When Jesus Takes the Wheel: An Investigation of Distraction in Autonomous Vehicles

Abstract

Autonomous vehicles have been suggested to be a solution to the problem of distracted driving. However, because autonomous vehicles are still developing, little is known about how drivers interact with them. Today’s autonomous vehicles still require drivers to be available to take control quickly. If drivers are engaged in secondary tasks, they are less able to safely take control or detect system notifications to take control. Using the delay discounting method, the cognitive underpinnings of the human decision-making process can be understood to inform us of the extent to which drivers are willing to engage with a distraction. The current work found a distinct group of high impulsivity group who were more willing to engage with distraction sooner, opposed to the low impulsive group. Regardless of impulsivity group, willingness to engage with distraction decreased after driving a partially autonomous vehicle. This timing effect was present in subsequent analyses for the high impulsive group but not the low impulsive group. However, there was an interaction for timing and vehicle driven among both the high and low groups in which the high impulsive group generally became less willing to engage with distraction after driving the most vehicles, and the low impulsive group became more willing to engage with distraction. Also, the overall group was less willing to engage with distraction when hypothetically driving a standard vehicle than a fully autonomous vehicle. This effect was also found among the high impulsivity group but not the low impulsivity group. Finally, only the low impulsivity group reported less willingness to respond when the message was on the phone’s screen rather than the vehicle voice system. However, there was an interaction with the message modality and timing. After driving, both the overall sample and low impulsivity group were less willing to respond to a message via the phone but more willing to respond via voice system. Examining driver behavior and cognitive demand in autonomous vehicles has critical implications for understanding how drivers interact with these vehicles. As autonomous vehicles become more mainstream, it becomes increasingly necessary for our safety to understand driver behavior in varying circumstances.