Adaptive backstepping control for induction motor based on neural networks and dynamic surface technique

This paper focuses on adaptive control problem of induction motors using neural networks (NNs). By introducing the dynamic surface technique that reduces the complexity of the controller designed by usual backstepping methods, a new adaptive backstepping design method is presented using radial basis function (RDF) NNs as nonlinear approximators. In addition, a new approach is proposed to avoid the controller singularity problem. Using Lyapunov stability analysis, the uniformly ultimate boundedness of closed-loop adaptive systems is proven. A simple simulation is included to illustrate the effectiveness of the proposed approach.