Development and validation of a driver distraction alert system for highway driving using a truck driving simulator

Abstract

Driver distraction is one of the major causes of vehicle accidents, injuries, fatalities and economical losses in Canada and across the world. The main objective of this thesis is to develop a real time driver distraction alert system that can generate early warning signals and prevent vehicle accidents. The proposed system is applied to a truck driving simulator by monitoring the driver response through the steering wheel and accelerator pedal during different driving experiments. In order to ensure an accurate prediction of the driver distraction, the simulation output of the truck driving simulator is compared to that of the commercial TruckSim® software using identical parameters of a given tractor semitrailer during similar driving conditions. Consequently, the driving speed and steering angle profiles for both simulation environments are then compared during standard test maneuvers and found to be in good agreement. Several distraction indicators are proposed in order to predict the driver distraction namely; the jerk profile, spikiness index and the rate of change for both the steering wheel angle and the accelerator pedal position. Several driving experiments are performed considering both the offline and real time assessment of a driver subjected to different types of distraction. The major finding of the thesis emphasizes the effectiveness of using steering wheel rate to estimate driver distraction. Moreover, in comparison to other techniques for predicting driver vigilance, the presented work requires low computational power and has a great potential for developing a real time simple system that can be affordable and reliable.