The Neural Networks Smooth Tracking Control For The Permanent Magnet Synchronous Motor

Permanent-magnet synchronous motor (PMSM) is a typically nonlinear, multi- variable and strongly coupled system. Servo control methods of PMSM are classified into linear PID control and the nonlinear control. Though the former strategy is simple to be implemented, it can't satisfy the requirement for higher dynamic performance. The latter one can obtain perfect dynamic performance and strong robustness of the system, but usually it's too complex to be realized.In the paper, the back-stepping control method of neural networks and the theory of the two-order nonlinear smooth trajectory filter are illustrated. Firstly, based on the mathematic model of PMSM, the motive system can be equivalent to the two-order system under the current linear closed loop control. Then, considering the excellent performance of the neural networks tracking the nonlinear function as well as the online weight-tuning scheme based on Lyapunov function, global asymptotic stability of the system is guaranteed. Finally, the nonlinear smooth trajectory tracking strategy is applied into the position servo control.In addition, we design a simple and effective load platform for the servo system, and obtain the load property through rigorous derivation.The experimental results demonstrate the effectiveness of the proposed method and the reliability of the load platform.