Tim Johns

MathWorks Consultant

Consultant at MathWorks. See also LinkedIn.

PhD Thesis: The Effects of Cognitive Workload on a Racing Driver's Steering and Speed Control

In Formula One racing, lap simulations are used to optimise vehicle design to achieve the minimum possible lap time, with such simulation schemes assuming a ‘perfect’ driver to close the loop with the vehicle. A real racing driver is then used in driverin-the-loop simulations to test how a given design performs. There is therefore a gap between the perfect driver of the pure simulation work, and the real driver of the driverin-the-loop simulations. This gap would benefit from being bridged with a simulation that uses a driver model more representative of a real driver.

This work examines the effect that cognitive workload has on a driver, hypothesising that there is a trade-off between a driver’s cognitive workload and the lap time that they can achieve. The aim of this research is to improve the mathematical tools available to predict the dynamic interaction between drivers and vehicles, with particular attention given to the cognitive process of a racing driver controlling a vehicle at the limit of adhesion.

A series of experiments are conducted that seek to measure the changes in a driver’s lateral and longitudinal control with cognitive workload. It is found that lap time increases as a driver is made to put more effort into an additional task to that of driving, primarily due to reducing the speed of the vehicle whilst maintaining a constant lateral deviation about the mean path. This contrasts an existing hypothesis in the literature. Whilst literature exists characterising a real human operator, the ideas have not been incorporated into a control model of a driver for use in lap simulation schemes. Compensatory steering controllers are therefore designed that include a representation of a driver’s finite cognitive capacity through the idea of intermittent control for a straight line, constant speed driving task, and experiments are conducted to validate the control models. An intermittent controller that utilises serial ballistic sequences of controls represents the experimental data best. This differs from established practice of only using the first control of a sequence.

The compensatory controller is incorporated into a circuit driving environment where speed is allowed to vary. It is found that an increase in lap time with increasing cognitive workload can be represented by increasing the length of the intermittency period of the controller, and then slowing the vehicle such that the maximum variation in the lateral displacement of the vehicle about the racing line remains constant.

Download full thesis [204 MB]