Intelligent Fuzzy Control And Its Software Development Platform

Fuzzy control is an important branch of intelllgent contfol theory. Itprovides a simple and convenient way to merge the control exPerience andreasoning process of human into automatic control strategy.The imperfect static precision is a weakness of classical fuzzy contrOl.This paper discusses three improving schemes: the compound controlmethod of fuzzy control and integration control, the fUzzy control withon-line parameter self-correction, and the non-quantized fuzzy control. Allof these three schemes can improve the static performance of forzy contrOlsystem notably. The fuzzy control method with on-line parameterself-correction can also improve the system's dynamic performanceeffectively, and the non-quantized fUzzy control method with analyticexpression is convenient for the adaptive adjustment of the fUzzy rules.A self optimizing control method based on the Smith predictioncontrol structure is presented for the control problem of time lag, non-fiXedand nonlinear system. It combines the fuzzy control with artificial neuralnetWorks. It is showed by simulation results that this method is good fortime lag, nonlinear time-variant plants.In order to reduce the programming complexity of practical fuzzycontrol systems, tWo MATLAB application program interface ways tocapitalize on the resources of MATLAB and its fuzzy logic toolbox in Cprogramming environment is discussed. One is the MAW enginefUnction library, by which C language starts a MATLAB course running inthe background. Another is the stand-alone C-code fuzzy inference engine,by which C language establlshes a fuzzy inference system completelybased on the C. Both of these can reduce the programming loads ofpractical fUzzy control system obviously. The former is powerfUl andconvenient, but it cannot mn without MATLAl3 softWare system. Thelatter can be aPplied into C language environment independently andembedded in other external applications...