Presentation
Connecting AV Behavior to Future Adoption: An On-Road Robotaxi Study
DescriptionIt is too often assumed in technology development that observing and measuring human interaction with that technology is not possible until after a prototype is created. We show here a counter-example, of measuring human responses to simulated vehicle behavior in a real autonomous vehicle in public traffic. We used Behavioral Intention to extrapolate from a specific in-car experience to the likelihood of future repeat ridership, based on its a validated link to future behavior. This method allows a lightweight measure, collectable in nearly any circumstances, to compare the adoption or revenue impact of multiple experiences before those experiences have been implemented. In this case, we discovered substantially more patience with in-traffic AV stops than we anticipated, with Behavioral Intention not decreasing substantially until stop duration was over 90 seconds. We used a private validation data set and internal assumptions about ride value and frequency to create financial projections comparing these different stop duration experiences to support decision-making around work prioritization and resources as well as vehicle behavior criteria. We suggest that Behavioral Intention and its sibling measure, Behavioral Expectation, can be used in robotics and automation more broadly to understand future usage behavior and guide system development.
Contributor
Event Type
Lecture
TimeTuesday, October 20th3:40pm - 4pm PDT
Location
Similar Presentations