A Comparative Study of Driver Drowsiness Detection Techniques

The number of motor vehicles has increased steadily in emerging countries during the last ten years. Traffic accident reports indicate that risky driving behaviours, such as drunken driving or fatigued driving, are responsible for most accidents. According to studies, sleepy driving is a contributing factor in 20% of all accidents. Drowsiness is a condition when the level of consciousness is lowered as a result of fatigue or sleep deprivation, and it can make a driver fall asleep silently. Drowsy driving results in a loss of control by the driver, which can cause the vehicle to drift off the road, hit an obstacle, or overturn. In order to extract and synthesise the methods and features that have been utilized in drowsiness detection process, this study has conducted a Systematic Literature Review (SLR) for this work and found 100 relevant studies using the defined search parameters, of which 30 studies are chosen for additional research. This study has thoroughly examined these works, evaluated the methods and features applied, and identified the research gap as well. According to the comparative analysis, the most used features are facial expressions like yawning, closing of the eyes, and head motions. These insights are based on an analysis of 30 papers and show that the Eye Aspect Ratio, Haar Cascade, Support Vector Machine (SVM), and dlib library are frequently utilised techniques.