Support vector machine fuzzy self-learning control with self-adaptive chaotic optimal learning algorithm for induction machines

In this paper, because the induction machines (IM) are described as the plants of highly nonlinear and parameters time-varying, to obtain excellent control performances of IM and overcome the shortcomings of the fast modified variable metric optimal learning algorithm (MDFP) and back propagation (BP) learning algorithm of neural network, such as requiring derivation in the process of learning and system identification, using a self-adaptive chaotic optimal learning algorithm (SAC), a support vector machine fuzzy self-learning control strategy for IM is presented based on the rotor field oriented motion model of IM. The fuzzy self-learning controller incorporated into the support vector machine fuzzy inference system (SVM-FIS) and a support vector machine identifier (SVMI) for IM adjustable speed system are designed. Simulation results show that the proposed control strategy is of the feasibility, correctness and effectiveness.