Adaptive and robust controls with application to vehicle traction control

A comprehensive adaptive vehicle traction control strategy, which includes anti-skid braking and anti-spin acceleration, has been proposed, implemented and tested. It integrates a robust tracking control algorithm with a tire/road surface identification process, and it results in good performance regardless of vehicle environments. The design of this control strategy is based on the modelling and the analysis of vehicle dynamics and control methodologies, followed by system developments and validations.; The robust tracking control algorithm proposed in this thesis is a discrete-time single input/single output tracking control scheme, which is robust against system uncertainties and disturbances. This scheme adopts "derivative feedback" to compensate the "nominal" plant uncertainties, and the synthesis of an external input to overcome the variation of these uncertainties during one sampling period. The scheme is superior to the sliding model control when the system uncertainties change only slightly during any single sampling period yet may accumulate over a certain period of time. For a system with such uncertainties, the chattering resulting from the new scheme can be significantly reduced compared with that from the sliding mode control. This merit has been illustrated in the case of vehicle traction control.; A new approach to the adaptive optimal control for certain nonlinear systems has also been suggested in the thesis. With such an approach, first, a certain performance index is formulated from the system output. The optimal operating point for the output is estimated based on this index. Since the system is nonlinear, the initial estimate usually does not agree with the true optimal point. However, if the estimate provides the direction of the true optimal point relative to the present output, the system output can be steered in this direction. As the system output approaches toward the true optimal point, the index itself is adaptively updated and the estimated optimal point converges to the true one. This approach is realized by the proposed adaptive vehicle traction control strategy, which employs a weighted least squared estimation and an adjustment algorithm to estimate the optimal operating slip as well as the discrete-time nonlinear control/sliding mode control to track the estimated optimal slip.