Assessment of Driver Fatigue
Between 10 and 30 % of road traffic accidents are attributed to driver fatigue. There is an increasing awareness of the considerable risks due to driver fatigue when driving at night outside built-up areas. But nodding off at the wheel, the so-called micro-sleep, can also occur during day-time. And long before micro-sleep episodes occur driving performance, information processing and the ability of vehicle guidance notably suffer. Today, there are a number of different methods for assessing driver drowsiness available: The analysis of lid closure behavior, physiological measures such as EEG-, ECG- and EDA-analyses, video-based expert ratings of driver behavior or analyses of steering behavior. Since some new car models have been equipped with driver drowsiness warning systems as standard, some of the above-mentioned methods are already applied in daily traffic.
These various methods of assessing driver drowsiness show different characteristics: They differ in their reliability and validity of the drowsiness assessment, in their sensitivity and their ability for the differentiation of beginning, medium and severe drowsiness, and in their susceptibility to interfering person- and environment-related influences. Drowsiness warning systems additionally require the definition, recognition and communication of the appropriate warning threshold.
Commissioned by the Federal Highway Research Institute HFC carries out an extensive review of methods and systems for drowsiness detection. This is complemented by expert interviews (Delphi-Study) in research and development and with manufacturers and users of drowsiness detection and warning systems. The goal of the project is to evaluate existing methods and systems of driver drowsiness assessment, to describe their suitability for different use cases and fields of applications and to identify further research needs.