A Review of Driver State Monitoring Systems in the Context of Automated Driving

Abstract

Conditionally automated driving (CAD) will lead to a paradigm shift in the field of driver state monitoring systems. High underload and the possibility of engaging in non-driving related activities will greatly influence the driver state. Level 3 also requires drivers to act as a fallback level in a take-over situation. Drivers have to get back in the loop and regain control with possible challenges due to their state. Therefore, driver state assessment will gain importance in order to ensure a safe and comfortable hand-over. The purpose of this paper is to provide an overview of driver state models and monitoring systems in the context of automated driving. Based on three driver state models, we focus on the commonly used driver state constructs fatigue, attention and workload. As part of this review, different definitions are summarized and possible metrics to operationalize these constructs were identified and critically reviewed. When reviewing the literature, it became apparent that driver state and the different constructs lack a common definition. Overall, eye-tracking is the technology with the most potential, but it needs further development to increase reliability. EEG lacks practicability and subjective measures are prone to misjudgement and may counteract extreme levels of fatigue.