Artificial Neural Network-Based Controller for Permanent Magnet Synchronous Motor Servo System

This paper implement an online training of dynamic neural networks (NNs) for identification and control of permanent magnet synchronous motor (PMSM) servo system. Utilizing two multilayer feed-forward NNs, it makes no such assumptions. The two networks work in tandem to simultaneously achieve system identification and adaptive control. The proposed control system is designed and its effectiveness in tracking application is verified by simulations. The ability of the controller to achieve the tracking process with a high degree of accuracy, even in the presence of external disturbance is also demonstrated. The simulation results clearly demonstrate the success of the proposed control structure.