Abstract
Drowsy driving is one of the causes of serious accident. In order to prevent such accidents, it is necessary to detect not only drowsy driving but also driver's increasing drowsiness at an early stage. In precious researches, increasing drowsiness was detected by using driver's bio-signals. However, acquiring these bio-signals on an actual vehicle is difficult because driver needs to wear the sensor and camera needs to be set up for capturing driver's face. Accordingly, this research aims at early detection of driver's drowsiness by using vehicle operation data.We examined the relation between driver's drowsiness and vehicle operation data obtained from the driving simulator. Subjects operated on the driving simulator until they drowse. Drowsiness level of each subject was evaluated based on the facial expression. From the driving simulator, we obtained vehicle operation data such as acceleration and brake pedal operation, speed and acceleration of vehicle, and lateral movement distance from the traffic lane center. We verified correlation between driver's drowsiness and vehicle operation data. As a result, we specified operation data that have relation with driver's drowsiness.
