Emergency Autonomous to Manual Takeover in a Driving Simulator: Teen vs. Adult Drivers – A Pilot Study

摘要

Autonomous and/or automated vehicles offer a host of future opportunities but leave many questions unanswered regarding their impact on crash avoidance or the ability of drivers to effectively scan and re-engage from self-driving mode when necessary to avoid crash scenarios. Considering a 16-year-old is several times more likely to die in an automobile crash than other licensed drivers, it was crucial to test both teenage drivers and adults to determine head-on collision avoidance abilities when subjected to a failing autopilot in a simulated autonomous vehicle. In this study, eight teenagers ages 16-19 and four experienced adults underwent four simulated drives (one manual practice drive and three simulated autonomous drives) using a hi-fidelity, Real Time Technologies SimDriver Simulator to represent being in a self-driving vehicle. When exposed to a head-on crash event where the opposing vehicle crosses the dividing line and drives towards the subject, most subjects successfully swerved to avoid the car albeit with several oscillations before stabilization.