Driver fatigue is a contributing factor in an estimated 10-20% of road accidents. A huge number of different methods is available to assess driver fatigue. The project’s goal was to describe and compare the strengths and weaknesses of different methods to assess driver fatigue and to present an overview of the existing fatigue detection and fatigue warning systems. The evaluation of these methods and systems is based on literature analysis and interviews with experts and users. The following procedures were taken: • Literature analysis and overview of physiological and performance-based methods to assess driver fatigue. • Literature analysis of fatigue detection and warning systems, interviews with their manufacturers and overview of the identified systems. • Interviews with 20 users of currently in middle-and upper-class cars implemented fatigue warning systems to assess the perceived quality of fatigue detection, driver acceptance and compliance. • Development of a list of criteria to assess and compare the quality of the most valid methods to assess driver fatigue. • Two-stage survey (Delphi-Study) with 12 experts from academic and industrial research areas. It was aimed at the selection and evaluation of the most valid methods to assess driver fatigue, their subsequent re-evaluation providing explanatory statements in case of contradicting opinions, and the assessment of the methods’ suitability for different fields of application. • Workshop with the experts that participated in the Delphi-Study to discuss and complement the obtained results. 70 fatigue detection systems were identified. They differ in their underlying methods to assess fatigue, their availability and design of driver feedback, and in their prevalence. Convincing evidence of their validity or the percentage of false and missed alarms are rarely reported. The fatigue warning systems that are more and more implemented in middle- and upper-class cars often fail to provide a satisfying fatigue detection performance from a users’ point of view. Several drivers continue to drive even despite of a system’s warning that they perceive as accurate. About half of them perceive fatigue warning systems as a gain in safety. The most valid methods to assess driver fatigue selected by the 12 experts are the assessment of driving performance (steering behaviour and lane keeping) and eyelid closure, video-based expert assessments, EEG and (quite controversial) the Pupillographic Sleepiness Test. Their differing strengths and weaknesses are extensively described in the report at hand. There is still no gold standard of fatigue assessment. Depending on the fields of application all six of the selected methods (research & development), only a few (fatigue warning systems) or none of them (traffic controls) are applicable. When choosing a method the goal and context of fatigue assessment and the strengths and weaknesses of the respective methods should be considered. The combination of at least two different methods is recommended to achieve a valid assessment of driver fatigue. Fatigue detection systems and the most promising methods to assess driver fatigue should be further optimized and accompanying studies need to be conducted. Further measures to effectively reduce the number of accidents were discussed during the workshop and presented in the report at hand. (Author/publisher)
Samenvatting