Abstract
One of the expectations towards the automated driving technology is dramatic reduction of road
crashes. The background hypothesis is that replacing the human driver by a computer will eliminate crashes that
are currently caused by human errors. However, complication of the systems and new tasks imposed on the driver
may generate new risks, system-induced problems. The main focus of human factors research in automated driving
is to reduce the system-induced problems. The research project was conducted from FY2016 to FY2018 with
the funding awarded by Cabinet Office and SIP-adus Phase 1. One of the three tasks of the project aimed at
understanding effects of cognitive states of the driver on his/her takeover performance and extracting metrics of
the influential driver states for driver monitoring systems. It was found that different driver states influenced
takeover performance in different ways. Some metrics for the influencing states were identified. These findings
are being discussed and applied to International Standards.
