Behavioral and physiological responses to takeovers in different scenarios during conditionally automated driving

A variety of takeover scenarios will happen in conditionally automated driving. Previous studies presented mixed results regarding the effects of scenarios on takeover performance. According to drivers' strategies for takeover requests, this study selected eight representative takeover scenarios and categorized them into lane keeping and lane changing scenarios. To investigate the effects of scenario type and road environment (highway vs. urban) on drivers' takeover performance and physiological responses, a driving simulation study was conducted as a mixed design with 40 participants (average age = 22.8 years). The results showed that in lane changing scenarios, with the same sensing capability, drivers on highways had deteriorated takeover performance in the form of harsher takeover maneuvers and higher collision risk, as well as higher arousal and stress, compared to urban areas. However, such effects disappeared or even reversed in lane keeping scenarios on the curves, where drivers on highways had smoother takeover maneuvers and lower arousal and stress. These findings will help us understand the vital roles scenario type and road environment play during takeover transitions. Our findings have implications for the design of advanced driver-assistance systems and will improve driving safety in conditionally automated driving.