EEG-Based Driver Drowsiness Detection Using the Dynamic Time Dependency Method

The increasing number of traffic accidents caused by drowsy drivinghas drawn much attention for detecting driver’s status and alarming drowsydriving. Existing research indicates that the changes in the physiological char?acteristics can reflect fatigue status, particularly brain activities. Nowadays, theresearch on brain science has made significant progress, such as the analysis ofEEG signal to provide technical supports for real world applications. In thispaper, we analyze drivers’ EEG data sets based on the self-adjusting DynamicTime Dependency (DTD) method for detecting drowsy driving. The proposedmodel, i.e. SEGAPA, incorporates the time window moving method and clusterprobability distribution for detecting drivers’ status. The preliminary experi?mental results indicates the efficiency of the proposed method.