A New Approach for a Real-Time Non-Invasive Fatigue Assessment System for Automated Driving

Abstract

Vehicles with conditional automation will be introduced to the market in the next few years. However, the effect of fatigue as one component of the driver state on the take-over performance still needs to be quantified. To examine this question, a valid, real-time capable and preferably non-invasive method for assessing fatigue while driving automatically is required. For this purpose, we developed an objective driver fatigue assessment system based on the data of a commercial remote eye-tracking system. The fatigue assessment system fuses various metrics based on eyelid opening and head movement. In a validation study with 12 participants in a driving simulator, the fatigue assessment system achieved a sensitivity of 90.0 % and a specificity of 99.2 %. This approach makes a fatigue-state-dependent study design possible and can also provide a basis for advancing existing fatigue assessment systems in automated vehicles.