Nn Method Of Fuzzy Control

In recent years, with the increase of complexity of control objects and the improvement of the demand of the control effect, the both situations engender the sharp conflict. The increase of complexity of control objects means it is more and more difficult to construct the precise math model of controlled objects. Classical control theory and Modern control theory just depend on the maths model of controlled objects. In addition, if the maths model of controlled objects is changed by outside or itself factors after adjusting all sorts of parameters of control systems, the whole parameters perhaps could not match the control requirement.The three methods of Intelligent control are fuzzy control, neural network, expert control. The biggest characteristics is it does not depend on the maths model of controlled objects which could solve the referred sharp conflicts. Expert knowledge is virtually limit for the more and more complex control systems. Therefore, the status of fuzzy control and neural network control become increasingly importance. Fuzzy control and neural network control have the merits and flaws themselves. Combining them could reach the mutually complementary effect.The importance of the thesis is the simulation of fuzzy and neural network control systems based on Matlab and the method discussion of fuzzy and neural network control systems based on DSP and the author chooses the TMS320LC/LF2407 as the discussional carrier of realized method of DSP.In the process of simulation, this thesis mainly focuses on the collection and treatment of training samples, the realization of the two kinds learning arithmetic, the control effects of phase step inputs and phase step inputs with interfere, the control effects of the changes of controlled object model and the realization of fuzzy neural network online self-adaptation. The realization of DSP is focused on the A/D conversion, D/A conversion and the compile and debug of each subprogram, finally, fully realize the fuzzy and neural network control systems simulation and the programs compiled of fuzzy and neural network control systems in DSP. The results of simulation shows that fuzzy and neural network control systems owns the better robustness and the control effect also could reach the requirement.