Modeling Users’ Behavior Toward Automated Vehicles and Mobility Services Using Revealed and Stated Preference Data

Abstract

Emerging technologies in transportation, such as automated vehicles (AVs) and mobility services, are expected to impact travelers’ behavior and choices. However, due to many uncertainties surrounding these new technologies, the magnitude and direction of this impact remain a mystery. Literature on AVs identifies several crucial questions and issues surrounding automated vehicles and new mobility services including: (1) potential induced demand, (2) trust in technology and its effect on adoption, (3) AVs as a mobility-as-a-service enabler. In this dissertation, I aimed to tackle these issues by quantifying value of travel time as a determinant of induced demand, study trust in AV technology as a key determinant of adoption, and modeling within tour inter-dependencies as determinant of multimodal travel and MaaS adoption. First, I use the data of actual mode choices between ridehailing and free-float carsharing to build models of mode choices to inform analyses of the prospective change in time valuation and travel behavior when riding in future highly AVs. Then, I discuss the design and implementation of a choice survey based on users’ revealed trip diary that overcomes shortcomings of revealed preference data. Next, I use the data from the choice survey to build an integrated choice and latent variable (ICLV) model that quantifies the impact of psychological constructs such as AVs safety perception on trip-based mode choices, specifically choices involving privately-owned AVs and driverless ridehailing services. Finally, I build tour-based mode choice models that allow capturing interdependencies among trips within a tour and explore potential for multimodal trip. My results from analyzing revealed preference data shows that riding in a car versus driving one reduces the value of travel time (VoTT) by $23/hour which confirms a significant time savings benefit in eliminating the burden of driving for travelers. While AVs potentially provide time saving benefits, based on current public’s assessment of the technology’s safety, market share of AVs remain small. However, improvements in users’ perception of AVs’ safety can considerably grow the market share for privately-owned AVs to the point that it hinders market share of driverless ridehailing. Another avenue for AVs to affect transportation system is enabling multimodal travel. Using tour-based mode choice modeling, I found that people preferences to use unimodal tours when using AVs are about the same as any other modes and I identified strong inclination among our sample to use unimodal tours despite the mode of travel. The findings of this dissertation highlight the potential for increases in VMT and as a result increases in induced demand and GHG emissions, as it is expected that people’s value of travel time considerably drops in AVs and market share of AVs grow substantially when users perceive them safe. Also, as highlighted in this dissertation, even with AVs and driverless ridehailing mode inertia is high among users, and solely introducing these new modes would not contribute to multimodal travels. This dissertation illustrates that the adoption of AVs cannot solve many of the pressing transportation issues if they are introduced to the current system without any changes to the system. There is a need for policies and plans in place to make sure the new technologies potential is directed toward a more sustainable future.