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

Abstract

The increasing number of traffic accidents caused by drowsy driving has drawn much attention for detecting driver’s status and alarming drowsy driving. Existing research indicates that the changes in the physiological characteristics can reflect fatigue status, particularly brain activities. Nowadays, the research on brain science has made significant progress, such as the analysis of EEG signal to provide technical supports for real world applications. In this paper, we analyze drivers’ EEG data sets based on the self-adjusting Dynamic Time Dependency (DTD) method for detecting drowsy driving. The proposed model, i.e. SEGAPA, incorporates the time window moving method and cluster probability distribution for detecting drivers’ status. The preliminary experimental results indicates the efficiency of the proposed method.