Drowsy Driver Detection Using Discriminant Analysis

Publisher Website
Google Scholar
Abstract
Driver drowsiness represents a serious hazard, and methods need to be developed to detect and counteract its effects. In this research, 14 dependent driving variables were investigated for their potential use in predicting driver drowsiness and, in particular to this research, eyelid closure. A computer-controlled automobile simulator was used to simulate a nighttime highway driving scenario for 20 drivers in both a rested and a partially sleep-deprived condition. Included in the simulation approximately every minute was one of two types of driving stimuli, torque or displacement. The responses of the drivers to these stimuli and the general driving characteristics of the drivers between stimuli were recorded. These data were subsequently analyzed using linear discriminant analyses. The discriminant analyses indicated that a number of dependent variables contributed to linear discriminant functions, which classified “alert” and “drowsy” observations with relatively low false-alarm and miss rates.

This publication has 2 references indexed in Scilit: