Parameter-dependent actuator fault estimation for vehicle active suspension systems based on RBFNN

Abstract

The actuator fault estimation (FE) problem is addressed in this study for the quarter-car active suspension system (ASS) with consideration of the sprung mass variation. Firstly, the ASS is modeled as a parameter-dependent system with actuator fault and external disturbance input. Then, a parameter-dependent FE observer is designed by using the radial basis function neural network (RBFNN) to approximate the actuator fault. In addition, the design conditions are turned into a linear matrix inequality (LMI) problem which can be easily solved with the aid of LMI toolbox. Finally, simulation and comparison results are given to show the accuracy and rapidity of the proposed FE method, as well as good adaptability against the sprung mass variation. Moreover, a simple FE-based active fault-tolerant control (AFTC) strategy is provided to further demonstrate the effectiveness and applicability of the proposed FE method.