ROBUST LATERAL VEHICLE CONTROL USING QUANTITATIVE FEEDBACK THEORY (QFT)

This paper presents a new robust controller for autonomous lateral control of a highway vehicle. The controller is designed using Quantitative Feedback Theory (QFT), which is a frequency based control design methodology developed to explicitly address plant parameter uncertainties, as well as external disturbances and sensor noise. In a practical implementation, there will be significant variations in the vehicle dynamic parameters for the various vehicle types to be controlled on the highway. For example, variations in inertia, cornering behavior, and vehicle speed will all lead to significant changes in the dynamic response of a vehicle to a steering input. Traditional controllers, such as PID control, will not successfully provide robust optimal performance over the entire range of dynamic variation, and in some cases, cannot even guarantee controller stability. Clearly, such performance degradation and instability is not acceptable in a deployable system, and must be addressed during the controller design. Here, the authors present a QFT-based controller design which meets given performance specifications in spite of a wide variation in the vehicle dynamic parameters. Simulation results for curve tracking are presented for this controller for a roadway vehicle subject to a wide range of parametric uncertainty, illustrating the inherent robustness of the QFT design. The results illustrate that the QFT-based controller achieves specified performance in the presence of significant parametric uncertainty, and remains stable over the entire range of variation.