Variable Structure Control Based on the Fuzzy Neural Networks

In this paper, the fuzzy neural network is applied in the sliding mode control. The fuzzy neural network is of the fuzzy logic reasoning ability and the neural network learning ability. The physics meaning of the fuzzy neural network is clear and intelligible. According to the Sugeno reasoning method and the connecting thought, the fuzzy neural network is designed. For the sliding mode control characteristic, the construction of whole control system is designed by means of combining the fuzzy neural network applied in control and the fuzzy neural network applied in identification. The learning algorithm for neural network is also given. The chattering is efficaciously eliminated, because the control signal may be smoothed by the fuzzy neural network sliding mode controller. And its robustness is stronger than that of the common sliding mode variable structure controller. At the same time, the bounded of system uncertainty and disturbance needn´t to be known in designing of the fuzzy neural network sliding mode controller. The stability of system is investigated by constructing Lyapunov function. Theorem of system global stability is summarized