Predictive Position Control for Precision Motion Systems Using Intelligent Prediction Model With Nonlinear Disturbance

In this article, a predictive position control method based on an intelligent prediction model with nonlinear disturbance is proposed to improve the position tracking performance of precision motion systems. The intelligent prediction model is constructed employing an optimized neural network structure. This model takes the motor state, control, and disturbance sequences as inputs, producing predictive position sequences as outputs. The disturbance sequence related to the reference speed sequence is initially unknown and requires determination. To enhance the model accuracy and the practical applicability of control applications, the model structure is optimized into a linear form with nonlinear disturbances, improving its practical applicability for controlling precision motion systems. The unknown model parameters are determined through a designed algorithm using the backpropagation method and experimental data. Subsequently, the intelligent prediction model is utilized to develop a predictive position controller. Moreover, an explicitly analytical control law is derived to achieve high-precision and robust position tracking while reducing energy consumption to the greatest extend. The developed controller comprises state feedback control, feedforward control, and disturbance feedforward compensation, leading to a more streamlined and compact control configuration. Finally, the effectiveness of the proposed method is validated via the comprehensive experiment.