UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Undersampled differential phase shift on-off keying for optical camera communications with phase error… Liu, Niu 2016

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

Item Metadata


24-ubc_2017_february_Liu_Niu.pdf [ 1.58MB ]
JSON: 24-1.0340495.json
JSON-LD: 24-1.0340495-ld.json
RDF/XML (Pretty): 24-1.0340495-rdf.xml
RDF/JSON: 24-1.0340495-rdf.json
Turtle: 24-1.0340495-turtle.txt
N-Triples: 24-1.0340495-rdf-ntriples.txt
Original Record: 24-1.0340495-source.json
Full Text

Full Text

UNDERSAMPLED DIFFERENTIALPHASE SHIFT ON-OFF KEYINGFOR OPTICAL CAMERACOMMUNICATIONS WITH PHASEERROR DETECTIONbyNIU LIUB.Eng., Nanjing University of Aeronautics and Astronautics, P. R. China, 2010M.Eng., Lanzhou University, P. R. China, 2012A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF APPLIED SCIENCEinTHE COLLEGE OF GRADUATE STUDIES(Electrical Engineering)THE UNIVERSITY OF BRITISH COLUMBIA(Okanagan)December 2016c© NIU LIU, 2016The undersigned certify that they have read, and recommend to the Col-lege of Graduate Studies for acceptance, a thesis entitled: UNDERSAM-PLED DIFFERENTIAL PHASE SHIFT ON-OFF KEYING FOROPTICAL CAMERA COMMUNICATIONS WITH PHASE ERRORDETECTION submitted by NIU LIU in partial fulfilment of the require-ments of the degree of MASTER OF APPLIED SCIENCEJulian Cheng, Faculty of Applied Science/School of EngineeringSupervisor, Professor (please print name and faculty/school above the line)Ayman Elnaggar, Faculty of Applied Science/School of EngineeringSupervisory Committee Member, Professor (please print name and faculty/school abovethe line)Jonathan Holzman, Faculty of Applied Science/School of EngineeringSupervisory Committee Member, Professor (please print name and faculty/school abovethe line)Zheng Liu, Faculty of Applied Science/School of EngineeringUniversity Examiner, Professor (please print name and faculty/school above the line)External Examiner, Professor (please print name and faculty/school above the line)December 19, 2016(Date Submitted to Grad Studies)Additional Committee Members include:(please print name and faculty/school above the line)(please print name and faculty/school above the line)iiAbstractThis thesis introduces the design and implementation of an optical ca-mera communication (OCC) system. Phase uncertainty and phase slippingcaused by camera sampling are the two major challenges for OCC. In thisthesis, we propose a novel modulation scheme to overcome these problems.The undersampled differential phase shift on-off keying is capable of enco-ding binary data bits without exhibiting any flicker to human eyes. Thephase difference between two consecutive samples conveys one-bit informa-tion which can be decoded by a low frame rate camera receiver. Errordetection techniques are also introduced in the thesis to enhance the reli-ability of the system. Furthermore, we present the hardware and softwaredesign of the proposed system. This low-cost communication system hasbeen implemented with a Xilinx FPGA and a Logitech commercial camera.Experimental results demonstrate that a bit-error rate of 10−5 can be achie-ved with 7.15 microwatts received signal power over a link distance of 15centimeters.iiiPrefaceThis thesis is original, unpublished, independent work by the author,Niu Liu.ivTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . vList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ixList of Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . xiAcknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . xiiiChapter 1: Introduction . . . . . . . . . . . . . . . . . . . . . . . 11.1 Background and Motivation . . . . . . . . . . . . . . . . . . . 11.1.1 Applications of VLC . . . . . . . . . . . . . . . . . . . 21.1.2 Motivation . . . . . . . . . . . . . . . . . . . . . . . . 51.2 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . 71.3 Thesis Organization and Contributions . . . . . . . . . . . . . 11Chapter 2: Image Sensors and OCC . . . . . . . . . . . . . . . 132.1 Image Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . 13vTABLE OF CONTENTS2.1.1 CCD Image Sensors . . . . . . . . . . . . . . . . . . . 142.1.2 CMOS Image Sensors . . . . . . . . . . . . . . . . . . 142.2 Optical Communications for Cameras . . . . . . . . . . . . . 172.2.1 Advantages of OCC . . . . . . . . . . . . . . . . . . . 172.2.2 Design Requirements . . . . . . . . . . . . . . . . . . . 182.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Chapter 3: Modulation/Demodulation . . . . . . . . . . . . . . 213.1 UDPSOOK . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213.2 UDPSOOK with Error Detection . . . . . . . . . . . . . . . . 273.3 Framing Structure . . . . . . . . . . . . . . . . . . . . . . . . 283.4 MIMO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31Chapter 4: Experimental Setup . . . . . . . . . . . . . . . . . . 324.1 Hardware Design . . . . . . . . . . . . . . . . . . . . . . . . . 324.2 Transmitter . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324.2.1 Modulator . . . . . . . . . . . . . . . . . . . . . . . . . 344.2.2 LED Driver Circuit . . . . . . . . . . . . . . . . . . . 354.3 Receiver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37Chapter 5: Experimental Results . . . . . . . . . . . . . . . . . 405.1 Received Optical Power . . . . . . . . . . . . . . . . . . . . . 405.2 BER and Bit Rate . . . . . . . . . . . . . . . . . . . . . . . . 425.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47Chapter 6: Conclusions . . . . . . . . . . . . . . . . . . . . . . . 48viTABLE OF CONTENTS6.1 Summary of Accomplished Work . . . . . . . . . . . . . . . . 486.2 Suggested Future Work . . . . . . . . . . . . . . . . . . . . . 49Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56Appendix A: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57viiList of TablesTable 5.1 Received optical intensity for the proposed OCC system 41Table 5.2 Key experimental parameters . . . . . . . . . . . . . . 44Table 5.3 Experimental results I . . . . . . . . . . . . . . . . . . 46Table 5.4 Experimental results II . . . . . . . . . . . . . . . . . . 46viiiList of FiguresFigure 1.1 Indoor VLC. . . . . . . . . . . . . . . . . . . . . . . . 3Figure 1.2 A VLC application. . . . . . . . . . . . . . . . . . . . 4Figure 1.3 Block diagram of a VLC system that uses image sen-sor receivers. . . . . . . . . . . . . . . . . . . . . . . . 8Figure 2.1 A CCD image sensor. . . . . . . . . . . . . . . . . . . 15Figure 2.2 A CMOS image sensor. . . . . . . . . . . . . . . . . . 16Figure 2.3 A multiple-LED OCC system. . . . . . . . . . . . . . 17Figure 3.1 Modulated UDPSOOK symbols. . . . . . . . . . . . . 23Figure 3.2 UDPSOOK demodulation. . . . . . . . . . . . . . . . 24Figure 3.3 Different sampling timing for UDPSOOK. . . . . . . 24Figure 3.4 Non-50% duty cycle demodulation . . . . . . . . . . . 26Figure 3.5 UDPSOOKED sampling . . . . . . . . . . . . . . . . 28Figure 3.6 Frame structure of the proposed OCC system . . . . 29Figure 3.7 A possible MIMO scheme for UDPSOOKED . . . . . 30Figure 4.1 Experimental setup. . . . . . . . . . . . . . . . . . . . 33Figure 4.2 Block diagram of the transmitter. . . . . . . . . . . . 34Figure 4.3 The transmission control finite state machine. . . . . 35Figure 4.4 The simulation results in Xilinx Vivado simulator. . . 36Figure 4.5 The PmodLED LED module . . . . . . . . . . . . . . 37ixLIST OF FIGURESFigure 4.6 A flow chart of the UDPSOOK receiver. . . . . . . . 38Figure 5.1 Optical power measurement setup. . . . . . . . . . . . 41Figure 5.2 DET36A responsivity . . . . . . . . . . . . . . . . . . 42Figure 5.3 Captured waveforms from the logic analyzer. . . . . . 43Figure 5.4 UDPSOOKED symbols captured by camera. . . . . . 45xList of AcronymsAcronyms DefinitionsADC Analog-to-Digital ConverterAR Augmented RealityBER Bit-Error Ratebps Bits Per SecondCCD Charge Coupled DeviceCFF Critical Flicker FrequencyCMOS Complementary Metal Oxide SemiconductorDC Direct CurrentDPSK Differential Phase-Shift KeyingFPGA Field-Programmable Gate ArrayFSM Finite State MachineHDL Hardware Description LanguageI2V Infrastructure-to-VehicleI/O Input/OutputxiList of AcronymsITS Intelligent Transport SystemsLED Light Emitting DiodeMIMO Multiple-Input Multiple-OutputOOK On-Off KeyingPAM Pulse Amplitude ModulationPAR Project Authorization RequestPIN Positive-Intrinsic-NegativeRTL Register-Transfer LevelPSK Phase-Shift KeyingQAM Quadrature Amplitude ModulationRF Radio FrequencyUDPSOOK Undersampled Differential Phase Shift On-Off KeyingUFSOOK Undersampled Frequency Shift On-Off KeyingUPSOOK Undersampled Phase Shift On-Off KeyingVICS Vehicle Information and Communication SystemVLC Visible Light CommunicationVLCC Visual Light Communication ConsortiumV2I Vehicle-to-InfrastructureV2V Vehicle-to-VehiclexiiAcknowledgementsI would like to thank my thesis supervisor Dr. Julian Cheng for hisguidance, advice, and encouragement. I could not accomplish my graduatestudy without his support.I would also like to thank Dr. Jonathan Holzman, and Dr. AymanElnaggar for their willingness to serve on the committee. Thanks to Dr.Zheng Liu for serving as the university examiner in the defense. I reallyappreciate their questions and constructive comments on my thesis.Special thanks to Dr. Xian Jin and Dr. Luanxia Yang for their help andfriendship during the three years.I would also like to thank Min He and her family for their love andcontinuing encouragement.I would like to thank UBC, where I started a brand new life.I would like to thank all the graduate students in our lab for their gen-erosity and support. They really encouraged me when I was in frustrationduring this journey.Finally, I would like to thank my parents for their love over all theseyears. All my achievements would not have been possible without them.xiiiChapter 1Introduction1.1 Background and MotivationVisible light communication (VLC) is a data communication technologybased on the visible light spectrum (between 400 and 800 THz), which istraditionally used for illumination. The history of using visible light forcommunication dates back to thousands of years ago when ancient Chinesepeople used beacon towers on the Great Wall to deliver military information.In 1867, Captain Philip Colomb invented signal lamps, which employeddifferent flashing patterns to communicate over a long distance. On June3rd 1880, Alexander Graham Bell transmitted the world’s first VLC messageon his invention, the “Photophone” [1]. During World War I, German armyused optical Morse transmitters as a secure communication technique, sincewired communication was often cut off [2]. Even though radio frequency(RF) had a dominant position in the past 100 years, a great deal of efforthas been made towards VLC. It is widely believed that VLC has greatpotential to be a complementary technology for RF communication.In recent years, with the increasing concerns regarding energy consump-tion and environmental protection, the light emitting diode (LED) has be-come the choice for sustainable illumination. LEDs are now gradually takingthe place of incandescent lamps. This green technology offers more bright-11.1. Background and Motivationness, lower energy consumption, and longer lifespan, making it well-suitedfor architectural decoration, traffic control, general lighting etc.The development of illumination technology offers great opportunities fortransmitting data by LEDs. Relevant research originated in Japan around2003 with the establishment of the Visual Light Communication Consortium(VLCC). After that, VLC has been attracting growing interest worldwide.In comparison with RF communication, VLC is more friendly to theenvironment. Using an illumination device for communication can reduceenergy consumption and greenhouse emissions. It is estimated that by repla-cing all the existing lighting sources with LEDs, the world electricity energyconsumption can be dramatically reduced by 50% [3]. Another motivationto use VLC is the increasing demand for high-speed wireless connectivity.With the emergence of smart devices, the requirement for big data trans-fer is growing exponentially. VLC can be a complementary technology fortraditional wireless communication systems in many application scenarios.1.1.1 Applications of VLCVLC technology applies to a number of scenarios. Some applications ofVLC are described as follows.Indoor Communication: Nowadays, as people are carrying multiple wire-less devices, WiFi has become a necessity in every step of our lives. Internetservice users demand not only faster communication speed, but also betterconnection quality, especially for high-definition videos and real-time onlinegames. When it comes to urban areas or an office building, where users tendto spend most of their time indoors, VLC is a powerful technique to enhancethe capacity of the existing network infrastructure. Figure 1.1 illustrates a21.1. Background and Motivationtypical indoor VLC environment. LED bulbs provide internet access to lap-tops, smart phones, TVs and other devices. In this case, VLC can eitherperform as a complement to WiFi networks, or provide wireless connectionsindependently when other communication links are not available.Figure 1.1: Indoor VLC.Inter-Vehicle Communication: The concept of intelligent transport sys-tems (ITS) has been proposed in recent years. It aims at improving roadutilization and providing innovative services to traffic participants [4]. Thekey approach is to establish wireless connectivity between vehicles and roadinfrastructure such as traffic lights, speed limit signs, and electronic roadpricing gantries etc. In Japan, a vehicle information and communication31.1. Background and Motivationsystem (VICS) was developed as an application of ITS [5]. Real-time trafficinformation was broadcasted by infrared beacons and radio beacons instal-led on road shoulders [6]. Such a communication system is of high cost andconsumes huge quantities of electric energy. Nowadays, LEDs are widelyused in automotive headlights and traffic signals as an energy-saving andreliable light source. As shown in Fig. 1.2, VLC is adopted for vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I) and infrastructure-to-vehicle(I2V) communications. Traffic and weather information can be efficientlybroadcasted via traffic lights and billboards. Collision avoidance systemscan also be implemented by setting up VLC links between vehicles.Figure 1.2: A VLC application.Wireless Communication in Hospitals: Modern medical communities areequipped with a variety of state-of-the-art technology. Medical equipmentis critical to our healthcare delivery system. However, diagnostic and treat-ment devices are sensitive to radio wave signals. Some of these devices relyon WiFi and cellular networks to transfer vital data, which could be inter-fered by nearby RF communications. Thus, there are certain areas in thehospital where wireless communication devices are strictly forbidden. As areliable solution to the electromagnetic interference problem [7], VLC worksin the range of the visible light frequency spectrum and does not interfere41.1. Background and Motivationwith other electrical equipment.Information Display and Communication: LEDs are used extensively inbuilding decoration and advertising. Those billboards, which are often madeup of hundreds of LEDs, are perfect VLC transmitters [8]. For example, theoutdoor sign of a restaurant may send food menus to pedestrians, and adepartment store may wish to push new electronic coupons to potentialbuyers. In this situation, a camera from the user’s cellphone could be usedas a VLC receiver. By detecting the intensity change of the outdoor LEDs,customers are able to receive promotion messages without entering a grocerystore or a shopping center.Toys and Entertainment: Another interesting application is that VLCcan be used in toys, theme parks, and public entertainment facilities. Toymanufacturer Disney [9] has already devoted itself to connecting toys tosmart devices wirelessly via LEDs. Existing light sources in theme parksalso provide ample opportunities for using VLC as part of the tourist loca-tion system and entertainment networks [10]. Since high-speed data commu-nication is not required for this application, hardware and communicationprotocols have lower complexity than the aforementioned applications.1.1.2 MotivationVLC systems can be divided into two categories by different types ofreceivers: non-imaging receiver and imaging receiver. Non-imaging receiversare widely used in VLC systems. For example, photodiode receivers areoften employed when high data rate or high sensitivity is required, andfor a bidirectional communication system requiring low complexity and lowcost, an LED can be simply used for both transmission and reception [11].51.1. Background and MotivationHowever, researchers have shown the limitations of non-imaging receivers.It has been proved that imaging receivers can provide better performancewhen multiple LEDs are deployed [12].An imaging receiver, namely an image sensor, which consists of hundredsof photodiodes (i.e. pixels), is also capable of receiving data from LEDs bytaking continuous images. Nowadays cameras can be found not only onlaptops and tablets but also in wearable devices such as smartwatches andsmart glasses. Optical communications for cameras (OCC) is proposed inrecent years to employ these embedded cameras as VLC receivers. Forinstance, in grocery stores, product information and coupons can be easilyobtained by a smartphone camera or an image sensor installed on shoppingcarts. Since no expensive hardware is required, camera receivers have shownadvantages over photodiodes in many applications.However, challenges and technical problems still exist for OCC. Firstly,the sample rate of a commercial camera is typically 30 frames per second(fps). As LED lighting flicker may induce biological human response, it issuggested that flickering in the 3 Hz to 70 Hz range should be avoided [13].So a flicker-free OCC transmitter has to operate at frequencies that exceedthe camera’s frame rate, and the receiver has to undersample the transmittedsignals, which cannot be easily reconstructed. In addition, the transmittersand the receivers are not strictly synchronized, and it is difficult to knowthe phase difference between the transmitted signal and the camera receiver,so the received on-off keying (OOK) waveform slowly slips with respect tothe sampling point [14]. All OCC systems suffer from the phase slippingproblem [14], which can cause more error bits and degenerate the systemperformance. For these reasons, demodulation of OCC is more challengingthan that of non-imaging systems.61.2. Literature ReviewThe objective of this research is to develop a low-cost VLC system thatuses camera receivers. The target system must realize data transmission wit-hout causing any LED flicker to human eyes. A new modulation techniquefor OCC needs to be proposed in order to overcome the phase uncertaintyand phase slipping problems.1.2 Literature ReviewSmart devices have been changing our lives dramatically in the pastdecade. With the existing billions of devices that carry image sensors, itwas proposed years ago that an image sensor can be used as a receiverfor communication. In fact, as early as the year 2004, researchers fromJapan [15] explored the possibility of using a camera as a VLC receiver.In the following years, cameras became popular in the topics related tointelligent traffic systems (ITS). For example, the authors in [16] and [17]proposed to establish connectivity between road infrastructure and vehiclesvia traffic lights and high-speed cameras, and the system was also knownas I2V-VLC. The authors in [18] introduced V2V-VLC using tail lights andcameras equipped on vehicles.Initiated in 2008 and completed in 2011, IEEE 802.15.7 was the firstVLC standard. After years of waiting, there are few commercialized VLCproducts on the market. Some researchers [19] believed that this slow adop-tion was due to not taking image sensor communication into considerationin the standard. Due to the increasing popularity of portable smart devices,camera communication seems to be much closer to the market than photo-diodes. More interest has been attracted to utilize embedded commercialcameras as VLC receivers [20].71.2. Literature ReviewFigure 1.3: Block diagram of a VLC system that uses image sensor receivers.Considering camera data reception is a necessary and significant amend-ment to the VLC standard, a revision to 802.15.7 is currently being un-dertaken by a task group called IEEE 802.15.7r1 OWC TG. The name ofOCC has been officially proposed in the IEEE P802.15.7 Project Authori-zation Request (PAR). This revision is expected to be published by 2018.Engineers are looking forward to extending OCC technology to billions ofexisting smart devices with the new VLC standard, without requiring anyhardware modification.Figure 1.3 shows the block diagram of a typical VLC system using LEDtransmitters and camera receivers. The data stream generated by the infor-mation source is modulated electrically by a modulator. In order to providea constant quantity of power to the LED transmitters, an LED driver circuitis applied prior to the transmitters. A camera receiver captures the intensity81.2. Literature Reviewchange of the LEDs at its frame rate. It converts the intensity into digitalluminance values which are then demodulated by the demodulator.The Nyquist-Shannon theorem states that to reconstruct an analog sig-nal, the sampling rate should be at least two times its maximum frequencycomponent [21]. Assuming that the frame rate of an image sensor is Rs,the frequency of the transmitted signal should be no greater than Rs/2 toavoid aliasing. To put it another way, if we have a camera receiver whichsamples at 30 fps, the corresponding transmitted carrier has an upper limitof 15 Hz in order to meet the sampling theorem. Operating at such a lowfrequency can be easily sensed by human eyes as flicker, which is not desi-rable in any OCC system. Another challenge for OCC is that the camerareceiver samples at a fixed frame rate, which cannot be synchronized withthe transmitters. This gives rise to an unknown sampling phase to the de-modulator. For these reasons, a simple OOK modulation cannot be used inOCC systems. Several new modulation techniques have been proposed forOCC in recent years [22–26].Undersampled frequency shift on-off keying (UFSOOK) [23] uses a pairof discrete OOK frequencies to transmit logic ones and zeros respectively.Different symbols present different patterns to the camera. The demodulatoruses two image frames to decode one-bit information. Because of this, thebit rate could not exceed half of the camera sample rate. To be more specific,if UFSOOK is applied to a 30 fps camera receiver, the data rate was limitedto 15 bits per second (bps).The authors in [24] and [25] introduced pulse amplitude modulation(PAM) and quadrature amplitude modulation (QAM) into OCC. Imagesensors are not linear devices. The operation of encoding and decoding lu-minance values, which is known as gamma correction, is nonlinear [27]. As91.2. Literature Reviewmore than two amplitude levels are used in [24] and [25], the nonlinearity in-troduced by gamma correction had to be taken into consideration. Receivedimages need to be compensated before demodulation in order to mitigate thenon-linear effect. A special frame head sequence was introduced to detectthe camera sample phase and obtain the gamma curve.Undersampled phase shift on-off keying (UPSOOK) [26] is a dual-LEDmodulation technique that can reach a higher transmission speed. The mo-dulation relies on different combinations of the ON/OFF status of two LEDs.Thus, the demodulator is required to know the exact sample phase of thecamera. To do so, the authors designed a special frame head which is exa-mined before every frame is demodulated. If an incorrect sample phase isdetected, all sampled values in the same data frame have to be inverted.OCC is sensitive to ambient light. Sunlight or other background noisecould degrade the system performance. In addition, a disturbing problemthat can increase the system bit-error rate (BER) for OCC is that the trans-mitters and receivers operate at asynchronous frequencies. To the best ofour knowledge, only a small number of error detection techniques have beenproposed for OCC. The authors in [14] proposed a space-time forward errorcorrection scheme based on USFOOK when dimming control was required.Multiple-phase sampling was employed to detect and correct the erroneoussample phase. This requires the transmitter to repeat the same data framethree times, either spatially or temporally, which will reduce the communi-cation efficiency.101.3. Thesis Organization and Contributions1.3 Thesis Organization and ContributionsThis thesis consists of six chapters. A summary of each chapter and itscontributions are presented as follows:In Chapter 1, we introduce necessary backgrounds and applications ofVLC and OCC. As an attractive technology, VLC can be applied to a greatnumber of application scenarios. We provide a comprehensive literaturereview of OCC in this chapter. The motivation and objective of this researchare also stated.In Chapter 2, essential technical background for this thesis is provi-ded. We review some fundamental principles of image sensors as well asimage sensor communication. After an overview of charge coupled de-vice (CCD) image sensors and complementary metal oxide semiconductor(CMOS) image sensors, we briefly introduce the image sensor as a VLC re-ceiver. The major challenges and requirements of designing an OCC systemare also presented.Chapter 3 focuses on the modulation and demodulation. A novel mo-dulation scheme undersampled differential phase shift on-off keying (UDPS-OOK) is proposed to overcome the phase uncertainty problem and increasethe data rate. The UDPSOOK demodulator does not need to know thesampling phase of the camera, and theoretically, it can double the data rateof UFSOOK. Furthermore, a novel error detection mechanism is proposedto help reduce the error bits caused by phase slipping. In addition, thedata frame structure and multiple-input and multiple-output (MIMO) arediscussed in Chapter 3.Chapter 4 and Chapter 5 focus on the experiments. Chapter 4 describesthe experimental setups for the proposed system. We discuss the finite state111.3. Thesis Organization and Contributionsmachine of the transmitter and the software flow chart of the receiver. De-tails of implementation are included in this chapter. Chapter 5 presents theexperimental results. Data rates and received optical power are measured.It is shown that UDPSOOK has doubled the data rate of UFSOOK. TheBER performance demonstrates that UDPSOOKED can effectively detectphase errors and can considerably enhance the reliability of UDPSOOK.Finally, we summarize the thesis in Chapter 6 and suggest some futureresearch work that may further improve the performance of the proposedcamera communication system.12Chapter 2Image Sensors and OCCNowadays image sensors are widely equipped in a variety of media, me-dical and other electronic devices. The applications for image sensors arechanging rapidly. Drones, robots, vehicles and augmented reality (AR) areamong the fastest growing market shares for image sensors. In addition,image sensors can also be used for communication purposes. In this chap-ter, we will give a brief introduction to CCD and CMOS image sensors aswell as optical camera communications.2.1 Image SensorsAn image sensor is a device that detects and conveys information of anoptical image by converting attenuation of light waves into electrical signals.An image sensor contains an array of pixels. The size of the array variesfrom 100× 100 (10,000 pixels) to 15700× 18000 (2,826,000,000 pixels) [28].Each pixel consists of a photodiode and a readout circuit. The readoutcircuit determines how fast camera capturing can be performed, i.e. theframe rate. For commercial cameras, it is typically 30 frame per second.CCD and CMOS image sensors are the two major technologies for imagecapturing [29]. Both types are capable of converting light into electrical sig-nals. These two technologies are comparable to each other in most respects.132.1. Image SensorsNeither of them has a compelling advantage over the other in image quality.However, as CMOS sensors cost less for fabrication and consume less powerthan CCD, they have been increasingly used in recent years. We will give abrief introduction to both of them.2.1.1 CCD Image SensorsIn a CCD sensor, pixels are organized in a two-dimensional array. Char-ges are accumulated in each photosite when light strikes. Those chargesmust be transferred to a readout node before they get measured. A shiftmechanism is used to sequentially transfer charges towards the readout stagewhere they are converted into voltages by an amplifier. As shown in Fig.2.1, the transfer circuit functions similar to a shift register. Since the num-ber of readout nodes is usually limited, the CCD is overshadowed by CMOStechnology when it comes to the capturing speed.2.1.2 CMOS Image SensorsCMOS image sensors have a different readout structure. Each pixelin a CMOS image sensor has its own individual amplifier integrated in-side. As shown in Fig. 2.2, at each photosite, there is a PD to carry outthe photon-to-electron conversion along with extra circuitry to convert thecharges immediately into a voltage. At the output stage, there is usually ananalog-to-digital converter (ADC) so that the image sensor is able to outputdigital values. CMOS sensors allow parallel operations and provide muchfaster readout speeds at the cost of a higher complexity of circuit design.Nowadays, the CMOS sensor has become more popular than its competitorin the fast-growing market.142.1. Image SensorsFigure2.1:ACCDimagesensor.152.1. Image SensorsFigure2.2:ACMOSimagesensor.162.2. Optical Communications for Cameras2.2 Optical Communications for Cameras2.2.1 Advantages of OCCAn image sensor consists of a lens and a two-dimensional PD array. Asshown in Fig. 2.3, light first passes through a lens and then strikes a grid ofphotosites, where current has been generated. After the electron-to-voltageconversion, the luminance value of every pixel is obtained. In this way, a fullimage of the LEDs is presented. The image sensor refreshes all the pixels ata fixed frame rate.When it comes to data transmission, a single PD receiver can achieve ahigher data rate compared to a low frame rate image sensor. The readouttime is mainly responsible for the relatively slow operation of the imagesensor. Since no readout circuit is deployed, a single photodiode can obtainthe output voltage more easily. However, there are also many competitiveadvantages of image sensor receivers.Figure 2.3: A multiple-LED OCC system.172.2. Optical Communications for CamerasThe most important advantage is that cameras are capable of receivingand processing data from multiple transmitters. The lens can spatially se-parate multiple transmitted signals. Having thousands and even millions ofpixels, an image sensor is able to receive signals from multiple channels. Sotechnically, each pixel can be modulated independently. As shown in Fig.2.3, there are two LED transmitters in this OCC system sending differentinformation simultaneously. The data is received in parallel even when twotransmitters are using different modulation schemes.Another great advantage of OCC is the ability to transfer visible lightsignals along with the position information. Every pixel has a row andcolumn position, which can be used to identify itself. That means we canmodulate the data by using the position information. Some VLC locationand communication systems have been proposed using this feature [30],[31].2.2.2 Design RequirementsThe design of an OCC system is limited by many factors. For OCCtransmitters, the modulation is subject to users’ demand for illumination.For OCC receivers, the hardware should be no more than an unmodifiedcommercial embedded camera [23]. As a consequence, an OCC system mustmeet the following requirements.First, an OCC system cannot affect the illumination performance ofLEDs. As a common light source in daily life, LEDs are required to exhibitstable intensity to users. It is well-known that images updating at a suffi-ciently high frame rate can appear steady. This rate is known as the criticalflicker fusion rate (CFF) [32]. We are making good use of this importantphenomenon everywhere in everyday life. Experiments have shown that the182.3. Summarymaximum observable rate for humans is 50 to 90 Hz [33]. On the otherhand, the camera’s cutoff frequency ranges from 1/8000 to several secondslong [34], depending on the shutter speed setting. For an embedded commer-cial camera, the upper limit is in the vicinity of 1/1000 seconds. As a result,in order to avoid visible flicker to human eyes, the modulation frequency forOCC is in the range of 90 to 1,000 Hz.Second, the receiver hardware must be a smartphone camera running ata common commercial frame rate, e.g. 30fps, since cellphones are scarcelyequipped with extremely expensive high-speed cameras. As OCC aims atthe commercial electronics market, a high frame rate receiver will absolutelynarrow the market for this technology.Moreover, OCC should offer accessible dimming control to users. Whenan intermittent light source has a frequency above the CFF, the Talbot-Plateau Law takes effect. It states that above CFF, a flash sequence isperceived as a steady light source which has the exact same average bright-ness as the former [35]. That is, assuming a 50% duty cycle is applied toan OOK signal, the LED will exhibit half intensity to human eyes. In otherwords, in order to offer controllable dimming levels to users, the OCC mustbe tolerant to any deviation from the designed duty cycle.In the following chapters, we will discuss the design and implementationof such an OCC system.2.3 SummaryIn this chapter, we presented essential technical background knowledgefor the entire thesis. CCD and CMOS image sensors were introduced. Wealso provided a brief description of optical camera communications. Finally,192.3. Summarywe discussed the design requirements of OCC systems.20Chapter 3Modulation/DemodulationIn this chapter, a new modulation scheme termed undersampled diffe-rential phase shift on-off keying is proposed for OCC. The basic idea ofUDPSOOK is to modulate binary bits by changing the phase difference be-tween two consecutive frames. We also introduce an error detection techni-que to improve the bit-error rate of the system. The data frame structureof UDPSOOK is introduced at the end of the chapter.3.1 UDPSOOKTo prevent human eyes from photobiological hazard, OCC systems arerequired to exhibit no flicker to users with different dimming requirements.As discussed in Chapter 2, the typical operation frequency of an OCC systemis between 90 Hz and 1,000 Hz. Assuming the frame rate of a commercialcamera is fs and the frequency of the OOK square wave carrier is fc, wecan always find an integer n for UDPSOOK to ensurefc = n× fs, 90Hz < fc < 1000Hz. (3.1)The transmitted UDPSOOK signal s(t, θ) can be expressed ass(t, θ) = sgn[sin(fct+ θ)] (3.2)213.1. UDPSOOKwhere fc is chosen by (3.1), θ is the phase of the OOK carrier, and signfunction is defined as follows:sgn(x) =1 x≥00 x<0(3.3)Every 1/fs seconds the modulator changes the value of θ depending onthe binary bit to be transmitted. A bit “1” is transmitted by adding 180◦phase shift to the current signal, while a “0” is transmitted by adding 0◦to θ. In other words, one bit of information is represented by a change ofphase between two consecutive frames.Figure 3.1 illustrates the modulated waveform of binary sequence “110”.In the figure, we assume fs is the camera sample rate and n = 4 so thatfc = 4fs. Every frame consists of 4 cycles of the OOK carrier. Dashed linesrepresent 1/fs seconds interval. The phase of the square waveform has beentoggled twice in the first two 1/fs seconds in order to transmit two logicones. Similarly, the phase stays unchanged for the last frame to send a logiczero.The red lines represent the sampling moments of the receiver. At thedemodulator, received bits are determined by comparing the phase betweentwo consecutive samples. The demodulation rule can be simply expressedasθk − θk−1 = θ∆ =0 “0”pi “1”(3.4)where θk is the carrier phase of the kth sampling. This rule can be imple-223.1. UDPSOOKFigure3.1:ModulatedUDPSOOKsymbols.233.1. UDPSOOKmented by using an exclusive OR operation asbk−1 = sk ⊕ sk−1 (3.5)where sk is the kth sampled value of the camera. We map sk to “1” whenthe LED is on and sk to “0” when the LED is off. The structure of thedemodulator can be simplified as shown in Fig. 3.2.Figure 3.2: UDPSOOK demodulation.Figure 3.3: Different sampling timing for UDPSOOK.UDPSOOK transmits signals by controlling the phases difference of twoconsecutive samples, so that n camera frames carry n− 1 bits information.243.1. UDPSOOKThe theoretical maximum data rate Rmax can be obtained byRmax = fs × limn→∞n− 1n= fs (3.6)which has been doubled compared to the maximum achieved data rate in[23].On the other hand, in a practical OCC system, it is difficult to predict thesampling phase of the camera. Fig. 3.3 is an example of camera sampling. Inthe figure the horizontal axis indicates the time. The time difference betweentwo consecutive arrows with the same colour is 1/fs. As the sampling canoccur at any position regarding the timeline, most modulation schemes [24–26] require extra algorithms to detect the phase relation between the camerasampling and the received signal waveforms. But UDPSOOK is totallyimmune from the camera phase uncertainty since the demodulation is basedon the phase difference between two frames rather than when the camerasampling occurs. In Fig. 3.3, red arrows indicate one set of possible samplesfor the transmitted sequence “110”. If the sampling takes place at the greenpositions, which have a pi phase shift from the red arrows, the sampled valueswill be fully inverted. However, the UDPSOOK demodulator obtains thesame bit sequence “110” from the green samples because the phase differencebetween two consecutive frames does not change.UDPSOOK also supports dimming control. The LED brightness canbe changed by increasing or decreasing the duty cycle of the OOK signal.A fixed 50% duty cycle is not practicable in reality. But for some OCCsystems [23–25], non-50% duty cycle modulation is not feasible or causesimproper sampling phases and hence more error bits. Assuming a logic oneis transmitted by UFSOOK (Fig. 3.4(a)) and UDPSOOK (Fig. 3.4(b))253.1. UDPSOOK(a) non-50% duty cycle decoding for UFSOOK(b) non-50% duty cycle decoding for UDPSOOKFigure 3.4: Non-50% duty cycle demodulation263.2. UDPSOOK with Error Detectionrespectively, it is shown that UFSOOK is more likely to have an erroneoussampling result (green arrows) when the duty cycle is greater than 50%, butUDPSOOK can adapt to any different duty cycle.Another problem introduced by image sensors is phase slipping. Fora typical OCC system, as the receiver and the transmitter are not strictlysynchronized, the sampling points are gradually slipping regarding the OOKwaveforms. A sampling phase error occurs when a pi phase shift comes. Inour experiment, phase slipping is a major source of error bits.3.2 UDPSOOK with Error DetectionAs phase slipping degenerates the performance of OCC systems, we pro-pose an error detection technique for UDPSOOK to reduce the system BER.In this proposed system, two LEDs are employed. We use one LED to ac-complish data transmission and an extra LED to carry out error detection.The data transmission LED is a UDPSOOK transmitter as introduced inChapter 3.1, and the error detection LED keeps sending bit “0”s during thedata transmission. As an all “0” UDPSOOK bit sequence triggers no phasetoggling on the LED, the demodulator expects to see no phase differenceon received image samples. Since this LED suffers from the same channelnoise and phase slipping problem, if any sampling error occurs, it can pre-cisely indicate the incorrect samples by looking for a sudden phase change.Then the demodulator corrects the sampling errors by simply flipping theerroneous sampled values.As shown in Fig. 3.5, the red arrows indicate a correct sampled sequence[ON, OFF], which is decoded as a bit “1”. When a phase error occurs, thesecond red sampling point slips to the green arrow position and gives rise273.3. Framing Structureto an incorrect sampled sequence [ON, ON], which will be decoded as a bit“0”. This error can be observed by the receiver due to the phase change ofthe error detection LED. The demodulator will invert the second sampledvalue to get the correct decoded bit.Figure 3.5: UDPSOOKED sampling3.3 Framing StructureIn order to detect the start of a transmission, the UDPSOOK data ispreceded by a start frame delimiter (SFD). An extra bit is added to the endof the payload for parity check to further enhance the system reliability. Adata frame is shown in Fig. 3.6.The SFD consists of two parts. The first part is a high-frequency OOKsymbol that lasts for two frames. This high frequency is required to begreater than the cutoff frequency of the camera so that it can be recognizedneither as a fully ON nor as a fully OFF. Instead, the camera extracts an283.3. Framing StructureFigure 3.6: Frame structure of the proposed OCC systemaverage intensity information from it [23] and interprets this status as a halfON assuming a 50% duty cycle is used.The second part of SFD for a data transmission LED is a UDPSOOK bit“1” (Fig. 3.6). If a logic “1” is not observed during the demodulation of anSFD, the data frame might have been corrupted and should be discarded.The error detection LED has a different second part of SFD. After the first2/fs seconds, it starts to transmit a bit “0”, i.e. two frames of OOK signalswhich have identical phase. In this way, the SFD not only starts a framebut also helps the receiver distinguish between the two different functioningLEDs. If a receiver observes a logic “0” after the 2/fs seconds half ON, itwill recognize this LED as an error detection transmitter.Following the SFD is the payload of the data frame. The data frameends when the receiver detects another SFD. The last bit of a data frameis a parity bit used to detect the possible transmission errors. The receivercalculates the number of “1”s when a full data frame is received. When293.4. MIMOparity check fails, all the data in current frame will be discarded.3.4 MIMOFigure 3.7: A possible MIMO scheme for UDPSOOKEDUDPSOOKED doubles the overhead of a single LED OCC system byintroducing an extra transmitter. However, multiple LEDs can share anerror detection LED in a MIMO system. In this case, the efficiency can besignificantly improved compared to the case with a single LED. Fig. 3.7illustrates one of the possible MIMO schemes for UDPSOOKED. In thefigure, LED 1 to LED 8 act as data transmission LEDs. Instead of usinganother 8 LEDs to accomplish error detection, a common LED (LED 9) issufficient to offer a better quality of service. Those transmitters are requiredto operate at a synchronous frequency.303.5. Summary3.5 SummaryThis chapter focused on the modulation and demodulation scheme ofUDPSOOK. Two consecutive frames were used to determine one-bit infor-mation. By introducing an extra error detection transmitter, we were ableto detect the phase slipping errors. UDPSOOKED applies when a highlyreliable transmission link is required. MIMO and data frame structure werealso discussed in this chapter.31Chapter 4Experimental SetupWe provide details of our experimental setup in this chapter. At thetransmitter, a Xilinx Virtex-7 field-programmable gate array (FPGA) mo-dulates the low power consumption LEDs. At the receiver, images are cap-tured by a Logitech 30 fps camera. Video frames have been recorded fordemodulation and further performance evaluation.4.1 Hardware DesignFigure 4.1 demonstrates our experimental setup. A Xilinx VC707 evalu-ation board is used to modulate information and carry out logical operationsfor transmitters. As shown in the figure, a breadboard provides connecti-ons between the evaluation board and the Digilent PmodLED LED module,which is powered by 2mA 3.3V FPGA input/output (I/O) pins. A camerareceiver continually records videos on the other side. Image processing isthen performed off-line in Matlab. We will introduce the transmitter andthe receiver hardware in the following sections.4.2 TransmitterFigure 4.2 illustrates the block diagram of the transmitter. The data isinitially stored in a read-only memory (ROM) created by the Xilinx Block324.2. TransmitterFigure 4.1: Experimental setup.334.2. TransmitterMemory Generator core. A parallel-to-serial converter fetches byte datafrom the memory and converts them into a serial bit stream. The modulatoris responsible for generating the OOK carrier and encoding binary bits intoUDPSOOK symbols. The modulated signals are then sent to an LED drivercircuit which provides sufficient transmission power for driving the LEDs.Figure 4.2: Block diagram of the transmitter.4.2.1 ModulatorThe modulation algorithm is implemented on a Xilinx Virtex-7 FPGA.A finite state machine (FSM) is designed to control the transmission processas shown in Fig. 4.3. When transmission starts, the modulator first sendsa high frequency signal as part of the SFD. Followed is a UDPSOOK bit“1”, i.e. two frames of OOK with pi phase difference. The modulator thenchanges the phase of the carrier every 1/fs seconds according to the nextinput bit. At the end of the transmission, a parity bit will be calculatedand added to the end of the frame. This modulation FSM is implementedin Verilog hardware description language (HDL). Appendix A gives a smallsnippet of the FSM implementation. Xilinx Vivado Design Suite performs344.2. Transmitterregister-transfer level (RTL) synthesis and timing analysis etc., and the in-built simulator provides reliable behavioral verification before testing onhardware. The simulation result is shown in Fig. 4.4. An enRead pulsereads one bit from the memory and LED En enables the I/O ports on andoff according to the modulation rules. Note that the time scale has beenshrunk by 1000 times to improve simulation efficiency and the figure onlydemonstrates the initial 300 us after reset.Figure 4.3: The transmission control finite state machine.4.2.2 LED Driver CircuitThe LED driver circuit is a power supply whose output may vary tomatch the electrical characteristics of the LEDs. We select the DigilentPmodLED LED module as our solution for transmitters. The PmodLEDLED module integrates four high-bright monochrome red LEDs with the354.3. ReceiverFigure 4.4: The simulation results in Xilinx Vivado simulator.necessary driver circuits. Those low power consumption LEDs can be drivenby less than 1 mA current. The schematic of the LED driver circuit is shownin Fig. ReceiverAt the receiver, videos are recorded by a Logitech Pro 900 camera. Thecamera is then connected to a PC where captured images are processed viaMATLAB. Auto white balance and other optimization options need to bedisabled. With proper configurations, the influence of ambient light can bereduced, so that the camera can identify the transmitted image with lowererror rate.After frames being sampled, we need to map the LED status into binarybits. Demodulation is performed on the sampled sequence. Fig. 4.6 explainsthe UDPSOOK demodulation algorithm. It starts with monitoring the firstSFD, which indicates the beginning of a new data frame. After an SFDis found, the demodulator begins to compare two consecutive samples todetermine the transmitted bit. This process repeats until the next SFD is364.4. SummaryFigure 4.5: The PmodLED LED module [36].detected. We have to check the parity bit before proceeding to the nextdata frame.For UDPSOOKED, the algorithm presented in Fig. 4.6 needs to beperformed on both LEDs. Assuming the decoded bit sequences on twoLEDs are respectively {Rd0, Rd1, Rd2, ......Rdk} and {Re0, Re1, Re2, ......Rek},the final decoded sequence is given byBk = Rdk ⊕Rek. (4.1)4.4 SummaryIn this chapter, we introduced our experimental setups for the proposedOCC system. Hardware selection and software flow chart are described. Wedesigned the transmitter and the receiver in a very inexpensive way in order374.4. SummaryFigure 4.6: A flow chart of the UDPSOOK receiver.384.4. Summaryto lower the entry barrier to the market. We will present the experimentalresults in Chapter 5.39Chapter 5Experimental ResultsIn this chapter, we present the experimental results of the proposed OCCsystem. Data rates are given as well as the BER performance. The receivedoptical power is also measured.5.1 Received Optical PowerBefore we quantify the system performance by measuring the BER, itis important to specify the experimental parameters. As the LED electricalpower consumption is only on the order of mW, one of our central concernsis how much optical power has reached the camera. Fig. 5.1 shows theexperimental setup used to measure the optical power at the receiver. In thefigure, d is the transmission distance and f is the focal length of the one-inch-diameter lens. We employ a high-speed photodetector (Thorlabs DET36A)for the measurement. The voltage difference across the load resistor RL isdirectly measured in the experiment. We select a 10MΩ resistor since a lowgenerated photocurrent can be expected. Table 5.1 gives the voltage resultsover three different transmission distances.The responsivity curves of the selected photodetector can be obtainedfrom the product specification as shown in Fig. 5.2 [37]. The PmodLEDmodule only emits red light (wavelength in the vicinity of 650 nm), and405.1. Received Optical PowerFigure 5.1: Optical power measurement setup.Table 5.1: Received optical intensity for the proposed OCC systemDistance Voltage Current Incident power Intensity15 cm 2.36 V 2.36 µA 7.15 µW 1.41 µW/cm250 cm 220 mV 22 nA 66.67 nW 13.16nW/cm2100 cm 60.8 mV 6.08 nA 18.42 nW 3.64 nW/cm2accordingly, we select responsivity R = 0.33A/W . As the responsivity ofa photodiode is defined as a ratio of the photocurrent to the incident lightpower at a given wavelength, to calculate the received optical power, we usePinc =IPDR(5.1)where Pinc is the incident power and IPD is the photocurrent, which can becalculated fromIPD =VLRL(5.2)where VL is the voltage across the load resistor RL. The result is also shownin Table 5.1. As expected, the received power dramatically drops when weincrease the transmission distance.Furthermore, the area of a one-inch-diameter lens is pi × (25.4mm/2)2,so the signal intensity at the receiver can be estimated. Over the distance415.2. BER and Bit RateFigure 5.2: DET36A responsivity [37].of 1 meter we have signal intensityIntd=1m =(Pinc)d=1mA=18.42nWpi × (25.4mm/2)2 ≈ 3.64nW/cm2. (5.3)Similarly, we can calculate the signal intensity for 15 cm and 50 cm, which isshown in Table 5.1. As the ambient intensity is on the order of 0.1nW/cm2[38], we have approximately 10 dB of dynamic range.5.2 BER and Bit RateThe modulator has been implemented on a Xilinx VC707 Evaluationboard. We use a logic analyzer to verify the modulated signals. The capturedwaveforms are shown in Fig. 5.3. The blue waveform represents the datatransmission channel and the yellow waveform represents the error detection425.2. BER and Bit RateFigure 5.3: Captured waveforms from the logic analyzer. 435.2. BER and Bit Ratechannel. In the figure, a new data frame starts at 296 ms with a 66 ms 12KHz high-frequency OOK. A bit “1” has been transmitted between 362msand 429 ms on the blue channel as part of the SFD, which is followed bymore UDPSOOK symbols. As we can see from the figure, there is no phasechange on the error detection channel.Then we use the modulated electrical signals to drive the LED circuits.Digilent PmodLED is a high-brightness LED module with low power requi-rements. In our experiment, for each LED the electrical power consumptionis 3.30V × 2.45mA = 8.10mW .At the receiver, videos have been recorded by a Logitech Pro 900 camera.The camera has been set up in a normal indoor environment with noise fromsunlight outside of the room. We set the frame rate as 30 fps and all collectedimages are processed by MATLAB. In the experiment, the auto focus andauto white balance function of the camera are disabled. Key experimentalparameters are shown in Table 5.2. We collect approximate 20,000 framesfor each measurement. Fig. 5.4 demonstrates different LED status capturedTable 5.2: Key experimental parametersParameter ValueCamera frame rate 30 fpsResolution 640×480Saturation 0LED DC offset 1.65 VLED peak-to-peak voltage 3.3 Vby our camera within a data transmission. On the left hand side is a datatransmission LED and on the right hand side is an error detection LED. Fig.445.2. BER and Bit Rate5.4(c) reflects the status of both LEDs when they are modulated by the high-frequency signal in an SFD frame head, which can only be interpreted byits average intensity. Fig. 5.4(a) and Fig. 5.4(b) present the fully ON andfully OFF status of the data transmission LED respectively, which composea UDPSOOK symbol “1”. As one might expect, the status of the errordetection LED stays unchanged during the bit transmission.(a) ON (b) OFF(c) SFDFigure 5.4: UDPSOOKED symbols captured by camera.We have tested three different communication distances for both UDPS-OOK and UDPSOOKED. The BER results are shown in Table 5.3. It is455.2. BER and Bit Ratesuggested that UDPSOOK provides acceptable BER performance under allthree distances considered, and UDPSOOKED can lower the BER efficiently,especially for the distances of 15 cm and 50 cm. The result also shows thatwhen transmission distance increases, the system BER increases.Table 5.3: Experimental results IDistance Modulation BER15 cm UDPSOOK 2.48×10−415 cm UDPSOOKED 4.96×10−550 cm UDPSOOK 4.25×10−450 cm UDPSOOKED 6.08×10−51 m UDPSOOK 8.15×10−31 m UDPSOOKED 1.88×10−3We have also implemented UFSOOK in the same experimental environ-ment. The data rate comparison is shown in Table 5.4. The experimentalresults demonstrate that UDPSOOK has indeed doubled the data rate ofUFSOOK as expected in Chapter 3.1.Table 5.4: Experimental results IIDistance Modulation Data Rate (bps)15 cm UFSOOK 11.0615 cm UDPSOOK 23.02465.3. Summary5.3 SummaryIn this chapter, we presented the experimental results of the proposedOCC system. The received optical power has been measured. By comparingthe BER performance of UDPSOOK and UDPSOOKED, we can concludethat the second LED can efficiently enhance the reliability of the proposedsystem. The data rate has doubled compared to the UFSOOK system.47Chapter 6ConclusionsIn this chapter, we summarize the accomplished work and propose somefuture research topics.6.1 Summary of Accomplished WorkIn this thesis, we have proposed a novel modulation technique calledUDPSOOK for OCC. By introducing this new scheme, the theoretical com-munication data rate has doubled compared with [23]. We have also designedan error detection scheme for UDPSOOK, i.e. to use a second LED as a de-tector of phase slipping errors. This method mitigates the asynchronizationproblem of OCC that other researchers are concerned with.On the other hand, an experimental communication link has been esta-blished by using monochrome LEDs and an inexpensive commercial camera.Experiments have demonstrated that a BER of 10−5 has been achieved bysuch low overheads. Experimental work shows that UDPSOOKED can pro-vide excellent robustness and reliability for OCC systems.In Chapter 3, we proposed UDPSOOK and UDPSOOKED for OCC. Weprovided the details on modulation and demodulation. In order to reducethe BER of the system, an error detection LED was employed. A MIMOscheme was also introduced in this chapter.486.2. Suggested Future WorkIn Chapter 4, we introduced our experimental setups for the proposedOCC system. Both hardware selection and software flow chart were detailed.We have designed the transmitter and the receiver in a very inexpensive wayin order to lower the entry barrier to the consumer electronics market.Chapter 5 presented the experimental results of the proposed OCC sy-stem. By comparing the BER performance of UDPSOOK and UDPS-OOKED, we found the proposed error detection scheme can reduce theerror bits efficiently. The communication data rate was doubled comparedto UFSOOK systems.6.2 Suggested Future WorkThe low data rate is still the bottleneck for the OCC technology dueto the relatively low frame rate of the receiver. However, there are stillsome approaches that we can use to improve the communication data rate.MIMO can be one of the major solutions. Another potential future initiativeis to design a real-time OCC system. Currently, we use captured videos fordemodulation in our experiments. In real use cases, users will be unsatisfiedwith the delay of video capturing. To this end, an easy-to-use smartphoneapplication is necessary. How to select appropriate hardware for real-timeOCC would be another interesting topic in the future.49Bibliography[1] A. G. Bell, “On the production and reproduction of speech by light,”American Journal of Science, vol. 20, no. 118, pp. 305–324, Oct. 1880.→ pages 1[2] Wikipedia, “Optical communication,” 2016. [Online]. Available:https://en.wikipedia.org/wiki/Optical communication → pages 1[3] I. Ahemen, D. K. De, and A. Amah, “A review of solid state white lightemitting diode and its potentials for replacing conventional lightingtechnologies in developing countries,” Applied Physics Research, vol. 6,p. 95, Feb. 2014. → pages 2[4] Wikipedia, “Intelligent transportation system,” 2016. [Online].Available: https://en.wikipedia.org/wiki/Intelligent transportationsystem → pages 3[5] K. Tamura and M. Hirayama, “Toward realization of VICS - VehicleInformation and Communication System,” in Vehicle Navigation andInformation Systems Conference, 1993., Proceedings of the IEEE-IEE,Ottawa, Canada, Oct. 1993, pp. 72–77. → pages 4[6] S. Kitano, S. Haruyama, and M. Nakagawa, “LED road illumina-tion communications system,” in Vehicular Technology Conference, Or-lando, Florida, Oct. 2003, pp. 3346–3350. → pages 450Bibliography[7] J. Song, W. Ding, F. Yang, H. Yang, J. Wang, X. Wang, and X. Zhang,“Indoor hospital communication systems: An integrated solution basedon power line and visible light communication,” in Faible Tension FaibleConsommation (FTFC), Monaco, Monaco, May 2014, pp. 1–6.→ pages4[8] N. Loureno, D. Terra, N. Kumar, L. N. Alves, and R. L. Aguiar, “Visi-ble light communication system for outdoor applications,” in Internati-onal Symposium on Communication Systems, Networks Digital SignalProcessing (CSNDSP), Poznan, Poland, Jul. 2012, pp. 1–6. → pages 5[9] G. Corbellini, K. Aksit, S. Schmid, S. Mangold, and T. Gross, “Con-necting networks of toys and smartphones with visible light commu-nication,” Communications Magazine, IEEE, vol. 52, pp. 72–78, Jul.2014. → pages 5[10] S. Mangold, “Visible light communications for entertainment networ-king,” in Photonics Society Summer Topical Meeting Series, Seattle,WA, Jul. 2012, pp. 100–101. → pages 5[11] P. Dietz, W. Yerazunis, and D. Leigh, “Very low-cost sensing and com-munication using bidirectional LEDs,” in Proc. International Confe-rence on Ubiquitous Computing, Seattle, Washington, Oct. 2003, pp.175–191. → pages 5[12] L. Zeng, D. C. O’Brien, H. Minh, G. E. Faulkner, K. Lee, D. Jung,Y. Oh, and E. T. Won, “High data rate multiple input multiple output(MIMO) optical wireless communications using white LED lighting,”IEEE Journal on Selected Areas in Communications, vol. 27, no. 9, pp.1654–1662, Dec. 2009. → pages 651Bibliography[13] A. Wilkins, J. Veitch, and B. Lehman, “LED lighting flicker and poten-tial health concerns: IEEE standard PAR1789 update,” in 2010 IEEEEnergy Conversion Congress and Exposition, Atlanta, GA, Sep. 2010,pp. 171–178. → pages 6[14] R. D. Roberts, “Space-time forward error correction for dimmable un-dersampled frequency shift ON-OFF keying camera communications(CamCom),” in Proc. International Conference on Ubiquitous and Fu-ture Networks (ICUFN), Da Nang, Vietnam, July 2013, pp. 459–464.→ pages 6, 10[15] S. Miyauchi, T. Komine, T. Ushiro, S. Yoshimura, S. Haruyama, andM. Nakagawa, “Parallel wireless optical communication using highspeed CMOS image sensor,” in Proc. International symposium in infor-mation Theory and its applications (ISITA), Parma, Italy, Oct. 2004,pp. 72–77. → pages 7[16] M. Wada, T. Yendo, T. Fujii, and M. Tanimoto, “Road-to-vehicle com-munication using LED traffic light,” in Proc. Intelligent Vehicles Sym-posium, Las Vegas, NV, Jun. 2005, pp. 601–606. → pages 7[17] H. C. N. Premachandra, T. Yendo, T. Yamasato, T. Fujii, M. Tanimoto,and Y. Kimura, “Detection of LED traffic light by image processing forvisible light communication system,” in Intelligent Vehicles Symposium,Xi’an, China, Jun. 2009, pp. 179–184. → pages 7[18] T. Yamazato, I. Takai, H. Okada, T. Fujii, T. Yendo, S. Arai, M. An-doh, T. Harada, K. Yasutomi, K. Kagawa et al., “Image-sensor-basedvisible light communication for automotive applications,” IEEE Com-munications Magazine, vol. 52, pp. 88–97, Jul. 2014. → pages 752Bibliography[19] R. D. Roberts, “Kick-starting the VLC market via optical camera com-munications,” in Proc. International Conference and Exhibition on Vi-sible Light Communications, Yokohama, Japan, Oct. 2015. → pages7[20] G. Corbellini, K. Aksit, S. Schmid, S. Mangold, and T. R. Gross, “Con-necting networks of toys and smartphones with visible light communi-cation,” IEEE Communications Magazine, vol. 52, pp. 72–78, Jul. 2014.→ pages 7[21] C. E. Shannon, “Communication in the presence of noise,” Proceedingsof the IEEE, vol. 72, pp. 1192–1201, Sep. 1984. → pages 9[22] C. Danakis, M. Afgani, G. Povey, I. Underwood, and H. Haas, “Using aCMOS camera sensor for visible light communication,” in Proc. IEEEGlobecom Workshops, Anaheim, CA, Dec. 2012, pp. 1244–1248. →pages 9[23] R. D. Roberts, “Undersampled frequency shift ON-OFF keying (UF-SOOK) for camera communications (CamCom),” in Proc. Wireless andOptical Communication Conference, Chongqing, China, May 2013, pp.645–648. → pages 9, 18, 25, 29, 48[24] P. Luo, Z. Ghassemlooy, H. L. Minh, H.-M. Tsai, and X. Tang,“Undersampled-PAM with subcarrier modulation for camera commu-nications,” in Proc. Opto-Electronics and Communications Conference(OECC), 2015, Shanghai, China, Jun. 2015, pp. 1–3. → pages 9, 10,25[25] P. Luo, M. Zhang, Z. Ghassemlooy, H. L. Minh, H. M. Tsai, X. Tang,53Bibliographyand D. Han, “Experimental demonstration of a 1024-QAM opticalcamera communication system,” IEEE Photonics Technology Letters,vol. 28, no. 2, pp. 139–142, Jan. 2016. → pages 9, 10, 25[26] P. Luo, Z. Ghassemlooy, H. Le Minh, X. Tang, and H.-M. Tsai, “Un-dersampled phase shift ON-OFF keying for camera communication,” inProc. International Conference on Wireless Communications and Sig-nal Processing (WCSP), Hefei, China, Oct. 2014, pp. 1–6. → pages 9,10, 25[27] C. Poynton, Digital Video and HDTV Algorithms and Interfaces. SanFrancisco: Morgan Kaufmann, 2002. → pages 9[28] Wikipedia, “Image sensor,” 2016. [Online]. Available: https://en.wikipedia.org/wiki/Image sensor → pages 13[29] S. Haruyama and T. Yamazato, “Image sensor based visible light com-munication,” in Visible Light Communication, S. Arnon, Ed. Cam-bridge: Cambridge University Press, 2015, pp. 181–205. → pages 13[30] H. C. N. Premachandra, T. Yendo, M. P. Tehrani, T. Yamazato,H. Okada, T. Fujii, and M. Tanimoto, “High-speed-camera image pro-cessing based LED traffic light detection for road-to-vehicle visible lightcommunication,” in Intelligent Vehicles Symposium, San Diego, CA,June 2010, pp. 793–798. → pages 18[31] S. Nishimoto, T. Nagura, T. Yamazato, T. Yendo, T. Fujii, H. Okada,and S. Arai, “Overlay coding for road-to-vehicle visible light commu-nication using LED array and high-speed camera,” in International54BibliographyConference on Intelligent Transportation Systems (ITSC), Washington,DC, Oct. 2011, pp. 1704–1709. → pages 18[32] Wikipedia, “Flicker fusion threshold,” 2016. [Online]. Available:https://en.wikipedia.org/wiki/Flicker fusion threshold → pages 18[33] J. Davis, Y.-H. Hsieh, and H.-C. Lee, “Humans perceive flicker artifactsat 500 Hz,” Scientific Reports, vol. 5, pp. 7861–7861, Feb. 2015.→ pages19[34] D. Doeffinger, Creative Shutter Speed: Master the Art of Motion Cap-ture. Indianapolis: Wiley, 2009. → pages 19[35] L. Pinneo and R. Heath, “Human visual system activity and percep-tion of intermittent light stimuli,” Journal of the Neurological Sciences,vol. 5, pp. 303–314, Sep. 1967. → pages 19[36] Digilent Inc., “PmodLED Reference Manual,” Pullman, WA, Mar.2016. → pages 37[37] Thorlabs Inc., “DET36A Si biased detector user guide,” Newton, NJ,Mar. 2015. → pages 40, 42[38] M. H. Bergen, A. Arafa, X. Jin, R. Klukas, and J. F. Holzman, “Charac-teristics of angular precision and dilution of precision for optical wirelesspositioning,” Journal of Lightwave Technology, vol. 33, pp. 4253–4260,Oct. 2015. → pages 4255Appendix56Appendix ATo implement the FSM described in Chapter 4.2, we have the followingcode snippet. This is a three-always-block style Verilog DHL implementationwith registered outputs, which is consistent with industry practice./∗ ∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗fsm fo r t r a n s im i t t e r∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗∗ ∗/parameter TX IDLE = 2 ’ b00 ;parameter TX SFD = 2 ’ b01 ;parameter TXMARK = 2 ’ b10 ;parameter TX SPACE = 2 ’ b11 ;reg [ 1 : 0 ] TxPreStatus ;reg [ 1 : 0 ] TxNxtStatus ;always @ (TxPreStatus or isSFD or pCounter or Counter orf S i z e or Nxt Bit )begincase ( TxPreStatus )TX IDLE :beginTxNxtStatus = TX SFD;endTX SFD:begini f ( isSFD & ( pCounter == SFD LEN − 1 ’ b1 )&& (Counter == TMS SFD −1’b1 ) )TxNxtStatus = TXMARK;57Appendix A.elseTxNxtStatus = TX SFD;endTXMARK:begini f ( ( f S i z e >= FRAME SIZE − 1 ’ b1 ) && ( pCounter >= BIT LEN −1 ’ b1 ) && (Counter >= TMS BIT −1’d1 ) )TxNxtStatus = TX SFD;else i f ( ( pCounter >= BIT LEN − 1 ’ b1 )&& (Counter >= TMS BIT −1’d1 ) & Nxt Bit )TxNxtStatus = TXMARK;else i f ( ( pCounter >= BIT LEN − 1 ’ b1 )&& (Counter >= TMS BIT −1’d1 ) & ˜Nxt Bit )TxNxtStatus = TX SPACE;elseTxNxtStatus = TXMARK;endTX SPACE:begini f ( ( f S i z e >= FRAME SIZE − 1 ’ b1 ) && ( pCounter >= BIT LEN −1 ’ b1 ) && (Counter >= TMS BIT −1’d1 ) )TxNxtStatus = TX SFD;else i f ( ( pCounter >= BIT LEN − 1 ’ b1 )&& (Counter >= TMS BIT −1’d1 ) & Nxt Bit )TxNxtStatus = TXMARK;else i f ( ( pCounter == BIT LEN − 1 ’ b1 )&& (Counter >= TMS BIT −1’d1 ) & ˜Nxt Bit )TxNxtStatus = TX SPACE;elseTxNxtStatus = TX SPACE;enddefault : TxNxtStatus = TX SFD;endcase58Appendix A.endalways @ (posedge CLK or posedge RST)begini f (RST)TxPreStatus <= TX IDLE ;elseTxPreStatus <= TxNxtStatus ;endalways @ (posedge CLK)begincase ( TxPreStatus )TX IDLE :beginCounter <= 24 ’ b0 ;pCounter <= 24 ’ d1 ;isSFD <= 1 ’ b1 ;c b i t <= 1 ’ b0 ;f S i z e <= 8 ’ d0 ;i sOver <= 1 ’ b0 ;l e d s t a t e <= 1 ’ b0 ;endTX SFD:beginf S i z e <= 8 ’ d0 ;l e d s t a t e <= 1 ’ b0 ;i f ( isSFD )begini f ( ( pCounter >= SFD LEN − 1 ’ b1 ) &&(Counter >= TMS SFD −1’b1 ) )beginCounter <= 24 ’ b0 ;59Appendix A.pCounter <= 24 ’ d0 ;c b i t <= 1 ’ b1 ;isSFD <= 1 ’ b0 ;endelse i f ( Counter >= TMS SFD −1’b1 )beginCounter <= 24 ’ b0 ;pCounter <= pCounter + 1 ’ b1 ;endelseCounter <= Counter + 1 ’ b1 ;endendTXMARK, TX SPACE:begini f ( ( f S i z e >= FRAME SIZE − 1 ’ b1 ) && ( pCounter >= BIT LEN −1 ’ b1 ) && (Counter >= TMS BIT −1’d1 ) )beginf S i z e <= 8 ’ d0 ;isSFD <= 1 ’ b1 ;Counter <= 24 ’ b0 ;pCounter <= 24 ’ d0 ;c b i t <= Nxt Bit ;i sOver <= 1 ’ b1 ;i f ( Nxt Bit )l e d s t a t e <= ˜ l e d s t a t e ;endelse i f ( ( pCounter >= BIT LEN − 1 ’ b1 )&& (Counter >= TMS BIT −1’b1 ) )beginCounter <= 24 ’ b0 ;pCounter <= 24 ’ d0 ;i sOver <= 1 ’ b1 ;60Appendix A.cb i t <= Nxt Bit ;f S i z e <= fS i z e + 1 ’ b1 ;i f ( Nxt Bit )l e d s t a t e <= ˜ l e d s t a t e ;endelse i f ( Counter >= TMS BIT −1’b1 )beginCounter <= 24 ’ b0 ;pCounter <= pCounter + 1 ’ b1 ;endelsebeginCounter <= Counter + 1 ’ b1 ;i sOver <= 1 ’ b0 ;endendendcaseend61


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items