Evaluation of a Smart Algorithm for Commercial Vehicle Driver Drowsiness Detection

Abstract

Drowsiness is a safety hazard in commercial vehicle driving. The conditions to which truck drivers are exposed put them at higher risk as compared to passenger car drivers. Unobtrusive drowsiness detection methods can avoid catastrophic crashes by warning or assisting the drivers. This paper describes an experimental analysis of commercially licensed drivers who were subjected to drowsiness conditions in a truck driving simulator and evaluates the performance of a neural network based algorithm which monitors only the drivers' steering input. Correlations are found between the change in steering and the state of drowsiness. The results show steering signals differences can be used effectively for detection.