Study on Control System of Permanent Magnet Synchronous Motor with Fixed Active Disturbance Rejection by RBF Neural Network

In the position control of permanent magnet synchronous motor (PMSM) servo system, active disturbance rejection controller (ADRC) has strong robustness and anti-interference ability compared with traditional PID controller. In this paper, a method based on radial basis neural network (RBF) to adjust the parameters of the state observer (ESO) is proposed to realize the real-time estimation and accurate compensation of the internal and external disturbances in the PMSM system. The simulation results show that the control precision, response speed and anti-disturbance ability of the ADRC controller with RBF on-line setting are improved compared with the conventional setting.