Track: Sensors and Sensing

Abstract

Driver impairment is yet another human factor that is likely to cause dangerous driving behavior (aka distracted driving), leading to safety hazards on the roads. Assuring acceptable drivers' cognitive state can improve safe driving performance. Behavioral and self-report measures are used to assess drivers' cognitive states, but psychophysiological measures can add more considerable value in the real-world driving context. Hence, monitoring, detecting, and responding to drivers' psychophysiological state compose a safety countermeasure that can reduce the risk of distracted driving and thus eliminate or mitigate safety hazards. The research investigates the recent used measures to detect drivers' psychophysiological state, in addition to the drivers' psychophysiological state classification using machine learning. Eye movement, respiration rate, electrocardiogram, electroencephalography, and electromyography are all examples of measures used to deduce drivers' overall psychophysiological state. Tracking and assessing these measures rely on several different metrics, methods, and technologies. For example, drowsiness, as a well-known psychophysiological state, can be assessed and tracked by the eye aspect ratio metric. This paper reviews the recent state-of-the-art measures used for detecting the drivers' overall psychophysiological state. The review considers different psychophysiological states such as fatigue, drowsiness, distraction, and stress. The paper also discusses the strengths and limitations of each reviewed measure, its usage, and its real-time applications. The result showed that the most common used measures are eyes and mouth state measures. Also, several machine learning methods have proven their reliability in classifying drivers' psychophysiological states, especially SVM and CNN as they have wider applications.