RBF-neural-network-based sliding mode controller of automotive Steer-by-Wire systems

This study proposes a robust steering controller for Steer-by-Wire systems using neural network. The proposed control consists of a nominal control and a nonsingular terminal sliding mode compensator where a radial basis function neural network (RBFNN) is utilized to adaptively learn the uncertainty bound in the Lyapunov sense and thus the uncertainty effects are effectively eliminated. Using the proposed neural controller, not only the robust steering performance against parameter variations and road disturbances is obtained, but also both the control gain and the control design complexity are greatly reduced due to the use of the RBFNN. Simulation results are demonstrated to validate the superior control performance of the proposed control as compared with other controllers.