Development of Warning Methods for Planned and Unplanned Takeover Requests in a Simulated Automated Driving Vehicle

Abstract

In several circumstances, a level-three automated vehicle cannot continue driving in an automated driving mode and requests a human driver to take over. In this study, a series of experiments to examine how to provide a TOR was conducted. First, for forty-one persons, a HUD icon, earcon, seat vibration, and combinations were compared. The results indicated that the HUD icon-earcon and HUD icon-seat vibration were the most effective. Second, the combinations of A-pillar LED light and cluster icon (visual), earcon and speech message (auditory), and presence/absence of seat vibration (haptic) were compared. Thirty-six volunteers participated in the ADS failure and forty in the highway exit experiment. In the ADS failure, the combination of A-pillar LED light and seat vibration (AH) reduced the RT but can induce stress. In the highway exit, a speech message is recommended due to control stability, and the AH is not recommended due to longitudinal instability.