Adaptive fuzzy radial basis function neural network integral sliding mode tracking control for heavy vehicle electro-hydraulic power steering systems

Abstract

Due to parametric uncertainties, unknown nonlinearities, and dynamic external disturbances, it is a challenging and valuable task for heavy vehicle electro-hydraulic power steering systems to realize high-precision tracking control. To cope with this complex nonlinear tracking control problem, the integral sliding mode control is an extremely potential control method, which has strong robustness to model uncertainties and unknown disturbances, and can effectively reduce the steady-state error in tracking control process. However, the inherent chattering phenomenon of integral sliding mode control seriously affects its control performance. In order to suppress the chattering while ensuring robustness, adaptive fuzzy technique is adopted as an effective auxiliary means, which can not only deal with the inherent chattering problem of integral sliding mode control and a priori knowledge of the disturbance upper bound in controller design but also dynamically adjust the parameters in the fuzzy rules. Moreover, the designed adaptive fuzzy–integral sliding mode control scheme still needs the precise mathematical models of the control systems. But it is difficult to obtain the model for heavy vehicle electro-hydraulic power steering systems with highly complex and coupling properties. Therefore, to further improve the method, this paper presents a novel adaptive fuzzy–radial basis function neural network–integral sliding mode control method for the complex systems to achieve timely and accurate steering angle tracking control. In addition to the advantages of adaptive fuzzy–integral sliding mode control, the modified controller no longer requires the precise mathematical models of heavy vehicle electro-hydraulic power steering systems and realizes the continuous adaptive updating of weights. Finally, the effectiveness and superiority of the proposed control scheme is illustrated by comparisons and extensive simulations.