A review on the measures and techniques adapted for the detection of driver drowsiness

Abstract

Driver drowsiness is one of the recent reasons behind accidents that can cause serious death, injury, and economic loss. Driving for long hours, sleepiness, medication, sleep disorders, and health conditions can cause drowsiness in a driver. It is of social concern since many lives including the passengers, drivers and wayfarers are at high risk due to drowsy driving. Detection of drowsiness and alarming the drivers can prevent a large number of accidents and thus the precious life can be saved. Input parameters like heart rate, pulse rate, vehicle steering movement, lane change, head movement, yawning, eye-closure can help to detect the drowsiness in advance. In the past, much research has been carried out to design an efficient driver drowsiness detection system using various measures to determine the drowsiness of the driver. In this paper firstly, we have reviewed the measures attempted by many researchers which are grouped under physiological, vehicle-based, and behavioral-based measures. Secondly, a detailed review of the deep learning approaches used is carried out along with the accuracy level achieved by each author. This detailed review will give a better insight for the young researchers to carry out prospective research in the specific field.