Predicting reduced driver alertness on monotonous highways

Abstract

Impaired driver alertness increases the likelihood of a driver making mistakes and reacting too late to unexpected events. This is a particular concern on monotonous roads, where a drivers attention can decrease rapidly. Although effective countermeasures dont currently exist, the development of in-vehicle sensors opens avenues for monitoring driving behavior in real time. The aim of this study is to predict driver alertness levels using surrogate measures collected from in-vehicle sensors. Electroencephalographic activity is used as a reference to evaluate alertness. Based on a sample of 25 drivers, the authors collected data in a driving simulator instrumented with an eye-tracking system, a heart-rate monitor, and an electrodermal activity device. They tested various classification models, from linear regressions to Bayesians and data mining techniques. Results indicate that neural networks were the most efficient model for detecting lapses in alertness. Findings also show that reduced alertness can be predicted up to five minutes in advance with 90 percent accuracy using surrogate measures such as time to line crossing, blink frequency, and skin conductance level. Such a method could be used to warn drivers of their alertness levels through the development of an in-vehicle device that monitors, in real time, driver behavior on highways.