Testing the Effectiveness of Objective and Subjective Predictors of Driving Behavior under Fatigue

Abstract

There is growing evidence that driver fatigue is a major road safety problem, causing crashes that frequently involve fatalities and severe injuries. Professional drivers are among the road users indicated by literature to be at high risk of involvement in a fatigue-related crash. The goal of this paper is to investigate the effectiveness of objective and subjective data collection and analysis in predicting driving behavior under fatigue. Toward this goal, the impact of a fatigue-training program addressing professional drivers was assessed through a naturalistic study on data collected by trip recorders (Geographical Positioning System devices). Analysis of these data was supplemented with the collection of self-reported data obtained through a questionnaire survey to investigate potential threats, deficiencies, and bias of the approaches. Findings revealed that there is some correlation between the two approaches measuring driving behavior under fatigue. Focusing on self-reported data, it was indicated that the training program affected positively the direction of change in the behavior of the professional drivers, addressed by an increase in the proportion of drivers who stop and rest when tired. The naturalistic approach and the testing of potential differences before and after the training program with reference to speed and average stop time showed that the training program encouraged drivers to reduce their speed and increase the time that they stop and rest.

Extraction of Optimal Measurements for Drowsy Driving Detection considering Driver Fingerprinting Differences

“…e research in the field experiments found that the drowsy driving state generally lasted 15-75 seconds, while the typical drowsy operation characteristics duration was generally 5-20 seconds [30]. When measurements are calculated by data in the entire drowsy state duration, the calculating average effect covers the drowsy operating characteristics [1,18]. us, the doublelayer sliding time window was proposed to divide raw data into samples and calculate measurements.…”

“…Drowsy driving is a typically dangerous driving behavior, which seriously threatens traffic safety and causes substantial financial costs to individuals and society [1][2][3]. Drowsy driving is a common phenomenon that results from multiple underlying reasons, including prolonged driving hours and sleep deprivation [3,4].…”