Fuzzy Control System Of Structured Analysis

In recent years, although intelligent control theory including fuzzy control, neural network control, and knowledge based expert control has provided an efficient approach to realize the modeling and control of complex systems with the uncertainties(structure uncertainty and parameter uncertainty), time-variant and nonlinear, multi-variable etc., at present we lack of efficient tools to design and analysis intelligent control systems. In this dissertation, enlightened by the traditional control theory (classical control theory and modem control theory) and neural network techniques, an extensive research on the systematic design theory, an important direction of intelligent control, has been performed. Having been developed for over thirty years, fuzzy control has achieved a lot of research progress in theory and applications, but it has become evident that many basic issues remain to be further addressed. Systematic design is among the most important issue for fuzzy control systems. Because linguistic variable and fuzzy inference based fuzzy control is essentially nonlinear control, it is difficult to construct a universal methodology for systematic design. Moreover, fuzzy control is mostly applied to the plants without their exact mathematical models, so it is even more difficult to design fuzzy control system. In order to cope with these basic problems, the systematic design methodology of fuzzy control systems is deeply discussed in this thesis by using traditional control theory and neural network technique. The discussing contents emphasis on such aspects as parameter design of fuzzy controller, essential research of TS fuzzy model and parameter identification of TS model by using the combination of fuzzy control and neural network. The whole thesis includes 6 chapters, and the main contents and conclusions are summarized as follows: In chapter 1, we review the basic theory and the basic method of intelligent control and we remark three important directions of intelligent control including fuzzy control, neural network technique and knowledge based expert control. We specially describe fuzzy control, and detailedly introduce the merit and flaw of fuzzy III control and the class of fuzzy control. In chapter 2, we put forward new method for the modeling and control of TS model system, specially research on its identification. In chapter 3, because the traditional TS fuzzy controller has too many parameters to be identified, the thesis introduces a kind of simplified TS fuzzy controller, so the number of parameter will be decreased. We essentially research on it, If it adopts three input variables, we can know that it is a kind of nonlinear parameter-variable PID controller, whose parameters?characteristics can be influenced by the parameters of the fuzzy model. If we reasonably choose the parameters of TS fuzzy model, the controller will have superior characteristics. In chapter 4, we inference the analytical formulate of the typical fuzzy controller having two inputs and one output and specially analysis the circs in which the input variables of the fuzzy controller adopt five fuzzy numbers. On the basis of these, we put forward a king of systematic design method for fuzzy control system. We use the old design method of P1/PD to design the parameters of the fuzzy controller. In chapter 5,we introduce a kind of hybrid neural network algorithm on the basis...