Robust Integral of Sign of Error and Neural Network Control for Servo System with Continuous Friction

In this paper, a novel robust controller is proposed for servo mechanisms with nonlinear friction and external disturbance. First, a continuously differentiable friction model is used to represent the nonlinear friction, and neural network (NN) is employed to approximate the nonlinear friction and external disturabance. Then, a novel robust controller is designed by using robust integral of the sign of the error (RISE) term. In order to reduce the measure noise, a desired compensation method is utilized in controller design, in which the model compensation term depends on the reference signal only. The stability of closed-loop is proved based on Lyapunov stability theory, and all signal are proved to be bounded simultaneously. Finally, comparative simulations based on a turnable servo system are implemented to validate the efficacy of the proposed method.