A study of effects of driver’s sleepiness on driver’s subsidiary behaviors

Traffic accidents caused by inattentive driving, including drowsiness, still occur frequently. Thus, as one of the approaches to prevent traffic accidents, developing a method for the early detection of low-arousal driving is critical. Our goal is to develop a driver status monitoring system for detecting a decrease in driver arousal by using small and low-cost wearable devices, such as wrist-worn accelerometers. As a basis for achieving this goal, in this paper, we designed and conducted an experiment to measure the frequency of subsidiary behaviors, including yawning, swaying of the head, and arm and hand activities associated with changes in sleepiness level. The correlation coefficient between the frequency of subsidiary behavior and sleepiness level obtained from the facial expression evaluation was then examined. In two of the four participants, the frequency of subsidiary behaviors increased as the sleepiness level increased, with correlation coefficients of 0.786 and 0.601. These results suggest the possibility of detecting a change in the driver’s sleepiness level from the frequency of subsidiary behaviors, including arm and hand activities.