DRIVER DROWSINESS DETECTION USING PHYSIOLOGICAL SIGNALS

Driver drowsiness is considered to be a very critical issue causing many fatal
accidents, injuries and property damages. Therefore it has been an area of intensive
research during recent years. In this study, a driver drowsiness detection system is
proposed, that integrates Electrocardiogram (ECG) and Electroencephalogram (EEG)
features to detect drowsiness. The proposed system is based upon the idea to observe
the physiological variations that occurs from alert to drowsy state. For physiological
data acquisition, a driving simulator based study was conducted in a monotonous
driving environment to get the driver drowsy naturally. Twenty two healthy university
students voluntarily participated in the experiment after strict inclusion criteria.
Features extracted from the EEG signal were absolute power, relative power, and
coherence in alert and drowsy state.