A Model to Determine Sleepiness of Drivers

Abstract

Drowsiness detection has been grabbing the attention of many areas, which involves critical working environments during the past few decades. For example traffic accidents have been increased significantly because of the drivers who falls into sleep at the wheels. Nightshift workers special security personals were unable to pay complete attention to their work because of the influence of drowsiness. This shows the necessity of drowsiness detection and warning system to make a promising contribution to active working and safety. On the way to a reliable monitoring device, valuable steps have been taken towards the development of detection algorithms from driving behavior and eyelid parameters. However, individual differences in human features and performance has been a difficult challenge to face on this way to an optimal and effective solution.

The major goal of this project work is to develop a model that determines the sleepiness of the driver/human operator based on a video stream of his/her eyes. Drowsiness will be predicted by monitoring the eyelid movements of drivers/operator. Results were expected to produce in real-time. Hence, solution was expected to consist of simple processing and analyzing mechanisms as much as possible while retaining the significant level of performance. To demonstrate the feasibility of the approach taken a prototype has been implemented.