A New Fuzzy Inference System And The State Prediction Inverse Control

Fuzzy logic and neural network have been widely applied in many fields. Fuzzy neural network, having strong learning and expressional ability, turns out to be more flexible than the conventional methods when modeling object and acting as a controller in the field of control engineering.In chapter 1-2, I review the relevant knowledge during the last three years. The content of chapter 3~6 is the results through my study. This thesis provides sections as follows:Chapter 1-2: Introduction to the ABC of fuzzy logic and neural network.Chapter 3: On the basis of chapter 1, I present some new conceptions including distance function, distance membership function and fuzzy set of variable vector. Then new fuzzy inference systems are constructed: Nmamdani FIS and NTS FIS.Chapter 4: Concentrating on the discussion of the general method when using NTS FIS to model object and the deduction of some algorithms when identifying NTS FIS. Besides, I also bring forward a kind of variable vector distance membership function梘auss distance membership function. For the convenience of using the distance membership function, I also get the general formula of partial derivative according to the parameter. Chapter 5: Using NTS FIS to model the cart-pole system and compare the result when using ANFIS.Chapter 6: Putting forward a new kind of control method-state predictive inverse control and using it to control cart-pole system.