Complex and critical task handover in Automated Vehicles

This project explored the safety issues arising from the introduction of Automated Vehicle (AV) as the driver changes from an active controller to a passive supervisor, with associated risks of inattention, reduced situational awareness, and human errors. The project built on the ongoing Cooperative and Highly Automated Driving (CHAD) Safety Study, a Department of Transport and Main Roads (Queensland) project delivered in partnership with Motor Accident Insurance Commission, QUT and iMOVE Australia.

Existing research in this area, while plenty, is limited by a reliance on using driving simulators, or is performed by vehicle manufacturers but not released publicly. Only a few field studies have been conducted using participants in SAE Level 3 or 4 AVs, and to our knowledge, none of those studies allowed the participant to seat in the driver’s seat. The ongoing CHAD Safety Study used the first research L3 and L4 AV authorised to drive on public roads in Australia (ZOE2) to start covering this gap in the research.

This project capitalised on the strong technical and research baselines established by the CHAD to further investigate other dimensions of the handover task in a real AV. It improved our understanding of the drivers’ performance in complex handover conditions, which had remained largely unknown. In particular, the methodology of the CHAD trials was expanded to include other dimensions compared to the existing research:

• Duration of takeover phase
• Distractions prior to/during the takeover phase
• Interruptions prior to the takeover phase
• System transparency
• HMI modalities including visual, auditive, and haptic feedback
• Traffic density