基于神經網絡和模糊系統(tǒng)的非線性隨機控制.pdf

1、華中科技大學碩士學位論文基于神經網絡和模糊系統(tǒng)的非線性隨機控制姓名：孔明申請學位級別：碩士專業(yè)：控制理論與控制工程指導教師：王永驥20050510II華中科技大學碩士學位論文AbstractThetraditionalmodelbasedstochasticcontrolmethodneedsaprecisesystemmodelbutit’sdifficulttobuildupprecisemathematicalmodelofpla

2、ntswithhighlynonlinearityuncertainty.Therearemanytroublesdifficulttocontrolthiskindofnonlinearstochasticsystems.sowestudiedsomecontrolstrategybasedontheneuralwksfuzzysystems.Atfirstsomecontrolmethodsbasedontheplant’sneur

3、alwkfwardmodelinversemodelsuchasinversemodelcontrolnonlinearpredictivecontrolareimplementedusedinthesimulationofnonlinearstochasticsystemscontrol.Thenalinearmodelisextractedfromanonlinearneuralwkmodelwecanusetheminimumva

4、riancecontrolmethodpoleplacementmethodonthislinearmodeltorealizetheselftuningcontrollerfnonlinearsystem.Thisextendthelineardesignmethodtononlinearsystem.AkindofminimumvariancecontrolmethodbasedonnonlinearTakagiSugenofuzz

5、ymodelisproposedinthispaperthismethodextendtheminimumvariancecontrolfmlinearsystemtononlinearsystem.Fuzzycontrollerhavestrongdisturbancerejectionabilitybutitsrulesmemberfunctionsarebasedontheexpert’sexperienceANFISnotonl

6、yhavethelearningabilityofneuralwksbutalsohavetheknowledgerepresentationabilityoffuzzyinferencesystemssowesuggestuseANFIStolearntheexpert’sexperienceANFIScanautomaticallygeneratethefuzzyrulesmemberfunctionsrealizetheslefl

7、earningcontroller.ThispaperimplementedthesecontrolmethodsbasedonMatlabNNSYSIDfuzzylogicothertoolboxthesimulationresultsshowtheefficiencyofthesemethods.Keywds:neuralwkfuzzysystemnonlinearstochasticcontrolTSfuzzymodelANFIS