Adaptive Recurrent Wavelet Neural Network Integral Backstepping Control of Permanent Magnet Linear Synchronous Motor Drive

Due to a lot of uncertainties such as parameter variations, external load disturbance and friction force, the control performance of the permanent magnet linear synchronous motor (PMLSM) drive system has seriously influenced. These uncertainties are difficult to establish accurate models for the nonlinear uncertainty and timevarying characteristics in the actual PMLSM drive system. An adaptive recurrent wavelet neural network (RWNN) integral backstepping control system is developed to raise robustness for PMLSM drive system in this paper. The proposed control strategy combined integral backstepping control with adaptive RWNN uncertainty observer. According to Lyapunov function, an adaptive rule of the RWNN uncertainty observer is employed to on-line adjust the weights of a mother wavelet Gaussian functions by using the gradient descent method and the backpropagation algorithm. The effectiveness of the proposed control scheme is verified by experimental results.