基于觀測器的立方體系統(tǒng)控制研究.pdf

1、南京理工大學(xué)碩士學(xué)位論文基于觀測器的立方體系統(tǒng)控制研究姓名：邱振彬申請學(xué)位級(jí)別：碩士專業(yè)：控制理論與控制工程指導(dǎo)教師：吳曉蓓20090608Abstract碩士論文AbstractThecubesystemisasimpleexperimentalapparatusButitiscomplicatedasacontrolledplantItismultivariable，nonlinearstrong—couplingandstruct

2、urallyinstableSothecubecallbestabilizedonlybyadoptingeffectivecontrolmethodThisapparatusCanbeusedtostudytheperformanceandcontrolmethodofthesystems，whicharenonlinearmultivariableandstrongcouplingThecubeusedinthispaperispr

3、ovidedbythecompanyofQuanserincalladaInthispaperthemathematicalmodelofthecubeisestablishedwithLagrangeequation，andthenitislinearizedThecubeisprovedtobecontrollableandobservablebyanalyzingthemathematicalmodelThentheLQRcont

4、rollawisobtainedaccordingtothelinearmodelBecauseoneofthestatevariablesCan’tbemeasured，areduced—orderobserverisdesignedInthispaperacontrolmethodcombiningRBFneuralnetworkswithgeneticalgorithmsisadoptedtocontrolthecubesyste

5、mGeneticalgorithmsareusedtooptimizetheparametersofRBFNNcontrollerThentheweightsofRBFNNareadjustedonlinewithRLSalgorithmtoenhancetherobustnessofthesystemTheRBFNNcontrollerisconstrainedbytheconditionofthelinearcontrollerde

6、isgnSOthattherearelessparameterstobeoptimizedBoththeLQRcontrollerandtheRBFNNcontrollerareusedtocontrolthecubeThesimulationiscompletedusingMATLAB／SimulinkAndtheexperimentsaredoneonthecubewithWinCon，whichisareal—timecontro

7、lsoftwaredevelopedbyQuancerTheresultsshowthattheobserversandthecontrollershavegoodperformanceInthispaperthedesignoftheLipschitznonlinearstateobserverisdiscussed，accordingtothecharacteristicsofthecubeAreduced—orderobserve

8、rforLipschitznonlinearsystemsisstudiedindetailsItisbasedonthelinearmatrixinequalityThenitisusedtodesignthereducedorderobserverforthecubesystemKeywords：cube，reduced—orderobserver，neuralnetwork(RBFNN)，geneticalgorithm，IlLQ