Driver Drowsiness Detection Using Subsidiary Task and Driving Performance Measures

Six male and six female subjects operated a moving-base automobile simulator for 2 1/2 hours after having been kept continuously awake for approximately 19 hours past their normal wake-up times. During this session, subjects drove both while performing three subsidiary tasks (auditory, visual, and tactual) which required a tactual output but which otherwise differed only in terms of stimulus input modality, and while performing no task. The driving performance, behavioral, physiological, and subsidiary task response measures collected during the experiment were subjected to a series of linear discriminant analyses, and the results indicated that the detection of driver impairment due to drowsiness may indeed be possible by monitoring combinations of subsidiary task and physiological measures.