Theory Of Dynamic Tracking Fuzzy Neural Network Control System

Study about the dynamic system and the fuzzy neural network steady-state control based on dynamic system is a very important branch in the intelligent control field. For the steady state control problem in dynamic systems, we characterized the control process by fuzzy neural network, established the dynamic tracking fuzzy dual neural network model based on dynamic system. Against the dynamic changes exist in control process, we also proposed the learning algorithm and supervised switching mechanism based on dynamic tracking fuzzy dual neural network model. This work has great significance on the theory and applications of intelligent control, is the forefront research in the intelligent control field.This dissertation discussed the theoretical foundation and main approaches of intelligent modeling and control based on fuzzy logic and artificial neural networks. Taking advantage of the learning algorithms and other intelligent computation methods, the dissertation tries to lay its emphasis on the development of theory of dynamic tracking fuzzy neural network control system.The major objective of this dissertation is to examine the neural network structure, to investigate techniques to combine both fuzzy logic systems and neural networks into an integrated system, to present a neural network methodology and a strategy of control, and to propose a learning algorithm base on knowledge acquisition and using this proposed DTFN network to complete an adaptive fuzzy neural control system for dynamic system. The dissertation is divided into four major parts, the theory of artificial neural network and fuzzy logic system (Chapter 2), the dynamic system and new fuzzy neural network architecture, the DTFN methodology and strategy of identification problems, and the DTFN control scheme (Chapter 3 & 4), the DTFN learning algorithm as automation knowledge acquisition approach (Chapter 5), the DTFN control applications in computer program (Chapter 6).The major innovation points of this dissertation are:(1) For steady-state control problem in the dynamic system, characterized by fuzzy neural network in control process, established the model of fuzzy tracking dual neural network in the dynamic system. Proposed the structure and training method of new fuzzy neural network architecture, based on the dynamic tracking fuzzy dual neural network (DTFN) system. (Section 3.4-3.5)(2) Established the steady-state controller based on the model of fuzzy tracking dual neural network. Combined with control model of the fuzzy neural network and learning model, proposed the framework of dynamic tracking fuzzy dual neural network (DTFN) system. (Section 4.2-4.3)(3) For the dynamic change exist in the control process, proposed the learning algorithm and the mechanisms for supervisory and switching process, based on the model of fuzzy tracking dual neural network. Proposed the learning algorithm based on the knowledge acquisition of the dynamic tracking fuzzy dual neural network (DTFN) system. (Section 4.4,5.4) In this dissertation, the idea of fuzzy dual neural network is applied to the study of dynamic system deeply. The research methods are benefit to enrich the ideas of intelligent control system, also helps to advance the basic theory research of neural network. The research results are benefit to provide a better theory for the steady state control problem in dynamic system, as well as to provide a theoretical framework for dynamic tracking dual fuzzy neural network model, which can be a kind of reference to ideas of realization the intelligent control.