Mathematical Modeling of Driver Speed Control With Individual Differences

The quantitative prediction and understanding of a driver's speed control is an essential component in preventing speeding and designing of vehicle systems. Driver speed control is a complex behavior of longitudinal vehicle control consisting of speed perception, decision making, motor control, vehicle dynamics modeling, and individual driver differences. However, there are few existing models that can integrate all of these aspects in a cohesive manner. To address this problem, this paper introduces a mathematical model for a driver's speed control with analytical solutions based on human cognitive mechanisms in driving. This model includes an integrated queuing network-model human processor structure and the rule-based decision field theory. This new model consequently can predict several aspects of driver speed control behavior at the same time, such as driving speed, throttle/brake pedal angle, acceleration, and the frequency of speedometer inspection. A laboratory session involving a driving simulator is conducted to validate the current model. The model accounted for over 99% of the experimental speed of the average driver, and over 95% of the experimental speed for the majority of individual drivers.