Research on Nonlinear Control Strategy of Servo System using RBF Neural Network and Composite RISE Control

This paper presents a nonlinear control strategy that combines RBF neural network and composite robust integral of the sign of the error (RISE) control, which is applied to a two-axis servo system. Firstly, an electromechanical coupling dynamic model is established, considering the impact of nonlinear factors. Next, the RBF neural network is employed to approximate the unknown nonlinearities, with the online adjustment of neural network weights using an adaptive law derived from the Lyapunov method. Additionally, RISE is utilized to counteract the effects of uncertain nonlinearity and disturbance on the system. To address the issue of the “differential explosion” phenomenon, an improved dynamic surface control strategy is proposed. It replaces the traditional “first-order filter” with a “third-order filter”. Finally, through experimental comparisons, it is demonstrated that the proposed control scheme effectively mitigates the influence of nonlinearity and external disturbance on the system, leading to higher tracking accuracy.