Augmented reality versus classical HUD to take over from automated driving: An aid to smooth reactions and to anticipate maneuvers

摘要

The takeover phase from highly automated systems, when the transition phase is short (here 10s), is critical for the driver when he has been out of the driving loop. We conducted a simulator study on 26 drivers to compare three conditions: manual mode (MD), transition from automated driving (AD) to MD with a classical head up display (HUD), transition from (AD) to MD with an augmented reality (AR) HUD. In our scenarios, the driver has to take over to respect the navigation request for a lane change, and has to face slowing traffic in the destination lane. Compared to MD, driving behavior is altered in the transition phase up to 30s: the maximum speed on the brake pedal is higher, resulting acceleration being thus increased. AR reduces both these effects, thus improving driving comfort while facing slow traffic. AR also helps better anticipate the lane change maneuver: it increases the distance to the maneuver limit point. As such, the situational awareness seems to be improved by AR. Further analysis of eye tracking data and verbatim is needed to confirm this assumption based on driving data.