Advances in the development of a virtual car driver

Fundamental approaches to modeling the control of a car by a driver are reviewed briefly. The context of the work presented is explained. Then previous research on the application of optimal linear preview control to aspects of driving road vehicles is extended. This prior research treated the tracking of a roadway by a vehicle and driver at constant forward speed and the tracking of a speed demand while running straight. The two previously separate problems are now combined, so that longitudinal and lateral path demands are considered in parallel. A new feature is that low-pass filters are included in the driver modeling to represent driver bandwidth limitations. This feature enables the finding of the influence of the driver’s control bandwidth on the optimal strategies and on the closed-loop system performance, by way of frequency-response calculations. A new optimal preview control toolbox is employed. Simulations of the virtual driver-controlled car are shown to demonstrate the closed-loop system following longitudinal and lateral position demands.