The Optimization Design Of Fuzzy Neural Network Controller

Fuzzy neural network (FNN) is an active branch in the intelligent control. It is composed of the neural network and fuzzy logic system organically. And it is good at self-learning and self-tuning. So the theory of FNN is very important for the intelligent control. But there are still some problems in it. In this dissertation, the following theories and application of FNN are discussed.1. The large account of compute work of updating weights and long training time usually discourage the FNN' s on-line application in industry. Moreover, when it is trained on-line to adapt to plant variations, the over-tuned may cause system oscillate extensively. In this dissertation, two kinds of optimization, methods are proposed. Firstly, only these linking weights corresponding to the control rules that affect the control performance significantly are updated in order to reduce the compute works and speed up the training progress. Secondly, the updating step is adjusted adaptively in accordance with the error and the change of error of the system based on the T-S model to get better performance. Simulation results show that the training time is reduced greatly and the convergence velocity is speed up.2. It is difficult to design a traditional controller for some nonlinearsystems, as some parts of the system are unknown. In this dissertation, the FNN and the traditional controller are combined to design two kinds of FNN adaptive controllers for a class of nonlinear plant. They are stable indirect adaptive controller based on fuzzy basic function network and stable direct adaptive controller based on T-S fuzzy neural network. The FNN in the system is used to approximate the unknown parts of the system. Then according to the Lyapunov's stability theory, the adaptive laws of parameters are designed. The parameters of network are adjusted on line and the system satisfies the Lyapunov's stability. The simulation results show that these two kinds of adaptive controllers can realize tracing the input signal commendably and satisfy the request of stability simultaneously.