The Research of Three-Layer Fuzzy Neural Network with Variable Structure Control System

Abstract

The three-layer fuzzy neural network with variable structure applies the overall variable structure algorithm. The learning ability of fuzzy neural network with variable structure helps to extract fuzzy rules from the data automatically. Fuzzy neural network with variable structure changes the structure in training, and it improves the learning speed and convergence performance of the neural network. The fuzzy neural network with variable structure achieves systematic design of the fuzzy system by changing structure in training. It avoids the randomness of the fuzzy system design. Variable structure algorithm applies to the control system can effectively improve the convergence speed of fuzzy neural network. Simulation results show that the fuzzy neural network with variable control system has the advantage of fast convergence in training fuzzy control rules. Mean square error and error rate converges rapidly than traditional fuzzy neural network.