A Novel Approach for Real-Time Drowsiness Detection and Alert to Driver

Abstract

Drowsiness is the situation just before sleep. It could be due to lack of sleep, continuous long working hours, time of day and physical and mental state. It limits the ability to concentrate while driving. It leads to additional symptoms, such as forgetfulness or falling asleep at inappropriate times. Drowsiness at the wheel has been a severe problem which can be controlled using drowsiness detection system. This project is based on drowsiness detection using behavioral measures which include physical traits of human body such as facial expressions and head movement. This paper proposes an efficient and accurate system to detect driver’s drowsiness by using three effective algorithms simultaneously, i.e., eye blinking, PERCLOS and head tilt techniques. A final triggering variable is obtained as resultant of contributing three algorithms. Each algorithm is set to produce the value of an intermediate variable up to a defined value, i.e., threshold value. These intermediate variables affect the value of final triggering variable to reach its threshold value. If the final variable reaches its threshold, it can be used to generate any type of alert.