Few attempts have been made to use physical and physiological driver characteristics to predict driver drowsiness. As a result, a reliable drowsy driver detection system has yet to be devised. Thus, the primary objectives of this research were to determine whether driving characteristics and response variables could be used to detect eyelid closure associated with drowsiness, and. to provide ‘potential measures of driver· drowsiness. In. the study, eyelid closure was defined as the measurement standard of drowsiness. Eyelid closure, in studies conducted at Duke University, was a reliable measure of drowsiness.

A computer simulated nighttime driving task introduced 90 minutes of typical highway driving to twenty driver/subjects seated ixx a moving-base driving simulator. Each driver/subject drove under two conditions--rested and after 19 hours of being awake. During the 90 minutes of driving, two types of low-level stimuli, steering wheel torque and front wheel displacement, were applied to the simulation. Responses to these stimuli as well as driving I measures from the intervals between stimuli were analyzed for variations associated with eyelid closure. Seventeen dependent variables were investigated.