Open Collections

Bardutz, Emily

University of British Columbia

This thesis aims to inform fully autonomous self-driving vehicle (SDV) policy with respect to pedestrian comfort and safety. To pursue new insights, the novel approach of Structural Topic Modeling is used to organize open-response comments into prominent topics, or themes, and explore the factors affecting the prevalence and content of topics. This thesis is comprised of two parts. In both parts, participants in an online survey rated the comfort and safety of pedestrian interactions shown in short video clips and provided open-response comments clarifying their ratings. In part 1, the video clips showed pedestrian interactions with human-driven vehicles (HDVs) and bicycles. In Part 2, the video clips only showed interactions with HDVs, except participants were told that half of the videos involved SDVS. This deception-based survey design allowed for isolation of the intrinsic effect of SDV technology on comfort and safety perceptions across a broad sample of the population, separate from potential operating differences from HDVs. Results indicate that pedestrian interactions with each of the three types of road users were perceived as fundamentally different and should be treated as such in future research and analysis. Interactions with bicycles elicited topics mainly concerned with yielding; interactions with HDVs elicited topics mainly concerned with risk; and interactions with SDVs elicited topics mainly concerned with rule compliance. Participants increasingly ascribed more responsibility to the pedestrian as the degree of detachment (physically and conceptually) between the pedestrian and the interacting road user increased – from cyclist to HDV to SDV. To ensure that the integration of SDVs into the existing transportation system does not undermine efforts to prioritize and promote active transportation, SDVs should be programmed to strictly comply with road rules.

Text

2022-08-16

Attribution-NonCommercial-NoDerivatives 4.0 International

http://hdl.handle.net/2429/82374

Planning

University of British Columbia

UBCV

http://creativecommons.org/licenses/by-nc-nd/4.0/