A human factors perspective on how to keep SAE Level 3 conditional automated driving safe

The Society of Automotive Engineers (SAE) defines Conditional Automated Driving (CAD) or SAE level 3 as the next step in the transition to higher automated driving. A human factors challenge at this level is to keep the fallback-ready user aware of the driving situation. The problem with the widely accepted terminology of SAE is that it does not consider the human factors of achieving this safety-critical transition and lacks in defining a required or appropriate state of fallback readiness and how a human reaches this state. This paper aims to understand how research and industry currently perceive the problem and how they counteract it. We conducted exploratory expert interviews with N = 15 subject matter experts from universities, research centres and vehicle manufacturers. Based on the findings of the expert interviews, we discuss how fallback-ready users might be kept in the loop. The results show that the user should remain physically and sensory in the state of regular drivers to allow a safe transition, which narrows down the problem to maintaining an appropriate cognitive state, level of arousal, and motivational conditions. Lastly, the experts identify that the NDRA should be (a) limited to the onboard-entertainment system and (b) context-aware and (temporarily) restricted or interrupted. We provide concrete design recommendations grounded in (or drawing upon) theoretical frameworks and current human factors expert views and highlight specific challenges that require more research.