	 the	 Experimental	 seat	decreased	 the	 head-to-head	 restraint	 contact	 times,	 peak	 head	 accelerations,	 upper	 neck					98	forces	and	moments,	and	neck	 injury	criteria	as	well	as	head	movement	relative	 to	 the	 torso	(θhead	and	RX).	Head	restraint	geometry,	however,	also	differed	between	seats	and	could	have	influenced	peak	ATD	 responses.	Based	on	additional	 analyses	using	 the	No	Motion	 condition	(without	 seatback	 deformation),	 the	 effect	 of	 seatback	 cushion	 deformation	 alone	 was	estimated	 to	 account	 for	 potentially	 33%	 of	 the	 reduction	 observed	 between	 Control	 and	Experimental	seats	(67%	due	to	head	restraint	geometry	and	the	different	seats).	In	comparison	to	 the	No	Motion	Condition,	 the	Experimental	 seat	yielded	a	 significant	decrease	 in	11	of	15	peak	ATD	responses	and	in	2	of	3	neck	injury	criteria	despite	observing	a	similar	acceleration	at	T1.		Hence,	the	added	decreases	in	peak	head	accelerations,	upper	neck	forces	and	moments,	and	neck	injury	criteria	may	be	a	result	of	the	deeper	pocketing	of	the	occupant	into	the	rigid	seatback	 due	 to	 the	 unspooling	 of	 the	 straps	 attached	 to	 the	 middle	 and	 lower	 seatback	motors.	The	motion	of	 the	seatback	cushion	deformation	allowed	 for	more	 time	 to	gradually	propagate	the	acceleration	and	forces	from	the	sled	to	the	occupant.			Peak	 linear	 forward	 acceleration	 at	 T1	 (aX-T1)	 was	 the	 only	 ATD	 response	 that	 was	 not	reduced	by	 the	 Experimental	 seat	 at	 the	 higher	 speed	 changes.	 	Many	 anti-whiplash	 devices	(e.g.,	Volvo	Whiplash	Injury	Prevention	Seat	–	WHIPS,	Saab’s	Active	Head	Restraint	–	SAHR,	and	the	GMHR)	and	the	neck	 injury	criteria	 (NICmax)	have	been	designed	with	 the	reduction	of	T1	acceleration	as	one	of	the	key	components	(Bostrom	et	al.,	1996,	Jakobsson	et	al.,	2008,	Viano,	2003d,	Viano	and	Olsen,	2001,	Viano	and	Parenteau,	2015).		One	possible	explanation	as	to	why	aX-T1	was	not	reduced	by	the	Experimental	seat	was	because	of	the	rigid	support	frame	installed					99	to	 prevent	 seat	 hinge	 rotation.	 On	 the	 Control	 seat,	 the	 seat	 hinge	 is	 naturally	 allowed	 to	deflect	 rearward	 to	 potentially	 reduce	 aX-T1.	 No	 seat	 hinge	 rotation	 was	 allowed	 on	 the	Experimental	seat	causing	earlier	time	of	peak	aX-T1	responses	at	similar	peak	amplitude	to	sled	acceleration	(Figure	4.4).	Further	research	is	required	to	understand	why	the	lack	of	a	reduction	in	the	horizontal	acceleration	of	T1	(aX-T1),	a	variable	that	 is	currently	used	as	a	test	metric	 in	evaluating	the	performance	of	anti-whiplash	seats,	was	the	only	variable	not	attenuated	by	the	seatback-deformation	mechanism.		The	 results	 from	 the	present	and	previous	 studies	 (Viano,	2003a,	Viano	and	Parenteau,	2015,	Hofinger	et	al.,	1999)	suggested	that	increased	occupant	penetration	into	the	seatback	by	either	using	softer	seatback	cushion	properties	or	dynamically	deforming	the	seatback	reduced	backset	distance	and	other	ATD	responses.	Hofinger	et	al.	(1999)	used	different	types	of	foam	mounted	on	a	stiff	seatback	and	a	pre-impact	dbackset	of	100	mm	to	determine	the	influence	of	seatback	 cushioning.	 They	 suggested	 that	 the	 combination	 of	 stiffer	 foam	 at	 the	 pelvis	 and	softer	 foam	 at	 the	 torso	 best	 reduced	 occupant	 acceleration	 by	 initiating	 earlier	 torso	 and	pelvis	rotation	and	decreasing	the	backset.	When	Hofinger	et	al.	(1999)	tested	the	combination	of	 stiffer	 foam	 at	 the	 torso	 and	 softer	 foam	 at	 the	 pelvis,	 they	 observed	 large	 backset	 and	higher	head	accelerations.	The	seatback	cushion	deformation	profile	was	akin	to	having	a	soft	foam	 at	 both	 the	 torso	 and	 the	 pelvis	 to	 delay	 the	 onset	 of	 torso	 acceleration,	 increase	occupant	 penetration	 into	 the	 seatback	 and	 decrease	 backset.	 One	 key	 difference	 between	these	 two	 experiments	 was	 that	 Hofinger	 et	 al.	 (1999)	 used	 a	 rigid	 solid	 seatback	 panel	 to					100	support	 the	 different	 foam	 combination	while	 the	 Experimental	 anti-whiplash	 seat	 utilized	 a	perimeter	 frame	with	 a	 compliant	 center.	 The	 rigid	 perimeter	 seat	 frame	 and	 the	 compliant	seatback	suspension	of	the	Experimental	seat	was	designed	to	create	softer	seatback	cushion	properties	 and	 increase	 occupant	 retention	 during	 a	 collision	 (Viano,	 2003a,	 Viano	 and	Parenteau,	 2015).	 The	 two	 components	work	 in	 unison	 as	 the	 yielding	 center	 section	 of	 the	seatback	 absorbs	 energy	 and	 gradually	 decelerates	 the	 torso	 during	 the	 collision,	 while	 the	stiffer	 seatback	 frame	 enhances	 occupant	 retention	 and	 prevents	 seatback	 deflection.	Regardless	of	the	seatback	deformation	mechanism	(i.e.	foam,	springs	or	motors),	these	results	suggested	 that	 a	 soft,	 yielding	 seat	 center	 that	 allows	 for	 decreased	 backset	 and	 deep	penetration	(dpenetration:	80	–	100	mm)	of	an	occupant’s	torso	and	pelvis	into	the	seatback	could	potentially	reduce	the	risk	of	whiplash	injury	following	low-speed,	rear-end	collisions.	The	repeatability	of	the	ATD	responses	on	the	Control	seat	was	mostly	acceptable	or	good	(COV	<	10%)	with	only	head	restraint	height	(dheight)	receiving	a	poor	rating	at	both	the	8	km/m	(COV	 =	 20.4%)	 and	 the	 12	 km/h	 (COV	 =	 11.6).	 For	 the	 Experimental	 seat,	 the	 COVs	 were	typically	larger,	with	5	good,	5	acceptable	and	8	poor	parameters	(a	total	of	18	parameters)	at	the	8	km/h	level,	and	4	good,	2	acceptable,	and	12	poor	parameters	at	the	12	km/h	level.	A	key	reason	 for	 the	 higher	 COV	 values	 in	 the	 Experimental	 seat	 is	 the	 lower	 mean	 responses	observed	 in	 this	 seat	 (Rhule	 et	 al.,	 2005).	 By	 comparing	 the	 variability	 of	 the	 SD	 values,	 the	similarities	 between	 the	 Control	 and	 Experimental	 SD	 values	 for	 the	 8	 km/h	 collision	 speed	suggested	 that	 the	 COV	 values	 considered	 poor	 (>	 10%)	were	more	 likely	 due	 to	 decreased					101	mean	responses	than	due	to	increased	variability	in	the	ATD	responses.	The	larger	SD	values	in	the	ATD	 responses	on	 the	 Experimental	 seat	 at	 the	 12	 km/h	 speed	 change	 compared	 to	 the	Control	seat	may	be	due	to	the	additional	dynamics	and	degrees	of	freedom	introduced	by	the	active	 seatback	 deformation.	 Regardless,	 the	 increased	 SD	 values	 in	 the	 Experimental	 seat	remained	 smaller	 than	previously	observed	SD	 (see	Table	2.2	 in	Chapter	2;	Davidsson,	1999,	Moorhouse	 et	 al.,	 2012).	 In	 addition,	 the	mean	 responses	 close	 to	 zero,	 as	 observed	 in	 pre-impact	 head	 restraint	 height	 (dheight:	 Control	 =	 -30.4	 mm	 and	 Experimental	 =	 -1.1	 mm),	generated	a	 large	COV	above	 the	acceptable	 range	 (>	10%)	with	a	small	SD	value.	Thus,	COV	values	from	the	Experimental	seat	have	to	be	interpreted	cautiously	due	to	the	reduced	mean	responses	by	the	addition	of	the	seatback	deformation	mechanism.		The	additional	assessment	of	 the	SD	values	has	shown	that	 the	ATD	seated	on	both	 the	Control	and	Experimental	 seats	have	sufficient	repeatability	to	justify	the	comparisons	made	in	this	study.											The	 speed	 changes	used	 in	 this	 study	were	 lower	 than	 the	16	 km/h	 speed	 change	 and	longer	 than	 the	 91	 ms	 collision	 pulse	 duration	 used	 by	 RCAR/IIWPG	 (Insurance	 Institute	 of	Highway	Safety,	2008b,	Insurance	Institute	of	Highway	Safety,	2008a).	Nevertheless,	the	pulses	used	here	provided	a	graded	range	of	speed	changes	that	represent	some	real-world	collisions	that	 cause	 whiplash	 injury	 (Bartsch	 et	 al.,	 2008,	 Krafft	 et	 al.,	 2005).	 The	 weight	 of	 the	Experimental	seat	with	the	seatback	motors	and	rigid	frame	(total	weight	=	204	kg)	limited	the	peak	 accelerations	 that	 could	 be	 achieved	 by	 the	 test	 sled.	 These	 seatback	 motors	 were	selected	to	allow	for	testing	a	wide	range	of	seatback	motor	parameters	and	explore	different					102	seatback	cushion	deformation	profiles.		A	collision	pulse	duration	of	Dt	=	148	ms	was	selected	for	collision	speeds	up	to	8	km/h	because	 it	was	achievable	and	not	dissimilar	 to	 the	135	ms	observed	 in	 prior	 vehicle-to-vehicle	 rear-end	 crashes	 (Brault	 et	 al.,	 2000,	 Brault	 et	 al.,	 1998,	Siegmund	et	al.,	2000,	Siegmund	et	al.,	1997).		However,	a	collision	pulse	duration	of	Dt	=	194	ms	was	needed	to	reach	a	12	km/h	collision	speed.	This	longer	duration	is	longer	than	those	of	many	modern	vehicles	(Linder	et	al.,	2003,	Linder	et	al.,	2001,	Stigson	et	al.,	2006)	but	shorter	than	 those	 observed	 in	 some	 older	 vehicles	 with	 bumper	 isolators	 (Siegmund	 et	 al.,	 1994).		Nevertheless,	the	results	of	this	study	provided	a	better	understanding	of	how	to	dynamically	control	 seatback	cushion	deformation	 to	minimize	ATD	 responses	during	a	 rear-end	collision.		Future	work	should	aim	to	refine	the	methods	for	deforming	the	seatback	and	to	conduct	tests	at	higher	speed	changes	and	shorter	pulse	durations.		Another	 possible	 limitation	 of	 this	 study	 was	 the	 structural	 differences	 between	 the	Control	 and	 Experimental	 seats.	 	 A	 rigid	 metal	 brace	 was	 attached	 to	 the	 back	 of	 the	Experimental	 seat	 to	 prevent	 the	 seat	 hinge	 from	 influencing	 ATD	 kinematic	 and	 kinetic	responses;	whereas,	the	Control	seat	was	an	unmodified	GMHR	seat	that	was	allowed	to	rotate	freely	about	the	seat	hinge.		In	the	Experimental	seat,	the	onsets	of	ATD	responses	were	earlier	than	those	seen	on	the	Control	seat	as	 the	rigid	seat	 reduced	the	ability	of	 the	seat	hinge	to	absorb	and	delay	the	transfer	of	acceleration	from	the	seat	to	the	ATD.	Finally,	the	seats	were	not	 instrumented	 to	 determine	 the	 different	 forces	 applied	 by	 the	 seat	 to	 an	 occupant	 on	either	 the	Control	or	Experimental	 seats.	 	Thus,	 further	work	 is	needed	 to	better	understand					103	the	 pressure	 distribution	 and	 interactions	 between	 an	 occupant’s	 back	 and	 the	 seatback	 to	develop	a	seat	that	better	reduces	the	forces	applied	to	the	occupant.					4.5 Conclusion	This	study	highlights	the	potential	importance	of	a	dynamic	seatback	cushion	deformation	safety	 mechanism	 to	 reduce	 the	 risk	 of	 whiplash	 injuries	 during	 rear-end	 impacts.	Approximately	 67%	 of	 the	 observed	 reduction	 in	 ATD	 responses	 between	 the	 Control	 and	Experimental	seats	was	determined	to	be	due	to	the	differences	 in	pre-impact	head	restraint	geometry.	 	 	 The	 remaining	 33%	 of	 the	 observed	 reductions	 in	 ATD	 responses	 could	 be	attributed	 to	 the	 addition	of	 a	 dynamic	 seatback	deformation	on	 the	 Experimental	 seat.	 The	Experimental	 anti-whiplash	 seat	 with	 the	 seatback	 cushion	 deformation	 profile	 increased	occupant	 penetration	 into	 the	 seatback	 to	 decrease	 backset	 and	 help	 maintain	 pre-impact	alignment	of	the	head,	neck	and	torso.	Further	development	of	the	Experimental	anti-whiplash	seat	could,	potentially,	yield	a	new	device	that	can	reduce	the	risk	of	whiplash	injuries	following	low-speed,	rear-end	collision.		 					104	Chapter	5. Combined	 Effect	 of	 Seat	 Hinge	 Rotation	 and	 Seatback	Cushion	Deformation	on	ATD	Responses	5.1 Introduction		 Whiplash	injuries	are	the	most	common	type	of	injuries	during	rear-end	collision	with	an	annual	 incident	 rate	 in	 the	 western	 world	 ranging	 from	 28	 to	 834	 per	 100,000	 inhabitants	(Cassidy	et	al.,	2000,	Holm	et	al.,	2008,	Jakobsson	et	al.,	2000,	National	Highway	Traffic	Safety	Administration,	 2014,	 Otremski	 et	 al.,	 1989).	Whiplash	 injuries	 are	 classified	 as	 a	 soft	 tissue	injury	 to	 the	cervical	 spine	most	commonly	 resulting	 from	a	sudden	acceleration	of	 the	head	relative	to	the	torso,	but	the	exact	injury	mechanism	remains	uncertain	(Kaneoka	et	al.,	1999,	McConnell	 et	 al.,	 1995,	 Svensson	 et	 al.,	 2000).	 In	 vehicles,	 the	 automotive	 seat	 and	 head	restraint	 are	 the	 primary	 safety	 devices	 for	 protecting	 against	 whiplash	 injuries	 during	 low-speed,	rear-end	collisions.	Current	anti-whiplash	seats,	such	as	General	Motors’	High	Retention	seat	(GMHR),	Volvo’s	Whiplash	Injury	Prevention	seat	(WHIPS)	and	Saab’s	Active	Head	Restraint	seat	 (SAHR),	 attempt	 to	 reduce	 the	 risk	 of	 whiplash	 injuries	 by	 reducing	 head	 and	 torso	acceleration	of	the	occupant	and/or	by	minimizing	the	relative	motion	between	the	head	and	upper	torso	(Lundell	et	al.,	1998a,	Lundell	et	al.,	1998b).	Epidemiological	studies	have	estimated	that	 these	 anti-whiplash	 seats	 have	 reduced	 the	 risk	 of	 whiplash	 injury	 by	 about	 50%	(Jakobsson	et	al.,	2008,	Kullgren	and	Krafft,	2010,	Viano	and	Olsen,	2001).	Further	research	and	development	 are	 required	 to	 understand	 why	 current	 anti-whiplash	 seats	 are	 only	 partially	effective	in	preventing	whiplash	injuries.							105		 According	 to	 the	 Research	 Council	 for	 Automobile	 Repairs/International	 Insurance	Whiplash	 Prevention	 Group	 (RCAR/IIWPG)	 seat	 and	 head	 restraint	 rating	 protocol,	 anti-whiplash	 seats	 can	 be	 rated	 as	 good	 (e.g.	 WHIPS	 and	 SAHR)	 or	 poor	 (e.g.	 GMHR)	 in	 their	abilities	 to	 reduce	 the	 risk	 of	 whiplash	 injury	 in	 low-/moderate-speed,	 rear-end	 collisions	(Insurance	 Institute	of	Highway	Safety,	 2008b,	 Insurance	 Institute	of	Highway	Safety,	 2008a).	Based	on	real-world	insurance	claims,	good-rated	seats	appear	to	better	 lower	occupant	neck	injury	 risk	 by	 11	 to	 15%	 compared	 to	 poor-rated	 seats	 (Farmer	 et	 al.,	 2008,	 Trempel	 et	 al.,	2016).	The	apparent	minor	real-world	benefits	of	good-rated	seats	prompted	a	previous	study	to	 compare	 the	 performance	of	 good-rated	 (2005	Volvo	 S40	WHIPS,	 2004	Volvo	 S60	WHIPS,	2005	 Saab	 9.3	 SAHR)	 and	 poor-rated	 (2004	 Pontiac	 Grand	 Am	 GMHR)	 anti-whiplash	 front	passenger	 seats	 during	 low-speed,	 rear-end	 collisions	 (speed	 changes:	 Δv	 =	 2	 to	 14	 km/h,	collision	pulse	duration:	Δt	=	141ms;	see	Chapter	2).	The	good-rated	seats	only	attenuated	four	key	 occupant	 responses	 (peak	 upper	 neck	 shear	 and	 axial	 forces,	 flexion/extension	moment	and	rearward	head	retraction)	in	comparison	to	the	poor-rated	seat	at	speed	changes	greater	than	6	km/h.	The	 lack	of	attenuation	 in	other	peak	occupant	kinematic	and	kinetic	responses	suggests	that	current	anti-whiplash	seats	could	be	improved	to	better	protect	occupants	from	whiplash	injuries	during	low-speed,	rear-end	collisions.	An	Experimental	anti-whiplash	seat	that	actively	rotates	the	seat	hinge	(see	Chapter	3)	and	modulates	the	compliance	of	the	seatback	cushion	(see	Chapter	4)	during	a	rear-end	collision	to	change	the	apparent	structural	stiffness	of	the	seat	hinge	and	seatback	cushion	was	developed	to	address	this	need.	When	activated	in	isolation,	 seat	 hinge	 rotation	 and	 seatback	 cushion	 deformation	 reduced	 peak	 occupant					106	responses	for	collision	speed	changes	greater	than	4	km/h	compared	to	a	control	seat	by	15	–	85%;	however,	the	performance	of	the	anti-whiplash	seat	when	both	seat	hinge	and	seatback	deformation	 are	 co-activated	 and	 its	 performance	 relative	 to	 existing	 anti-whiplash	 seats	remain	unknown.	The	 primary	 goal	 of	 this	 study	was	 to	 compare	 the	 performance	 of	 the	 Experimental	anti-whiplash	 seat	 with	 the	 co-activation	 of	 the	 seat	 hinge	 rotation	 and	 seatback	 cushion	deformation	 to	 existing	 anti-whiplash	 seats	 (WHIPS,	 SAHR,	 and	 GMHR)	 in	 attenuating	 the	kinematic	and	kinetic	 responses	of	a	BioRID	 II	anthropomorphic	 test	device	 (ATD)	exposed	to	low-speed,	 rear-end	 perturbations	 (speed	 changes	 of	 2	 to	 12	 km/h).	 The	 seat	 hinge	 and	seatback	deformation	profiles	 identified	 in	Chapters	3	and	4	were	combined	and	the	 relative	timing	 between	 each	 profile	 was	 determined	 from	 preliminary	 experiments	 performed	 at	 a	single	speed	change	(Δv	=	12	km/h;	see	Appendix	D).	We	hypothesized	that	the	Experimental	anti-whiplash	seat	would	attenuate	most	peak	ATD	responses	and	neck	injury	criteria	across	all	tested	collision	severities	in	comparison	to	the	existing	anti-whiplash	seats.	A	secondary	goal	of	the	 experiments	was	 to	 quantify	 the	 repeatability	 of	 the	ATD	 responses	 evoked	on	 the	 anti-whiplash	seat.			5.2 Methods	5.2.1 Experimental	Anti-Whiplash	Seat		 The	 Experimental	 anti-whiplash	 seat	 design	 consisted	 of	 a	 modified	 GMHR	 and	 five	motors	 to	 control	 seat	 hinge	 rotation	 and	 seatback	 cushion	 deformation	 (Figure	 5.1).	 All	 5					107	motors	(2	seat	hinge	and	3	seatback)	were	controlled	by	separate	digital	servo	drives	(Servostar	600,	 Kollmorgen,	 Waltham,	 MA,	 USA)	 connected	 to	 two	 universal	 motion	 interfaces	 and	 a	motion	 controller	 (NI	 UMI	 7774	 and	 NI	 PXI	 7350,	 National	 Instruments	 Corporation,	 Austin,	Texas,	 USA).	 A	 custom	 LabVIEW	 program	 (National	 Instruments	 Corporation,	 Austin,	 Texas,	USA)	 was	 created	 to	 send	 commands	 to,	 monitor	 the	 status	 of,	 and	 record	 encoder	 data	directly	from	the	motors.				Figure	5.1.	Photograph	of	the	experimental	set-up	with	the	BioRID	II	ATD	on	the	Experimental	anti-whiplash	 automotive	 seat	 and	 laboratory	 reference	 frame	 (X,	 Z).	 	 The	 seat	 hinge	 and	seatback	motors	on	 the	 left	 side	of	 the	anti-whiplash	 seat	are	 labelled.	Additional	 seat	hinge	and	seatback	motors	are	 located	on	 the	 right	 side	of	 the	seat	 (not	 labelled).	 	 (Figure	5.1	 is	a	repeat	of	Figure	3.1B.)							108	The	seat	pan	and	head	restraint	used	in	the	Experimental	anti-whiplash	seat	were	from	a	 GMHR	 seat	 and	 remained	 unmodified.	 	 The	 seatback	 consisted	 of	 a	 rigid	 aluminum	 outer	frame	 with	 a	 yielding	 seatback	 center	 (i.e.,	 a	 modified	 GMHR	 upper	 seatback	 cushion	suspended	 within	 the	 outer	 aluminum	 frame).	 Two	 large	 rotational	 servomotors	 (AKM52K,	Kollmorgen,	Waltham,	MA,	USA)	connected	to	helical	right-angle	gearheads	(VTR014-035,	35:1	gear	ratio,	Thomson	Linear,	Radford,	VA,	USA)	were	mounted	on	either	side	of	the	outer	frame	at	the	seat	hinge	location	and	were	geared	to	rotate	in	unison	(one	in	a	positive	direction	and	the	other	in	a	negative	direction)	to	control	the	rotation	of	the	seat	hinge	through	predefined	rotation	 profiles.	 Seatback	 cushion	 deformations	 were	 controlled	 by	 three	 rotational	servomotors	 (AKM24D,	 Kollmorgen,	 Waltham,	 MA,	 USA)	 connected	 to	 helical	 right-angle	gearheads	(VTR006-008,	8:1	gear	ratio,	Thomson	Linear,	Radford,	VA,	USA)	staggered	on	either	side	of	the	rigid	outer	frame.	The	three	motors	were	placed	at	the	top,	middle	and	bottom	of	the	 seatback	 and	 attached	 to	 four	 47mm-wide	 seatbelt	 straps	 (with	 the	 middle	 motor	connecting	to	the	2	middle	straps)	that	spanned	horizontally	across	the	rigid	outer	frame	(see	Figure	4.2).				The	GMHR	upper	seatback	cushion	was	suspended	from	these	straps	and	the	rotation	of	each	motor	 separately	 controlled	 the	 seatback	 cushion	deformation	at	 the	upper-torso,	mid-torso	and	pelvis	 regions.	Tightening	 the	seatbelt	 straps	served	to	 limit	 the	seatback	cushion's	rearward	deformation	and	increase	the	apparent	stiffness	of	the	seatback;	whereas,	loosening	the	 straps	 allowed	 the	 seatback	 cushion	 to	 deform	 further	 rearward	 and	 allow	 for	 deeper					109	occupant	penetration	into	the	seatback.	The	initial	tension	in	all	the	straps	were	set	to	generate	an	initial	seatback	angle	of	27	deg	rearward	from	vertical	(Siegmund	et	al.,	2005a).	Through	a	series	of	preliminary	experiments	performed	at	a	12	km/h	speed	change	(see	Appendix	 D),	 a	 combined	 seat	 hinge	 rotation	 and	 seatback	 cushion	 deformation	 profile	was	identified	that	effectively	reduced	the	ATD	responses.		The	dynamic	seat	hinge	rotation	profile	(Figure	5.2A,	grey	line)	was	programmed	to	rotate	the	seatback	forward	commencing	at	90	ms	(tforward-onset)	 before	 the	 collision	 onset	 by	 5.6	 deg	 (θforward-peak)	 at	 an	 angular	 velocity	 of	 -33.4	deg/s	 (ωforward-int)	 and	 to	 then	 rotate	 the	 seatback	 rearward	 by	 5.7	 deg	 (θrearward-peak)	 at	 an	angular	velocity	 (ωrearward-int)	of	24.0	deg/s	beginning	70	ms	 (trearward-onset)	after	 collision	onset.	The	observed	output	seat	hinge	rotation	(Figure	5.2A,	red	line)	resulted	in	a	forward	rotation	of	4.6	deg	 followed	by	 a	 rearward	 rotation	of	 3.8	deg.	 The	 seatback	motor	deformation	profile	was	programmed	to	unspool	the	webbing	from	the	two	lower	seatback	motors	beginning	200	ms	 (tdeformation-onset)	before	 the	collision	onset	 to	a	peak	angle	 (θrearward-peak)	of	168	deg	and	an	initial	angular	velocity	(ωdeformation-int)	of	980	deg/s.		The	upper-torso	motor	was	maintained	in	a	stationary	 position,	 effectively	 creating	 a	 hinge	 point	 at	 the	 upper	 back	 and	 increased	 the	pocketing	of	the	pelvis	into	the	seat	during	the	collision.		A	motion	capture	system	(Optotrak	Certus,	Northern	Digital,	Waterloo,	ON,	Canada)	was	used	 to	 estimate	 the	 deformation	 of	 the	 seatback	 cushion	 during	 the	 perturbation	 because	slack	 in	 the	 straps	 induced	 by	 the	 two	 lower	 seatback	 motors	 did	 not	 correspond	 with	 the	physical	 penetration	 of	 the	 occupant.	 	 Maximum	 rearward	 penetration	 of	 the	 T1	 vertebra					110	(dpenetration-T1)	 and	 pelvis	 (dpenetration-pelvis)	 were	 used	 to	 describe	 the	 seatback	 cushion	deformation	and	were	defined	as	the	horizontal	displacement	(along	the	global	X-axis)	between	the	 infrared	 light	 emitting	 diode(IRED)	 markers	 mounted	 to	 the	 seatback	 frame	 and	 the	 T1	vertebra	and	pelvis,	respectively	(Figure	5.2).	The	initial	position	of	the	ATD	at	rest	was	set	to	zero	 (dpenetration-T1	 =	 0	 mm	 and	 dpenetration-pelvis	 =	 0	 mm)	 and	 rearward	 displacements	 into	 the	seatback	were	defined	as	negative	values.	For	a	12	km/h	collision,	 the	programmed	seatback	rotation	profile	allowed	maximum	occupant	penetration	into	the	seatback	of	62	mm	at	T1	and	94	mm	at	the	pelvis	(Figure	5.2B,	green	and	blue	lines,	respectively).			Figure	 5.2.	 Programmed	 inputs	 and	 resulting	 outputs	 for	 A.)	 seat	 hinge	 rotation	 and	 B.)	seatback	cushion	deformation	profiles	during	a	12	km/h	collision	(∆t	=	187	ms;	black	line).	Grey	lines	 represent	 input	A.	 rotation	 (θseat-hinge)	 and	B.	deformation	 (θseatback-motors)	 profiles	 to	 seat	hinge	 and	 seatback	motors,	 respectively.	 Theses	 programmed	 input	 seat	 hinge	 rotation	 and	seatback	deformation	profiles	were	used	across	all	collision	speeds.	Seat	hinge	rotation	(θseat-hinge-output:	 red	 line)	 as	 well	 as	 rearward	 penetration	 of	 ATD	 T1	 (dpenetration-T1:	 green	 line)	 and	pelvis	(dpenetration-pelvis:	blue	line)	into	the	seatback	indicate	resulting	outputs	to	their	respective	input	profiles.						111	5.2.2 Anthropomorphic	Test	Device	and	Instrumentation	A	 BioRID	 II	 ATD	 (Humanetics,	 Plymouth,	MI,	 USA)	was	 instrumented	 to	measure	 head,	torso	(T1),	and	pelvis	kinematics	and	kinetics	(Figure	5.1).	A	six-axis	load	cell	(Forces:	FX,	FY,	FZ;	and	Moments:	MX,	MY,	MZ;	Model	4949a,	Robert	A.	Denton,	Inc.,	Rochester	Hills,	MI,	USA)	was	mounted	 at	 the	 atlanto-occipital	 joint	 (AOJ)	 to	 measure	 upper	 neck	 forces	 and	 moments.		Linear	 forward	 accelerations	 of	 the	 head	 and	 T1	 were	 measured	 using	 two	 uni-axial	accelerometers	(7264C	sensors;	±500g,	Endevco,	San	Juan	Capistrano,	CA,	USA)	mounted	at	the	head	 center	 of	 mass	 and	 the	 T1	 vertebra.	 A	 uni-axial	 angular	 rate	 sensor	 (ARS-1500;	 ±26.2	rad/s,	 DTS,	 Seal	 Beach,	 CA,	 USA)	was	 also	mounted	 at	 the	 head	 center	 of	mass	 to	measure	angular	kinematics	in	the	sagittal	plane	(i.e.,	flexion	and	extension).			The	Optotrak	 system	was	used	 to	 track	 IRED	markers	affixed	 to	 the	head,	AOJ,	T1,	and	pelvis	of	the	ATD	to	measure	their	displacements.	 	Additional	 IRED	markers	were	mounted	to	the	seat	to	record	displacements	and	rotations	of	the	seatback	and	to	create	a	global	reference	frame	 (+X	 forward,	 +Y	 right,	 and	 +Z	 down,	 Figure	 5.1).	 	 Horizontal	 sled	 acceleration	 was	measured	with	a	uni-axial	accelerometer	 (2220-100;	±100g,	Silicon	Design	 Inc.,	 Issaquah,	WA,	USA)	 mounted	 directly	 to	 the	 base	 of	 the	 linear	 sled	 frame.	 	 Head	 restraint	 contact	 was	detected	with	 a	 force	 sensitive	 resistor	 (FSR;	Model	 406,	 Interlink	 Electronics,	 Camarillo,	 CA,	USA)	attached	to	the	front	of	the	head	restraint.		 All	 accelerometer,	 load	 cell	 and	 angular	 rate	 sensor	 signals	 were	 simultaneously	sampled	at	10kHz	using	a	National	Instruments	Data	Acquisition	(DAQ)	PXI	system	(PXI-4495	&					112	PXI-6289,	National	Instruments	Corporation,	Austin,	Texas,	USA)	and	a	custom-written	LabVIEW	virtual	instrument.	Optotrak	data	were	acquired	at	200	Hz	and	collection	was	triggered	by	the	DAQ	system	to	synchronize	the	data.	Subsequent	data	and	statistical	analyses	were	performed	using	Matlab	(R2017A,	Mathworks,	Natick,	MA,	USA).	All	data	channels	were	digitally	filtered	in	Matlab	and	 conformed	 to	 the	 SAE	 J211	 (Channel	Class	180	 for	 the	ATD	 sensors	 and	Channel	Class	60	for	the	sled	accelerometer)	(SAE,	1995).		5.2.3 Test	Procedures		 The	BioRID	II	ATD	was	clad	in	two	layers	of	 lycra	and	seated	on	the	Experimental	anti-whiplash	seat	(Figure	5.1)	or	one	of	four	different,	unmodified	front	passenger	seats:	A)	2005	Volvo	S40	WHIPS,	B)	2004	Volvo	S60	WHIPS,	C)	2005	Saab	9.3	SAHR,	and	D)	2004	Pontiac	Grand	Am	GMHR	 (Figure	 5.3).	 The	 seats	were	mounted	on	 a	 10	m	 long,	 feedback-controlled	 linear	sled	(Kollmorgen	IC55-100A7,	Waltham,	MA,	USA).	For	all	seats,	the	initial	seatback	angle	was	set	to	27	deg	rearward	from	vertical	(Siegmund	et	al.,	2005a).	To	maximize	the	repeatability	of	the	experiments,	the	initial	position	of	the	ATD	was	adjusted	to	a	pre-defined	posture	that	was	constant	within	each	 seat	and	as	 similar	 as	possible	between	 seats.	 This	pre-defined	posture	was	confirmed	using	the	3D	positions	measured	by	Optotrak	at	the	beginning	of	each	trial.	Seat	geometry	was	measured	before	and	after	each	collision	to	ensure	no	permanent	deformation	had	occurred.	After	each	test,	the	deformable	elements	in	the	WHIPS	seat	were	verified	to	have	not	 been	 damaged	 and	 the	 SAHR	 mechanism	 was	 verified	 to	 have	 returned	 to	 its	 original					113	position.	 The	 seatbelt	 was	 removed	 from	 the	 seat	 to	 prevent	 interactions	 that	 could	 affect	head,	neck	and	torso	kinematics.			 For	 all	 conditions,	 the	 ATD	 was	 exposed	 to	 five	 rear-end	 perturbations	 of	 increasing	speed	changes	(Δv	=	2,	4,	6,	and	8	km/h	with	a	Δt	of	147	ms,	and	12	km/h	with	a	Δt	of	187	ms)	(Figure	 5.3).	 Due	 to	 limitations	 in	 the	 force-generating	 ability	 of	 the	 linear	motors,	 a	 longer	pulse	duration	was	needed	for	the	12	km/h	speed	change.	All	of	the	pulses	used	here	have	a	longer	duration	than	the	RCAR/IIWPG	pulse	(Δt	=	91	ms)	(Insurance	Institute	of	Highway	Safety,	2008b).	The	sled	was	accelerated	 forward	 from	a	stationary	position	 for	 the	2,	4	and	6	km/h	speed	changes	and	 from	a	 constant	 rearward	 speed	of	6	 km/h	 for	 the	8	and	12	km/h	 speed	changes.	A	comparison	of	the	ATD	responses	for	an	8	km/h	collision	starting	from	a	stationary	position	to	an	8	km/h	collision	starting	from	a	6km/h	rearward	velocity	showed	less	than	10%	differences	 in	most	 of	 the	 occupant	 responses	 and	 neck	 injury	 criteria	 (see	 Chapter	 2).	 	 To	assess	the	repeatability	of	the	ATD	responses,	four	additional	repeated	trials	were	collected	at	the	8	and	12	km/h	collision	speed	changes	 for	both	 the	GMHR	and	Experimental	 seats	 (for	a	total	of	5	trials	for	each	seat	at	both	speed	changes).			 	For	 all	 seats	 the	horizontal	 position	 (dbackset)	 and	 vertical	 position	 (dheight)	 of	 the	head	relative	 to	 the	 head	 restraint	 before	 the	 collision	 onset	 were	 adjusted	 (SAHR,	 GMHR	 and	Experimental	seats:	midrange	of	its	vertical	adjustment	positions	and	most	rearward	horizontal	position;	WHIPS:	fixed	head	restraints)	and	rated	according	to	the	RCAR/IIWPG	seat	and	head	restraint	 evaluation	 protocol	 as	 good,	 acceptable,	 marginal	 or	 poor	 (Table	 5.1)	 (Insurance					114	Institute	of	Highway	Safety,	2008b,	Insurance	Institute	of	Highway	Safety,	2008a).			Both	WHIPS	seats	 had	 good-rated	 head-to-head-restraint	 geometry	 (S40	 WHIPS:	 dbackset	 =	 42.8	 mm	 and	dheight	=	-2.3	mm,	and	S60	WHIPS:	dbackset	=	69.2	mm	and	dheight	=	-42.2	mm)	and	the	SAHR’s	seat	(dbackset	=	89.8	mm	and	dheight	=	-44.7	mm)	had	an	acceptable-rated	geometry.	The	GMHR	seat	had	 a	marginal-rated	 head-to-head-restraint	 geometry	 (dbackset	 =	 93.5	mm	 and	 dheight	 =	 -26.9	mm)	 that	 would	 receive	 an	 overall	 RCAR/IIWPG	 rating	 of	 poor.	 	 The	 anti-whiplash	 seat	 was	rated	as	good	according	 to	 the	RCAR/IIWPG	criteria	with	a	pre-impact	dbackset	=	50.4	mm	and	dheight	=	1.5	mm.		Differences	in	pre-impact	dbackset	and	dheight	are	known	to	affect	the	occupant	responses	and	the	risk	of	whiplash	injuries	(Eriksson,	2005,	Nygren	et	al.,	1985,	Siegmund	et	al.,	1999,	Stemper	et	al.,	2006),	and	have	been	shown	to	potentially	confound	the	reduction	in	ATD	responses	 observed	on	 the	 anti-whiplash	 seat	 (see	 Chapter	 4).	 	 Thus,	 the	 comparison	of	 the	Experimental	anti-whiplash	seat	to	other	good-rated	seats	with	similar	head-to-head-restraint	geometry	(i.e.,	WHIPS	S40	and	S60)	will	be	used	to	determine	whether	the	observed	differences	in	ATD	responses	between	seats	are	caused	by	differences	in	head	restraint	geometry	or	by	the	active	control	of	seat	hinge	rotation	and	seatback	deformation.			 A	total	of	41	trials	were	collected:	13	trials	each	for	the	Experimental	and	GMHR	seats,	and	5	 trials	each	 for	 the	SAHR	and	two	WHIPS	seats.	 	Only	 the	results	of	 the	13	trials	on	the	GMHR	seat	have	been	previously	 reported	as	 the	Control	 condition	 in	Chapters	3	and	4.	The	trials	on	the	existing	anti-whiplash	seats	 (WHIPS,	SAHR,	and	GMHR	seats)	were	similar	 to	 the	trials	reported	in	Chapter	2,	but	utilized	different	collision	pulse	durations	for	the	same	speed					115	changes	(Chapter	2:	Δv	=	2	–	14	km/h	with	a	Δt	of	141	ms,	and	current	study:	Δv	=	2	–	8	km/h	with	 a	 Δt	 of	 147	ms	 and	 Δv	 =	 12	 km/h	with	 a	 Δt	 of	 187	ms).	 	 For	 test	 conditions	with	 five	repeated	 trials	 (i.e.,	 the	Δv	=	8	and	12	km/h	 trials	on	 the	GMHR	and	Experimental	 seat),	 the	average	of	the	peak	responses	across	the	five	trials	was	used	for	analysis.					Figure	5.3.	Photographs	of	the	experimental	set-up	with	the	BioRID	II	ATD	on	the	four	existing	anti-whiplash	seats:	A.)	2005	Volvo	S40	WHIPS,	B.)	2004	Volvo	S60	WHIPS,	C.)	2005	Saab	9.3	SAHR	and	D.)	2004	Pontiac	Grand	Am	GMHR		 		 	C.	D.	A.		B.					116		Figure	5.4.	Exemplar	sled	A.)	velocity	and	B.)	acceleration	pulses	for	increasing	collision	speeds	(Δv	of	2,	4,	6,	and	8	km/h	with	a	pulse	duration	(Δt)	of	147	ms,	and	Δv	of	12	km/h	with	a	Δt	of	187	ms;	lightest	to	darkest).	Collision	onset	occurred	at	time	=	0	ms.		This	figure	is	a	repeat	of	Figure	3.3.					 	Table	5.1.	Mean	(standard	deviation)	of	head	restraint	backset	(dbackset)	and	height	(dheight)	as	well	as	resulting	RCAR/IIWPG	static	seat	geometry	rating.		Seat	 n	 dbackset	(mm)	 dheight	(mm)	RCAR/IIWPG	Rating	Volvo	S40	WHIPS	 4	 42.8	(8.4)	 -2.3	(5.1)	 Good	Volvo	S60	WHIPS	 4	 69.2	(13.0)	 -42.2	(4.8)	 Good	Saab	9.3	SAHR	 3	 89.8	(6.0)	 -44.7	(3.6)	 Acceptable	Pontiac	Grand	Am	GMHR	 13	 93.5	(2.5)	 -26.9	(6.4)	 Marginal/Poor	Experimental	Anti-Whiplash	Seat	 13	 50.4	(3.1)	 1.5	(4.1)	 Good	Notes:	n	denotes	the	number	of	trials	for	each	seat	used	to	determine	mean	and	standard	deviation	backset	 and	 height	 values.	 Trials	 with	 missing	 Optotrak	 IRED	 markers	 were	 excluded	 from	 these	calculations.	Backset:	distance	between	the	back	of	the	head	to	the	front	face	of	the	head	restraint.		Height:	 distance	 between	 the	 top	 of	 the	 head	 to	 the	 top	 of	 the	 head	 restraint	 (negative	 values	indicated	the	top	of	the	head	restraint	was	lower	than	the	ATD	head).					117	5.2.4 Data	Analysis		 	Peak	 horizontal,	 forward,	 sled	 accelerations	 (aX-sled)	 were	 extracted	 directly	 from	 the	accelerometer	mounted	to	the	sled.		Head	and	torso	linear	accelerations	were	reported	in	local	ATD	head	and	T1	reference	frames	and	were	corrected	to	remove	the	earth’s	gravity	using	the	head	and	T1	orientations	determined	from	the	Optotrak	data.	Peak	linear	forward	accelerations	of	 the	head	(aX-head)	and	T1	vertebra	(aX-T1)	were	extracted	directly	 from	the	gravity-corrected	accelerometer	 data,	 and	 peak	 rotational	 velocities	 of	 the	 head	 (ωY-head)	 in	 the	 sagittal	 plane	were	extracted	directly	from	the	angular	rate	sensor.	Peak	upper	neck	shear	(FX)	and	axial	(FZ)	forces	and	the	flexion/extension	bending	moment	(MY)	were	determined	from	the	upper	neck	load	cell	and	reported	in	the	ATD	reference	frame	as	the	forces/moments	applied	by	the	neck	to	the	head.		Initial	head	angle	was	defined	as	the	average	angle	between	-300	ms	and	-200	ms	preceding	 the	 onset	 of	 aX-sled,	 and	 peak	 head	 extension	 angle	 (θhead)	 was	 defined	 as	 the	maximum	 rotation	of	 the	head	 into	extension	 relative	 to	 the	 initial	 head	angle.	A	 foreperiod	between	 -300	ms	 and	 -200	ms	 before	 the	 onset	 of	 forward	 aX-sled	 was	 used	 to	 define	 initial	values	because	the	pre-perturbation	deformation	of	the	seatback	was	initiated	200	ms	before	collision	 onset	 (tdeformation–onset	 =	 -200	ms).	 Peak	 retraction	 (Rx)	 was	 defined	 as	 the	maximum	horizontal	 displacement	 in	 the	 global	 reference	 frame	 of	 the	 AOJ	 with	 respect	 to	 the	 T1	vertebrae,	with	rearward	displacements	defined	as	negative	values.			 The	shape	of	 the	back	of	 the	ATD’s	head	and	front	of	 the	head	restraint	was	digitized	relative	 to	Optotrak	 IRED	markers.	These	 shapes	were	 then	used	 to	calculate	 the	 initial	head					118	restraint	backset	 (the	back	of	ATD’s	head	to	the	 front	 face	of	 the	head	restraint)	and	vertical	position	(the	top	of	ATD	head	to	the	top	of	the	head	restraint)	for	each	trial.	The	initial	backset	(dbackset)	 and	 height	 (dheight)	 of	 the	 head	 restraint	were	 defined	 as	 the	 average	 head-to-head-restraint	horizontal	and	vertical	distances	between	 -300	ms	and	 -200	ms	before	 the	onset	of	forward	aX-sled.	Negative	values	for	the	height	of	the	head	restraint	indicated	that	the	top	of	the	head	 restraint	 was	 lower	 than	 the	 top	 of	 the	 ATD	 head.	 The	 same	 head	 and	 head	 restraint	shapes	were	used	 to	calculate	 the	peak	 forward	rebound	 (drebound),	which	was	defined	as	 the	maximum	 forward	 head-to-head-restraint	 distance.	 Time-to-head-restraint	 contact	 (Δthead-contact)	was	 extracted	 from	 the	 time	 of	 force	 onset	 in	 the	 FSR	 attached	 to	 the	 head	 restraint	(onset	of	forward	sled	acceleration	defined	as	t	=	0	ms).	Onsets	of	sled	acceleration	and	head	restraint	 contact	 were	 determined	 when	 the	 accelerometer	 and	 FSR	 signals,	 respectively,	reached	1.5	times	the	peak	background	noise	level	present	between	-300	ms	to	-200	ms	before	the	onset	of	 the	collision	perturbation	 and	were	 confirmed	visually.	The	sled	accelerometer	 (aX-sled)	and	head	contact	FSR	signals	had	high	signal-to-noise	ratios	after	filtering	according	to	SAE	J211	Channel	class	180	and	did	not	require	further	manual	corrections	of	the	onsets.				 Three	neck	injury	criteria	(NICmax,	Nij,	and	Nkm)	were	computed	from	the	accelerometer	and	 load	 cell	 data.	 	 The	 Neck	 Injury	 Criterion	 (NICmax)	 was	 calculated	 from	 the	 relative	horizontal	 acceleration	 (corrected	 for	 gravity)	 and	 velocity	 in	 the	 global	 reference	 frame	between	 the	head	 center	 of	mass	 and	 the	 T1	 joint	 (Equation	 5.1;	 Bostrom	et	 al.,	 1996).	 The	Normalized	 Neck	 Injury	 Criterion	 (Nij)	 was	 calculated	 from	 the	 axial	 load	 (FZ)	 and	 the					119	flexion/extension	bending	moment	(MY)	measured	from	the	upper	neck	load	cell	(Equation	5.2;	critical	Fint	and	Mint	intercept	values	used	for	normalization:	Fint-tension	=	6806	N,	Fint-compression	=	-6160	N,	Mint-flexion	=	310	N,	and	Mint-extension	=	-135	N;	Eppinger	et	al.,	1999,	Eppinger	et	al.,	2000).	The	Neck	Protection	Criterion	 (Nkm)	was	 calculated	 from	 the	 sagittal	 shear	 force	 (FX)	 and	 the	flexion/extension	bending	moment	(MY)	measured	from	the	upper	neck	load	cell	(Equation	5.3;	critical	Fint	and	Mint	intercept	values	used	for	normalization:	Fint-shear-anterior	=	845	N,	Fint-shear-posterior	=	845	N,	Mint-extension	=	47.5	Nm,	and	Mint-flexion	=	88.1	Nm;	Schmitt	et	al.,	2001,	Schmitt	et	al.,	2002).	 	 Peak	 values	 of	 the	 three	 neck	 injury	 criteria	 were	 then	 extracted	 for	 analysis	 and	compared	to	proposed	injury	thresholds	(NIC:	15	m2/s2,	Eichberger	et	al.,	1998;	Nij	and	Nkm	:	a	normalized	value	of	one,	Schmitt	et	al.,	2001,	Schmitt	et	al.,	2002,	Eppinger	et	al.,	1999).		These	injury	thresholds	correspond	to	different	human	tolerance	levels	for	the	causation	of	whiplash-related	injuries	–	NIC:	long-term	whiplash-associated	disorders	(WAD)	levels	1-3	(Bostrom	et	al.,	2000,	Bostrom	et	al.,	1996),	Nij:	22%	risk	of	abbreviated	injury	scale	(AIS)	level	3	(i.e.,	fracture;	Eppinger	et	al.,	1999,	Eppinger	et	al.,	2000),	and	Nkm:	AIS	 level	1	 injury	causation	 (i.e.,	minor	injury;	Schmitt	et	al.,	2001,	Schmitt	et	al.,	2002).	NICmax t = maximumfirst	150	ms #$%& ' ∙ ) + (,$%&(')). 	 …	 Equation	5.1	N01(t) = 23(4)2567 + 89:;(4)8567 	 	 	 	 	 …	 Equation	5.2	N<=(t) = 2>(4)2567 + 89:;(4)8567 	 	 	 	 	 …		 Equation	5.3							120		 The	repeatability	of	 the	ATD	on	the	Experimental	seats	was	assessed	for	the	8	and	12	km/h	 speed	 changes	 using	 the	 Coefficient	 of	 Variation	 (COV),	 which	 was	 calculated	 as	 the	standard	 deviation	 (SD)	 divided	 by	 the	 mean	 and	 expressed	 as	 a	 percentage.	 The	 National	Highway	Traffic	Safety	Administration	(NHTSA)	defines	a	COV	of	5%	or	less	as	excellent,	a	COV	of	10%	or	less	as	acceptable	and	a	COV	greater	than	10%	as	poor	(Rhule	et	al.,	2005).	In	Chapter	3,	 the	repeatability	of	 the	ATD	on	the	Control	GMHR	seat	was	determined	to	be	excellent	or	acceptable	for	most	parameters	at	the	8	and	12	km/h	speed	changes,	except	for	head	restraint	height	(dheight:	COV	=	22.2%	and	11.3%	at	8	and	12	km/h	respectively)	(Table	5.2).		The	COV	of	all	ATD	responses	and	neck	injury	criteria	were	compared	to	determine	if	the	repeatability	was	different	between	the	Control	and	the	Experimental	seats.	Because	COV	 is	dependent	on	the	mean	 and	 SD	 of	 the	 responses,	 the	 expected	 decreases	 in	mean	 responses	 (which	 form	 the	denominator	in	the	COV	calculation)	for	the	Experimental	seat	may	generate	larger	COV	values.		A	comparison	of	the	SD	values	between	the	Control	and	the	Experimental	seats	would	provide	an	 alternate	 assessment	 of	 variability	 and	would	 help	 to	 determine	whether	 the	 larger	 COV	values	 for	 the	Experimental	 seat	were	 likely	due	 to	 the	 lower	mean	 responses	or	due	 to	 the	greater	 inherent	variability	 (i.e.,	SD	 in	the	ATD	responses.	The	SD	values	for	all	kinematic	and	kinetic	responses	as	well	as	neck	criteria	observed	at	both	the	8	and	12	km/h	collision	speeds	were	used	in	this	assessment	of	variability.	the	COV	and	SD	were	assumed	to	be	independent	observations	and	a	Wilcoxon	Signed-Rank	Test	was	performed	after	a	Kolmogorov-Smirnov	test	determined	 the	COV	and	 SD	 values	 (for	 a	 given	 collision	 severity	 or	 seat)	were	 not	 normally					121	distributed.	All	tests	were	performed	using	predefined	functions	(kstest	and	signrank)	in	Matlab	(R2017A,	Mathworks,	Natick,	MA,	USA)	at	a	significance	level	a	=	0.05.					To	compare	ATD	responses	between	the	Experimental	and	existing	anti-whiplash	seats	(GMHR,	SAHR	and	WHIPS),	99th	percentile	predictive	 intervals	were	 created	around	 the	peak	responses	 on	 the	 Experimental	 and	 GMHR	 seats	 using	 the	maximum	 variability	 (COV	 value)	observed	from	repeated	trials	at	either	the	8	or	12	km/h	collision	severity	(Horslen	et	al.,	2015).		Ninety-ninth	 (99th)	 percentile	 predictive	 intervals	 were	 not	 created	 around	 the	 WHIPS	 and	SAHR	anti-whiplash	seats	because	no	repeated	trials	were	collected	for	these	seats.	 	The	COV	value	 was	 then	 multiplied	 by	 2.58	 (z-score	 for	 two-tailed	 99th	 percentile)	 to	 estimate	 the	predictive	 limits.	 Since	 the	 COV	 values	 were	 defined	 as	 the	 maximum	 standard	 deviations	divided	by	the	mean	of	the	peak	ATD	responses,	these	predictive	limits	were	represented	as	a	percentage	 and	 then	multiplied	 by	 the	 peak	 responses	 to	 create	 predictive	 intervals	 at	 each	collision	 severity	 (2	 –	 12	 km/h).	 	 Any	ATD	 responses	 from	 the	 single-trial	 anti-whiplash	 seats	(i.e.,	 WHIPS	 and	 SAHR	 seats)	 located	 outside	 the	 99th	 percentile	 predictive	 intervals	 of	 the	Experimental	seat	were	assumed	to	be	statistically	different.	Statistical	differences	between	the	Experimental	 and	GMHR	 seats	were	 determined	 if	 their	 respective	 99th	 percentile	 predictive	intervals	 did	 not	 overlap.	 	 The	 99th	 percentile	 predictive	 limits	were	 used	 to	 account	 for	 the	numerous	 comparisons	 between	 the	 Experimental	 and	 existing	 anti-whiplash	 seats	 for	 each	ATD	 responses	 at	 each	 collision	 severity	 and	 to	provide	 a	 conservative	 estimate	of	 statistical	significance.					122	5.3 Results			 The	 combination	of	 the	 seat	hinge	 rotation	and	 seatback	 cushion	deformation	on	 the	Experimental	seat	generated	ATD	kinematic	and	kinetic	responses	that	were	different	from	the	other	 existing	 anti-whiplash	 seats.	 The	ATD	 responses	 on	 the	WHIPS,	 SAHR	 and	GMHR	 seats	exhibited	positively	graded	responses	to	increasing	speed	change	in	most	kinematic	(aX-head,	aX-T1,	ωY-head,	θhead,	RX,	dpenetration-T1,	dpenetration-pelvis	and	drebound)	and	kinetic	(FX,	FZ	and	MY)	responses	(Figures	5.5,	5.6	and	5.7,	see	also	Figures	E5.2	–	E5.5	in	the	Appendix	E	and	Chapter	2).	On	the	Experimental	 seat,	 with	 the	 exception	 of	 aX-T1,	 dpenetration-T1,	 and	 dpenetration-pelvis	 that	 showed	 a	positively	 graded	 response	 to	 collision	 severity,	 most	 kinematic	 and	 kinetic	 responses	 were	attenuated	to	a	near	constant	amplitude	within	each	response	across	all	severities	(Figure	5.5,	56	and	5.7,	 see	also	Figure	E5.1	 in	 the	Appendix	E).	The	pre-perturbation	forward	seat	hinge	rotation	(tforward-onset=	 -90	ms)	and	the	seatback	cushion	deformation	(tperturbation-onset=	 -200	ms)	shifted	the	head-to-head-restraint	contact	times	to	before	the	onset	of	collision	(Δthead-contact	=	-19.0	±	3.9	ms;	 Figure	5.6	and	 5.7G).	With	 the	head	 restraint	 supporting	 the	head	before	 the	collision,	peak	rearward	head	angle	(θhead	=	2.2	±	0.3	deg)	and	rearward	head	retraction	(RX	=	-13.4	±	3.4	mm)	remained	constant	across	all	collision	severities	(Figure	5.7H	and	5.7I).			 In	 comparison	 to	 the	poor-rated	GMHR	seats,	 there	was	no	overlap	between	 the	99th	percentile	predictive	intervals	of	the	kinematic	responses	in	the	GMHR	and	Experimental	seats	for	collision	speed	changes	of	4	km/h	or	greater.	Peak	 forward	head	 (aX-head)	and	 torso	 (aX-T1)	accelerations	were	50	–	77%	lower	and	34	–	65%	lower,	respectively,	in	the	Experimental	seat					123	than	in	the	GMHR	seat.	Peak	head	angular	velocity	(ωY-head)	was	similarly	71	–	73%	lower	in	the	Experimental	 seat	 (Figure	 5.7A	 –	 5.7C).	 	 Head-to-head-restraint	 contact	 time	occurred	 131	 –	182	ms	earlier	to	support	the	head	before	the	collision	onset	(Δthead-contact	=	-19.0	±	3.9	ms)	and	to	limit	the	displacements	of	the	head	and	neck	(θhead:	83	–	89%	lower,	RX:	70	–	82%	lower	and	drebound:	51	–	85%	lower,	Figure	5.7G	–	5.7I	and	57L).	The	kinetic	responses	on	the	Experimental	seat	 were	 significantly	 lower	 than	 the	 GMHR	 seat	 for	 collision	 speed	 changes	 of	 6	 km/h	 or	greater	(FX:	71	–	73%	lower,	FZ:	77	–	89%	lower,	and	MY:	77	–	81%	lower)	(Figure	5.7D	–	5.7F).			 For	collision	speeds	4	km/h	or	greater,	the	Experimental	seat	significantly	decreased	all	kinematic	 responses	 compared	 to	 the	 good-rated	 seats	 (WHIPS	and	 SAHR)	by	25	–	89%.	 The	observed	reductions	in	peak	kinematic	responses	were	55	–	78%	for	aX-head,	29	–	70%	for	aX-T1,	59	 –	 75%	 for	 ωY-head,	 78	 –	 89%	 for	 θhead,	 31	 –	 86%	 for	 drebound,	 and	 25	 –	 79%	 for	 RX.	 The	Experimental	seat	also	reduced	some	ATD	kinetic	responses	compared	the	good-rated	seats	for	collision	speeds	6	km/h	and	higher	(Figure	5.7E	and	5.7F):	FZ	was	58	–	84%	lower	and	MY	was	28	–	74%.	However,	the	MY	response	for	Volvo	S40	WHIPS	and	Experimental	seats	at	the	8	and	12	km/h	collision	severities	showed	no	differences.	For	the	8	and	12	km/h	speed	changes,	the	peak	upper	neck	shear	force	(FX)	was	larger	for	the	Experimental	seat	(57.8	and	76.1	N,	respectively)	than	for	the	WHIPS	and	SAHR	seats	(WHIPS	S40:	28.5	and	8.6	N,	WHIPS	S60:	1.8	and	2.1	N,	and	SAHR:	 19.4	 and	 1.7	 N,	 respectively),	 but	was	 still	 lower	 than	 for	 the	 GMHR	 seat	 (213.5	 and	264.4	N,	respectively;	Figure	5.7D).	 					124	Across	 all	 seats	 and	 collision	 severities,	 none	of	 the	 neck	 injury	 criteria	 (NICmax,	Nij	 and	Nkm)	 exceeded	 the	 proposed	 injury	 thresholds.	 Positively	 graded	 responses	 to	 increasing	collision	 severity	 for	 all	 three	neck	 injury	 criteria	were	observed	 for	 all	 existing	anti-whiplash	seats,	whereas,	the	responses	of	the	Experimental	seat	were	constant	across	collision	severities	(NICmax	=	2.79	±	0.65	m2/s2,	Nij	=	0.041	±	0.004,	and	Nkm	=	0.15	±	0.03;	Figure	5.7M	–	5.7O).	With	the	exception	of	Nij	(GMHR	and	S40	WHIPS	at	the	4	km/h	severity)	and	Nkm	(S60	WHIPS	at	the	8	km/h	severity),	the	three	neck	injury	criteria	for	the	Experimental	seat	were	9	–	73%	lower	for	collision	speeds	greater	than	2	km/h	(NICmax:	9	–	71%	lower,	Nij:	23	–	70%	lower,	and	Nkm:	16	–	73%	lower).						 The	repeatability	of	the	BioRID	II	ATD	responses	for	the	GMHR	seat	have	been	reported	in	 Chapter	 3	 and	 were	 excellent	 or	 acceptable	 at	 both	 speed	 changes	 for	most	 parameters	except	for	head	restraint	height	(dheight,	COV	=	22.2%	and	11.3%	at	8	and	12	km/h	respectively).	The	 COV	 values	 from	 the	 repeated	 trials	 on	 the	 Experimental	 seat	were	more	 variable	 than	those	previously	observed	on	 the	GMHR	seat.	COV	values	 for	peak	accelerations	of	 the	head	(aX-head),	 torso	 (aX-T1)	 and	 sled	 (aX-sled),	 and	 occupant	 penetration	 at	 both	 the	 T1	 and	 pelvis	(dpenetration-T1	and	dpenetration-pelvis,	respectively)	responses	were	considered	excellent	(COV	≤	5%);	whereas,	the	neck	shear	force	(FX),	initial	backset	(dbackset),	forward	rebound	(drebound),	and	neck	injury	criteria	Nij	and	Nkm	were	considered	acceptable	(5%	<	COV	≤	10%).	 	All	other	responses	had	COV	values	greater	10%	and	were	considered	poor	(Δthead-contact	=	54.3%,	FZ	=	25.6%,	MY	=	13.1%,	ωY-head	=	13.0%,	θhead	=	19.9%,	RX	=	10.8%,	NICmax	=	14.8%,	and	dheight	=	199.1%).	The	COV	values	on	the	Experimental	seat	were	significantly	higher	than	those	on	the	GMHR	seat	for	the					125	8	and	12	km/h	speed	changes	(Wilcoxon	Signed-Rank	Test;	8	km/h:	Z	=	3.29,	p	=	0.001,	and	12	km/h:	Z	=	3.72,	p	=	0.0002).	The	standard	deviations	(SD)	of	the	Experimental	and	GMHR	seats	were	not	significantly	different	(Wilcoxon	Signed-Rank	Test;	8	km/h:	Z	=	-0.89,	p	=	0.37,	and	12	km/h:	Z	=	0.72,	p	=	0.47).		The	larger	COV	values	on	the	Experimental	seat	with	similar	SD	values	to	the	GMHR	seat	suggested	that	the	COV	values	considered	poor	(>	10%)	were	more	likely	due	to	decreased	mean	responses	(which	form	the	denominator	in	the	COV	calculation)	than	due	to	increased	variability	in	the	ATD	responses	(i.e.	SD	values).		 					126		Figure	 5.5.	 Exemplar	 BioRID	 II	 ATD	 response	 data	 comparing	 the	 Experimental	 anti-whiplash	seat	 (black	 line),	 a	 RCAR/IIWPG	 good-rated	 (Volvo	 S40	WHIPS;	 red	 line)	 and	 a	 RCAR/IIWPG	poor-rated	(Pontiac	Grand	Am	GMHR;	blue	 line).	Each	column	represents	occupant	responses	while	exposed	to	various	collision	severities	(Δv	of	2,	4,	6	and	8	with	a	collision	pulse	duration	(Δt)	of	147	ms	and	Δv	of	12	km/h	with	a	Δt	=	187	ms).	Hollow	circles	 represent	 the	onset	of	head-to-head-restraint	contact	and	peak	responses	of	each	ATD	response	parameter	for	each	trial.	Similar	plots	for	all	seats	are	given	in	the	Appendix	E:	Figures	E5.1	–	E5.5.	 					127		Figure	5.6.	Exemplar	BioRID	 II	ATD	 response	data	comparing	 the	seat	hinge	 rotation,	backset	and	occupant	penetration	at	the	T1	and	pelvis	for	the	on	the	Experimental	anti-whiplash	seat	(black	 line),	 a	 RCAR/IIWPG	 good	 rated	 (Volvo	 S40	WHIPS;	 red	 line)	 and	 a	 RCAR/IIWPG	 poor	rated	(Pontiac	Grand	Am	GMHR;	blue	line).	Each	column	represents	occupant	responses	while	exposed	to	various	collision	severities	(Δv	of	2,	4,	6	and	8	with	a	collision	pulse	duration	(Δt)	of	147	ms	and	Δv	of	12	km/h	with	a	Δt	=	187	ms).	The	grey	 lines	represent	the	input	seat	hinge	rotation	and	seatback	motor	deformation	profiles	for	the	Experimental	anti-whiplash	seat.						128		Figure	5.7.	Experimental	peak	kinematic	and	kinetic	responses	for	the	BioRID	II	ATD	seated	on	the	WHIPS,	SAHR,	GMHR	and	Experimental	seats.	For	all	 the	graphs,	red	circles	represent	the	Volvo	 S40	WHIPS	 seat,	 blue	 triangles	 represent	 the	 Volvo	 S60	WHIPS	 seat,	 green	 diamonds	represent	the	SAHR	seat,	black	squares	represent	the	GMHR	seat	and	magenta	stars	represent	the	 Experimental	 seat.	 The	 faded	 black	 and	 magenta	 bars	 represent	 the	 99th	 percentile	predictive	intervals	for	the	GMHR	and	Experimental	seats,	respectively.	Y-axis	values	for	panels	6I,	6J	and	6K	have	been	inverted	for	visual	purposes	to	show	increasing	responses	from	bottom	to	top.						129		Table	5.2.	Mean	(standard	deviation)	and	Coefficient	of	Variation	(COV)	from	five	repeated	perturbations	on	the	GMHR	and	Experimental	seats	at	Δv	=	8	km/h,	Δt	=	147	ms	and	Δv	=	12	km/h,	Δt	=	187	ms.	The	underlined	COV	values	indicate	a	COV	rating	of	acceptable	(5%	≤	COV	<	10%),	the	bolded	COV	values	indicate	a	COV	rating	of	poor	(COV	>	10%),	and	all	other	values	are	rated	good	(COV	<	5%).	Automotive	Seat	 GM	Pontiac	Grand	AM	(GMHR)	 Experimental	Anti-Whiplash	Seat	Δv	(km/h)	 8	km/h,	Δt	=	147	ms	 12	km/h,	Δt	=	187	ms	 8	km/h,	Δt	=147	ms	 12	km/h,	Δt	=	187	ms	Parameter	 Mean	(SD)	 COV	(%)	 Mean	(SD)	 COV	(%)	 Mean	(SD)	 COV	(%)	 Mean	(SD)	 COV	(%)	Initial	dbackset	(mm)	 94.3	(1.9)	 2.0	 94.5	(2.6)	 2.8	 50.4	(1.8)	 3.5	 51.5	(4.8)	 9.3	Initial	dheight	(mm)	 -23.8	(5.3)	 22.2	 -25.0	(2.8)	 11.3	 5.3	(3.2)	 60.4	 -1.4	(2.8)	 199.1	aX-sled	(m/s2)	 22.8	(1.1)	 4.7	 31.6	(0.4)	 1.2	 21.5	(0.5)	 2.3	 31.6	(0.4)	 1.4	Δthead	contact	(ms)	 124.3	(1.1)	 0.9	 109.4	(1.6)	 1.4	 -16.5	(9.0)	 54.3	 -26.1	(5.7)	 5.7	dpenetration-T1	(mm)	 -58.1	(2.2)	 3.8	 -67.9	(1.0)	 1.5	 -52.2	(0.8)	 1.5	 -61.6	(1.2)	 2.0	dpenetration-pelvis	(mm)	 -38.9	(1.1)	 2.9	 -55.8	(1.2)	 2.2	 -78.8	(1.8)	 2.3	 -94.1	(2.8)	 3.0	aX-head	(m/s2)	 85.2	(1.0)	 1.1	 120.7	(1.7)	 1.4	 19.5	0.5)	 2.8	 35.1	(1.4)	 4.1	aX-T1	(m/s2)	 44.1	(0.5)	 1.0	 47.0	(0.3)	 0.6	 18.8	(0.8)	 4.1	 31.2	(1.3)	 4.1	FX	(N)	 213.5	(3.8)	 1.8	 264.4	(14.1)	 5.4	 57.8	(5.5)	 9.5	 76.1	(4.2)	 5.5	FZ	(N)	 414.4	(9.3)	 2.2	 586.1	(16.2)	 2.8	 72.5	(10.0)	 13.7	 63.5	(17.4)	 27.2	MY	(Nm)	 20.1	(1.2)	 6.1	 25.1	(0.1)	 0.2	 4.3	(0.5)	 10.9	 4.9	(0.6)	 13.1		ωY-head(deg/s)	 329.6	(13.5)	 4.1	 390.2	(11.1)	 2.8	 91.5	(11.9)	 13.0	 98.7	(11.6)	 11.8	θ	head	(deg)	 15.7	(0.4)	 2.3	 16.5	(0.6)	 3.8	 2.0	(0.3)	 12.3	 2.2	(0.4)	 19.9	RX	(mm)	 -61.8(1.1)	 1.8	 -64.4	(2.3)	 3.6	 -11.1	(1.0)	 9.0	 -19.2	(2.1)	 10.8	drebound	(mm)	 335.1	(10.4)	 3.5	 413.4	(7.7)	 1.9	 51.3	(2.0)	 3.8	 201.3	(15.2)	 7.5	NICmax	(m2/s2)	 6.9	(0.2)	 3.1	 9.8	(0.2)	 1.7	 2.7	(0.4)	 14.8	 3.6	(0.5)	 14.2	Nij	 0.12	(0.00)	 2.3	 0.16	(0.00)	 1.5	 0.04	(0.00)	 9.3	 0.05	(0.00)	 5.5	Nkm	 0.48	(0.02)	 3.4	 0.60	(0.02)	 2.9	 0.15	(0.01)	 6.6	 0.18	(0.01)	 5.8					130	5.4 Discussion		 The	goal	of	this	study	was	to	evaluate	the	performance	of	the	Experimental	seat	against	four	existing	anti-whiplash	seats	(WHIPS	S40,	WHIPS	S60,	SAHR,	and	GMHR)	across	a	range	of	speed	 changes.	 Existing	 anti-whiplash	 seats	 yielded	 positively	 graded	 ATD	 responses	 to	increased	collision	 severity	while	 the	Experimental	anti-whiplash	 seat	evoked	nearly	 constant	ATD	responses	across	 the	 tested	speed	changes	 (2	 to	12	km/h).	With	 the	exception	of	upper	neck	shear	force	(FX),	the	Experimental	seat	performed	better	than	existing	anti-whiplash	seats	at	reducing	ATD	kinematics,	kinetics,	and	neck	injury	criteria	for	collision	severities	between	4	and	12	km/h.	At	the	2km/h	speed	change,	the	Experimental	seat	lowered	θhead,	RX,	and	Δthead-contact	in	comparison	to	the	existing	anti-whiplash	seats,	but	the	other	ATD	responses	and	neck	injury	criteria	were	similar	in	amplitude	across	all	tested	seats.	Epidemiological	studies	have	not	shown	whiplash	 injury	at	 speed	changes	as	 low	as	2	km/h	 (Bartsch	et	al.,	 2008,	Krafft,	2002,	Krafft	et	al.,	2005)	and,	thus	the	lack	of	a	difference	between	the	seats	at	2km/h	collision	speed	was	not	considered	further.	The	results	of	this	study	showed	that	combining	seat	hinge	rotation	and	seatback	cushion	deformation	 reduces	peak	ATD	responses	and	potentially	mitigates	 the	risk	of	whiplash	injuries	following	low-speed,	rear-end	collisions	(Δv	=	4	–	12	km/h).				 The	 Experimental	 seat	 appeared	 to	 achieve	 its	 benefits	 in	 multiple	 ways.	 	 The	 pre-impact	 forward	 rotation	 of	 the	 seatback	 and	 the	 deformation	 of	 the	 seatback	 allowed	 the	occupant	to	pocket	into	the	seatback	while	rotating	the	head	restraint	towards	the	occupant’s	head	to	decrease	the	backset	distance.	Head	restraint	contact	occurred	before	collision	onset,					131	which	 reduced	 the	 relative	 movement	 of	 the	 head	 and	 neck,	 and	 attenuated	 the	 peak	kinematic	 and	 kinetic	 responses.	 Another	 benefit	 of	 the	 forward	 seat	 hinge	 rotation	was	 an	initial	 increase	 in	effective	 stiffness	of	 the	 seat	hinge	 to	 resist	 the	 rearward	deflection	of	 the	seatback	created	by	the	occupant’s	 inertia.	 	A	third	benefit	of	the	Experimental	seat	was	that	the	seat	hinge	and	seatback	motors	prolonged	the	ATD-seatback	interaction	and	did	not	return	to	 their	 original	 configurations	 after	 the	 impact.	 This	 seat	 behaviour	 reduced	 the	 energy	typically	returned	to	the	occupant	during	the	rebound	or	restitution	phase	of	the	occupant-seat	interaction.	Together	these	benefits	of	actively	controlling	the	seat	hinge	rotation	and	seatback	deformation	profiles	generated	robust	decreases	in	most	kinematic	and	kinetic	responses	and	could	potentially	reduce	the	risk	of	whiplash	injuries.					The	 Experimental	 seat	 attenuated	 all	 ATD	 responses	 and	 neck	 injury	 criteria	 in	comparison	 to	 the	 poor-rated	 GMHR	 seat,	 and	 most	 ATD	 responses	 (with	 the	 exception	 of	upper	neck	shear	force:	FX)	and	neck	injury	criteria	in	comparison	to	the	good-rated	WHIPS	and	SAHR	 seats.	 	 Epidemiological	 studies	 have	 estimated	 that	 current	 anti-whiplash	 seats	 (SAHR,	WHIPS	 and	 GMHR)	 reduce	 the	 risk	 of	 whiplash	 injuries	 following	 a	 rear-end	 collision	 by	approximately	50%	(Jakobsson	et	al.,	2008,	Kullgren	and	Krafft,	2010,	Viano	and	Olsen,	2001),	with	 the	 good-rated	 seats	 exhibiting	 a	 further	 11	 –	 15%	 decline	 in	 occupant	 neck	 injury	 risk	compared	 to	 poor-rated	 seats	 (Farmer	 et	 al.,	 2008,	 Trempel	 et	 al.,	 2016).	 In	 addition,	 no	whiplash	 associated	 disorders	 (WAD)	 symptoms	have	 been	 reported	 in	 field	 data	 at	 collision	severities	≤	4	km/h	while	a	20%	risk	of	WAD	symptoms	lasting	more	than	one	month	has	been					132	observed	for	8	km/h	collisions	(Krafft	et	al.,	2005,	Bartsch	et	al.,	2008).	Since	the	Experimental	seat	 reduced	most	 of	 the	 occupant	 responses	 at	 8	 km/h	 and	 12	 km/h	 speed	 change	 to	 the	levels	observed	at	2	and	4	km/h,	it	can	be	postulated	that	the	Experimental	anti-whiplash	seat	could	potentially	reduce	the	overall	20%	WAD	risk	at	8	km/h	collisions	to	near	zero.	Although	these	 results	 demonstrated	 the	 ability	 of	 the	 anti-whiplash	 seat	 to	 attenuate	 peak	 ATD	responses	(with	the	exception	of	FX)	better	than	current	anti-whiplash	seats,	real-world	testing	of	 the	 interventions	built	 into	 the	Experimental	anti-whiplash	seat	 is	 required	to	quantify	 the	actual	reduction	in	the	risk	of	whiplash	injuries	following	low-speed,	rear-end	collisions.		Across	all	tested	seats,	the	initial	head	restraint	backset	was	different,	with	RCAR/IIWPG	ratings	 ranging	 from	 good	 to	marginal:	 good	 (dbackset	 ≤	 70	mm):	 S40	WHIPS,	 S60	WHIPS	 and	Experimental	 seat,	 acceptable	 (70	 mm	 <	 dbackset	 ≤	 90	 mm):	 	 SAHR	 seat,	 and	 marginal/poor	(dbackset	>	90	mm):	GMHR.	In	comparison	to	the	good-rated	WHIPS	seats	(S40	WHIPS:	dbackset	=	42.8	mm	and	dheight	=	-2.3	mm,	and	S60	WHIPS:	dbackset	=	69.2	mm	and	dheight	=	-42.2	mm),	the	Experimental	seat	(dbackset	=	50.4	mm	and	dheight	=	1.5	mm)	decreased	most	ATD	response	(with	the	exception	of	FX)	by	25	–	89%	and	neck	injury	criteria	by	9	–	73%	for	collision	speeds	4	km/h	or	higher	despite	similar	initial	head	restraint	backsets.		The	decreased	ATD	responses	observed	on	 the	 Experimental	 seat	 may	 be	 due	 to	 the	 pre-impact	 forward	 rotation	 of	 the	 hinge	 and	seatback	 deformation	 to	 decrease	 the	 head-to-head-restraint	 distance,	 time-to-head-contact	and	the	relative	movement	between	the	head,	neck	and	torso.	 	The	head	contacted	the	head	restraint	19.0	±	3.9	ms	before	collision	onset.	This	 factor	 in	 isolation	could	 lead	 to	decreased					133	risk	of	whiplash	injuries.	Indeed,	the	risk	of	whiplash	injury	was	found	to	decrease	by	10%	for	every	25	mm	decrease	in	dbackset,	and	was	the	lowest	when	the	dbackset	=	0	mm	(Eriksson,	2005).	Thus,	decreasing	the	backset	from	50.4	mm	to	0	mm	prior	to	collision	onset	could	potentially	decrease	 the	 risk	 of	 neck	 injury	 by	 20%.	 Future	 testing	 is	 required	 to	 determine	 the	 relative	contribution	 of	 initial	 head-to-head-restraint	 contact	 prior	 to	 the	 collision,	 the	 seat	 hinge	rotation,	 and	 the	 seatback	 cushion	 deformation	 on	 the	 peak	 ATD	 responses	 to	 the	 risk	 of	whiplash	injury	following	rear-end	collisions.	 	All	 existing	 anti-whiplash	 seats	 generated	 positively	 graded	 responses	 to	 increasing	speed	changes	in	most	ATD	responses	(aX-head,	aX-T1,	ωY-head,	FX,	FZ,	MY,	θhead,	RX,	and	drebound)	and	neck	injury	criteria	(NICmax,	Nij,	Nkm).	Similar	to	a	previous	study	(Chapter	2),	the	ATD	responses	on	the	good-rated	seats	(WHIPS	and	SAHR)	attenuated	only	four	peak	responses	(FX:	47	-	99%	reduction,	FZ:	19	-	68%	reduction,	MY:	18	-	75%	reduction,	RX:	19	–	60%	reduction)	compared	to	the	RCAR/IIWPG	poor-rated	seat	(GMHR).	These	four	variables	(FX,	FZ,	MY	and	RX)	attenuated	by	the	WHIPS	and	SAHR	seats	had	the	most	overlap	with	the	99th	percentile	predictive	intervals	of	the	Experimental	seat	responses.	At	the	8	and	12	km/h	collision	speeds,	the	WHIPS	and	SAHR	seats	 attenuated	 the	 peak	 FX	 responses	 more	 than	 the	 Experimental	 seat.	 	 The	 possible	activation	 of	 the	 WHIPS	 and	 SAHR	 safety	 mechanisms	 at	 these	 higher	 collision	 severities	transitioned	 the	 FX	 responses	 from	 positive	 to	 negative	 peaks,	 reducing	 the	 rearward	 neck	loading	 and	 adopting	 a	 forward	 flexion	 position	 to	 reduce	 the	 risk	 of	 whiplash	 injuries.	 The	WHIPS	and	SAHR	seats	are	equipped	with	dynamic	anti-whiplash	devices	that	rely	on	occupant					134	loading	of	the	seatback	to	deform	a	recliner	mechanism	that	controls	seatback	translation	and	rotation	(WHIPS)	(Jakobsson	et	al.,	2008,	Jakobsson	et	al.,	2000)	or	to	move	the	head	restraint	upward	 and	 forward	 (SAHR)	 (Viano	 and	Olsen,	 2001).	 Thus,	 the	WHIPS	 and	 SAHR	 seats	may	require	 increased	 occupant	 loading	 of	 the	 seatback	 during	 higher	 collision	 speeds	 (Δv	 >	 12	km/h)	to	 fully	activate	their	 respective	safety	mechanism	and	reduce	the	FX	responses.	These	results	suggest	that	these	seats	may	have	been	optimized	for	higher	collision	pulses	closer	to	the	industry	standard	RCAR/IIWPG	16	km/h	test	pulse	(Δt	=	91	ms)	and	were	not	designed	to	attenuate	ATD	response	during	the	lower	speed	changes	used	in	this	study	(Δv	≤	12	km/h).	The	inability	of	the	WHIPS	and	SAHR	seats	to	attenuate	ATD	responses	at	lower	speed	changes	may	explain	why	existing	anti-whiplash	seats	are	currently	only	50%	effective	at	reducing	the	risk	of	whiplash	injury.			The	COV	values	of	the	BioRID	II	ATD	responses	were	used	to	test	the	repeatability	of	the	ATD	on	both	the	GMHR	and	Experimental	seats.	Compared	to	the	GMHR	seat,	the	COV	values	for	the	Experimental	seat	were	more	variable,	with	8	of	the	18	COV	values	rated	as	poor	(COV	>	10%)	 at	 the	 8	 and	 12	 km/h	 speed	 changes.	 Although	 the	 COV	 values	 were	 higher	 for	 the	Experimental	seat	than	for	the	GMHR	seat	at	both	the	8	and	12	km/h	speed	changes,	the	SD	values	were	at	similar	magnitudes	between	seats.		Consequently,	the	larger	COV	values	for	the	Experimental	seat	were	likely	due	to	lower	mean	values	(which	form	the	denominators	in	the	COV	calculation)	than	due	to	greater	variability	(i.e.,	SD	values)	in	the	ATD	responses.	Thus,	we					135	believe	that	the	Experimental	seat	had	sufficient	repeatability	to	justify	the	comparisons	made	to	the	other	anti-whiplash	seats	in	this	study.	The	 collision	 severities	 used	 in	 this	 study	 were	 less	 severe	 than	 the	 RCAR/IIWPG	standard	 crash	 test	 pulse	 (16	 km/h,	 Δt	 =	 91	 ms)	 but	 provided	 a	 graded	 range	 of	 collision	velocities	applicable	to	some	real-world	collisions	that	cause	whiplash	injury	(Krafft	et	al.,	2005,	Kullgren	et	al.,	2007).	A	collision	pulse	duration	of	Δt	=	147	ms	was	selected	for	collision	speeds	up	to	8	km/h	because	it	was	similar	to	the	135	ms	observed	in	prior	vehicle-to-vehicle	rear-end	crashes	(Brault	et	al.,	1998).		However,	a	collision	pulse	duration	of	Δt	=	187	ms	was	needed	to	reach	a	12	km/h	collision	speed	due	to	the	weight	of	the	Experimental	seat	(total	mass	=	191	kg).	 	 The	 Experimental	 seat	 has	 only	 been	 tested	 at	 collision	 speeds	 up	 to	 12	 km/h	 and	 it	remains	unclear	whether	 the	combined	seat	hinge	 rotation	and	seatback	deformation	profile	can	attenuate	ATD	responses	at	collision	speeds	greater	 than	12	km/h.	 	Future	testing	of	 the	Experimental	anti-whiplash	seat	is	required	at	the	RCAR/IIWPG	crash	pulse	(Δv	=	16	km/h,	Δt	=	91	ms)	 to	 compare	 the	 effectiveness	 of	 actively	 controlling	 seat	 hinge	 rotation	 and	 seatback	deformation	on	the	attenuation	of	ATD	responses	against	existing	anti-whiplash	seats.					 Another	limitation	of	the	present	study	was	that	the	automotive	floor	geometry	relative	to	 the	 seat	 base	 were	 not	 properly	 maintain	 between	 seats.	 The	 WHIPS	 and	 SAHR	 seats	contained	 electrical	 motors	 under	 the	 seat	 pan	 to	 adjust	 fore/aft	 position	 and	 seat	 pan	 tilt	angle.	To	mount	these	seats,	the	seat	base	was	elevated	to	allow	for	clearance	under	the	seat	pan	and	changed	the	vertical	distance	between	the	top	of	the	seat	pan	and	automotive	floor					136	(see	Chapter	2).	However,	the	relative	geometry	between	the	BioRID	II	ATD’s	head,	torso	and	pelvis	were	replicated	between	the	different	automotive	seats	to	ensure	that	the	ATD	was	in	a	similar	initial	position	and	confirmed	using	IRED	marker	positions	(Figure	5.1	and	5.3).	Despite	the	 resulting	 differences	 in	 ATD	 lower	 limb	 positions,	 the	 similarities	 observed	 in	 T1	accelerations	 across	 the	 four	 existing	 anti-whiplash	 seats	 in	 both	 a	 previous	 study	 and	 this	current	 study	 suggested	 that	 lower	 limb/foot	 position	 had	 little	 influence	 on	 peak	 ATD	kinematic	and	kinetic	responses.			 The	BioRID	ATD	used	 for	 this	 study	 represented	a	50th	 percentile	male	occupant,	 and	further	work	 is	 needed	 to	 evaluate	 the	 Experimental	 seat	 for	male	 and	 female	 occupants	 of	different	heights	and	weights.	Although	the	BioRID	ATD	was	designed	to	mimic	the	motion	of	human	occupants,	 including	their	active	muscle	response,	 further	work	 is	needed	to	evaluate	whether	 the	 pre-perturbation	 forward	 rotation	 of	 the	 seat	 hinge	 alters	 the	 neck	 muscle	response	 in	 humans	 and	 whether	 the	 BioRID	 ATD	 remains	 a	 valid	 surrogate	 for	 human	occupants	under	these	pre-crash	conditions.	Further	work	is	also	needed	on	how	to	affordably	and	 efficiently	 implement	 an	 active	 seat	 that	 behaves	 like	 the	 large,	 heavy	 and	 expensive	prototype	seat	used	for	this	study.			5.5 Conclusion		 Compared	to	the	current	anti-whiplash	seats	currently	on	the	market	(WHIPS,	SAHR	and	GMHR),	 the	 Experimental	 anti-whiplash	 seat	 attenuated	 all	 peak	 ATD	 responses	 (with	 the					137	exception	 of	 FX).	 The	 observed	 peak	 ATD	 responses	 on	 the	 Experimental	 seat	 remained	essentially	constant	for	collision	severities	ranging	from	2	to	12	km/h.	For	a	12	km/h	collision,	the	 peak	 ATD	 responses	 on	 the	 Experimental	 seat	 were	 similar	 or	 below	 those	 observed	 in	current	anti-whiplash	seats	for	a	4	km/h	collision.	The	results	of	this	study	suggested	that	the	Experimental	 seat	 dynamically	 controlling	 seat	 hinge	 rotation	 and	 seatback	 cushion	deformation	 could	 potentially	 reduce	 the	 risk	 of	 whiplash	 injury	 during	 low-speed,	 rear-end	collisions.			 					138	Chapter	6. 	General	Discussion	and	Conclusion	6.1 Summary	of	the	Research	and	Discussion	The	overall	objective	of	 these	experiments	presented	 in	 this	dissertation	was	 to	design,	build	and	 test	a	novel	Experimental	 anti-whiplash	 to	prevent	whiplash	 injuries	 following	 low-speed,	rear-end	collisions.		The	key	safety	features	of	the	anti-whiplash	seat	were	the	dynamic	control	 of	 seat	 hinge	 rotation	 and	 seatback	 cushion	 deformation	 before	 and	 during	 the	collision.	 A	 series	 of	 four	 experiments	 was	 conducted	 to	 determine	 the	 performance	 of	 the	Experimental	 anti-whiplash	 seat	 and	 to	 compare	 its	 performance	 to	 existing	 anti-whiplash	seats.	The	goal	of	Experiment	1	(Chapter	2)	was	to	gain	insight	into	how	current	anti-whiplash	seats	 perform	 during	 low-speed,	 rear-end	 collisions	 at	 collision	 severities	 lower	 than	 the	Research	 Council	 for	 Automobile	 Repairs/International	 Insurance	Whiplash	 Prevention	Group	(RCAR/IIWPG)	 standard	16	 km/h	 crash	pulse.	 The	 results	 of	 Experiment	 1	 revealed	 that	 anti-whiplash	seats	rated	as	good	according	to	the	RCAR/IIWPG	(2	Volvo	Whiplash	Injury	Prevention	seats:	 WHIPS	 and	 1	 Saab’s	 Active	 Head	 Restraint:	 SAHR)	 only	 attenuated	 four	 key	 ATD	responses	 (upper	 neck	 shear	 and	 compressive	 forces,	 upper	 neck	 flexion	 and	 extension	moment,	 and	 rearward	 retraction)	 compared	 to	 a	 poor-rated	 seat	 (General	 Motor’s	 High	Retention	seat:	GMHR).	The	goals	of	 the	next	two	experiments	 (Experiment	2	and	3)	were	to	determine	if	the	active	control	of	seat	hinge	rotation	or	of	seatback	cushion	deformation	could	attenuate	 ATD	 responses	 during	 low-speed,	 rear-end	 collisions.	 	 Active	 rotation	 of	 the	 seat					139	hinge	 and	active	modulation	of	 the	 seatback	decreased	most	ATD	 responses	 and	neck	 injury	criteria	 compared	 to	 a	 Control	 seat	 (Chapters	 3	 and	 4).	 The	 goals	 of	 Experiment	 4	 were	 to	determine	 if	 co-activation	 of	 both	 the	 seat	 hinge	 rotation	 and	 seatback	 deformation	 safety	mechanisms	on	the	Experimental	anti-whiplash	seat	could	attenuate	further	the	ATD	responses	and	to	compare	the	performance	of	the	Experimental	anti-whiplash	seat	against	existing	anti-whiplash	seats	(GMHR,	SAHR	and	WHIPS).	The	co-activation	of	both	the	seat	hinge	rotation	and	seatback	 deformation	 safety	mechanisms	 resulted	 in	 ATD	 responses	 that	 were	 not	 different	from	those	observed	when	only	the	seatback	deformation	mechanism	was	used.	In	comparison	to	 existing	 anti-whiplash	 seats,	 the	 Experimental	 seat	 decreased	 ATD	 kinematic	 and	 kinetic	responses	by	25	–	99%	and	decreased	neck	injury	criteria	by	9	–	73%	for	collision	speeds	of	4	km/h	or	greater.		Through	the	four	experiments	included	in	this	dissertation,	we	accomplished	the	 overall	 research	 objective	 of	 developing	 an	 Experimental	 anti-whiplash	 seat	 that	dynamically	modified	 the	 seat	 hinge	 and	 seatback	 cushion	 properties	 to	 attenuate	 occupant	kinematic	 and	 kinetic	 responses	 as	 well	 as	 neck	 injury	 criteria	 during	 low-speed,	 rear-end	collisions.			The	 safety	 features	of	 the	Experimental	 anti-whiplash	automotive	 seat	were	 the	use	of	motors	to	dynamically	control	seat	hinge	rotation	and	seatback	cushion	deformation	during	the	whiplash	 perturbation.	 By	 controlling	 seat	 hinge	 rotation	 and	 seatback	 cushion	 deformation,	the	 anti-whiplash	 seat	was	 designed	 to	 decrease	 the	 amplitude	 of	 the	 forces	 applied	 by	 the	vehicle	 to	 the	 occupant	 and	 to	 support	 the	 head	 earlier	 in	 the	 collision	 to	 minimize	 the					140	displacement	 of	 the	 head	 relative	 to	 the	 torso.	 The	 seat	 hinge	 profile	 rotated	 the	 seatback	forward	 prior	 to	 the	 collision	 onset	 to	 decrease	 the	 head-to-head-restraint	 distance	 before	actively	 rotating	 the	 seatback	 rearward	 as	 the	 occupant	 loaded	 the	 seat	 to	 attenuate	 the	acceleration	experienced	by	the	occupant.	In	addition	to	decreasing	head	restraint	backset,	the	active	rotation	of	the	seatback	may	have	increased	the	effective	stiffness	of	the	seat	hinge	to	resist	 the	 rearward	 seatback	 deflection	 created	 by	 the	 occupant’s	 inertia	 as	 the	 sled	 was	accelerated	 forward.	The	predictive	seatback-motor	deformation	profile	 introduced	slack	 into	the	webbing	 prior	 to	 the	 collision	 onset	 to	 allow	deeper	 occupant	 penetration	 into	 the	mid-torso	and	pelvis	regions	of	the	seatback	and	to	reduce	the	pre-impact	backset.	After	reaching	peak	 rearward	 rotation,	 both	 the	 seat	 hinge	 rotation	 and	 seatback	 deformation	 profiles	stopped	 in	 order	 to	 attenuate	 forward	 rebound	 of	 the	 seatback.	 This	 attenuated	 rebound	reduced	 the	 acceleration	 and	 speed	 change	experienced	by	 the	occupant.	 Consequently,	 the	changes	 in	 both	 the	 seat	 hinge	 rotation	 and	 seatback	 cushion	 deformation	 both	 before	 and	during	 the	 collision	 combined	 with	 the	 termination	 of	 the	 motion	 at	 peak	 hinge	 angle	 and	cushion	displacement	served	to	prolong	the	ATD-seatback	interaction	and	reduced	the	energy	typically	 returned	 to	 the	occupant	during	 the	 rebound	or	 restitution	phases	of	 the	occupant-seat	interaction.		When	 comparing	 the	 relative	 benefits	 of	 the	 seat	 hinge	 rotation	 and	 seatback	deformation	 safety	 mechanisms	 at	 each	 collision	 severity,	 the	 seat	 hinge	 rotation	 better	reduced	 peak	 acceleration	 of	 the	 head	 and	 torso	 (aX-head	 and	 aX-T1),	 whereas	 the	 seatback					141	deformation	 improved	 occupant	 penetration	 (dpenetration-T1	 and	 dpenetration-pelvis)	 and	 better	minimized	the	relative	motion	between	the	head	and	upper	torso	to	reduce	kinetic	responses	(FX,	FZ,	and	MY),	 linear	displacements	(RX	and	drebound),	angular	rotation	and	velocity	(θhead	and	ωY-head),	 head-to-head-restraint	 contact	 time	 (Δthead-contact),	 and	neck	 injury	 criteria	 (NICmax,	Nij	and	Nkm)	(Table	6.1).		Both	safety	mechanisms	decreased	most	ATD	responses	compared	to	the	Control	 seat,	but	 the	seatback	cushion	deformation	mechanism	generated	greater	 reductions	(average	 of	 -61%)	 than	 the	 seat	 hinge	 rotation	 mechanism	 (average	 of	 -46%)	 for	 the	 ATD	responses	in	the	bottom	portion	of	Table	6.1	(paired	t-test,	p=0.007).		The	combination	of	both	safety	mechanisms	 (Chapter	5)	 typically	generated	 the	 largest	reductions	in	ATD	responses	compared	to	the	Control	seat	(see	the	bold	percentages	in	Table	6.1),	 although	 these	 reductions	 (average	 of	 -73%)	 were	 not	 significantly	 different	 from	 the	seatback	deformation	mechanism	alone	(paired	t-test,	p=0.088).	The	lack	of	additional	benefits	from	 the	 combined	 activation	 of	 the	 seat	 safety	 mechanisms	 compared	 to	 the	 isolated	activation	of	the	seatback	mechanism	could	suggest	that	implementing	only	one	of	these	safety	features	is	sufficient	to	protect	an	occupant	during	a	rear-end	collision.	In	terms	of	the	future	use	of	these	mechanisms	as	potential	safety	devices,	the	seatback	deformation	mechanism	may	be	easier	to	implement	as	there	was	only	one	component	(i.e.,	introducing	slack	in	the	seatback	cushion	 support	 at	 the	 mid-torso	 and	 pelvis),	 whereas,	 the	 seat	 hinge	 rotation	 mechanism	would	require	two	components	(i.e.,	the	pre-impact	forward	rotation	of	the	seat	hinge	and	the	subsequent	 rearward	 rotation	 of	 the	 seat	 hinge).	 Further	work	 is	 needed	 to	 develop	 a	 cost-				142	effective	method	 of	 achieving	 this	 seatback	 cushion	 behavior.	 Further	 research	may	 also	 be	required	to	understand	why	the	lack	of	a	reduction	in	the	horizontal	acceleration	of	T1	(aX-T1),	a	variable	that	 is	currently	used	as	a	test	metric	 in	evaluating	the	performance	of	anti-whiplash	seats,	was	the	only	variable	not	attenuated	by	the	seatback-deformation	mechanism.		Across	all	experiments	presented	in	this	dissertation,	differences	in	initial	position	of	the	head	 restraint	 relative	 to	 the	 occupant’s	 head	 potentially	 confounded	 the	 observed	 ATD	kinematic	and	kinetic	responses.	The	horizontal	backset	(dbackset)	and	vertical	height	(dheight)	of	the	head	restraint	relative	to	the	head	are	known	to	affect	the	occupant	responses	and	the	risk	of	whiplash	injury	(Nygren	et	al.,	1985,	Siegmund	et	al.,	1999,	Stemper	et	al.,	2006,	Eichberger	et	 al.,	 1996,	 Eriksson,	 2005).	 For	 example,	 	 every	 25	 mm	 reduction	 in	 dbackset	 reportedly	decreases	the	risk	of	neck	injury	by	10%	(Eriksson,	2005).	The	Experimental	seat	was	designed	to	have	a	good-rated	initial	head	restraint	position	and	the	pre-impact	rotation	seat	hinge	and	deformation	of	 the	seatback	 further	 reduced	the	 initial	backset	distance	prior	 to	 the	collision	onset.	 In	Chapter	3,	 the	addition	of	the	seat	hinge	rotation	affected	the	ATD	responses	more	than	 differences	 in	 initial	 backset	 between	 the	 poor-rated	GMHR	 seat	 and	 the	 Experimental	seat.	In	comparison	to	good-rated	existing	anti-whiplash	seats	(e.g.,	WHIPS	and	SAHR;	Chapter	5),	 the	 Experimental	 seat	 with	 co-activation	 of	 the	 seat	 hinge	 rotation	 and	 seatback	deformation	 decreased	 all	 kinematic	 and	 most	 kinetic	 responses	 (with	 the	 exception	 of	 FX)	despite	having	a	similar	 initial	backset.	The	pre-impact	forward	rotation	of	the	seat	hinge	and	rearward	deformation	of	the	seatback	cushion	reduced	the	head-to-head-restraint	distance	and					143	caused	the	time-to-head-contact	to	occur	26.1	ms	before	the	collision	onset	(Table	6.1,	∆v	=	12	km/h).	Given	 that	approximately	30%	of	vehicles	observed	at	 intersections	have	marginal-	or	poor-rated	 head	 restraint	 positions	 (Romilly	 et	 al.,	 2011),	 the	 pre-impact	 activation	 of	 the	Experimental	anti-whiplash	seat	would	actively	change	the	positions	of	both	the	occupant	and	the	head	restraint	to	support	the	head	and	neck	during	the	collision	and	to	better	protect	the	occupants	of	these	vehicles.									144	Table	6.1.	Mean	(standard	deviation)	and	percent	change	(%)	from	Control	GMHR	seat	for	the	Experimental	seat	with	seat	hinge	rotation	only	 (Chapter	 3),	 seatback	 deformation	 only	 (Chapter	 4)	 and	 the	 combination	 of	 both	 seat	 hinge	 rotation	 and	 seatback	 deformation	(Chapter	5)	during	at	a	Δv	=	12	km/h	collision	speed.	 	Five	repeated	perturbations	were	collected	of	the	ATD	seated	on	the	GMHR	and	each	test	condition	of	the	Experimental	seat.	The	largest	beneficial	change	is	shown	in	bold.	Automotive	Seat	 GM	Pontiac	Grand	AM	(GMHR)	Control	Seat	Experimental	Chapter	3:	Seat	Hinge	Rotation	Experimental	Chapter	4:	Seatback	Deformation	Experimental	Chapter	5:	Rotation	and	Deformation	Δv	(km/h)	 12	km/h,	Δt	=	185	ms	 12	km/h,	Δt	=	185	ms	 12	km/h,	Δt	=194	ms	 12	km/h,	Δt	=	187	ms	Parameter	 Mean	(SD)	 Mean	(SD)	 Change	(%)	 Mean	(SD)	Change	(%)	 Mean	(SD)	Change	(%)	Initial	dbackset	(mm)	 94.5	(2.6)	 51.5	(4.8)	 -45.5	 52.0	(7.3)	 -45.0	 51.5	(4.8)	 -45.5	Initial	dheight	(mm)	 -25.0	(2.8)	 -1.6	(1.2)	 -93.6	 2.5	(7.1)	 -110.0	 -1.4	(2.8)	 -94.4	aX-sled	(m/s2)	 31.6	(0.4)	 30.2	(0.4)	 -4.4	 30.7	(0.5)	 -2.8	 31.6	(0.4)	 0.0	Δthead	contact	(ms)	 109.4	(1.6)	 82.8	(6.0)	 -24.3	 51.4	(10.2)	 -53.0	 -26.1	(5.7)	 -123.9	dpenetration-T1	(mm)	 -67.9	(1.0)	 -	 -	 -78.5	(1.7)	 15.6	 -61.6	(1.2)	 -9.3	dpenetration-pelvis	(mm)	 -55.8	(1.2)	 -	 -	 -99.6	(3.8)	 78.5	 -94.1	(2.8)	 68.6	aX-head	(m/s2)	 120.7	(1.7)	 49.3	(3.1)	 -59.1	 60.7	(4.6)	 -49.7	 35.1	(1.4)	 -70.9	aX-T1	(m/s2)	 47.0	(0.3)	 35.7	(1.0)	 -24.0	 47.1	(1.0)	 0.2	 31.2	(1.3)	 -33.6	FX	(N)	 264.4	(14.1)	 140.7	(5.7)	 -46.8	 87.8	(22.8)	 -66.8	 76.1	(4.2)	 -71.2	FZ	(N)	 586.1	(16.2)	 267.3	(4.4)	 -54.4	 134.7	(21.6)	 -77.0	 63.5	(17.4)	 -89.2	MY	(Nm)	 25.1	(0.1)	 11.9	(0.7)	 -52.6	 5.9	(1.2)	 -76.5	 4.9	(0.6)	 -80.5	ωY-head(deg/s)	 -390.2	(11.1)	 -189.4	(11.5)	 -51.5	 -147.8	(11.9)	 -62.1	 -98.7	(11.6)	 -74.7	θ	head	(deg)	 16.5	(0.6)	 8.9	(0.4)	 -46.1	 5.3	(1.7)	 -67.9	 2.2	(0.4)	 -86.7	RX	(mm)	 -64.4	(2.3)	 -35.8	(1.3)	 -44.4	 -32.6	(6.7)	 -49.4	 -19.2	(2.1)	 -70.2	drebound	(mm)	 413.4	(7.7)	 267.5	(7.0)	 -35.3	 126.0	(17.1)	 -69.5	 201.3	(15.2)	 -51.3	NICmax	(m2/s2)	 9.8	(0.2)	 4.0	(0.4)	 -59.2	 2.0	(0.9)	 -79.6	 3.6	(0.5)	 -63.3	Nij	 0.16	(0.00)	 0.09	(0.01)	 -43.8	 0.027	(0.006)	 -83.1	 0.05	(0.00)	 -68.8	Nkm	 0.60	(0.02)	 0.28	(0.02)	 -53.3	 0.23	(0.05)	 -61.7	 0.18	(0.01)	 -70.0					145	6.2 Implications	for	Whiplash	Injury	Research	Current	anti-whiplash	seats	are	tested	using	the	RCAR/IIWPG	test	pulse	with	a	speed	change	of	16	km/h.	Despite	53%	of	 rear-end	collisions	occurring	at	collision	severities	 less	than	15	km/h	(Hell	et	al.,	1998),	there	is	limited	research	into	occupant	responses	below	the	16	 km/h	 standard	 test	 pulse.	 The	 results	 of	 Chapter	 2	 revealed	 positively	 graded	 ATD	kinematic	 and	 kinetic	 responses	 to	 an	 increasing	 speed	 changes	 from	 2	 –	 14	 km/h	while	seated	 on	 existing	 anti-whiplash	 seats	 (e.g.,	 Volvo’s	 WHIPS,	 Saab’s	 SAHR	 and	 General	Motor’s	GMHR).	 The	 attenuation	of	 only	 four	 key	ATD	 responses	 (peak	upper	neck	 shear	and	 axial	 forces,	 flexion/extension	 moment	 and	 rearward	 retraction)	 by	 the	 good-rated	seats	 in	 comparison	 to	 a	 poor-rated	 seat	 demonstrated	 the	 limited	 low-speed,	 occupant	protection	 available	 in	 the	 tested	 seats.	 	 Since	 the	 exact	 injury	 mechanisms	 underlying	whiplash	injuries	remain	unclear,	reducing	occupant	accelerations	(head	and	torso)	and/or	minimizing	 the	 movement	 of	 the	 head	 relative	 to	 the	 upper	 torso,	 although	 potential	targets,	may	not	 completely	protect	occupants	 against	whiplash	 injuries	 (Siegmund	et	 al.,	2009).	 Future	 work	 is	 needed	 to	 identify	 the	 biomechanical	 factors	 leading	 to	 whiplash	injuries	to	help	the	development	of	anti-whiplash	automotive	seats.	The	main	difference	between	 the	Experimental	 anti-whiplash	 seat	 and	existing	 anti-whiplash	seats	is	that	the	Experimental	seat	dynamically	changes	its	properties	both	before	(predictive)	 and	 during	 (reactive)	 the	 collision	 to	 reduce	 ATD	 kinematic	 and	 kinetic	responses.	 Current	 anti-whiplash	 seats	 (WHIPS	 and	 SAHR)	 seats	 are	mostly	 reactive	 seats	that	 require	 the	 inertial	 mass	 of	 an	 occupant	 loading	 the	 seatback	 during	 the	 whiplash	perturbation	to	activate	their	respective	safety	mechanisms.	One	key	predictive	factor	that					146	the	 Experimental	 seat	 addressed	 is	 the	 pre-impact	 reduction	 of	 initial	 backset	 by	dynamically	 rotating	 the	 seatback	 forward	 to	 bring	 the	 head	 restraint	 towards	 the	 ATD’s	head	and/or	translating	the	seatback	cushion	rearward	to	bring	the	ATD’s	head	towards	the	head	 restraint	 prior	 to	 the	 collision	 onset.	 Head	 restraint	 backset	 in	 isolation	 reportedly	decreases	the	risk	of	whiplash	injury	by	10%	for	every	25	mm	decrease	in	backset,	with	the	lowest	 injury	 risk	 observed	 when	 the	 dbackset	 =	 0	 mm	 (Eriksson,	 2005).	 	 The	 pre-impact	reduction	 of	 backset	 initiated	 earlier	 head-to-head-restraint	 contact	 time,	 decreased	displacements	of	the	head	relative	to	the	upper	torso,	and	reduced	most	ATD	kinematic	and	kinetic	 responses	 to	 better	 reduce	 the	 risk	 of	 whiplash	 injuries	 following	 low-speed,	collision.	 Although	 the	 Experimental	 seat	 attenuated	 the	 ATD	 responses	 relative	 to	 the	Control	seat,	even	when	the	head	restraint	backset	was	similar,	it	remains	unclear	whether	similar	reductions	can	be	achieved	by	just	adjusting	the	initial	backset	to	zero.	Future	testing	is	 required	 to	 better	 quantify	 the	 relative	 contributions	 of	 initial	 head-to-head-restraint	contact	 prior	 to	 the	 collision	 compared	 to	 both	 the	 seat	 hinge	 rotation	 and	 the	 seatback	cushion	deformation	on	 the	peak	ATD	 responses	and	 the	 risk	of	whiplash	 injury	 following	rear-end	collisions.	 	The	 risk	 of	 developing	whiplash	 associated	 disorders	 (WAD)	 symptoms	 lasting	more	than	a	month	is	approximately	20%	following	an	8	km/h	rear-end	collision	and	no	WAD	have	been	 reported	 in	 field	data	 at	 collision	 speeds	 less	 than	or	 equal	 to	 4	 km/h	 (Krafft	 et	 al.,	2005,	Bartsch	et	al.,	2008).	In	contrast	to	current	anti-whiplash	seat,	the	ATD	responses	on	the	Experimental	anti-whiplash	seat	were	not	graded	with	 increasing	collision	severity	but	were	attenuated	to	a	similar	level	despite	an	increase	in	collision	speeds	(Chapter	5).	Thus,					147	the	 Experimental	 anti-whiplash	 seat,	 with	 both	 seat	 hinge	 rotation	 and	 seatback	 cushion	deformation,	attenuated	most	ATD	responses	at	the	8	km/h	and	12	km/h	speed	changes	to	levels	observed	at	a	4	km/h	speed	change	on	existing	anti-whiplash	seats	(Figure	6.1).	Based	on	these	data,	the	combination	of	seat	hinge	rotation	and	seatback	cushion	deformation	on	the	Experimental	anti-whiplash	seat	could	potentially	reduce	the	overall	20%	WAD	risk	at	an	8	km/h	collision	speed	to	near	zero,	 i.e.	the	risk	of	WAD	at	4	km/h	collision	speed	or	 less.	Future	experiments	at	higher	collision	speeds	are	required	to	determine	the	effectiveness	of	the	 Experimental	 anti-whiplash	 seat	 at	 higher	 collision	 speeds	 and	 future	 field	 tests	 of	 a	production	version	of	the	seat	are	needed	to	determine	the	potential	reduction	in	the	risk	of	whiplash	injuries	in	the	field.					 					148		Figure	6.1.	Exemplar	BioRID	II	ATD	response	data	comparing	the	Experimental	anti-whiplash	seat	 (black	 line)	 with	 the	 co-activation	 of	 both	 seat	 hinge	 rotation	 and	 seatback	 cushion	deformation	 during	 a	 12	 km/h	 collision	 (Δt	 =	 187	ms)	 to	 a	 RCAR/IIWPG	 good-rated	 seat	(Volvo	S40	WHIPS;	red	line)	and	a	RCAR/IIWPG	poor-rated	seat	(Pontiac	Grand	Am	GMHR;	blue	 line)	 during	 a	 4	 km/h	 collision	 (Δt	 =	 147	ms).	 Hollow	 circles	 represent	 the	 onset	 of	head-to-head-restraint	 contact	 and	 peak	 responses	 of	 each	 ATD	 response	 parameter	 for	each	trial.	 					149	6.3 General	Limitations	and	Future	Directions	All	 of	 the	 experiments	 conducted	 in	 this	 dissertation	 utilized	 a	 BioRID	 II	anthropomorphic	test	device	(ATD)	and	a	custom	linear	sled	 located	 in	a	 laboratory.	 	 	The	ATD	was	a	useful	model	 throughout	 the	experiments	as	 it	 allowed	 the	 testing	of	 collision	severities	 that	might	 cause	 injury	 to	human	volunteers	 (Brault	et	al.,	2000).	Now	that	 the	overall	effectiveness	of	the	anti-whiplash	seat	in	the	reduction	of	ATD	kinematic	and	kinetic	responses	has	been	established,	 its	 effectiveness	on	human	occupants	 can	be	 tested.	 For	example,	human	volunteers	exhibit	a	startle	response	during	rear-end	collisions	that	can	be	modified	by	a	pre-impact	stimuli	 (Mang	et	al.,	2015,	Mang	et	al.,	2012).	Whether	the	seat	hinge	rotation	and/or	seatback	cushion	deformation	of	the	Experimental	anti-whiplash	seat	elicits	a	startle	response	that	can	reduce	occupant	responses	remains	unknown.	In	addition,	all	experiments	were	conducted	on	an	ATD	that	represents	a	50th	percentile	male.	It	remains	unclear	whether	the	seat	hinge	rotation	and	seatback	cushion	deformation	profiles	will	also	attenuate	 the	 responses	 of	 male	 and	 female	 occupants	 with	 different	 anthropometric	measurements.	 Further	 testing	 should	 first	 be	 done	 using	 the	 50th	 percentile	 female	 and	95th	percentile	male	ATDs	to	verify	the	robustness	of	the	seat	hinge	rotation	and	seatback	deformation	profiles	for	occupants	with	different	anthropometry.	Then,	these	experiments	should	 be	 expanded	 in	 human	 volunteers	 of	 both	 sexes	 with	 a	 range	 of	 anthropometric	characteristics.	Initial	human	experimentations	could	begin	with	a	speed	change	of	2.8	km/h	(Dv	=	0.78	m/s)	and	a	peak	acceleration	of	2.1g	 (used	previously	with	human	volunteers),	although	little	or	no	reduction	in	the	ATD	responses	were	observed	at	these	low	levels.	After	these	initial	tests,	higher	speed	changes	could	be	explored	incrementally	given	the	expected					150	reduction	 in	 kinematics	 and	 kinetics	 responses	 associated	 with	 the	 Experimental	 anti-whiplash	seat.	The	GMHR	seat	was	initially	selected	due	to	the	availability	of	a	finite	element	model	of	the	GMHR	seat	developed	at	UBC	using	LS-Dyna	explicit	finite	element	solver	(Livermore	Software	 Technology	 Corp.,	 Livermore,	 CA,	 USA)	 (Romilly	 and	 Skipper,	 2005).	 The	 GMHR	seat	was	designed	with	a	rigid	perimeter	frame	and	a	yielding	center	to	promote	pocketing	of	 the	 pelvis	 to	 increase	 occupant	 retention	 (Viano,	 2003a,	 Viano	 and	 Parenteau,	 2015).		However,	 because	 the	 GMHR	 seat	 received	 a	 poor	 rating	 by	 the	 RCAR/IIWPG	 seat/head	restraint	 evaluation	 protocol,	 any	 comparisons	 between	 the	 GMHR	 seat	 and	 either	 the	good-rated	 seats	 (e.g.,	 WHIPS	 and	 SAHR)	 or	 the	 Experimental	 anti-whiplash	 seat	 will	potentially	 be	 confounded	 by	 differences	 in	 initial	 backset	 distance.	 The	 effect	 of	 initial	backset	was	addressed	in	Experiment	2	(Chapter	3)	and	showed	that	the	seat	hinge	rotation	had	a	larger	effect	on	ATD	responses	than	the	differences	in	initial	backset.	All	 the	 collision	 severities	 used	 in	 these	 experiments	 were	 less	 severe	 than	 the	industry-standard,	RCAR/IIWPG	collision	test	pulse	(16	km/h,	Dt	=	91	ms).	The	whiplash	test	sled	used	 in	 these	experiments	was	designed	 for	human	 testing	and	was	not	designed	 to	generate	the	RCAR/IIIWPG	collision	pulse.	Also,	the	weight	of	the	Experimental	seat	and	all	the	motors	added	an	additional	191	kg	to	the	load	carried	by	the	sled	and	limited	its	peak	acceleration	 and	 performance.	 	 Further	 testing	 at	 the	 standard	 16	 km/h	 test	 pulse	 is	required	to	evaluate	the	performance	of	the	Experimental	anti-whiplash	seat	according	to	the	dynamic	test	component	of	the	RCAR/IIWPG	seat	and	head	restraint	evaluation	protocol	(Insurance	 Institute	 of	 Highway	 Safety,	 2008b,	 Insurance	 Institute	 of	 Highway	 Safety,					151	2008a).		Additional	testing	at	the	RCAR/IIWPG	collision	test	pulse	will	require	a	redesign	of	the	Experimental	anti-whiplash	seat	to	transform	the	large,	heavy	and	expensive	prototype	seat	used	in	these	experiments	into	a	lightweight,	affordable	and	efficient	implementation	of	 the	 seat.	 Due	 to	 the	 robustness	 of	 the	 attenuated	 ATD	 responses	 observed	 in	 the	Experimental	anti-whiplash	seat,	we	anticipate	that	the	Experimental	anti-whiplash	seat	will	also	 attenuate	 ATD	 kinematic	 and	 kinetic	 responses	 at	 higher	 collision	 severities	 (≥	 16	km/h).			The	original	design	and	implementation	of	the	Experimental	anti-whiplash	seat	was	a	flexible	system	with	large	motors	to	allow	the	testing	of	a	wide	range	of	seat	hinge	rotation	and	 seatback	 cushion	 deformations	 profiles.	 The	 use	 of	 the	 large	 motors	 may	 not	 be	required	 to	 generate	 the	 small	 seat	 hinge	 rotations	 (<	 4	 deg)	 and	 seatback	 cushion	deformations	 (<	 10	 cm)	 that	 was	 observed	 to	 successfully	 attenuate	 the	 ATD	 responses.	Replacing	the	motors	in	the	final	design	of	the	anti-whiplash	seat	will	simplify	the	seat	hinge	rotation	 and	 seatback	 deformation	mechanisms	 as	 well	 as	 reduce	 the	 overall	 production	cost.	 Instead	of	motors,	future	redesigns	may	utilize	different	combinations	of	deformable	structures,	 springs	 and	 dampeners,	 ratchets,	 and	 pyrotechnic	 components	 to	 generate	predefined	seat	hinge	rotation	and	seatback	deformation	profiles.		Final	implementation	of	the	 anti-whiplash	 seat	 will	 also	 require	 interfacing	 with	 a	 radar-based	 collision-detection	system	 that	 triggers	 the	early	onsets	of	 the	 seat	hinge	 rotation	and	 the	 seatback	 cushion	deformation	 profiles.	 The	 final	 product	 could	 be	 a	 low-cost	 solution	 that	 could	 be	implemented	with	existing	collision	avoidance	systems	to	reduce	whiplash	injuries.								152	6.4 Conclusion	The	 experiments	 presented	 in	 this	 dissertation	 focused	 on	 the	 development	 and	testing	of	a	novel	Experimental	anti-whiplash	automotive	seat	to	address	whiplash	injuries	and	 its	 associated	 disorders	 during	 rear-end	 collisions.	 	 Through	 a	 series	 of	 experiments,	seat	 hinge	 rotation	 and	 seatback	 cushion	 deformation	 profiles	 were	 identified	 that	 best	attenuated	BioRID	II	ATD	kinematic,	kinetic	and	neck	injury	criteria	responses	across	a	range	of	 collision	 severities	 (2	 –	 12	 km/h).	 	 The	 Experimental	 anti-whiplash	 seat,	 with	 the	 co-activation	 of	 the	 seat	 hinge	 rotation	 and	 the	 seatback	 cushion	 deformation	 profiles,	attenuated	the	ATD	responses	during	a	12	km/h	speed	change	to	the	same	response	levels	observed	during	a	4	km/h	speed	change	while	seated	on	other	existing	anti-whiplash	seats	(GMHR,	 SAHR	 and	WHIPS).	 	 Thus,	 the	 results	 of	 the	 experimental	work	 presented	 in	 this	dissertation	confirmed	that	dynamic	control	of	the	seat	hinge	rotation	and	seatback	cushion	deformation	during	a	whiplash	collision	can	attenuate	peak	occupant	responses	and	could	potentially	reduce	the	risk	of	whiplash	injuries	following	low-speed,	rear-end	collisions.		 					153	Bibliography	ALDMAN, B. An Analytical Approach to the Impact Biomechanics of Head and Neck.  30th Annual AAAM Conference, 1986 Montreal, QC, Canada. 439-454. ANDERSON, J. S., HSU, A. W. & VASAVADA, A. N. 2005. Morphology, architecture, and biomechanics of human cervical multifidus. Spine, 30, E86-91. BARNSLEY, L., LORD, S. & BOGDUK, N. 1994. Whiplash injury. Pain, 58, 283-307. BARNSLEY, L., LORD, S. M., WALLIS, B. J. & BOGDUK, N. 1995. The prevalence of chronic cervical zygapophysial joint pain after whiplash. Spine (Phila Pa 1976), 20, 20-5; discussion 26. BARTSCH, A. J., GILBERSTON, L. G., PRAKASH, V., MORR, D. R. & WIECHEL, J. F. Minor Crashes and ‘Whiplash’ in the United States.  52nd Annals of Advances in Automotive Medicine Annual Conference, October 2008 San Diego, California, USA. BOGDUK, N. & YOGANANDAN, N. 2001. Biomechanics of the cervical spine Part 3: minor injuries. Clin Biomech (Bristol, Avon), 16, 267-75. BOSTROM, O., FREDRIKSSON, R., HALAND, Y., JAKOBSSON, L., KRAFFT, M., LOVSUND, P., MUSER, M. H. & SVENSSON, M. Y. 2000. Comparison of car seats in low speed rear-end impacts using the BioRID dummy and the new neck injury criterion (NIC). Accid Anal Prev, 32, 321-8. BOSTROM, O., SVENSSON, M., ALDMAN, B., HANSSON, H. A., HALAND, Y., LOVSUND, P., SEEMAN, T., SUNESON, A., SALJO, A. & ORTENGREN, R. A new neck injury criterion candidate based on injury findings in the cervical spine ganglia after experimental neck extension trauma.  1996 International IROCBI Conference on the Biomechanics of Impact, 1996 Dublin, Ireland. 123-136. BRAULT, J. R., SIEGMUND, G. P. & WHEELER, J. B. 2000. Cervical muscle response during whiplash: evidence of a lengthening muscle contraction. Clin Biomech, 15, 426-35. 				154	BRAULT, J. R., WHEELER, J. B., SIEGMUND, G. P. & BRAULT, E. J. 1998. Clinical response of human subjects to rear-end automobile collisions. Arch Phys Med Rehabil, 79, 72-80. BUITENHUIS, J., DE JONG, P. J., JASPERS, J. P. & GROOTHOFF, J. W. 2009. Work disability after whiplash: a prospective cohort study. Spine (Phila Pa 1976), 34, 262-7. CAPPON, H., PHILIPPENS, M. M. & WISMANS, J. A new test method for the assessment of neck injuries in rear-end collisions 16th International Technical Conference on the Enhanced Safety if Vehicles (ESV) 2001 Amsterdam, The Netherlands. CASSIDY, J. D., CARROLL, L. J., COTE, P., LEMSTRA, M., BERGLUND, A. & NYGREN, A. 2000. Effect of eliminating compensation for pain and suffering on the outcome of insurance claims for whiplash injury. N Engl J Med, 342, 1179-86. DAVIDSSON, J. 1999. BioRID II Final Report. Goteborg, Sweden: Chalmers University of Technology. EICHBERGER, A., GEIGL, B. C., MOSER, A., FACHBACH, B. & STEFFAN, H. 1996. Comparison of Difference Car Seats Regarding Head - Neck Kinematics of Volunteers during Rear End Impact. 1996 International IRCOBI Conference on the BIomechanics of Impacts. Dublin, Ireland. EICHBERGER, A., STEFFAN, H., GEIGL, B. C., SVENSSON, M., BOSTROM, O., LEINZINGER, P. E. & DAROK, M. Evaluation of the applicability of the neck injury criterion (NIC) in rear end impacts on the basis of human subject tests.  1998 International IRCOBI Conference on the Biomechanics of Impact, 1998 Goteborg, Sweden. 321-333. EPPINGER, R., SUN, E., BANDAK, F., HAFFNER, M., KHAEWPONG, N., MALTESE, M., KUPPA, S., NGUYEN, T., TAKHOUNTS, E., TANNOUS, R., ZHANG, A. & SAUL, R. 1999. Development of Improved Injury Criteria for the Assessment of Advanced Automotive Restraint Systems - II.: National Highway Traffic Safety Administration (NHTSA). EPPINGER, R., SUN, E., KUPPA, S. & SAUL, R. 2000. Supplement: Development of Improved Injury Criteria for the Assessment of Advanced Automotive 				155	Restraint Systems - II. National Highway Traffic Safety Administration (NHTSA). ERIKSSON, L. 2005. Influence of Head Restraint Position on Long-Term AIS 1 Neck Injury Risk. International Journal of Crashworthiness, 10, 545-555. EVANS, R. W. 1992. Some observations on whiplash injuries. Neurol Clin, 10, 975-97. EVANS, T. H., MAYER, T. G. & GATCHEL, R. J. 2001. Recurrent disabling work-related spinal disorders after prior injury claims in a chronic low back pain population. Spine J, 1, 183-9. FARMER, C. M., WELLS, J. K. & LUND, A. K. 2003. Effects of head restraint and seat redesign on neck injury risk in rear-end crashes. Traffic Inj Prev, 4, 83-90. FARMER, C. M., ZUBY, D. S., WELLS, J. K. & HELLINGA, L. A. 2008. Relationship of dynamic seat ratings to real-world neck injury rates. Traffic Inj Prev, 9, 561-7. FICE, J. B., BLOUIN, J. S. & SIEGMUND, G. P. Seatback rotational control reduces whiplash injury potential: a preliminary computational study.  Injury Biomechanics Research - Proceedngs of the Fourtienth International Workshop, October 28th, 2012 2012 Savanah, GA. HARDER, S., VEILLEUX, M. & SUISSA, S. 1998. The effect of socio-demographic and crash-related factors on the prognosis of whiplash. J Clin Epidemiol, 51, 377-84. HELL, W., LANGWIEDER, K. & WALZ, F. Reported Soft Tissue Neck Injuries After Rear-End Car Collisions.  1998 International IROCBI Conference on the Biomechanics of Impact, 1998 Gothenburg, Sweden 261 - 274. HOFINGER, M., MAYRHOFER, E., GEIGL, B. C., MOSER, A. & STEFFAN, H. Reduction of neck injuries by improving the occupant interaction with the seat back cushion.  1999 International IRCOBI Conference on Biomechanics of Impact, 1999 Sitges, Spain. 201-212. 				156	HOLM, L. W., CARROLL, L. J., CASSIDY, J. D., HOGG-JOHNSON, S., COTE, P., GUZMAN, J., PELOSO, P., NORDIN, M., HURWITZ, E., VAN DER VELDE, G., CARRAGEE, E. & HALDEMAN, S. 2008. The burden and determinants of neck pain in whiplash-associated disorders after traffic collisions: results of the Bone and Joint Decade 2000-2010 Task Force on Neck Pain and Its Associated Disorders. Spine, 33, S52-9. HORSLEN, B. C., DAKIN, C. J., INGLIS, J. T., BLOUIN, J.-S. & CARPENTER, M. G. 2015. Modulation of human vestibular reflexes with increased postural threat. J Physiol, 592, 3671-85. ICBC 2000. Annual Report. British Columbia: Insurance Company of British Columbia. ICBC 2006. Annual Report. British Columbia: Insurance Company of British Columbia. ICBC 2007. Traffic Collision Statistics: Police-attended Injury and Fatal Collisions. British Columbia: Insurance Company of British Columbia. INSURANCE INSTITUTE OF HIGHWAY SAFETY 2004. Status Report - Special Issue: Protecting Against Neck Injury in Rear Crashes. INSURANCE INSTITUTE OF HIGHWAY SAFETY 2008a. A Procedure for Evaluating Motor Vehicle Head Restraints (Issue 3). Ruckersville, VA, USA: Research Council for Automobile Repairs. INSURANCE INSTITUTE OF HIGHWAY SAFETY 2008b. RCAR-IIWPG Seat/head restraint evaluation protocol (version 3). Arlington, VA, USA: Research Council for Automobile Repairs & International Insurance Whiplash Prevention Group. IVANCIC, P. C. 2011. Does knowledge of seat design and whiplash injury mechanisms translate to understanding outcomes? Spine (Phila Pa 1976), 36, S187-93. JAKOBSSON, L., ISAKSSON-HELLMAN, I. & LINDMAN, M. 2008. WHIPS (Volvo Cars' Whiplash Protection System)—The Development and Real-World Performance. Traffic Injury Prevention, 9, 600-605. 				157	JAKOBSSON, L., LUNDELL, B., NORIN, H. & ISAKSSON-HELLMAN, I. 2000. WHIPS - Volvo's Whiplash Protection Study. Accid Anal Prev, 32, 307-19. KANEOKA, K., ONO, K., INAMI, S. & HAYASHI, K. 1999. Motion analysis of cervical vertebrae during whiplash loading. Spine, 24, 763-9; discussion 770. KRAFFT, M. 2002. When do AIS 1 neck injuries result in long-term consequences? Vehicle and human factors. Traffic Inj Prev, 3, 89-97. KRAFFT, M., KULLGREN, A., MALM, S. & YDENIUS, A. Influence of crash severity on various whiplash injury symptoms: A study based on real-life rear-end crashes with recorded crash pulses.  19th International Technical Conference on the Enhanced Safety of Vehicles (ESV), 2005 Washington, D.C., USA. KULLGREN, A. & KRAFFT, M. Gender analysis on whiplash seat effectiveness: results from real-world crashes.  2010 International IRCOBI Conference on the Biomechanics of Impact, 2010 Hanover, Germany. 17-28. KULLGREN, A., KRAFFT, M., LIE, A. & TINGVALL, C. The Effect of Whiplash Protection Systems in Real- Life Crashes and Their Correlation to Consumer Crash Test Programmes.  The 20th International Technical Conference on the Enhanced Safety of Vehicles Conference (ESV), 2007 Lyon, France. LANGWIEDER, K. & HELL, W. 2002. Proposal of an International Harmonized Dynamic Test Standard for Seal/Head Restraints. . Traffic Inj Prev, 3, 150-158. LINDER, A., AVERY, M., KRAFFT, M. & KULLGREN, A. Change of velocity and pulse characteristics in rear impacts - Real world and vehicle tests data.  The 18th International Technical Conference on the Enhanced Safety of Vehicles, 2003 Nagoya, Japan. Paper No. 285. LINDER, A., AVERY, M., KRAFFT, M., KULLGREN, A. & SVENSSON, M. Acceleration pulses and crash severity in low velocity rear impacts – Real world data and barrier test.  The 17th International Technical Conference on the Enhanced Safety of Vehicles, 2001 Amsterdam, The Netherlands. Paper No. 216. 				158	LUAN, F., YANG, K. H., DENG, B., BEGEMAN, P. C., TASHMAN, S. & KING, A. I. 2000. Qualitative analysis of neck kinematics during low-speed rear-end impact. Clin Biomech (Bristol, Avon), 15, 649-57. LUNDELL, B., JAKOBSSON, L., ALFREDSSON, B., JERNSTROM, C. & ISAKSSON-HELLMAN, I. Guidelines for and the desgin of a car seat concept for improved protection against neck injuries in rear end car impacts.  Internation Congress and Exposition, February 23-26, 1998 1998a Detriot, Michigan. Society of Automotive Engineers, Inc. LUNDELL, B., JAKOBSSON, L., ALFREDSSON, B., LINDSTROM, M. & SIMONSSON, L. The WHIPS seat - A car seat for improved protection against neck injuries in rear end impacts.  16th International Technical Conference on the Enhanced Safety of Vehicles, 1998b Windsor, Ontario, Canada. MACNAB, I. 1971. The "whiplash syndrome". Orthop Clin North Am, 2, 389-403. MANG, D. W., SIEGMUND, G. P., BROWN, H. J., GOONETILLEKE, S. C. & BLOUIN, J. S. 2015. Loud preimpact tones reduce the cervical multifidus muscle response during rear-end collisions: a potential method for reducing whiplash injuries. Spine J, 15, 153-61. MANG, D. W., SIEGMUND, G. P., INGLIS, T. J. & BLOUIN, J. S. 2012. The startle response during whiplash: a protective or harmful response? J Appl Physiol, 113, 532-40. MCCONNELL, W. E., HOWARD, R. P., GUZMAN, H. M., BOMAR, J. B., RADDIN, J. H., BENEDICT, J. V., SMITH, H. L. & HATSELL, C. P. Analysis of human test subject kinematic responses to low velocity rear end impacts (930889).  Proceedings of the 37th Stapp Car Crash Conference, 1993 San Antonio, Texas. Society of Automotive Engineers, 21-30. MCCONNELL, W. E., HOWARD, R. P., VAN POPPEL, J., KRAUSE, R., GUZMAN, H. M., BOMAR, J. B., RADDIN, J. H., BENEDICT, J. V. & HATSELL, C. P. Human head and neck kinematic after low velocity rear-end impacts - Understanding "whiplash" (952724).  Proceedings of the 39th Stapp Car Crash Conference, 1995 Coronado, CA. Society of Automotive Engineers, 215-238. 				159	MCGUCKIN, N. & FUCCI, A. 2018. Summary of Travel Trends: 2017 National Household Travel Survey. In: FEDERAL HIGHWAY ADMINISTRATION (ed.). Washington, DC, USA: U.S. Department of Transportation. MOORHOUSE, K., DONNELLY, B., KANG, Y.-S., BOLTE, J. H. & HERRIOTT, R. 2012. Evaluation of the internal and external biofidelity of current rear impact ATDs to response targets developed from moderate-speed rear impacts of PMHS. Stapp Car Crash J, 56, 171-229. MORDAKA, J. & GENTLE, C. R. Biomechanical analysis of whiplash injuries: Women are not scaled down men., 2003. 4th European LS-DYNA Users Conference. NATIONAL HIGHWAY TRAFFIC SAFETY ADMINISTRATION 2014. Traffic Safety Facts 2014: A compilation of motor vehicle crash data from the fatality analysis reporting system and the general estimates system. In: ANALYSIS, N. C. F. S. A. (ed.). United States Department of Transportation. NYGREN, A., GUSTAFSSON, H. & TINGVALL, C. 1985. Effect of different types of headrest in rear-end collisions. In: NATIONAL HIGHTWAY TRAFFIC SAFERY ADMINISTRATION (ed.) Tenth Experimental Safert Vehicle Conference. Washington, DC.: US Department of Transportation. ORTENGREN, T., HANSSON, H. A., LOVSUND, P., SVENSSON, M. Y., SUNESON, A. & SALJO, A. 1996. Membrane leakage in spinal ganglion nerve cells induced by experimental whiplash extension motion: a study in pigs. J Neurotrauma, 13, 171-80. OTREMSKI, I., MARSH, J. L., WILDE, B. R., MCLARDY SMITH, P. D. & NEWMAN, R. J. 1989. Soft tissue cervical spinal injuries in motor vehicle accidents. Injury, 20, 349-51. PEARSON, A. M., IVANCIC, P. C., ITO, S. & PANJABI, M. M. 2004. Facet joint kinematics and injury mechanisms during simulated whiplash. Spine, 29, 390-7. QUINLAN, K. P., ANNEST, J. L., MYERS, B., RYAN, G. & HILL, H. 2004. Neck strains and sprains among motor vehicle occupants-United States, 2000. Accid Anal Prev, 36, 21-7. 				160	RHULE, D., RHULE, H. & DONNELLY, B. 2005. The Process of Evaluation and Documentation of Crash Test Dummies for Part 572 of the Code of Federal Regulations. The 19th International Technical Conference on the Enhanced Safety of Vehicles (ESV). Washington, D.C., USA: Paper No. 05-0284. ROMILLY, D. P., LUK, H., XING, P., WHITE, M. & DESAPRIYA, E. Whiplash Prevention Campaign Initiative: Development of an Observational Protocol for Proper Head Restraint Use. In: PROFESSIONALS, C. A. O. R. S., ed. 21st Canadian Multidisciplinary Road Safety Conference, May 8 - 11, 2011 2011 Halifax, Nova Scoitia. ROMILLY, D. P. & SKIPPER, C. 2005. Seat structural design choices and the effect on occupant injusty potential in rear end collisions. SAE International Journal of Passenger Cars - Mechanical Systems, 114, 1512-1520. SAE 1995. Surface Vehicle Recommended Practice. Instrumentation for impact test - Part 1 - Electronic Instrumentation. Warendale, PA, USA: Society of Automotive Engineers. SCHMITT, K.-U., MUSER, M. H. & NIEDERER, P. F. A new injury criterion candidate for rear-end collisions taking into account shear forces and bending moments.  17th Enhanced Safety of Vehicles (ESV) Conference, 2001 Amsterdam, Netherlands. SCHMITT, K.-U., MUSER, M. H., WALZ, F. H. & NIEDERER, P. F. 2002. Nkm - A proposal for a neck protection criterion for low-speed rear-end impacts. Traffic Inj Prev, 3, 117-126. SEKIZUKA, M. Seat Designs for Whiplash Injury Lessening. In: NHTSA, ed. 16th International Technical Conference on the Enhansced Safety of Vehicles, May 31st - June 4th, 1998 1998 Windsor, ON, Canada. 1570-78. SEVERY, D., MATHEWSON, J. & CO, B. 1955. Controlled automovile rear-end collisions, an investigation of related engineering and medical phenomena. Can Serv Med J., 11, 727-759. SIEGMUND, G. P., BAILEY, M. N. & KING, D. J. Characteristics of Specific Automobile Bumpers in Low-Velocity Impacts. Paper #940916. In: ENGINEERS., S. O. A., ed. Proceedings of the Thirty Eighth Stapp Car Crash 				161	Conference, February 28 - March 4rd, 1994 1994 Detroit, Michigan, ISA. SAE Technical Paper Series. SIEGMUND, G. P., BLOUIN, J. S., CARPENTER, M. G., BRAULT, J. R. & INGLIS, J. T. 2008a. Are cervical multifidus muscles active during whiplash and startle? An initial experimental study. BMC Musculoskeletal Disorders, 9. SIEGMUND, G. P., BRAULT, J. R. & WHEELER, J. B. 2000. The relationship between clinical and kinematic responses from human subject testing in rear-end automobile collisions. Accid Anal Prev, 32, 207-17. SIEGMUND, G. P., DAVIS, M. B., QUINN, K. P., HINES, E., MYERS, B. S., EJIMA, S., ONO, K., KAMIJI, K., YASUKI, T. & WINKELSTEIN, B. A. 2008b. Head-turned postures increase the risk of cervical facet capsule injury during whiplash. Spine, 33, 1643-9. SIEGMUND, G. P., HEINRICHS, B. E., CHIMICH, D. D., DEMARCO, A. L. & BRAULT, J. R. 2005a. The effect of collision pulse properties on seven proposed whiplash injury criteria. Accid Anal Prev, 37, 275-85. SIEGMUND, G. P., HEINRICHS, B. E., CHIMICH, D. D. & LAWRENCE, J. 2005b. Variability in vehicle and dummy responses in rear-end collisions. Traffic Inj Prev, 6, 267-77. SIEGMUND, G. P., HEINRICHS, B. E. & WHEELER, J. B. 1999. The influence of head restraint and occupant factors on peak head/neck kinematics in low-speed rear-end collisions. Accid Anal Prev, 31, 393-407. SIEGMUND, G. P., KING, D. J., LAWRENCE, J. M., WHEELER, J. B., BRAULT, J. R. & SMITH, T. A. Head/neck kinematic response of human subjects in low-speed rear-end collisions (973341).  Proceedings of the 41st Stapp Car Crash Conference, 1997 Warrendale, PA. Society of Automotive Engineers, 357-385. SIEGMUND, G. P., MYERS, B. S., DAVIS, M. B., BOHNET, H. F. & WINKELSTEIN, B. A. 2001. Mechanical evidence of cervical facet capsule injury during whiplash: a cadaveric study using combined shear, compression, and extension loading. Spine, 26, 2095-101. 				162	SIEGMUND, G. P., WINKELSTEIN, B. A., IVANCIC, P. C., SVENSSON, M. Y. & VASAVADA, A. 2009. The anatomy and biomechanics of acute and chronic whiplash injury. Traffic Inj Prev, 10, 101-12. SPITZER, W. O., SKOVRON, M. L., SALMI, L. R., CASSIDY, J. D., DURANCEAU, J., SUISSA, S. & ZEISS, E. 1995. Scientific monograph of the Quebec Task Force on Whiplash-Associated Disorders: redefining "whiplash" and its management. Spine, 20, 1S-73S. STEMPER, B. D., YOGANANDAN, N. & PINTAR, F. A. 2006. Effect of head restraint backset on head-neck kinematics in whiplash. Accid Anal Prev, 38, 317-23. STERNER, Y. & GERDLE, B. 2004. Acute and chronic whiplash disorders--a review. J Rehabil Med, 36, 193-209; quiz 210. STIGSON, H., YDENIUS, A. & KULLGREN, A. 2006. Variation of Crash Severity and Injury Risk Depending on Collision with Different Vehicle Types and Objects. International Research Council on Bioechanics of Injury Madrid, Spain. SUISSA, S., HARDER, S. & VEILLEUX, M. 2001. The relation between initial symptoms and signs and the prognosis of whiplash. Eur Spine J, 10, 44-9. SVENSSON, M. Y., ALDMAN, B., LOVSUND, P., HANSSON, H. A., SEEMAN, T., SUNESON, A. & ORTENGREN, T. Pressure effects in the spinal canal during whiplash extension motion - a possible cause of injury to the cervical spinal ganglia.  1993 International IRCOBI Conference on the Biomechanics of Impact, 1993 Eindhoven, Netherlands. 189-200. SVENSSON, M. Y., BOSTROM, O., DAVIDSSON, J., HANSSON, H. A., HALAND, Y., LOVSUND, P., SUNESON, A. & SALJO, A. 2000. Neck injuries in car collisions--a review covering a possible injury mechanism and the development of a new rear-impact dummy. Accid Anal Prev, 32, 167-75. SVENSSON, M. Y., LOVSUND, P., HALAND, Y. & LARSSON, S. 1996. The influence of seat-back and head-restraint properties on the head-neck motion during rear-impact. Accid Anal Prev, 28, 221-7. TODMAN, D. H. 2007. Whiplash injuries: A historical review. The Internet Journal of Neurology, 8. 				163	TREMPEL, R. E., ZUBY, D. S. & EDWARDS, M. A. 2016. IIHS head restraint ratings and insurance injury claim rates. Traffic Inj Prev, 17, 590-6. VASAVADA, A. N., BRAULT, J. R. & SIEGMUND, G. P. 2007. Musculotendon and fascicle strains in anterior and posterior neck muscles during whiplash injury. Spine, 32, 756-65. VERSTEEGEN, G. J., KINGMA, J., MEIJLER, W. J. & TEN DUIS, H. J. 2000. Neck sprain after motor vehicle accidents in drivers and passengers. Eur Spine J, 9, 547-52. VIANO, D. C. 2002. Head restraint position during normal driving: implications to neck injury risks in rear crashes. Role of seat in rear crash safety. Warrendale, PA, USA: Society of Automotive Engineers, Inc. VIANO, D. C. 2003a. Effectiveness of High-Retention Seats in Preventing Fatality: Initial Results and Trends. Society of Automotive Engineers, SAE 2003-01-1351. VIANO, D. C. 2003b. Influence of seat properties on occupant dynamics in severe rear crashes. Traffic Inj Prev, 4, 324-36. VIANO, D. C. 2003c. Seat influences on female neck responses in rear crashes: a reason why women have higher whiplash rates. Traffic Inj Prev, 4, 228-39. VIANO, D. C. 2003d. Seat properties affecting neck responses in rear crashes: a reason why whiplash has increased. Traffic Inj Prev, 4, 214-27. VIANO, D. C. 2008. Seat design principles to reduce neck injuries in rear impacts. Traffic Inj Prev, 9, 552-60. VIANO, D. C. & OLSEN, S. 2001. The effectiveness of active head restraint in preventing whiplash. J Trauma, 51, 959-69. VIANO, D. C. & PARENTEAU, C. S. 2015. Update on the effectiveness of high retention seats in preventing fatal injury in rear impacts. Traffic Inj Prev, 16, 154-8. 				164	WIKLUND, K. & LARSSON, H. 1997. SAAB active head restraint (SAHR) - seat design to reduce the risk of neck injuries in rear impacts. Warrendale, PA, USA. WINKELSTEIN, B. A. & MYERS, B. S. 2000. Experimental and computational characterization of three-dimensional cervical spine flexibility. Stapp Car Crash J, 44, 139-58. WINKELSTEIN, B. A., NIGHTINGALE, R. W., RICHARDSON, W. J. & MYERS, B. S. 2000. The cervical facet capsule and its role in whiplash injury: a biomechanical investigation. Spine, 25, 1238-46. WINKELSTEIN, B. A. & SANTOS, D. G. 2008. An intact facet capsular ligament modulates behavioral sensitivity and spinal glial activation produced by cervical facet joint tension. Spine, 33, 856-62. ZADOR, P., KRAWCHUK, S. & VOAS, R. 2000. FInal Report - Automotive Collision Avoidance System (ACAS) Program. In: ADMINISTRATION, N. H. T. S. (ed.). Kokomo, IN, USA: Delphi-Delco Electronic Systems. 		 			 165	Appendix	A:	Experimental	Anti-Whiplash	Seat	Design		Proof-of-Concept	Seat	Hinge	Rotation	Computation	Study	In	 a	 preliminary	 computational	 study,	 Fice	 et	 al.	 (2013)	 modelled	 the	 feed-forward	dynamic	 control	 of	 seat	 hinge	 rotation	 to	 quantify	 the	 potential	 reduction	 in	 occupant	kinematics	for	the	Experimental	anti-whiplash	seat.		A	model	of	a	simplified	2002	Pontiac	Grand	Am	seat	was	combined	with	a	model	of	a	Hybrid	 III	anthropomorphic	test	device	(ATD)	 in	LS-Dyna	 (LSTC,	 Livermore,	 CA,	 USA)	 to	 simulate	 three	 rear	 impact	 speed	 changes:	 Δv	 =	 8km/h,	12km/h	 and	 16	 km/h	 (Fice	 et	 al.,	 2012,	 Romilly	 and	 Skipper,	 2005).	 	 The	 seat	 pan	 and	 the	seatback	 were	 de-coupled	 in	 the	 simulation	 to	 allow	 for	 feed-forward	 control	 of	 seat	 hinge	rotations.		The	input	seat	hinge	rotation	profiles	were	parameterized	and	optimized	to	find	the	minimum	 peak	 forward	 resultant	 acceleration	 of	 the	 first	 thoracic	 vertebrae	 (T1)	 (aresultant-T1)	using	LS-OPT	(LSTC,	Livermore,	CA,	USA).	When	compared	to	a	standard	unmodified	seat,	the	optimized	seat	hinge	rotational	profiles	reduced	aresultant-T1	of	the	ATD	by	58%,	57%,	and	59%	for	rear	impact	speed	changes	of	8km/h,	12	km/h	and	16	km/h,	respectively	(Fice	et	al.,	2012).	This	preliminary	 study	provided	a	proof-of-concept	 for	 the	 feed-forward	control	of	 the	 seat	hinge	rotation	 to	 reduce	 aresultant-T1	 responses	 and,	 potentially,	 reduce	 the	 risk	 of	 whiplash	 injuries	during	rear-end	collisions.	Due	to	differences	in	the	ATD	tested	and	the	simplified	model	of	the	automotive	seat,	 the	results	 from	this	preliminary	study	did	not	directly	transfer	to	testing	of	the	Experimental	anti-whiplash	seat	but	did	provide	insight	to	the	design	of	the	experiments.				 166	Experimental	Anti-Whiplash	Seat		The	active	Experimental	anti-whiplash	seat	consisted	of	a	modified	General	Motors	High	Retention	 seat	 (GMHR;	 front	 passenger	 seat	 from	 a	 2004	 Pontiac	 Grand	 Am)	 and	 several	motors	to	control	seat	hinge	rotation	and	seatback	cushion	deformation	(total	mass	=	191	kg,	Figure	A1	and	A2).	The	seat	pan	and	head	restraint	of	the	GMHR	seat	remained	unmodified	in	this	 design.	 	 The	 seatback	 consisted	 of	 a	 rigid	 aluminum	 outer	 frame.	 Depending	 on	 the	experiment,	 the	 seatback	 cushion	 from	 a	GMHR	 seatback	was	 either	 rigidly	mounted	 to	 this	outer	frame	(Chapter	3	and	Appendix	B)	or	suspended	across	the	outer	frame	from	side-to-side	by	47	mm	wide	seatbelt	straps	(see	Figure	4.2;	Chapter	4	and	5,	and	Appendix	C	and	D).	 	To	dynamically	 control	 seat	hinge	 rotations	 (Chapter	3	and	5),	 two	 large	 rotational	 servomotors	(AKM52K,	Kollmorgen,	Waltham,	MA,	USA)	connected	to	helical	right-angle	gearheads	(VTR014-035,	35:1	gear	 ratio,	Thomson	Linear,	Radford,	VA,	USA)	were	mounted	on	both	 sides	of	 the	seat	hinge.	These	motors	were	geared	to	rotate	in	unison	(one	in	a	positive	direction	and	the	other	 in	a	negative	direction)	and	had	a	maximum	rated	speed	of	157	revolutions	per	minute	(RPM)	 (942	deg/s)	with	 a	maximum	 rated	 torque	of	 3.90	Nm.	 	 Rotation	of	 the	 seatback	was	limited	 from	 vertical	 (0	 deg)	 to	 50	 deg	 rearward.	 	 Initial	 seatback	 angle	 was	 set	 to	 27	 deg	rearward	from	vertical	(Siegmund	et	al.,	2005a).	To	dynamically	control	seatback	cushion	deformation	(Chapter	4	and	5),	three	rotational	servomotors	 (AKM24D,	 Kollmorgen,	 Waltham,	 MA,	 USA)	 connected	 to	 helical	 right-angle	gearheads	 (VTR006-008,	 8:1	 gear	 ratio,	 Thomson	 Linear,	 Radford,	 VA,	 USA)	 were	 mounted	staggered	to	both	sides	of	the	rigid	seatback.		The	seatback	motors	were	capable	of	rotating	at			 167	speeds	 of	 1000	 RPM	 (6000	 deg/s)	 with	 a	maximum	 torque	 of	 1.11	 Nm.	 These	motors	were	placed	at	 the	 top,	middle	and	bottom	of	 the	 seatback	and	attached	 to	47	mm	wide	 seatbelt	straps	 that	 spanned	 across	 the	 perimeter	 frame	 to	 separately	 control	 the	 seatback	 cushion	deformation	at	the	upper-torso,	mid-torso	and	lower	pelvis	regions	(see	Figure	4.2).	Tightening	the	webbing	would	pull	the	suspended	GMHR	upper	seatback	cushion	towards	the	front	of	the	seat	(+X	direction)	and	increase	the	apparent	stiffness	of	the	seatback;	whereas,	loosening	the	webbing	 would	 allow	 the	 seatback	 cushion	 to	 translate	 towards	 the	 rear	 of	 the	 seat	 (-X	direction)	and	allow	for	deeper	occupant	penetration	into	the	seatback.	For	all	tests,	the	upper-torso	motor	was	 locked	 in	 a	 stationary	 position	 and	 effectively	 created	 a	 hinge	 point	 at	 the	upper	back	to	promote	pocketing	of	the	pelvis	into	the	seatback.	Rotation	of	the	mid-torso	and	pelvis	seatback	motors	were	physically	limited	from	a	taut	seatbelt	strap	position	of	0	deg	to	a	spooled	out	loose	webbing	position	of	200	deg.	Initial	seatback	position	was	set	with	the	mid-torso	 spooled	 out	 by	 22	 deg	 and	 pelvis	 motors	 spooled	 out	 by	 25	 deg	 to	 achieve	 an	 initial	seatback	angle	of	27	deg.		All	 5	motors	 (2	 seat	 hinge	 and	 3	 seatback)	were	 controlled	 independently	 by	 separate	digital	servo	drives	(Servostar	600,	Kollmorgen,	Waltham,	MA,	USA)	connected	to	two	NI	UMI	7774	 universal	 motion	 interface	 and	 a	 NI	 PXI	 7350	 motion	 controller	 (National	 Instruments	Corporation,	 Austin,	 Texas,	 USA).	 	 A	 custom	 LabVIEW	 program	 (National	 Instruments	Corporation,	Austin,	Texas,	USA)	was	created	to	send	commands	to,	monitor	the	status	of	and	record	encoder	data	directly	from	these	motors.					 168		Figure	A1.	Images	of	the	novel	active	anti-whiplash	automotive	seat.	A.)	preliminary	computer	aided	design	(CAD)	model	and	B.)	working	prototype.				Figure	A2.	Photograph	of	the	experimental	set-up	with	the	BioRID	II	ATD	on	the	experimental	anti-whiplash	automotive	seat	and	laboratory	reference	frame	(X,	Z).	 	Motors	mounted	to	the	seat	hinge	control	seatback	hinge	rotations;	whereas,	motors	mounted	to	the	seatback	control	seatback	cushion	deformations.	 	A.	 	B.			 169	Summary			 An	 in-depth	 literature	 review	 revealed	 a	 limited	 number	 of	 published	 studies	 or	research	into	the	development	of	automotive	seats	safety	devices	for	injury	prevention.		Thus,	there	 is	 a	 need	 to	 understand	 and	 determine	 how	 different	 seat	 hinge	 rotation	 or	 seatback	cushion	deformation	parameters	affected	occupant	responses.	Combining	parameters	such	as	peak	amplitude,	initial	angular	velocity,	and	onset	will	create	rotation	and	deformation	profiles	to	define	how	fast,	how	far	and	when	either	the	seat	hinge	will	rotate	or	the	seatback	cushion	will	deform.	Other	rotation	and	deformation	profile	parameters	such	as	profile	shape	and	the	pressence	 of	 a	 “pre-perturbation	 supportive”	 pulse	 should	 also	 be	 investigated.	 These	parameters	of	the	seat	hinge	rotation	and	seatback	deformation	will	be	further	investigated	in	subsequent	 Appendices	 to	 define	 rotation	 and	 deformation	 profiles	 to	 be	 used	 throughout	experiments	presented	in	this	dissertation.										 170	Appendix	B:		Creating	the	Seat	Hinge	Rotation	Profile	The	goal	of	 this	appendix	was	 to	develop	a	 seat	hinge	 rotation	profile	 that	best	 reduce	whiplash-evoked	ATD	responses.	This	appendix	summarizes	the	key	steps	undertaken	to	create	the	active	seat	hinge	rotation	profiles	used	in	Experiment	2	(Chapter	3).	The	natural	seat	hinge	response	observed	from	the	Control	General	Motor’s	High	Retention	seat	(GMHR)	during	a	12	km/h	 perturbation	 was	 used	 as	 a	 template	 for	 the	 preliminary	 seat	 hinge	 rotation	 profile	(Figure	 B1).	 This	 preliminary	 profile	 was	 then	modified	 to	 only	 include	 the	 retraction	 phase	(onset	 of	 seat	 hinge	 rotation	 to	 peak	 rearward	 head	 angle)	 and	 to	 prevent	 the	 seat	 from	rebounding.	 	 The	 timing	 of	 the	 input	 seat	 hinge	 rotation	 profile	 was	 delayed	 by	 20	 ms	 to	account	for	the	internal	backlash	of	the	motors	and	gearheads.	The	resulting	input	seat	hinge	rotation	profile	(Figure	B1)	had	a	peak	rearward	angle	(θrearward-peak)	of	5.7	deg,	an	initial	angular	velocity	(ωrearward-int)	of	8.5	deg/s,	and	an	onset	of	seat	hinge	rotation	(trearward-onset)	30	ms	after	the	collision	onset.	This	initial	input	rotation	profile	was	used	as	the	starting	point	from	which	the	 parameters	 for	 a	 seat	 hinge	 rotation	 profile	 were	 isolated	 and	 methodically	 evaluated	through	a	series	of	five	experiments.		For	all	five	experiments,	a	standard	GMHR	seatback	was	rigidly	mounted	to	the	outer	seatback	frame	to	isolate	the	effects	of	varying	seat	hinge	rotation	on	ATD	 responses	during	 a	 12	 km/h	 collision	 speed	with	 a	pulse	duration	 (Δt)	 of	 208	ms.	 	A	Control	trial	on	a	standard	GMHR	seat	was	included	to	provide	baseline	ATD	responses	and	to	compare	against	the	Experimental	profiles.		To	 quantify	 the	ATD	 responses,	 the	 focus	was	 on	 nine	 peak	ATD	 kinematic	 and	 kinetic	responses	(aX-head,	aX-T1,	ωY-head,	FX,	FZ,	MY,	θhead,	RX	and	Δthead-contact)	and	three	peak	neck	injury			 171	criteria	(NIC,	Nkm,	and	Nij)	responses.		Within	each	response	variable,	the	different	profiles	were	ranked,	and	then	the	profile	with	the	lowest	averaged	rank	across	all	12	variables	was	identified	as	 the	best	profile.	A	Friedman	test	determined	 if	 there	was	a	significant	difference	between	the	average	rank	of	the	different	profiles	(chi-squared	value:	χ2,	degrees	of	freedom:	df,	and	the	significance	level:	p).	A	post-hoc,	pair-wise	multiple	comparison	test	was	then	performed	on	the	results	of	the	Friedman	test	to	determine	differences	between	the	individual	profiles.		All	tests	were	performed	using	predefined	functions	(friedman	and	multcompare)	in	Matlab	using	a	=	0.05.						Figure	B1.	Rearward	seat	hinge	rotations	(θseat-hinge)	for	the	Control	(blue	line)	and	Experimental	seat	 (red	 solid	 line)	 during	 a	 12	 km/h	 collision	 severity	 perturbation	 with	 a	 collision	 pulse	duration	 (Δt)	 of	 208	 ms.	 The	 dotted	 red	 line	 illustrates	 the	 input	 θseat-hinge	 profile	 to	 the	Experimental	seat	to	generate	the	output	θseat-hinge	response	(solid	red	 line)	used	as	the	 initial	seat	hinge	rotation	profile	for	Part	1.				 172	Experiment	1:	Effects	of	Individual	Seat	Hinge	Parameters	In	 this	 first	 experiment,	 the	 effects	 of	 three	 seat	 hinge	 rotation	 (θseat-hinge)	 parameters	(peak	 angle,	 initial	 angular	 velocity	 and	 onset)	 were	 isolated	 and	 evaluated	 to	 determine	specific	parameters	that	best	reduced	ATD	responses.		Peak	rearward	angle	(θrearward-peak),	initial	angular	 velocity	 (ωrearward-int)	 and	 onset	 (trearward-onset)	 of	 seat	 hinge	 rotation	 defined	 the	maximum	rearward	angle	(negative	rotation	about	the	laboratory	Y-axis	reference	frame),	the	speed	at	which	the	seat	hinge	initially	rotated	and	the	timing	of	when	the	seat	hinge	started	to	rotate,	respectively.		When	varying	the	effects	of	one	of	the	seat	hinge	rotation	parameters,	the	other	two	rotation	parameters	remained	unchanged	at	their	initial	values	(Figure	B1).			Peak	Seat	Hinge	Angle			Peak	seat	hinge	angle	was	varied	from	2.5	to	52	deg	from	the	initial	seating	position	of	27	deg	rearward	 (Figure	B2A	&	B2B).	 	To	achieve	these	peak	angles,	 the	amplitude	of	 the	 initial	seat	 hinge	 rotation	 profile	 was	 scaled	 by	 50,	 100,	 300,	 500,	 700	 and	 900%	 and	 resulted	 in	θrearward-peak	of	2.5,	5.7,	17,	29,	41	and	52	deg.	Exemplar	experimental	data	from	ATD	head	and	torso	acceleration	(aX-head	and	aX-T1:	Figure	B2D	&	B2E,	respectively)	and	neck	injury	criteria	(NIC:	Figure	B2F)	as	well	as	their	respective	peak	ATD	response	(Figure	B2G	–	B2I)	showed	elevated	responses	 for	 larger	 peak	 seat	 hinge	 angle	 trials.	 A	 Friedman	 test	 revealed	 a	 significant	difference	 between	 trials	 (1	 Control	 and	 6	 Experimental,	 χ2	 =	 46.29,	 df	 =	 6,	 p	 <	 0.001).	 	 The	Experimental	 trial	 with	 a	 θrearward-peak	 =	 5	 deg	 had	 the	 lowest	mean	 ranked	 score	 for	 all	 ATD	responses	and	neck	injury	criteria.	This	reduction,	however,	was	not	significantly	different	from	the	Control	trial,	but	was	significantly	different	from	larger	rearward	rotations	(θrearward-peak	≥	29			 173	deg,	p	<	0.001)	(Figure	B2C).	These	results	suggested	that	future	implementations	and	studies	do	not	require	large	peak	rearward	rotation	of	the	seatback.		Smaller	rearward	rotation	angles	prevent	the	seat	from	impinging	into	the	cabin	space	of	the	rear	passengers	and	lowers	the	risk	of	occupants	ramping	up	and	ejecting	out	of	the	automotive	seat.	Thus,	a	peak	θrearward-peak	of	5.7	 deg	 was	 selected	 for	 subsequent	 testing	 in	 the	 development	 of	 the	 Experimental	 anti-whiplash	seat.				 174		Figure	B2.	ATD	responses	to	the	Control	seat	 (blue	 lines	and	markers)	and	the	Experimental	seat	with	various	amplitudes	of	seat	hinge	 rotation	 (θseat-hinge;	 grey	and	 red	 lines	and	markers)	during	a	12	km/h	collision	 severity	 (Δt	=	208	ms,	black	 lines,	 right	axis,	Panels	A	&	B).	The	red	lines	and	markers	indicate	the	seat	hinge	peak	angle	parameter	selected	for	further	testing.	Panels	(A)	and	(B),	 respectively,	show	the	programmed	input	and	resulting	output	seat	hinge	rotations	for	the	Experimental	seat	with	 increasing	θrearward-peak.	 The	 results	of	 the	post-hoc	 test	 (C)	 determined	 the	mean	 rank	of	each	 trial	 as	well	 as	 any	 significant	differences	 (*)	between	 the	 selected	 seat	 hinge	 rotation	 parameter	 trial	 (red)	 and	 the	 Control	 or	 other	 Experimental	 trials.	 	 	 Exemplar	 ATD	responses	(D	–	F)	and	their	corresponding	peak	responses	(G-I)	for	head	and	torso	accelerations	in	the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	criteria	(NIC),	respectively.					 175	Initial	Seat	Hinge	Angular	Velocity		 Initial	seat	hinge	rotation	velocity	was	important	because	if	the	rotation	was	too	slow,	the	ATD	would	hit	 the	seatback	cushion	and	rebound	forward.	Conversely	 if	 the	rotation	was	too	fast,	an	air	gap	may	form	between	the	seat	and	the	ATD	reducing	the	support	provided	by	the	seatback.	Initial	seat	hinge	rotation	velocities	varied	from	3.8	to	12.2	deg/s	by	expanding	or	contracting	 the	 initial	 seat	 hinge	 rotation	 profile	 in	 time.	 The	 time-stretched	 or	 -shortened	rotation	profiles	 resulted	 in	ωrearward-int	of	3.8,	5.1,	6.2,	8.5,	9.5,	and	12.2	deg/s	 (Figure	B3A	&	B3B).	 	 Exemplar	 raw	 data	 and	 peak	 responses	 of	 aX-head,	 aX-T1,	 and	 NIC	 showed	 lower	 peak	responses	with	slower	initial	seat	hinge	rotation	velocities	(Figure	B3D	–	B3I).	The	Friedman	test	revealed	a	 significant	difference	between	all	 tested	 trials	 (1	Control	 and	6	Experimental,	 χ2	 =	38.76,	df	=	6,	p	<	0.001).	 	 In	comparison	to	the	Control	trial,	ωrearward-int	=	3.8	deg/s	trial	had	a	lower	 mean	 rank	 across	 all	 ATD	 responses	 and	 neck	 injury	 criteria	 (p	 =	 0.001,	 Figure	 B3C).		Slower	 initial	seat	hinge	rotation	velocity	may	be	more	beneficial	as	 it	allows	the	occupant	to	penetrate	 deeper	 into	 the	 seatback	 to	 further	 dissipate	 the	 forces	 and	 accelerations	experienced	 by	 the	 occupant.	 Thus,	 an	 initial	 angular	 velocity	 of	 3.8	 deg/s	 was	 selected	 for	subsequent	testing	as	an	effective	initial	velocity.				 176		Figure	B3.	ATD	responses	to	the	Control	seat	(blue	lines	and	markers)	and	the	Experimental	seat	with	various	initial	velocities	of	seat	hinge	 rotation	 (θseat-hinge;	 grey	and	 red	 lines	and	markers)	during	a	12	km/h	collision	 severity	 (Δt	=	208	ms,	black	 lines,	 right	axis,	Panels	A	&	B).	The	red	lines	and	markers	indicate	the	seat	hinge	initial	angular	velocity	parameter	selected	for	further	testing.	Panels	(A)	 and	 (B),	 respectively,	 show	 the	 programmed	 input	 and	 resulting	 output	 seat	 hinge	 rotations	 for	 the	 Experimental	 seat	with	increasing	 ωrearward-int.	 The	 results	 of	 the	 post-hoc	 test	 (C)	 determined	 the	 mean	 rank	 of	 each	 trial	 as	 well	 as	 any	 significant	differences	 (*)	 between	 the	 selected	 seat	 hinge	 rotation	 parameter	 trial	 (red)	 and	 the	 Control	 or	 other	 Experimental	 trials.			Exemplar	ATD	responses	(D	–	F)	and	their	corresponding	peak	responses	(G-I)	for	head	and	torso	accelerations	in	the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	criteria	(NIC),	respectively.					 177	Seat	Hinge	Rotation	Onset		 The	seat	hinge	 rotation	onset	determined	whether	 the	seat	hinge	whiplash	mitigation	system	would	be	a	predictive	(seat	hinge	onset	before	the	onset	of	the	whiplash	perturbation:	trearward-onset	<	0	ms)	or	a	reactive	(seat	hinge	onset	after	the	onset	of	the	whiplash	perturbation:	trearward-onset	 >	 0	ms)	 system.	 	 Similar	 to	 the	 effects	 of	 initial	 seat	 hinge	 angular	 velocity,	 pilot	trials	showed	that	 trearward-onset	occurring	before	the	collision	onset	 (trearward-onset	<	0	ms)	would	create	an	air	gap	between	the	seat	and	ATD;	whereas,	a	delayed	trearward-onset	 (trearward-onset	>	80	ms)	would	cause	the	ATD	to	rebound	prematurely	away	from	the	seatback.		Thus,	the	trearward-onset	parameters	were	tested	through	a	narrow	range	of	onsets	between	10	ms	to	70	ms	after	whiplash	collision	by	shifting	the	initial	seat	hinge	rotation	profile	(trearward-onset	=	10,	30,	50	and	70	ms;	Figure	B4A	&	B4B).		Exemplar	data	and	peak	responses	of	aX-head,	aX-T1,	and	NIC	showed	little	differences	between	Control	and	Experimental	trials	(Figure	B4D	–	B4I.	The	Friedman	test,	however,	 revealed	 a	 significant	 difference	 between	 all	 tested	 trials	 (1	 Control	 and	 4	Experimental,	χ2	=	9.96,	df	=	4,	p	=	0.04),	but	the	post-hoc	test	showed	no	differences	between	any	of	the	trials	(p	>	0.12)	(Figure	B4C).		Consequently,	a	trearward-onset	=	30	ms,	the	original	time	delay	for	the	unmodified	Control	GMHR	seat,	was	selected	for	subsequent	experiments.			 			 178		Figure	B4.	ATD	responses	to	the	Control	seat	(blue	lines	and	markers)	and	the	Experimental	seat	with	various	onsets	of	seat	hinge	rotation	(θseat-hinge;	grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	208	ms,	black	lines,	right	axis,	Panels	A	&	B).	 The	 red	 lines	 and	markers	 indicate	 the	 seat	 hinge	 rotation	 onset	 parameter	 selected	 for	 further	 testing.	 Panels	 (A)	 and	 (B),	respectively,	 show	 the	 programmed	 input	 and	 resulting	 output	 seat	 hinge	 rotations	 for	 the	 Experimental	 seat	 with	 increasing	trearward-onset.	 The	 results	of	 the	post-hoc	 test	 (C)	 determined	 the	mean	 rank	of	each	 trial	 as	well	 as	 any	 significant	differences	 (*)	between	 the	 selected	 seat	 hinge	 rotation	 parameter	 trial	 (red)	 and	 the	 Control	 or	 other	 Experimental	 trials.	 	 	 Exemplar	 ATD	responses	(D	–	F)	and	their	corresponding	peak	responses	(G-I)	for	head	and	torso	accelerations	in	the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	criteria	(NIC),	respectively.					 179	Experiment	2:	Verification	of	Ideal	Seat	Hinge	Rotation	Profile	The	goal	of	Experiment	2	was	to	verify	that	the	seat	hinge	rotation	profile	determined	in	Experiment	1	(θrearward-peak	=	5.7	deg,	ωrearward-int	=	3.8	deg/s,	and	trearward-onset	=	30	ms)	was	a	local	minimum	that	best	reduced	ATD	responses.	In	this	experiment,	the	Control	trial	(blue	lines	and	markers;	 Figure	 B5)	 and	 the	 seat	 hinge	 rotation	 profile	 from	 Experiment	 1	 (red	 lines	 and	markers;	Figure	B5)	were	compared	to	24	additional	profile	with	different	combination	of	seat	hinge	rotation	parameters	tested	in	Experiment	1	(θrearward-peak	=	2.4,	5.7,	17,	29,	41	and	52	deg,	ωrearward-int	 =	 3.8,	 5.1,	 6.2,	 8.5,	 9.5,	 and	 12.2	 deg/s	 and	 trearward-onset	 =	 10,	 30,	 50,	 and	 70	ms;	Figure	 B5A	&	 B5B).	 	 	 The	 Friedman	 test	 revealed	 a	 significant	 difference	 between	 all	 tested	trials	(1	Control	and	25	Experimental,	χ2	=	139.67.,	df	=	25,	p	<	0.001).	The	multiple	comparison	test	 further	 confirmed	 that	 the	 mean	 rank	 of	 the	 trial	 with	 seat	 hinge	 rotation	 profile	determined	 in	 Experiment	 1	 (red	 lines	 and	markers)	was	 ranked	 significantly	 lower	 than	 the	Control	and	Experimental	trials	with	larger	and	faster	seat	hinge	rotation	profiles	(Figure	B5C).	This	experiment	showed	that	the	current	seat	hinge	rotation	profile	determined	in	Experiment	1	best	reduced	ATD	responses	following	low-speed,	rear-end	collisions	(Dv	=	12	km/h,	Dt	=208	ms).								 180		Figure	B5.	ATD	responses	to	the	Control	seat	(blue	 lines	and	markers)	and	the	Experimental	seat	with	various	seat	hinge	rotation	profiles	(θseat-hinge;	grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	208	ms,	black	lines,	right	axis,	Panels	A	&	B).	The	red	 lines	and	markers	 indicate	the	seat	hinge	rotation	profile	selected	for	further	testing.	Panels	(A)	and	(B),	 respectively,	show	the	programmed	 input	and	 resulting	output	 seat	hinge	 rotations	 for	all	 the	different	Experimental	 trials.	The	 results	of	 the	post-hoc	test	(C)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	differences	(*)	between	the	selected	seat	hinge	rotation	parameter	trial	(red)	and	the	Control	or	other	Experimental	trials.			Exemplar	ATD	responses	(D	–	F)	and	their	corresponding	peak	responses	(G-I)	for	head	and	torso	accelerations	in	the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	criteria	(NIC),	respectively.					 181	Experiment	3.	Effect	of	Pulse	Shape	The	seat	hinge	rotation	profile	was	initially	programmed	to	mimic	the	retraction	phase	of	 the	 seat	 hinge	 rotation	 observed	 from	 an	 unmodified	 Control	 GMHR	 seat	 (Figure	 B1).	 To	simplify	 future	 implementations	of	 the	 seatback	 rotation	profile,	 the	 shape	of	 the	 seat	hinge	rearward	 rotation	 profile	was	 converted	 into	 a	 sigmoidal	 curve	while	 preserving	 θrearward-peak,	ωrearward-int,	 and	 trearward-onset	 parameters	 (Figure	B6).	 	 Twenty-four	additional	 sigmoidal-shaped	profiles	were	then	tested	and	compared	to	the	Control	trial	(blue	lines	and	markers)	as	well	as	the	 seat	 hinge	 rearward	 rotation	 profile	 from	 previous	 Experiments	 1	 &	 2	 (green	 lines	 and	markers;	θrearward-peak	=	5.7	deg,	ωrearward-int	=	3.8	deg/s,	and	trearward-onset	=	30	ms)	(Figure	B7A	&	B7B).	The	Friedman	test	showed	a	significant	difference	between	all	tested	trials	(1	Control	and	25	Experimental,	χ2	=	217.94.,	df	=	25,	p	<	0.001).	However,	the	post-hoc	test	showed	that	the	simplified	sigmoidal	curve	of	the	seat	hinge	rotation	profile	from	Experiment	1	&	2	(Figure	B6)	was	 not	 ranked	 significantly	 different	 from	 either	 the	 Control	 trial	 or	 the	 previously	 tested	profile	in	Experiment	1	&	2	(Figure	B7C).	Despite	the	new	sigmoidal	profile	not	reducing	overall	ATD	responses	and	neck	injury	criteria	as	well	as	the	previously	test	seat	hinge	rotation	profile,	the	 sigmoidal	 curve	 would	 be	 easier	 to	 implement	 in	 future	 implementations	 of	 the	Experimental	anti-whiplash	 seat.	 	 Thus,	 this	new	sigmoidal-shaped	seat	hinge	 rotation	profile	was	chosen	for	future	experiment.			 			 182		Figure	 B6.	 Seat	 hinge	 rotation	 (θseat-hinge)	 input	 signals	 for	 the	 seat	 hinge	 rotation	 profile	determined	previously	(green	line,	Experiments	1	&	2)	and	the	new	simplified	sigmoidal	shape	curve	(red	line)	during	a	12	km/h	collision	severity	(Δt	=	208	ms,	black	lines,	right	axis).												 183		Figure	B7.	ATD	responses	to	the	Control	seat	(blue	lines	and	markers),	the	Experimental	seat	with	the	seat	hinge	rotation	(θseat-hinge)	profile	determined	previously	 in	 Experiments	1	&	2	 (green	 lines	 and	markers),	 and	 the	Experimental	 seat	with	 various	 simplified	sigmoidal	seat	hinge	rotation	profiles	(grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	208	ms,	black	lines,	right	axis,	Panels	A	&	B).	The	red	 lines	and	markers	 indicate	the	sigmoidal	seat	hinge	rotation	profile	selected	for	 further	testing.	Panels	 (A)	 and	 (B),	 respectively,	 show	 the	 programmed	 input	 and	 resulting	 output	 seat	 hinge	 rotations	 for	 all	 the	 different	Experimental	trials.	The	results	of	the	post-hoc	test	(C)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	differences	(*)	 between	 the	 selected	 seat	 hinge	 rotation	 parameter	 trial	 (red)	 and	 the	 Control	 or	 other	 Experimental	 trials.	 	 	 Exemplar	 ATD	responses	(D	–	F)	and	their	corresponding	peak	responses	(G-I)	for	head	and	torso	accelerations	in	the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	criteria	(NIC),	respectively.					 184	Experiment	4:	Pre-Perturbation	Forward	Rotation	To	 better	 maintain	 proper	 head	 and	 neck	 alignment,	 we	 investigated	 whether	dynamically	rotating	the	seat	forward	towards	the	occupant	(a	negative	rotation	about	the	Y-axis)	prior	to	the	collision	would	engage	the	occupant	earlier	and	reduce	their	responses.		The	pre-perturbation	forward	rotation	of	the	seatback	was	elicited	prior	to	the	collision	to	ensure	that	 the	 occupant	was	 supported	 as	 early	 as	 possible.	 By	 rotating	 the	 seatback	 forward,	 the	head-to-head-restraint	distance	would	potentially	decrease,	and	the	occupant’s	torso	would	be	better	coupled	with	the	seatback	throughout	the	collision.	The	pre-perturbation	rotation	onset	(tforward-onset)	was	set	at	-90	ms	prior	to	the	collision.	Previous	analysis	into	the	biomechanics	of	rear-end	collisions	showed	that	the	ATD	and	the	seat	remained	relatively	stationary	for	the	first	60	 to	 80	ms	 of	 the	 collision	 following	 the	 onset	 of	 bumper	 contact	 (McConnell	 et	 al.,	 1993,	McConnell	 et	 al.,	 1995,	 Severy	 et	 al.,	 1955).	 By	 accounting	 for	 the	 detection	 and	 processing	time	of	existing	pre-crash	avoidance	systems	(i.e.	radar	sensors	with	an	operating	range	of	5	to	200	m,	 range	 rate	 limits	 of	 37	 to	 70	m/s,	 and	 range	 rate	 accuracy	 of	 0.25	m/s;	 Zador	 et	 al.,	2000)),	 the	 Experimental	 anti-whiplash	 seat	 should	 be	 able	 to	 activate	 the	 pre-perturbation	rotation	at	a	tforward-onset	=	-90	ms	prior	to	onset	of	the	collision.			 To	determine	the	effect	of	the	pre-perturbation	forward	rotation	only,	the	Control	trial	(blue	 lines	 and	 markers)	 were	 compared	 against	 the	 sigmoidal-shaped	 rearward	 seat	 hinge	rotation	only	profile	 from	 the	previous	experiment	 (green	 lines	and	markers)	and	 three	peak	forward	rotation	profiles	(θforward-peak	=	-2.9,	-4.7,	and	-8.1	deg,	tforward-onset	=	-90	ms;	Figure	B8A	&	B8B).	During	the	collision,	the	Experimental	seat	was	programmed	to	stop	at	the	maximum			 185	forward	angle	and	was	only	allowed	to	return	to	the	normal	resting	position	after	the	collision.	The	 Friedman	 test	 showed	 a	 significant	 difference	 between	 all	 tested	 trials	 (1	 Control	 and	 4	Experimental,	χ2	=	46.2.,	df	=	4,	p	<	0.001).	Post-hoc	analysis	revealed	that	both	the	-4.7	and	-8.1	deg	forward	pre-perturbation	rotations	generated	ATD	responses	that	were	ranked	significantly	lower	than	both	the	Control	 trial	and	the	rearward	only	rotation	profile	 (Figure	B8C).	Thus,	a	θforward-peak	of	-4.7	deg	would	be	a	beneficial	addition	to	the	seat	hinge	rotation	profile	as	it	did	not	push	the	ATD	too	far	forward	and,	yet,	reduced	peak	ATD	responses	(Figure	B8D	–	B8I).					 186		Figure	B8.	ATD	responses	to	the	Control	seat	(blue	lines	and	markers),	the	Experimental	seat	with	the	sigmoidal	rearward	rotation	profile	(green	lines	and	markers,	Experiment	3),	and	the	Experimental	seat	with	increasing	angles	of	pre-perturbation	only	forward	seat	hinge	rotation	(θseat-hinge;	grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	208	ms,	black	lines,	right	axis,	Panels	A	 &	B).	 The	 red	 lines	 and	markers	 indicate	 the	 pre-perturbation	 seat	 hinge	 rotation	 parameter	 selected	 for	 subsequent	testing.	Panels	(A)	and	(B),	respectively,	show	the	programmed	input	and	resulting	output	seat	hinge	rotations	for	the	Experimental	seat	with	increasing	forward	pre-perturbation	θseat-hinge	angles.	The	results	of	the	post-hoc	test	(C)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	difference	between	 the	selected	seat	hinge	 rotation	parameter	 trial	 (red)	and	 the	Control	or	other	Experimental	trials.			Exemplar	ATD	responses	(D	–	F)	and	their	corresponding	peak	responses	(G-I)	for	head	and	torso	accelerations	in	the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	criteria	(NIC),	respectively.					 187	Experiment	5:	Final	seat	hinge	profile	combination		 In	 this	 last	 experiment,	 the	 pre-perturbation	 rotation	 profile	 was	 combined	 with	 the	sigmoidal	 rearward	 rotation	 profile	 determined	 from	 Experiment	 3.	 	 Peak	 forward	 angles	(θforward-peak	 =	 -2.9,	 -4.7,	 -5.6,	 -6.5,	 -7.2	 and	 -8.1	deg,	 tforward-onset	 =	 -90	ms,	Figure	B9A	&	B9B)	combined	with	the	rearward	rotation	profile	were	compared	to	the	Control	seat	(blue	lines	and	markers)	 and	 the	 rearward	 seat	 hinge	 rotation	 only	 profile	 (green	 lines	 and	 markers).	 In	combining	the	rearward	rotation	profile	with	the	forward	pre-perturbation	rotation,	the	onset	of	the	rearward	rotation	was	delayed	by	60	ms	to	a	trearward-onset	of	90ms	after	the	onset	of	sled	acceleration.	 The	 exemplar	 data	 and	 peak	 responses	 of	 aX-head,	 aX-T1,	 and	NIC	 showed	 graded	decreasing	 ATD	 responses	 with	 increase	 forward	 pre-perturbation	 seat	 hinge	 angle	 (Figure	B93D	 –	 B9I).	 	 The	 Friedman	 test	 showed	 a	 significant	 difference	 between	 all	 tested	 trials	 (1	Control	and	7	Experimental,	χ2	=	70.58.,	df	=	7,	p	<	0.001).	Post-hoc	comparison	showed	that	forward	pre-perturbations	of	-5.6	deg	and	greater	generated	ATD	responses	that	were	ranked	significantly	 lower	 than	 both	 the	 Control	 trial	 and	 the	 rearward	 only	 rotation	 profile	 (Figure	B9C).	 Rotating	 the	 seat	 forward	 supported	 the	 occupant	 earlier	 in	 the	 collision	 and	 the	subsequent	rearward	rotation	further	helped	to	reduce	the	accelerations	and	forces	applied	to	the	 ATD’s	 head	 and	 torso.	 	 This	 combined	 movement	 of	 forward	 and	 backward	 seat	 hinge	rotation	reduced	head	and	torso	accelerations,	upper	neck	forces	and	moments,	head	rearward	rotation	 angle	 and	 retraction	 as	 well	 as	 all	 neck	 injury	 criteria	 between	 23%	 and	 48%	 in	comparison	to	the	Control	seat.	A	θforward-peak	of	-5.6	deg	(red	lines	and	markers,	Figure	B9)	was	selected	 for	 further	 testing	because	 this	peak	angle	was	 the	 lowest	 forward	angle	 to	 show	a	significant	 difference	 in	 mean	 rank	 from	 the	 Control	 and	 rearward	 seat	 hinge	 rotation	 only			 188	trials.		This	lower	peak	θforward-peak	may	also	be	easier	to	implement	in	the	future	development	of	the	anti-whiplash	seat.				 189		Figure	B9.	ATD	responses	to	the	Control	seat	(blue	lines	and	markers),	the	Experimental	seat	with	the	sigmoidal	rearward	rotation	profile	(θseat-hinge;	green	lines	and	markers,	Experiment	3),	and	the	Experimental	seat	with	a	combination	of	increasing	amplitudes	of	pre-perturbation	 forward	 seat	 hinge	 angle	 and	 the	 rearward	 rotation	 profile	 (grey	 and	 red	 lines	 and	markers)	 during	 a	 12	 km/h	collision	severity	 (Δt	=	208	ms,	black	 lines,	 right	axis,	Panels	A	&	B).	The	red	 lines	and	markers	 indicate	the	pre-perturbation	seat	hinge	 rotation	 parameter	 selected	 for	 subsequent	 testing.	 Panels	 (A)	 and	 (B),	 respectively,	 show	 the	 programmed	 input	 and	resulting	output	seat	hinge	rotations	for	the	Experimental	seat	with	 increasing	forward	pre-perturbation	θseat-hinge	amplitudes.	The	results	of	the	post-hoc	test	(C)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	difference	between	the	selected	seat	hinge	 rotation	 parameter	 trial	 (red)	 and	 the	 Control	 or	 other	 Experimental	 trials.	 	 	 Exemplar	 ATD	 responses	 (D	 –	 F)	 and	 their	corresponding	peak	responses	(G-I)	for	head	and	torso	accelerations	in	the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	criteria	(NIC),	respectively.					 190	Summary	A	seat	hinge	rotation	profile	has	been	defined	to	investigate	the	effects	of	dynamic	seat	hinge	 rotation	during	 low-speed,	 rear-end	collisions.	This	dynamic	seat	hinge	 rotation	 (Figure	B10,	grey	line)	was	programmed	to	rotate	the	seatback	forward	(tforward-onset)	-90	ms	before	the	collision	to	a	peak	pre-perturbation	forward	angle	(θforward-peak)	of	-5.6	deg	and	then	rotate	the	seatback	rearward	to	a	peak	rearward	angle	(θrearward-peak)	of	5.7	deg	at	an	initial	angular	velocity	(ωrearward-int)	of	3.8	deg/s	occurring	90	ms	(trearward-onset)	after	the	collision	onset.	 	The	observed	output	seat	hinge	rotation	(Figure	B10,	red	 line)	resulted	in	a	pre-perturbation	θforward-peak	of	-3.6	deg	occurring	at	a	tforward-onset	of	-90	ms	followed	by	the	rearward	seat	hinge	rotation	with	a	θrearward-peak	of	3.6	deg	occurring	at	a	trearward-onset	of	40	ms.		The	seat	hinge	rotation	profile	 includes	two	components:	a	 ‘pre-perturbation’	 forward	rotation	 followed	 by	 a	 rearward	 rotation.	 	 To	 determine	 the	 individual	 contribution	 of	 each	component	and	the	combined	effect	of	the	components,	the	peak	responses	to	the	following	seat	hinge	rotation	pulses	were	compared:	1.	‘Rearward	Rotation	Only’	(Experiment	3:		θrearward-peak	 =	5.7	deg,	ωrearward-int	 =	3.8	deg/s,	 and	 trearward-onset	 =	30	ms),	 2.	 ‘Pre-Perturbation	Rotation	Only’	 (Experiment	4:	θforward-peak	=	 -4.7	deg,	 tforward-onset	=	 -90	ms),	and	3.	 the	 ‘Combination’	 	of	the	selected	pre-perturbation	forward	rotation,	 followed	by	rearward	rotation	(Experiment	5;	θforward-peak	 =	 -5.6	deg,	 tforward-onset	 =	 -90	ms,	θrearward-peak	 =	 5.7	deg,	ωrearward-int	 =	 3.8	deg/s,	 and	trearward-onset	=	90	ms).	In	the	‘Pre-Perturbation	Rotation	Only’	condition,	the	selected	θrearward-peak	=	-5.7	deg	amplitude	was	not	tested	and,	thus,	the	θrearward-peak	=	-4.7	deg	trial	was	used	in	this	comparison.	 	All	peak	responses	were	normalized	to	the	Control	seat	to	allow	for	comparison			 191	between	 seat	 hinge	 rotation	 profiles	 (Table	 B1).	 	 In	 comparison	 to	 the	 Control	 trial,	 the	‘Rearward	 Rotation	 Only’	 profile	 generated	 2%	 -	 19%	 larger	 peak	 responses	 in	 most	 ATD	responses	and	only	decreased	aX-T1,	Nij,	Nkm,	and	Dthead-restraint	peak	responses	by	-0.8%	to	-9.6%.	from	Control.	The	 ‘Pre-Perturbation	Rotation	Only’	and	the	 ‘Combination’	 seat	hinge	rotation	profiles	 decreased	 peak	 responses	 by	 11%	 to	 53%	 and	 23%	 to	 48%,	 respectively,	 from	 the	Control	trial	(Figure	B8	and	Figure	B9).	 	These	larger	decreases	observed	from	Control	seat	 in	trials	 that	 included	 a	 pre-perturbation	 forward	 seat	 hinge	 rotation	 suggested	 the	 pre-perturbation	rotation	may	be	the	key	component	generating	these	reduced	ATD	responses	and	neck	injury	criteria	responses.						 192		Figure	B10.	The	programed	input	seat	hinge	rotation	profile	(grey	line)	and	the	observed	output	seat	 hinge	 rotation	 (red	 line)	 of	 the	 anti-whiplash	 seat	 during	 a	 12	 km/h	perturbation	 (black	line,	right	axis).		The	cyan	lines	illustrated	measurements	of	each	seat	hinge	rotation	parameter	used	 to	 define	 the	 input	 rotation	 profile	 (onset	 delay	 of	 pre-perturbation	 forward	 rotation:	tforward-onset,	peak	pre-perturbation	forward	angle:	θforward-peak,	peak	rearward	angle:	θrearward-peak,	initial	rearward	angular	velocity:	ωrearward-int,	and	onset	delay	of	rearward	rotation:	trearward-onset).	This	figure	is	a	repeat	of	Figure	3.2.		 			 193				 	Table	B1.	Comparing	percentage	decreases	of	normalized	Experimental	conditions	from	Control	seat	for	‘Rearward	Rotation	Only’,	forward	‘Pre-Perturbation	Rotation	Only’	and	the	‘Combination’	of	the	two	different	rotational	components.	Condition	 aX-head	 aX-T1	 FX	 FZ	 MY	 ωY-head	 θhead	 NIC	 Nij	 Nkm	 RX	Δthead-	contact	Rearward	Rotation	Only	 9.8	 -9.6	 2.4	 18.6	 4.7	 -5.8	 5.8	 2.3	 -4.3	 -4.4	 4.2	 -0.8	Pre-Perturbation	Rotation	Only	 -36.1	 -34.7	 -34.4	 -45.4	 -47.1	 44.5	 -52.7	 -40.8	 -46.4	 -41.7	 -41.7	 -10.5	Combination	 --43.2	 -23.3	 -33.3	 -35.2	 -43.2	 45.0	 -41.4	 -47.8	 -29.4	 -41.3	 -39.3	 -33.5			 194	Appendix	C:	Creating	the	Seatback	Cushion	Deformation	Profile	The	goal	of	this	appendix	was	to	develop	a	seatback	cushion	deformation	profile	that	best	reduce	 whiplash-evoked	 ATD	 responses.	 This	 appendix	 summarizes	 a	 series	 of	 three	experiments	 undertaken	 to	 create	 the	 active	 seatback	 cushion	 deformation	 profiles	 used	 in	Experiment	3	 (Chapters	4).	 In	all	 three	experiments,	 the	ATD	was	seated	 in	 the	Experimental	anti-whiplash	seat	and	exposed	to	a	series	of	perturbations	with	a	collision	severity	of	12	km/h	(Δt	=	195	ms).	Rigid	metal	braces	were	installed	between	the	seatback	and	the	base	of	the	sled	to	 prevent	 the	 seatback	 from	 rotating	 and	 to	 isolate	 only	 the	 effects	 of	 seatback	 cushion	deformation	 (Figure	 C1).	 For	 these	 experiments,	 a	 standard	 GMHR	 seatback	was	 suspended	across	the	rigid	seatback	by	seatbelt	straps	attached	to	the	three	seatback	motors	(upper-torso,	mid-torso	and	pelvis).	For	all	 tests,	 the	upper-torso	motor	was	 locked	 in	a	stationary	position	and	effectively	created	a	hinge	point	at	the	upper	back	to	promote	pocketing	of	the	pelvis	into	the	seatback.		A	motion	capture	system	(Optotrak	Certus,	Northern	Digital,	Waterloo,	ON,	Canada)	was	used	 to	 estimate	 the	 deformation	 of	 the	 seatback	 cushion	 during	 the	 perturbation	 because	slack	 in	 the	 straps	 induced	 by	 the	 two	 lower	 seatback	 motors	 did	 not	 correspond	 with	 the	physical	 penetration	 of	 the	 occupant.	 	 Maximum	 rearward	 penetration	 of	 the	 ATD	 T1	(dpenetration-T1)	and	pelvis	(dpenetration-pelvis)	were	defined	as	the	horizontal	displacement	(along	the	global	 X-axis,	 Figure	 C1)	 of	 the	 T1	 vertebra	 and	 pelvis	 infrared	 light	 emitting	 diode	 (IRED)	markers	relative	to	the	seatback.	Rearward	displacement	deeper	into	the	seatback	was	defined			 195	as	 negative	 values	 and	 forward	 rebound	 away	 from	 the	 seat	was	 defined	 as	 positive	 values.			The	initial	position	of	the	ATD	was	set	as	dpenetration-T1	=	0	mm	and	dpenetration-pelvis	=	0	mm.	A	sigmoidal-shaped	profile	was	used	as	a	template	for	the	seatback	motors.	The	 input	seatback	cushion	deformation	profile	rotated	the	seatback	motors	and	gear	head	combination	to	a	peak	angle	 (θdeformation-peak)	of	56	deg	with	an	 initial	angular	velocity	 (ωdeformation-int)	of	228	deg/s.	 Preliminary	 experiments	 showed	 little	 effect	 of	 seatback	 cushion	 deformation	 on	ATD	responses	 when	 the	 seatback	 motors	 onset	 time	 (tdeformation-onset)	 occurred	 30	 ms	 after	 the	collision	 onset.	 	 However,	 decreased	 ATD	 responses	were	 observed	when	 tdeformation-onset	 was	shifted	to	-130	ms	before	the	collision	onset.		The	initial	seatback	cushion	deformation	profile	to	 the	 motors	 (θdeformation-peak	 =	 56	 deg,	 ωdeformation-int	 =	 228	 deg/s,	 tdeformation-onset	 =	 -130ms)	generated	occupant	penetration	into	the	seatback	of	dpenetration-T1	=	-78	mm,	and	dpenetration-pelvis	=	-75	mm.		This	rotation	profile	was	used	as	the	starting	point	from	which	the	parameters	for	a	seatback	 cushion	 deformation	 profile	 were	 isolated	 and	 evaluated	 through	 a	 series	 of	 three	experiments.	In	addition	to	the	Control	GMHR	seat	used	in	Appendix	B,	a	No	Motion	condition	which	 consisted	 of	 the	 Experimental	 seat	 with	 rigid	 metal	 braces	 and	 the	 seat	 hinge	 and	seatback	motors	locked	in	their	initial	position	was	also	tested.		Nine	peak	kinematic	and	kinetic	(aX-head,	aX-T1,	ωY-head,	FX,	FZ,	MY,	θhead,	RX	and	Δthead-contact)	and	three	peak	neck	injury	criteria	(NIC,	Nkm,	and	Nij)	responses	were	selected	to	quantify	the	ATD	responses.		Within	each	response	variable,	the	different	profiles	were	ranked,	and	then	the	profile	with	the	lowest	averaged	rank	across	all	12	variables	was	identified	as	the	best	profile.	A	Friedman	test	determined	if	there	was	a	significant	difference	between	the	average	rank	of	the			 196	different	profiles	(chi-squared	value:	χ2,	degrees	of	freedom:	df,	and	the	significance	level:	p).	A	post-hoc,	 pair-wise	 multiple	 comparison	 test	 was	 then	 performed	 on	 the	 results	 of	 the	Friedman	test	to	determine	differences	between	individual	profiles.	 	All	tests	were	performed	using	predefined	functions	(friedman	and	multcompare)	in	Matlab	using	a	=	0.05.							Figure	C1.	Photograph	of	the	experimental	set-up	with	the	BioRID	II	ATD	on	the	Experimental	anti-whiplash	automotive	seat	and	rigid	metal	braces	to	prevent	seat	hinge	rotation.	The	global	laboratory	 reference	 frame	 is	 illustrated	 with	 positive	 X-axis	 forward	 and	 positive	 Z-axis	downwards.	The	seat	hinge	and	seatback	motors	on	the	left	side	of	the	Experimental	seat	are	labelled.	Additional	 seat	hinge	and	 seatback	motors	are	 located	on	 the	 right	 side	of	 the	 seat	(not	 labelled).	 	 The	 seat	 hinge	motors	 are	 not	 used	 in	 this	 current	 appendix.	 This	 figure	 is	 a	repeat	of	Figure	4.1.					 197	Experiment	1:	Effects	of	Individual	Seatback	Parameters	The	effects	of	three	seatback	motor	rotation	parameters	(peak	amplitude,	initial	angular	velocity	 and	 onset)	were	 evaluated	 to	 determine	 specific	 parameters	 that	 best	 reduced	ATD	responses.	Each	parameter	of	the	initial	profile	was	sequentially	varied	to	determine	its	effect	on	the	ATD	responses.		In	these	experiments,	the	seatback	motor	rotation	parameters	defined	the	characteristics	of	a	motor	pulley	that	controlled	the	deformation	of	the	seatback	cushion.		Thus,	a	motion	capture	system	(Optotrak	Certus,	Northern	Digital,	Waterloo,	ON,	Canada)	was	used	to	estimate	the	deformation	of	the	seatback	cushion	during	the	perturbation	because	the	rotation	 of	 the	 seatback	 motors	 did	 not	 correspond	 with	 the	 physical	 penetration	 of	 the	occupant.	 However,	 these	 seatback	 motor	 rotation	 parameters	 were	 used	 to	 define	 the	seatback	deformation	profile	because	they	were	manipulated	to	provide	a	repeatable	profile.	Peak	Seatback	Motor	Angle		 The	first	parameter	varied	was	peak	angle	of	the	seatback	motors	from	28	to	168	deg	from	the	initial	resting	position	(mid-torso	motor:	22	deg	and	pelvis	motor:	25	deg,	where	0	deg	is	defined	as	a	taut	seatbelt)	(Figure	C2A).	 	To	achieve	these	peak	rotations,	the	amplitude	of	the	initial	seatback	motor	profile	was	scaled	by	50,	100,	150,	200,	250	and	300%	and	resulted	in	θdeformation-peak	 of	 28,	 56,	 84,	 112,	 140	 and	 168	 deg.	 Increasing	 θdeformation-peak	 led	 to	 increased	peak	dpenetration-pelvis	and	had	 little	effect	on	peak	dpenetration-T1.	Exemplar	ATD	head	acceleration	(aX-head)	 and	neck	 injury	 criteria	 (NIC)	data	 showed	decreasing	peak	 responses	with	 increased	θdeformation-peak;	whereas,	peak	T1	acceleration	(aX-T1)	remained	similar	across	all	test	conditions	(Figure	C3).	A	Friedman	test	showed	a	significant	difference	between	all	tested	trials	(1	Control			 198	seat,	1	No	Motion	and	6	Experimental	conditions,	χ2	=	53.43,	df	=	7,	p	<	0.001).		The	profile	with	a	θdeformation-peak	=	168	deg,	which	generated	occupant	penetration	at	T1	 (dpenetration-T1)	of	 -81.6	mm	and	at	the	pelvis	(dpenetration-pelvis)	of	-93.7	mm,	received	the	lowest	mean	rank	score	for	all	ATD	responses	and	neck	injury	criteria	(Figure	C2C	&	C2D).	Post-hoc	comparisons	showed	that	this	 ranked	score	was	significantly	 lower	 than	 the	Control	 seat,	 the	No	Motion	condition	and	θdeformation-peak	≤	84	deg	profiles	(multiple	p	≤	0.024)	(Figure	C2B).	These	results	suggested	that	larger	 θdeformation-peak	 better	 reduced	 occupant	 responses	 despite	 similar	 aX-T1	 levels	 across	 all	conditions.		Deeper	penetration	of	the	occupant	into	the	seatback	may	delay	the	onset	of	torso	acceleration	and	decrease	the	head-to-head-restraint	distance	to	better	support	the	head	and	neck	 during	 the	 collision.	 	 Subsequent	 experiments	 to	 determine	 the	 seatback	 motor	deformation	profile	will	be	conducted	at	the	largest	tested	seatback	motor	angle	of	θdeformation-peak	=	168	deg.				 			 199		Figure	 C2.	 Seatback	 motor	 profiles	 and	 ATD	 responses	 to	 the	 Control	 seat	 (blue	 lines	 and	markers),	the	No	Motion	condition	(magenta	lines	and	markers)	and	the	Experimental	seat	with	various	peak	angles	of	seatback	motor	rotations	(θdeformation-peak;	grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	195	ms,	black	lines,	Panel	A,	right	axis).	The	red	lines	and	 markers	 indicate	 the	 seatback	 motor	 parameter	 (θdeformation-peak	 =	 168	 deg)	 selected	 for	further	 testing.	 Panel	 A	 shows	 the	 programmed	 input	 seatback	 motor	 rotation	 for	 the	Experimental	 seat	 with	 increasing	 	 	 θdeformation-peak.	 The	 results	 of	 the	 post-hoc	 test	 (B)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	differences	(*)	between	the	selected	 seatback	 motor	 parameter	 trial	 (red)	 and	 the	 Control,	 the	 No	 Motion	 or	 other	Experimental	 trials.	Panels	C	 and	D	 show	the	 resulting	 seatback	deformation	at	 the	T1	 spinal	level	and	at	the	pelvis	(dpenetration-T1	and	dpenetration-pelvis,	respectively).				 200		Figure	C3.	Exemplar	ATD	responses	for	the	Control	seat	(blue	lines	and	markers),	the	No	Motion	condition	(magenta	lines	and	markers),	and	the	Experimental	seat	with	various	peak	angles	of	the	seatback	motors	(θdeformation-peak;	grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	195	ms).	Calibrated	responses	(A	–	C)	and	their	corresponding	peak	responses	(D	–	F)	for	head	and	torso	accelerations	 in	the	X-direction	(aX-head	&	aX-T1)	and	neck	 injury	criteria	(NIC),	 respectively.	 	 The	 red	 lines	 and	 markers	 indicate	 the	 seatback	 motor	 parameter	(θdeformation-peak	=	168	deg)	selected	for	further	testing.					 201	Initial	Seatback	Motor	Angular	Velocity		 Initial	 seatback	 motor	 angular	 velocity	 controlled	 the	 rate	 at	 which	 the	 motors	unspooled	 the	 seatbelt	webbing	 to	allow	occupant	penetration	 into	 the	 seatback.	 	The	 initial	angular	velocities	varied	from	268	to	1201	deg/s	and	were	created	by	expanding	or	contracting	the	 initial	 seatback	 motor	 rotational	 profiles	 in	 time.	 The	 resulting	 profiles	 consisted	 of	θdeformation-peak	=	168	deg,	tdeformation-onset	=	-130	ms,	and	ωdeformation-int	of	268,	351,	683,	826,	980,	and	1201	deg/s	 (Figure	C4A).	Exemplar	peak	ATD	head	and	T1	acceleration	 responses	 (aX-head	and	 aX-T1:	 Figure	 C5A	&	 C5B)	 did	 not	 vary	 with	 the	 different	 profiles,	 but	 peak	 neck	 injury	criteria	(NIC:	Figure	C5C)	decreased	with	increased	ωdeformation-int.	The	Friedman	test	revealed	a	significant	 difference	 between	 all	 tested	 trials	 (1	 Control,	 1	 No	 Motion	 condition,	 and	 6	Experimental,	 χ2	 =	 60.07,	 df	 =	 7,	 p	 <	 0.001).	 	 Post-hoc	 comparisons	 showed	 that	 the	Experimental	 trial	with	ωdeformation-int	=	980	deg/s	had	 the	 lowest	mean	rank	score	and	ranked	lower	than	the	Control	seat,	the	No	Motion	condition	and	profiles	with	ωdeformation-int	≤	351	deg/s	(p	≤	0.014,	Figure	C4B).	The	Experimental	trial	with	ωdeformation-int	=	980	deg/s	resulted	 in	peak	penetration	 of	 -83.2	 mm	 at	 T1	 (dpenetration-T1)	 and	 of	 -99.8	 mm	 at	 the	 pelvis	 (dpenetration-pelvis).		Subsequent	 experiments	 to	 determine	 the	 seatback	 motor	 deformation	 profile	 will	 be	conducted	with	an	ωdeformation-int	=	980	deg/s	to	the	θdeformation-peak	=	168	deg.					 202			Figure	 C4.	 Seatback	 motor	 profiles	 and	 ATD	 responses	 to	 the	 Control	 seat	 (blue	 lines	 and	markers),	 the	 No	Motion	 condition	 (magenta	 lines	 and	markers),	 and	 the	 Experimental	 seat	with	various	 initial	angular	velocities	of	 the	seatback	motors	 (ωdeformation-int;	grey	and	red	 lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	195	ms,	black	lines,	Panel	A,	right	axis).	The	 red	 lines	 and	markers	 indicate	 the	 initial	 angular	 velocity	 parameter	 (ωdeformation-int	 =	 980	deg/s)	 of	 the	 seatback	motors	 selected	 for	 further	 testing.	 Panel	A	 shows	 the	 programmed	input	of	the	seatback	motors	for	the	Experimental	seat	with	increasing	ωdeformation-int.	The	results	of	 the	 post-hoc	 test	 (B)	 determined	 the	 mean	 rank	 of	 each	 trial	 as	 well	 as	 any	 significant	differences	(*)	between	the	selected	seatback	motor	parameter	trial	(red)	and	the	Control	seat,	the	 No	 Motion	 condition	 or	 other	 Experimental	 trials.	 Panels	 C	 and	 D	 show	 the	 resulting	seatback	 deformation	 at	 the	 T1	 spinal	 level	 and	 at	 the	 pelvis	 (dpenetration-T1	 and	 dpenetration-pelvis,	respectively).		 			 203			Figure	C5.	Exemplar	ATD	responses	for	the	Control	seat	(blue	lines	and	markers),	the	No	Motion	condition	 (magenta	 lines	and	markers),	and	the	Experimental	 seat	with	various	 initial	angular	velocity	of	the	seatback	motor	(ωdeformation-int;	grey	and	red	lines	and	markers)	during	a	12	km/h	collision	 severity	 (Δt	 =	 195	 ms).	 Calibrated	 responses	 (A	 –	 C)	 and	 their	 corresponding	 peak	responses	 (D	–	F)	 for	head	and	torso	accelerations	 in	the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	criteria	(NIC),	respectively.		The	red	lines	and	markers	indicate	the	initial	seatback	motor	rotation	velocity	parameter	(ωdeformation-int	=	980	deg/s)	selected	for	further	testing.			 			 204	Seatback	Cushion	Deformation	Onset		 The	 seatback	 cushion	 deformation	 onset	 (tdeformation-onset)	 defined	 when	 the	 seatback	motors	start	to	unspool.		In	this	experiment,	tdeformation-onset	was	varied	between	-250	ms	before	to	200	ms	after	the	collision	onset	(tdeformation-onset	=	-250,	-200,	-150,	-130,	-100,	-50,	0,	50,	100,	150	or	200	ms)	(Figure	C6A).	For	tdeformation-onset	=	-200	ms	or	earlier,	the	aX-head	and	NIC	showed	lower	peak	responses	than	both	the	Control	seat	and	the	No	Motion	condition	(Figure	C7D	&	C7F);	whereas	peak	aX-T1	responses	were	similar	across	all	trials	(Figure	C7E).	The	Friedman	test	revealed	a	significant	difference	between	trials	(1	Control	seat,	1	No	Motion	condition,	and	11	Experimental	 conditions,	 χ2	 =	89.18,	df	=	12,	p	<	0.001).	 	 	Post-hoc	comparisons	 showed	 that	tdeformation-onset	 	of	 -250	and	-200	ms	were	ranked	significantly	 lower	than	the	Control	seat	and	the	No	Motion	 condition	 as	well	 as	 tdeformation-onset	 =	 100	 and	200	ms	 trials	 (p	 ≤	 0.016,	Figure	C6B).	For	the	profile	with	tdeformation-onset	=	-200	ms,	the	ATD	reached	a	peak	penetration	of	-80.1	mm	 at	 T1	 (dpenetration-T1)	 and	 -99.4	mm	 at	 the	 pelvis	 (dpenetration-pelvis)	 (Figure	 C6C	&	C6D).	 The	combination	of	 existing	pre-crash	avoidance	 systems	as	well	 as	 the	 time	 for	 the	acceleration	pulse	 to	 propagate	 through	 the	 vehicle	 to	 the	 seat	 would	 be	 sufficient	 to	 activate	 the	deformation	of	the	seatback	-200	ms	prior	to	the	onset	of	the	collision	(McConnell	et	al.,	1993,	McConnell	et	al.,	1995,	Severy	et	al.,	1955,	Zador	et	al.,	2000).	Thus,	a	tdeformation-onset	=	-200	ms	was	selected	for	subsequent	experiments	with	the	seatback	motor	deformation	pulse.			 205		Figure	 C6.	 Seatback	 motor	 profiles	 and	 ATD	 responses	 to	 the	 Control	 seat	 (blue	 lines	 and	markers),	 the	 No	Motion	 condition	 (magenta	 lines	 and	markers),	 and	 the	 Experimental	 seat	with	 various	onsets	of	 the	 seatback	motor	 rotation	 (tdeformation-onset;	 grey,	 green,	 and	 red	 lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	195	ms,	black	lines,	Panel	A,	right	axis).	The	green	lines	and	markers	indicate	the	initial	seatback	motor	onset	used	in	the	previous	two	experiments	 (tdeformation-onset	 =	 -130	ms)	 and	 the	 red	 lines	 and	markers	 indicate	 the	 seatback	motor	onset	parameter	 (tdeformation-onset	 =	 -200	ms)	 selected	 for	 further	 testing.	Panel	A	 shows	the	 programmed	 input	 of	 the	 seatback	 motors	 for	 the	 Experimental	 seat	 with	 increasing	tdeformation-onset.	 The	 results	of	 the	post-hoc	 test	 (B)	determined	 the	mean	 rank	of	each	 trial	 as	well	 as	 any	 significant	 differences	 (*)	 between	 the	 selected	 seatback	 motor	 parameter	 trial	(red)	and	the	Control	seat,	the	No	Motion	condition	or	other	Experimental	trials.	Panels	C	and	D	show	the	resulting	seatback	deformation	at	the	T1	spinal	level	and	at	the	pelvis	(dpenetration-T1	and	dpenetration-pelvis,	respectively).		 			 206		Figure	C7.	Exemplar	ATD	responses	for	the	Control	seat	(blue	lines	and	markers),	the	No	Motion	condition	 (magenta	 lines	and	markers),	 and	 the	Experimental	 seat	with	various	onsets	of	 the	seatback	motor	 rotation	 (tdeformation-onset;	 grey,	 green,	 and	 red	 lines	 and	markers)	 during	 a	 12	km/h	collision	severity	(Δt	=	195	ms).	Calibrated	responses	(A	–	C)	and	their	corresponding	peak	responses	 (D	–	F)	 for	head	and	torso	accelerations	 in	the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	 criteria	 (NIC),	 respectively.	 	 The	 green	 lines	 and	 markers	 indicate	 the	 initial	 seatback	motor	onset	used	in	the	previous	two	experiments	(tdeformation-onset	=	-130	ms)	and	the	red	lines	and	markers	 indicate	the	seatback	motor	onset	parameter	 (tdeformation-onset	=	 -200	ms)	selected	for	further	testing.			 			 207	Experiment	2:	Verification	of	Ideal	Seatback	Deformation	Profile	The	 goal	 of	 Experiment	 2	 was	 to	 verify	 that	 the	 seatback	 deformation	 profile	determined	in	Experiment	1	(θdeformation-peak	=	168	deg,	ωdeformation-int	=	980	deg/s,	and	tdeformation-onset	=	-200	ms)	was	a	local	minimum	that	best	reduced	ATD	responses.	In	this	experiment,	the	Control	 seat	 (blue	 lines	and	markers),	 the	No	Motion	condition	 (magenta	 lines	and	markers),	and	 the	 seatback	 deformation	 profile	 from	 Experiment	 1	 (red	 lines	 and	 markers)	 were	compared	 against	 12	 additional	 profiles	 with	 different	 combinations	 of	 seatback	 motor	parameters.		The	tdeformation-onset	parameter	was	kept	constant	at	-200	ms,	but	θdeformation-peak	and	ωdeformation-int	were	varied	(θdeformation-peak	=	56,	112	and	168	deg	and	ωdeformation-int	=	176,	683,	980	and	1485	deg/s;	Figure	C8A).			The	Friedman	test	revealed	a	significant	difference	between	all	tested	trials	(1	Control,	1	No	Motion	and	12	Experimental,	χ2	=	100.59,	df	=	13,	p	<	0.001).	Post-hoc	 comparisons	 showed	 that	 the	 mean	 rank	 of	 the	 trial	 with	 seatback	 motor	 deformation	profile	determined	 in	Experiment	1	 (red	plot;	dpenetration-T1	=	 -84.3	mm	&	dpenetration-pelvis	=	 -96.8	mm)	was	ranked	third	lowest	and	was	ranked	significantly	lower	than	the	Control	seat,	the	No	Motion	condition,	and	two	other	Experimental	trials	(Figure	C8B).	The	other	two	Experimental	conditions	that	ranked	lower	than	the	seatback	motor	profile	both	had	θdeformation-peak	=	168	deg	with	different	ωdeformation-int	=	683	and	1485	deg/s,	but	 these	conditions	were	not	 significantly	different	from	the	selected	seatback	profile	condition.		These	results	suggest	that	the	θdeformation-peak	 parameter	 may	 have	 a	 larger	 influence	 on	 the	 reduction	 of	 ATD	 responses	 than	 the	ωdeformation-int	parameter.		This	experiment	showed	that	the	current	seatback	motor	deformation	profile	determined	in	Experiment	1	is	a	potentially	effective	solution	to	reducing	ATD	responses	following	low-speed,	rear-end	collisions	(Dv	=	12	km/h,	Dt	=195	ms).					 208						Figure	 C8.	 Seatback	 motor	 profiles	 and	 ATD	 responses	 to	 the	 Control	 seat	 (blue	 lines	 and	markers),	 the	 No	Motion	 condition	 (magenta	 lines	 and	markers),	 and	 the	 Experimental	 seat	with	various	combinations	of	θdeformation-peak	and	ωdeformation-int	parameters	at	a	constant	tdeformation-onset	=	-200	ms	(grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	195	ms,	black	 lines,	Panel	A,	right	axis).	The	red	 lines	and	markers	 indicate	the	seatback	motor	profile	(θdeformation-peak	=	168	deg,	ωdeformation-int	 =	980	deg/s,	 and	 tdeformation-onset	 =	 -200	ms)	determined	from	 Experiment	 1	 and	 selected	 for	 further	 testing.	 The	 results	 of	 the	 post-hoc	 test	 (B)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	differences	(*)	between	the	selected	 seatback	 motor	 parameter	 trial	 (red)	 and	 the	 Control,	 the	 No	 Motion	 or	 other	Experimental	 trials.	Panels	C	 and	D	 show	the	 resulting	 seatback	deformation	at	 the	T1	 spinal	level	and	at	the	pelvis	(dpenetration-T1	and	dpenetration-pelvis,	respectively).				 209			Figure	C9.	Exemplar	ATD	responses	for	the	Control	seat	(blue	lines	and	markers),	the	No	Motion	condition	(magenta	lines	and	markers),	and	the	Experimental	seat	with	various	combinations	of	θdeformation-peak	and	ωdeformation-int	parameters	at	a	constant	tdeformation-onset	=	-200	ms	(grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	195	ms).	Calibrated	responses	(A	–	C)	and	 their	 corresponding	peak	 responses	 (D	–	F)	 for	head	and	 torso	accelerations	 in	 the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	criteria	(NIC),	respectively.		The	red	lines	and	markers	indicate	 the	 seatback	motor	 profile	 (θdeformation-peak	 =	 168	 deg,	 ωdeformation-int	 =	 980	 deg/s,	 and	tdeformation-onset	=	-200	ms)	determined	from	Experiment	1	and	selected	for	further	testing.				 			 210	Experiment	3:	Forward	Pre-Perturbation	Seatback	Deformation	The	results	 from	the	seat	hinge	rotation	experiments	suggested	that	 the	addition	of	a	forward	 pre-perturbation	 rotation	 of	 the	 seatback	 can	 significantly	 reduce	 ATD	 responses	during	 the	 whiplash	 perturbation	 (see	 Appendix	 B:	 Experiment	 4).	 Thus,	 we	 investigated	whether	 a	 forward	 (positive	 X-direction)	 deformation	 of	 the	 seatback	 toward	 the	 occupant	prior	 to	 the	 perturbation	 would	 further	 reduce	 ATD	 responses.	 A	 forward	 pre-perturbation	seatback	 deformation	 was	 added	 to	 the	 seatback	 motor	 deformation	 profile	 determined	 in	Experiment	 2.	 	 Nine	 different	 combinations	 of	 peak	 forward	 deformation	 amplitudes	(θdeformation-forward	=	-9.1,	-15.8,	and	-21.4	deg)	and	onsets	(tdeformation-forward	=	-700,	-500	and	-300	ms,	Figure	C10A)	were	tested.		These	nine	Experimental	conditions	were	then	compared	to	the	Control	 seat	 (blue	 lines	 and	markers),	 the	No	Motion	 condition	 (magenta	 lines	 and	markers)	and	 the	 ‘rearward	 only	 seatback’	 deformation	 profile	 (red	 lines	 and	markers).	 	 None	 of	 the	profiles	generated	markedly	lower	ATD	peak	aX-head	and	NIC	responses	than	the	rearward-only	seatback	deformation	profile	(Figure	C11).	The	Friedman	test	revealed	a	significant	difference	between	all	trials	(1	Control,	1	No	Motion	and	10	Experimental,	χ2	=	72.25.,	df	=	11,	p	<	0.001).	Post-hoc	comparisons	showed	that	only	two	pre-perturbation	conditions	and	the	rearward-only	seatback	deformation	profile	were	 ranked	 significantly	 lower	 than	both	 the	Control	 seat	 and	the	 No	 Motion	 condition	 (p	 ≤	 0.037,	 Figure	 C10B).	 The	 mean	 rank	 of	 these	 three	 lowest	conditions	were	not	significantly	different	from	each	other.	Given	that	the	addition	of	the	pre-perturbation	 seatback	deformation	did	not	 significantly	 further	 reduce	ATD	 responses,	a	pre-perturbation	 component	 of	 the	 seatback	 deformation	 was	 excluded	 from	 the	 final	 seatback	deformation	profile.					 211			Figure	 C10.	 Seatback	 motor	 profiles	 and	 ATD	 responses	 to	 the	 Control	 seat	 (blue	 lines	 and	markers),	 the	 No	Motion	 condition	 (magenta	 lines	 and	markers),	 and	 the	 Experimental	 seat	with	 either	 different	 combinations	 of	 pre-perturbation	 forward	 seatback	 deformation	 pulses	parameters	 (θdeformation-forward	 and	 tdeformation-forward,	 grey	 lines	 and	 markers)	 or	 without	 a	 pre-perturbation	 (red	 lines	 and	markers)	 during	 a	 12	 km/h	 collision	 severity	 (Δt	 =	 195	ms,	 black	lines,	 Panel	 A,	 right	 axis).	 The	 red	 lines	 and	 markers	 indicate	 the	 seatback	 motor	 profile	(θdeformation-peak	=	168	deg,	ωdeformation-int	 =	980	deg/s,	 and	 tdeformation-onset	 =	 -200	ms)	determined	from	 Experiment	 1	&	 2	 and	 selected	 for	 further	 testing.	 The	 results	 of	 the	 post-hoc	 test	 (B)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	differences	(*)	between	the	selected	seatback	motor	parameter	trial	(red)	and	the	Control	seat,	the	No	Motion	condition	or	other	 Experimental	 trials.	 Panels	C	 and	D	 show	 the	 resulting	 seatback	deformation	at	 the	T1	spinal	level	and	at	the	pelvis	(dpenetration-T1	and	dpenetration-pelvis,	respectively).						 212			Figure	 C11.	 Exemplar	 ATD	 responses	 to	 the	 Control	 seat	 (blue	 lines	 and	 markers),	 the	 No	Motion	condition	(magenta	lines	and	markers),	and	the	Experimental	seat	with	either	different	combinations	of	pre-perturbation	forward	seatback	deformation	pulses	parameters	(θdeformation-forward	and	tdeformation-forward,	grey	lines	and	markers)	or	without	a	pre-perturbation	(red	lines	and	markers)	during	a	12	 km/h	 collision	 severity	 (Δt	 =	195	ms).	 Calibrated	 responses	 (A	–	C)	 and	their	corresponding	peak	responses	(D	–	F)	for	head	and	torso	accelerations	in	the	X-direction	(aX-head	&	aX-T1)	and	neck	 injury	criteria	 (NIC),	 respectively.	 	The	red	 lines	and	markers	 indicate	the	 seatback	motor	profile	 (θdeformation-peak	=	168	deg,	ωdeformation-int	 =	980	deg/s,	and	 tdeformation-onset	=	-200	ms)	determined	from	Experiment	1	&	2	and	selected	for	further	testing.			 			 213	Summary	A	 seatback	motor	 deformation	 profile	 has	 been	 defined	 to	 investigate	 the	 effects	 of	seatback	cushion	deformation	on	ATD	responses	to	low-speed,	rear-end	collisions.	The	dynamic	seatback	 motor	 deformation	 profile	 (Figure	 C12,	 grey	 line)	 was	 programmed	 to	 rotate	 the	seatback	motors	 (tdeformation-onset)	 -200	ms	before	 the	collision	onset	 to	a	peak	seatback	motor	angle	(θdeformation-peak)	of	168	deg	at	an	initial	angular	velocity	(ωdeformation-int)	of	980	deg/s.	 	The	resulting	ATD	penetration	into	the	seatback	was	-84.3	mm	at	T1	(dpenetration-T1,	green	line)	and	-96.8	 mm	 at	 the	 pelvis	 (dpenetration-pelvis,	 blue	 line)	 (Figure	 C12).	 	 With	 the	 exception	 of	 linear	forward	 acceleration	 of	 the	 T1	 vertebrae	 (aX-T1),	 the	 dynamic	 seatback	 cushion	 deformation	decreased	the	other	eight	peak	kinematic	and	kinetic	responses	(aX-head,	ωY-head,	FX,	FZ,	MY,	θhead,	RX	and	Δthead-contact)	and	three	peak	neck	injury	criteria	(NIC,	Nkm,	and	Nij)	from	the	Control	trial.					 214		Figure	C12.	Horizontal	seatback	cushion	deformation	of	the	T1	vertebra	and	pelvis	(dpenetration-T1,	green	line	&	dpenetration-pelvis,	blue	line)	in	response	to	the	seatback	motor	rotation	(θseatback-motor,	grey	 line)	 with	 θdeformation-peak	 =	 168	 deg,	 ωdeformation-int	 =	 980	 deg/s,	 and	 tdeformation-onset	 =	 -200	(deformation	profile	parameters,	cyan	lines)	during	a	12	km/h	collision	severity	(aX-sled,	Δt	=	195	ms,	black	line,	right	axis).				 			 215	Appendix	 D:	 	 Combining	 the	 Seat	 Hinge	 Rotation	 and	 the	 Seatback	Cushion	Deformation	Profiles		In	 this	 appendix,	 the	 combined	 effects	 of	 the	 seat	 hinge	 rotation	 and	 the	 seatback	cushion	deformation	profiles	identified	in	Appendix	B	and	C	were	evaluated	through	a	series	of	three	 experiments	 varying	 the	 relative	 onset	 of	 the	 seat	 hinge	 and	 seatback	 deformation	profiles.	 An	 ATD	 seated	 on	 the	 Experimental	 anti-whiplash	 seat	 was	 exposed	 to	 a	 series	 of	perturbations	with	a	collision	severity	of	12	km/h	(Δt	=	187	ms)	 (Figure	D1).	An	Optotrak	3-D	motion	 capturing	 system	 was	 used	 to	 quantify	 the	 seatback	 deformation	 induced	 by	 the	rotations	of	the	seatback	motors.	Maximum	rearward	penetration	of	the	ATD	T1	(dpenetration-T1)	and	pelvis	(dpenetration-pelvis)	were	defined	as	the	horizontal	displacement	(along	the	global	X-axis,	see	Figure	A4	in	Appendix	A)	of	the	T1	vertebra	and	pelvis	infrared	light	emitting	diode	(IRED)	markers	relative	to	the	seatback.	Rearward	displacement	deeper	into	the	seatback	was	defined	as	 negative	 values	 and	 forward	 rebound	 away	 from	 the	 seat	was	 defined	 as	 positive	 values.		The	initial	position	of	the	ATD	was	set	as	dpenetration-T1	=	0	mm	and	dpenetration-pelvis	=	0	mm.	The	dynamic	seat	hinge	rotation	profile	(Figure	D1A,	grey	line)	was	programmed	to	rotate	the	seatback	forward	commencing	at	90	ms	(tforward-onset)	before	the	collision	onset	by	5.6	deg	(θforward-peak)	 at	 an	angular	 velocity	of	 -33.4	deg/s	 (ωforward-int)	 and	 to	 then	 rotate	 the	 seatback	rearward	by	5.7	deg	(θrearward-peak)	at	an	angular	velocity	(ωrearward-int)	of	24.0	deg/s	beginning	70	ms	(trearward-onset)	after	collision	onset.	The	observed	output	seat	hinge	rotation	(Figure	5.2A,	red	line)	resulted	in	a	forward	rotation	of	4.6	deg	followed	by	a	rearward	seat	hinge	rotation	of	3.8	deg.	The	seatback	motor	deformation	profile	was	programmed	to	unspool	 the	webbing	 from			 216	the	two	lower	seatback	motors	beginning	200	ms	(tdeformation-onset)	before	the	collision	onset	to	a	peak	angle	 (θrearward-peak)	of	168	deg	 (angular	velocity	ωdeformation-int	 =	980	deg/s).	 The	 resulting	ATD	penetration	into	the	seatback	was	-61.6	mm	at	T1	(dpenetration-T1,	green	line)	and	-94.1	mm	at	 the	pelvis	 (dpenetration-pelvis,	blue	 line)	 (Figure	D1B).	 	These	profiles	were	used	as	 the	 relative	timing	between	the	seat	hinge	rotation	(trotation-onset)	and	seatback	deformation	(tdeformation-onset)	profiles	 were	 manipulated	 through	 a	 series	 of	 three	 experiments	 varying:	 1)	 trotation-onset	 2)	tdeformation-onset,	and	3)	both	trotation-onset	and	tdeformation-onset.	For	comparison,	the	ATD	was	exposed	to	the	same	collision	(12	km/h;	Δt	=	187	ms)	while	seated	on	the	Control	GMHR	seat	and	on	the	Experimental	 seat	but	with	 the	hinge	and	seatback	motors	 locked	 in	 their	 initial	position	 (No	Motion	condition).		To	quantify	the	ATD	responses,	nine	peak	kinematic	and	kinetic	(aX-head,	aX-T1,	ωY-head,	FX,	FZ,	MY,	θhead,	RX	and	Δthead-contact)	and	three	peak	neck	injury	criteria	(NIC,	Nkm,	and	Nij)	responses	were	selected	for	analysis.	 	Within	each	response	variable,	the	different	profiles	were	ranked,	and	then	the	profile	with	the	lowest	averaged	rank	across	all	12	variables	was	identified	as	the	best	 profile.	 A	 Friedman	 test	 determined	 if	 there	 was	 a	 significant	 difference	 between	 the	average	 rank	of	 the	different	profiles	 (chi-squared	value:	 χ2,	degrees	of	 freedom:	df,	 and	 the	significance	level:	p).	A	post-hoc,	pair-wise	multiple	comparison	test	was	then	performed	on	the	results	of	the	Friedman	test	to	determine	differences	between	individual	profiles.	All	tests	were	performed	using	predefined	 functions	 (friedman	 and	multcompare)	 in	Matlab	using	a	 =	0.05.							 217		Figure	 D1.	 Programmed	 inputs	 and	 resulting	 outputs	 for	 A.)	 seat	 hinge	 rotation	 and	 B.)	seatback	cushion	deformation	profiles	during	a	12	km/h	collision	(∆t	=	187	ms;	black	line).		Grey	lines	represent	 input	rotation	(θseat-hinge)	and	deformation	(θseatback-motors)	profiles	 to	seat	hinge	and	 seatback	 motors,	 respectively.	 Seat	 hinge	 rotation	 (θseat-hinge-output:	 red	 line)	 as	 well	 as	rearward	penetration	of	ATD	T1	 (dpenetration-T1:	 green	 line)	 and	pelvis	 (dpenetration-pelvis:	 blue	 line)	into	the	seatback	indicate	resulting	outputs	to	their	respective	input	profiles.		Experiment	1:	Varying	Onset	of	Seat	Hinge	Rotation			 In	this	first	experiment,	the	effects	of	shifting	the	seat	hinge	rotation	onset	from	-170	to	-10	 ms	 (trotation-onset	 =	 -170,	 -150,	 -130,	 -110,	 -90,	 -70,	 -50,	 -30	 and	 -10	 ms)	 relative	 to	 the	seatback	deformation	(tdeformation-onset	=	-200ms)	were	tested	(Figure	D2A).	Exemplar	ATD	head	acceleration	(aX-head)	data	showed	lower	peak	responses	for	trotation-onset	between	-110	to	-50	ms	from	 the	 Control	 seat	 and	 the	 No	 Motion	 condition,	 whereas,	 peak	 T1	 acceleration	 (aX-T1)	decreased	only	trotation-onset	=	-90ms	(Figure	D3).	A	Friedman	test	confirmed	these	observations,	showing	 a	 significant	 difference	 between	 all	 tested	 trials	 (1	 Control,	 1	 No	 Motion	 and	 9	Experimental,	χ2	=	88.61,	df	=	10,	p	<	0.001).		Post-hoc	comparison	showed	that	trotation-onset	of	-90	 and	 -70	 ms	 were	 ranked	 significantly	 lower	 than	 the	 Control	 seat	 and	 the	 No	 Motion	condition	as	well	as	trotation-onset	=	-170,	-150	and	-10	ms	trials	(multiple	p	≤	0.044;	Figure	D2C).			 218	Since	there	was	no	significant	difference	between	trotation-onset	=	-90	and	-70	ms	trials,	a	trotation-onset	=	-90	ms	was	selected	for	subsequent	experimental	testing.	Overall,	these	results	suggested	that	trotation-onset	had	a	significant	effect	on	peak	ATD	responses	and	the	selected	pulse	timing	(-90	ms)	was	a	local	minimum.	The	profile	with	a	trotation-onset	=	-90	ms	and	a	tdeformation-onset	=	-200	ms	generated	occupant	penetrations	of	dpenetration-T1	=	 -61.6	mm	and	dpenetration-pelvis	=	 -94.1mm	(Figure	D2E	and	D2F).						 			 219		Figure	D2.	Seat	hinge	and	seatback	motor	profiles	as	well	as	ATD	responses	to	the	Control	seat	(blue	 lines	 and	 markers),	 the	 No	 Motion	 condition	 (magenta	 lines	 and	 markers)	 and	 the	Experimental	 seat	with	 various	 seat	 hinge	motor	 onsets	 at	 a	 constant	 seatback	motor	 onset	(trotation-onset	 &	 tdeformation-onset	 =	 -200	 ms;	 grey	 and	 red	 lines	 and	 markers)	 during	 a	 12	 km/h	collision	 severity	 (Δt	=	187	ms,	black	 lines,	 right	axis).	The	 red	 lines	and	markers	 indicate	 the	seat	hinge	onset	 (trotation-onset	 =	 -90	ms)	 selected	 for	 further	 testing.	Panels	A	and	B	 show	 the	programmed	 input	 and	 output	 seat	 hinge	 rotation	 profiles.	 Panel	D	 shows	 the	 programmed	input	 seatback	 motor	 profiles	 with	 resulting	 seatback	 deformation	 at	 the	 T1	 spinal	 level	(dpenetration-T1)	and	at	the	pelvis	(dpenetration-pelvis)	in	Panels	E	and	F,	respectively.	The	results	of	the	post-hoc	test	(C)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	differences	(*)	between	the	selected	seatback	motor	parameter	trial	(red)	and	the	Control,	No	Motion	or	other	experimental	trials.						 220		Figure	 D3.	 Exemplar	 ATD	 responses	 for	 the	 Control	 seat	 (blue	 lines	 and	 markers),	 the	 No	Motion	 condition	 (magenta	 lines	 and	markers),	 and	 the	 Experimental	 seat	with	 various	 seat	hinge	motor	onsets	and	a	constant	seatback	motor	onset	(trotation-onset	&	tdeformation-onset	=	-200	ms;	grey	and	 red	 lines	and	markers)	during	a	12	km/h	collision	severity	 (Δt	=	187	ms).	Calibrated	responses	 (A	 –	 C)	 and	 their	 corresponding	 peak	 responses	 (D	 –	 F)	 for	 head	 and	 torso	accelerations	in	the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	criteria	(NIC),	respectively.		The	red	lines	and	markers	indicate	the	seatback	motor	parameter	(trotation-onset	=	-90	ms)	selected	for	further	testing.				 221	Experiment	2:	Varying	Onset	of	Seatback	Deformation			 In	 this	experiment,	 the	 seatback	deformation	onset	was	 shifted	 from	 -280	 to	 -120	ms	(tdeformation-onset	=	-280,	-260,	-240,	-220,	-200,	-180,	-160,	-140	and	-120	ms)	before	the	collision	relative	to	a	fixed	seat	hinge	rotation	onset	(trotation-onset	=	-90	ms)	to	determine	the	influence	of	varying	 seatback	deformation	onset	 (Figure	D4D).	 	 Exemplar	ATD	 responses	 (aX-head,	 aX-T1,	 and	NIC)	for	the	Experimental	trials	were	lower	than	the	Control	seat	and	the	No	Motion	condition	but	remained	similar	across	all	Experimental	trials	despite	varying	tdeformation-onset	 (Figure	D5).	A	Friedman	test	showed	a	significant	difference	between	all	tested	trials	(1	Control,	1	No	Motion	and	 9	 Experimental,	 χ2	 =	 62.82,	 df	 =	 10,	 p	 <	 0.001).	 	 Post-hoc	 comparison	 confirmed	 all	Experimental	 conditions	 were	 significantly	 lower	 than	 the	 Control	 seat	 and	 the	 No	 Motion	condition	 (p	 ≤	 0.044,	 Figure	 D4C)	 but	 no	 significant	 differences	 between	 Experimental	conditions	were	observed.	Thus,	the	seatback	tdeformation-onset	at	-200	ms	was	selected	for	future	testing	(Figure	D4C).	These	results	suggested	that	changes	to	tdeformation-onset	between	the	range	of	-280	to	-120	ms	before	the	collision	onset	had	little	effect	on	peak	ATD	responses	between	trials.	 	 The	 selected	 onsets	 for	 seat	 hinge	 rotation	 (trotation-onset	 =	 -90	 ms)	 and	 seatback	deformation	 (tdeformation-onset	 =	 -200	 ms)	 were	 the	 same	 from	 Experiment	 1	 and	 generated	occupant	 penetration	 at	 T1	 (dpenetration-T1)	 of	 -61.6	 mm	 and	 at	 the	 pelvis	 (dpenetration-pelvis)	 of	 -94.1mm	(Figure	D4E	and	D4F).				 222			Figure	D4.	Seat	hinge	and	seatback	motor	profiles	as	well	as	ATD	responses	to	the	Control	seat	(blue	 lines	 and	 markers),	 the	 No	 Motion	 condition	 (magenta	 lines	 and	 markers)	 and	 the	Experimental	 seat	with	 various	 seatback	motor	 onsets	 at	 a	 constant	 seat	 hinge	motor	 onset	(tdeformation-onset	&	trotation-onset	=	-90	ms;	grey	and	red	lines	and	markers)	during	a	12	km/h	collision	severity	(Δt	=	187	ms,	black	lines,	right	axis).	The	red	lines	and	markers	indicate	the	seat	hinge	onset	 (tdeformation-onset	 =	 -200	 ms)	 selected	 for	 further	 testing.	 Panels	 A	 and	 B	 show	 the	programmed	 input	 and	 output	 seat	 hinge	 rotation	 profiles.	 Panel	D	 shows	 the	 programmed	input	 seatback	 motor	 profiles	 with	 resulting	 seatback	 deformation	 at	 the	 T1	 spinal	 level	(dpenetration-T1)	and	at	the	pelvis	(dpenetration-pelvis)	in	Panels	E	and	F,	respectively.	The	results	of	the	post-hoc	test	(C)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	differences	(*)	between	the	selected	seatback	motor	parameter	trial	(red)	and	the	Control,	No	Motion	or	other	Experimental	trials.			 			 223			Figure	 D5.	 Exemplar	 ATD	 responses	 for	 the	 Control	 seat	 (blue	 lines	 and	 markers),	 the	 No	Motion	 condition	 (magenta	 lines	 and	 markers),	 and	 the	 Experimental	 seat	 with	 various	seatback	motor	onsets	and	a	constant	seat	hinge	motor	onset	(tdeformation-onset	&	trotation-onset	=	-90	ms;	 grey	 and	 red	 lines	 and	 markers)	 during	 a	 12	 km/h	 collision	 severity	 (Δt	 =	 187	 ms).		Calibrated	responses	(A	–	C)	and	their	corresponding	peak	responses	(D	–	F)	for	head	and	torso	accelerations	in	the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	criteria	(NIC),	respectively.		The	red	 lines	 and	 markers	 indicate	 the	 seatback	 motor	 parameter	 (tdeformation-onset	 =	 -200	 ms)	selected	for	further	testing.				 			 224	Experiment	3:	Varying	both	Seat	Hinge	Rotation	and	Seatback	Deformation	Onsets			 The	results	 from	the	previous	 two	experiments	confirmed	that	a	combined	seat	hinge	rotation	(trotation-onset	=	-90ms)	and	seatback	deformation	(tdeformation-onset	=	-200ms)	could	reduce	ATD	responses	following	a	rear-end	collision.	 	 In	this	experiment,	the	relative	timing	between	the	two	profiles	(∆tonset	=	110	ms)	was	preserved	and	the	timing	of	both	the	seat	hinge	rotation	and	 the	 seatback	 deformation	 onsets	 were	 shifted	 in	 unison.	 The	 tested	 trotation-onset	 and	tdeformation-onset	were	identical	to	the	onset	values	used	in	the	previous	two	experiments	(trotation-onset	=	-170,	-150,	-130,	-110,	-90,	-70,	-50,	-30	and	-10	ms;	tdeformation-onset	=	-280,	-260,	-240,	-220,	-200,	-180,	-160,	-140,	and	-120	ms;	Figure	D6A	and	D6D).		Hence,	the	earliest	seat	hinge	and	seatback	onsets	were	trotation-onset	=	-170	ms	and	tdeformation-onset	=	-280	ms,	while	the	latest	onsets	were	trotation-onset	=	-10	ms	and	tdeformation-onset	=	-120	ms.			Exemplar	ATD	responses	(aX-head,	aX-T1,	and	NIC,	Figure	D7)	showed	similar	peak	responses	to	just	changing	trotation-onset	in	Experiment	1.	ATD	head	acceleration	(aX-head)	showed	lower	peak	responses	for	trotation-onset	between	-110	to	-70	 ms	 and	 tdeformation-onset	 between	 -220	 to	 -180	 ms	 from	 the	 Control	 and	 the	 No	 Motion	conditions,	whereas	peak	T1	acceleration	(aX-T1)	decreased	only	for	the	trotation-onset	=	-90	ms	and	tdeformation-onset	 =	 -200	 ms	 condition.	 	 Neck	 injury	 criterion	 (NIC)	 decreased	 compared	 to	 the	Control	seat	and	the	No	Motion	condition	in	all	experimental	trials.	The	Friedman	test	revealed	a	significant	difference	between	trials	(1	Control,	1	No	Motion,	and	9	Experimental,	χ2	=	82.26,	df	=	10,	p	<	0.001).	Post-hoc	comparisons	revealed	that	trotation-onset	=	-90	ms	and	tdeformation-onset	=	-200	ms	was	ranked	significantly	 lower	than	the	Control	seat	and	the	No	Motion	condition	as	well	as	two	conditions	with	trotation-onset	≤	-	150	ms	and	tdeformation-onset	≤	-260	ms	(p	≤	0.036,	Figure	D6C).	The	similarities	in	the	results	of	this	experiment	and	Experiment	1	further	suggested	that			 225	changing	 seat	 hinge	 trotation-onset	 had	 a	 greater	 effect	 on	 peak	 ATD	 responses	 than	 changing	seatback	tdeformation-onset.	Despite	showing	the	second	 lowest	rank,	the	trotation-onset	=	 -90	ms	and	tdeformation-onset	 =	 -200	 ms	 profile	 timing	 was	 selected	 for	 future	 experiments	 because	 1)	 it	decreased	 peak	 T1	 acceleration	 (aX-T1)	 and	 2)	 an	 onset	 of	 200	ms	 prior	 to	 the	 collision	 was	reasoned	to	be	the	upper	limit	possible	for	impending	impact	detection	and	subsequent	active	deformation	of	the	automotive	seat.			 			 226		Figure	D6.	Seat	hinge	and	seatback	motor	profiles	as	well	as	ATD	responses	to	the	Control	seat	(blue	 lines	 and	 markers),	 the	 No	 Motion	 condition	 (magenta	 lines	 and	 markers)	 and	 the	Experimental	seat	with	shifted	seat	hinge	and	seatback	motor	onsets	 (trotation-onset	&	tdeformation-onset;	 grey	 and	 red	 lines	 and	markers)	 during	 a	 12	 km/h	 collision	 severity	 (Δt	 =	 187	ms,	 black	lines,	right	axis).	The	red	lines	and	markers	indicate	the	seat	hinge	onset	(trotation-onset	=	-90	ms	and	tdeformation-onset	=	-200	ms)	selected	for	further	testing.	Panels	A	and	B	show	the	programmed	input	and	output	seat	hinge	rotation	profiles.	Panel	D	 shows	the	programmed	 input	seatback	motor	profiles	with	 resulting	 seatback	deformation	at	 the	T1	spinal	 level	 (dpenetration-T1)	and	at	the	pelvis	 (dpenetration-pelvis)	 in	 Panels	E	and	F,	 respectively.	 The	 results	 of	 the	post-hoc	 test	 (C)	determined	the	mean	rank	of	each	trial	as	well	as	any	significant	differences	(*)	between	the	selected	 seatback	 motor	 parameter	 trial	 (red)	 and	 the	 Control,	 No	 Motion	 or	 other	experimental	trials.						 227			Figure	 D7.	 Exemplar	 ATD	 responses	 for	 the	 Control	 seat	 (blue	 lines	 and	 markers),	 the	 No	Motion	 condition	 (magenta	 lines	 and	markers),	 and	 the	 Experimental	 seat	 with	 shifted	 seat	hinge	and	seatback	motors	onsets	(trotation-onset	&	tdeformation-onset;	grey	and	red	lines	and	markers)	during	 a	 12	 km/h	 collision	 severity	 (Δt	 =	 187	 ms).	 Calibrated	 responses	 (A	 –	 C)	 and	 their	corresponding	peak	responses	(D	–	F)	for	head	and	torso	accelerations	in	the	X-direction	(aX-head	&	aX-T1)	and	neck	injury	criteria	(NIC),	respectively.		The	red	lines	and	markers	indicate	the	seat	hinge	 and	 seatback	 motor	 parameters	 (trotation-onset	 =	 -90	 ms	 and	 tdeformation-onset	 =	 -200	 ms)	selected	for	further	testing.				 228	Summary	The	 onsets	 for	 the	 seat	 hinge	 rotation	 (trotation-onset	 =	 -90ms)	 and	 the	 seatback	deformation	(tdeformation-onset	=	-200ms)	profiles	have	been	optimally	combined	into	one	dynamic	movement	 on	 the	 Experimental	 anti-whiplash	 seat.	 	 This	 optimal	 solution	 was	 a	 direct	combination	 of	 the	 seat	 hinge	 rotation	 profile	 determined	 in	Appendix	 B	 and	 the	 seatback	deformation	profile	determined	in	Appendix	C	(Figure	D1).		Between	the	Control	seat	and	the	No	Motion	 (seat	hinge	and	seatback	motors	held	 in	 initial	position,	but	 the	seatback	was	not	rigidly	 braced	 against	 the	 sled)	 condition,	 the	 No	Motion	 decreased	 seven	 of	 the	 nine	 peak	responses	 and	all	 three	peak	neck	 injury	 criteria	 by	 -17%	 to	 -58%.	 	Only	peak	 linear	 forward	acceleration	of	the	T1	vertebra	(aX-T1)	and	time-to-head-restraint	contact	(Δthead-contact)	remained	the	 same	 as	 the	 Control	 seat.	 	 In	 comparison	 to	 the	 Control	 seat,	 the	 combined	 seat	 hinge	rotation	 and	 seatback	 deformation	 profile	 reduced	 all	 nine	 peak	 kinematic	 and	 kinetic	responses	(aX-head:	-71%,	aX-T1:	-34%,	FX:	-71%,	FZ:	-89%,	MY:	-81%,	ωY-head:	-75%,	θhead:	-87%,	RX:	-70%	and	Δthead-contact:	-136	ms)	and	three	peak	neck	injury	criteria	(NIC:	-90%,	Nkm:	-70%,	and	Nij:	-70%).	 	 Thus,	 the	 combined	 seat	 hinge	 rotation	 and	 seatback	 deformation	 profile	may	 be	 a	potentially	 effective	 solution	 to	 reducing	 ATD	 responses	 following	 low-speed,	 rear-end	collisions.		 			 229	Appendix	E:	Supplementary	Material	Supplementary	Material	for	Chapter	2			Table	 E2.1.	 Mean	 (Standard	 Deviation)	 and	 Coefficient	 of	 Variation	 (COV)	 from	 five	repeated	whiplash-like	perturbation	at	a	Dv	=	4,	8	and	12	km/h	with	a	Dt	=	141	ms	on	the	2004	GM	Pontiac	Grand	AM	GMHR	seat.	Automotive	Seat	 GM	Pontiac	Grand	AM	(GMHR)	Dv	(km/h)	 4	km/h,	Dt	=	141	ms	 8	km/h,	Dt	=	141	ms	 12	km/h,	Dt	=	141	ms	Parameter	 Mean	(SD)	 COV	(%)	Mean	(SD)	 COV	(%)	Mean	(SD)	 COV	(%)	aX-sled	(m/s2)	 11.0	(0.7)	 6.2*	 20.5	(0.2)	 0.8	 28.6	(0.1)	 0.4	Dthead-contact	(ms)	159.6	(1.7)	 1.1	122.8	(2.7)	 1.3	108.9	(1.5)	 1.4*	aX-head	(m/s2)	 37.7	(0.4)	 0.9	 86.0	(1.6)	 1.9*	130.4	(1.4)	 1.1	aX-T1	(m/s2)	 22.0	(1.2)	 5.3*	 42.1	(0.7)	 1.6	 51.3	(0.5)	 1.0	FX	(N)	113.3	(4.6)	 4.0	222.9	(18.1)	 8.1*	220.9	(17.5)	 7.9	FZ	(N)	 85.4	(3.8)	 4.5	562.5	(35.2)	 6.3*	684.7	(39.5)	 5.8	MY	(Nm)	 1.6	(0.2)	 11.2*	 13.6	(0.6)	 4.2	 14.0	(0.6)	 4.2	w	Y-head(deg/s)	238.6	(21.8)	 9.2*	315.9	(11.6)	 3.7	433.1	(12.7)	 2.9	wY-T1	(deg/s)	160.1	(13.4)	 8.4*	305.1	(16.6)	 5.4	443.0	(20.6)	 4.7	qhead	(deg)	 14.4	(0.3)	 2.4	 15.2	(0.9)	 5.7*	 17.4	(0.5)	 2.8	RX	(mm)	 -48.3	(0.8)	 1.7	 -51.0	(2.8)	 5.5*	 -54.9	(1.7)	 3.0	NIC	(m2/s2)	 3.5	(0.2)	 6.9*	 8.1	(0.4)	 4.3	 11.0	(0.3)	 3.1	Nij	0.05	(0.00)	 4.3	0.13	(0.01)	 5.3	0.14	(0.01)	 5.4*	Nkm	0.25	(0.01)	 2.5	0.42	(0.03)	 6.7*	0.42	(0.03)	 6.4	Notes:		The	underlined	COV	values	indicate	a	COV	rating	of	acceptable	(5%	≤	COV	<	10%),	the	bolded	COV	values	indicate	a	COV	rating	of	poor	(COV	>	10%),	and	all	other	values	are	rated	 good	 (COV	 <	 5%).	 *	 indicates	 the	maximum	COV	 value	 for	 each	 experimental	 ATD	parameter.					 230	Table	 E2.2.	 Mean	 (standard	 deviation)	 and	 Coefficient	 of	 Variation	 (COV)	 from	 five	 repeated	whiplash-like	perturbation	at	a	Dv	=	8	km/h	with	a	Dt	=	141	ms	for	the	2005	Saab	9.3	SAHR,	2005	Volvo	S40	WHIPS	and	2004	Volvo	S60	WHIPS	seats.		Parameter	 Mean	(SD)	 COV	(%)	 Mean	(SD)	 COV	(%)	 Mean	(SD)	 COV	(%)	Automotive	Seat	 Saab	9.3	SAHR	 Volvo	S40	WHIPS	 Volvo	S60	WHIPS	Dv	(km/h)	 8	km/h,	Dt	=	141	ms	 8	km/h,	Dt	=	141	ms	 8	km/h,	Dt	=	141	ms	aX-sled	(m/s2)	 19.9	(0.1)	 0.6	 20.4	(0.2)	 1.0	 20.3	(0.1)	 0.3	Dthead-contact	(ms)	 95.1	(1.1)	 1.2	 96.7	(2.2)	 2.3	 104.5	(3.8)	 3.6	aX-head	(m/s2)	 68.4	(1.6)	 2.4	 87.2	(3.4)	 3.9	 87.9	(6.1)	 7.0	aX-T1	(m/s2)	 42.5	(0.7)	 5.3	 44.6	(1.1)	 2.5	 46.5	(0.7)	 1.5	FX	(N)	 16.5	(1.8)	 11.1	 38.1	(6.1)	 15.9	 2.2	(1.0)	 44.1	FZ	(N)	155.1	(13.6)	 8.8	170.6	(19.9)	 11.7	241.5	(37.6)	 15.6	MY	(Nm)	 5.8	(0.0)	 0.8	 4.5	(0.4)	 9.8	 5.0	(0.4)	 7.5	w	Y-head(deg/s)	 251.1	(9.3)	 3.7	220.9	(22.9)	 10.4	266.3	(21.5)	 8.1	wY-T1	(deg/s)	264.3	(23.4)	 8.8	262.8	(15.8)	 6.0	292.9	(26.0)	 8.9	qhead	(deg)	 11.1	(0.2)	 2.1	 10.6	(0.6)	 5.3	 12.4	(0.6)	 5.2	RX	(mm)	 -25.7	(1.9)	 7.6	 -28.1	(2.2)	 7.9	 -20.9	(2.3)	 10.9	NIC	(m2/s2)	 5.4	(0.2)	 4.1	 5.7	(0.4)	 6.1	 6.2	(0.4)	 5.6	Nij	 0.15	(0.01)	 9.3	 0.09	(0.01)	 8.5	 0.05	(0.00)	 8.2	Nkm	 0.48	(0.03)	 6.7	 0.23	(0.02)	 10.0	 0.16	(0.01)	 5.9	Notes:	 	 The	underlined	COV	values	 indicate	a	COV	 rating	of	 acceptable	 (5%	≤	COV	<	10%),	 the	bolded	COV	values	indicate	a	COV	rating	of	poor	(COV	>	10%),	and	all	other	values	are	rated	good	(COV	<	5%).				 231		Figure	E2.1.	 Experimental	data	 for	 the	BioRID	 II	ATD	 seated	on	a	2005	Volvo	S40	WHIPS	 seat.	 Each	 column	 represents	occupant	responses	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6,	8,	10,	12	and	14	km/h	with	a	collision	pulse	duration	(Δt)	of	141	ms).	Hollow	circles	represent	the	onset	of	head-to-head-restraint	contact	and	peak	responses	of	each	ATD	response	parameter	for	each	trial.			 232		Figure	E2.2.	 Experimental	data	 for	 the	BioRID	 II	ATD	 seated	on	a	2004	Volvo	S60	WHIPS	 seat.	 Each	 column	 represents	occupant	responses	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6,	8,	10,	12	and	14	km/h	with	a	collision	pulse	duration	(Δt)	of	141	ms).	Hollow	circles	represent	the	onset	of	head-to-head-restraint	contact	and	peak	responses	of	each	ATD	response	parameter	for	each	trial.	 			 233		Figure	 E2.3.	 Experimental	 data	 for	 the	 BioRID	 II	 ATD	 seated	 on	 a	 2005	 Saab	 9.3	 SAHR	 seat.	 Each	 column	 represents	 occupant	responses	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6,	8,	10,	12	and	14	km/h	with	a	collision	pulse	duration	(Δt)	of	141	ms).	Hollow	circles	represent	the	onset	of	head-to-head-restraint	contact	and	peak	responses	of	each	ATD	response	parameter	for	each	trial.			 234		Figure	 E2.4.	 Experimental	 data	 for	 the	 BioRID	 II	 ATD	 seated	 on	 a	 2004	 Pontiac	 Grand	 Am	GMHR	 seat.	 Each	 column	 represents	occupant	responses	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6,	8,	10,	12	and	14	km/h	with	a	collision	pulse	duration	(Δt)	 of	 141	ms).	Hollow	 circles	 represent	 the	 onset	 of	 head-to-head-restraint	 contact	 and	 peak	 responses	 of	 each	ATD	 response	parameter	for	each	trial.				 235	Supplementary	Material	for	Chapter	5		Figure	E5.1.	Experimental	data	for	the	BioRID	II	ATD	seated	on	the	Experimental	seat	utilizing	both	 seat	hinge	 rotation	and	 seatback	 cushion	deformation	 safety	mechanisms.	Each	column	represents	occupant	responses	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6	and	8	km/h,	Δt	=	147	ms	and	Δv	=	12	km/h,	Δt	=	187	ms).	Hollow	circles	represent	the	onset	of	head-to-head-restraint	contact	and	peak	responses	of	each	ATD	response	parameter	 for	each	trial.	Whiplash	perturbation	onset	occurred	at	t	=	0	ms.					 236		Figure	E5.2.	Experimental	data	for	the	BioRID	II	ATD	seated	on	a	2005	Volvo	S40	WHIPS	seat.	Each	column	represents	occupant	responses	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6	and	8	km/h,	Δt	=	147	ms	and	Δv	=	12	km/h,	Δt	=	187	ms).	Hollow	circles	represent	the	onset	of	head-to-head-restraint	contact	and	peak	responses	of	each	ATD	response	parameter	for	each	trial.	Whiplash	perturbation	onset	occurred	at	t	=	0	ms.				 237		Figure	E5.3.	Experimental	data	for	the	BioRID	II	ATD	seated	on	a	2004	Volvo	S60	WHIPS	seat.	Each	column	represents	occupant	responses	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6	and	8	km/h,	Δt	=	147	ms	and	Δv	=	12	km/h,	Δt	=	187	ms).	Hollow	circles	represent	the	onset	of	head-to-head-restraint	contact	and	peak	responses	of	each	ATD	response	parameter	for	each	trial.	Whiplash	perturbation	onset	occurred	at	t	=	0	ms.				 238			Figure	E5.4.	Experimental	data	for	the	BioRID	II	ATD	seated	on	a	2005	Saab	9.3	SAHR	seat.	Each	column	represents	occupant	responses	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6	and	8	km/h,	Δt	=	147	ms	and	Δv	=	12	km/h,	Δt	=	187	ms).	Hollow	circles	represent	the	onset	of	head-to-head-restraint	contact	and	peak	responses	of	each	ATD	response	parameter	for	each	trial.	Whiplash	perturbation	onset	occurred	at	t	=	0	ms.						 239				Figure	E5.5.	Experimental	data	for	the	BioRID	II	ATD	seated	on	a	2004	Pontiac	Grand	Am	GMHR	seat.	Each	column	represents	occupant	responses	while	exposed	to	various	collision	severities	(Δv	=	2,	4,	6	and	8	km/h	with	a	collision	pulse	duration	(Δt)	=	147	ms	and	Δv	=	12	km/h	with	a	Δt	=	 187	 ms).	 Hollow	 circles	 represent	 the	 onset	 of	 head-to-head-restraint	 contact	 and	 peak	responses	 of	 each	 ATD	 response	 parameter	 for	 each	 trial.	 Whiplash	 perturbation	 onset	occurred	at	t	=	0	ms.		

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0384553/manifest

Comment

Related Items