Effects of Visual Planning Support on Take-over Performance in Conditional Automated Driving

Abstract

L3 automated driving system allows the driver to leave from driving tasks within its Operational Design Domain (ODD), while the driver has to take over control of the vehicle once the system reaches the boundary of its functions. It will be a challenge for the Human-Machine Interface (HMI) to issue a Take-over Request (TOR) to get the driver back in the driving loop. This study investigated the effects of different types of visual planning support on take-over performance via a driving simulator experiment with 24 participants. We evaluated three types of visual planning support: Normal, Text, and Icon during emergency scenarios. Moreover, two Non-Driving Related Tasks (NRDTs) were hypothesized to affect the driver's performance. Based on a total of 8 objective metrics, the results showed that the type of visual planning support did not influence the reaction time; however, it had considerable effects on take-over performance. Compared with normal type which only included vocal alarm, the success rate under visual planning support was higher and the lateral control of vehicle was also more stable, which indicated an improvement of take-over performance.