一類(lèi)非均勻采樣數(shù)據(jù)系統(tǒng)的遞推最小二乘類(lèi)辨識(shí).pdf

1、江南大學(xué)碩士學(xué)位論文一類(lèi)非均勻采樣數(shù)據(jù)系統(tǒng)的遞推最小二乘類(lèi)辨識(shí)姓名：劉艷君申請(qǐng)學(xué)位級(jí)別：碩士專(zhuān)業(yè)：控制理論與控制工程指導(dǎo)教師：丁鋒20090501AbstractAbstractNonuniformlysampleddatasystemsexistwidelyinmodernindustryAccordingtothesamplingtheoryperiodicallynonuniformsamplingismoreflexibleth

2、anuniformsamplingInchemicalprocesscontrol，samplingdataforsomevariablesmaybenonuniformbecausethelimitsofhardwaressuchassensorsetcwhichresultsinthenon—uniformlysampleddata(multirate)systemsForexample，somevariableswhichcann

3、otbemeasuredonlineareoftenobtainedbylaboratoryanalysis，whichisusuallynonuniformintimeTherefore，itissignificanttostudytheidentificationproblemsofnon—uniformlysampledsystemsbothintheoryandapplicationsInthisthesis，therecurs

4、iveleastsquarestypeidentificationmethodsforaclassofnon—uniformlysampled—datasystemsarederivedbasedontheleastsquaresprincipleThemainworkofthisthesiscontainfollows：1Fortheperiodicallynonuniformupdatinganduniformsamplingsch

5、emeandfordifferentnoisedisturbances，theidentificationmodelsforaclassofnon—uniformlysampleddatasystemsarederivedbyusingthemodeltransformationmethods，andaredividedintotwocategories：equationerrortypenonuniformlysampledsyste

6、msandoutputerrortypenonuniformlysampledsystems2Referringtothederivationoftherecursiveleastsquaresalgorithm，definingdifferentinformationvectorandparametervector，andusingtheideaofreplacingtheunmeasurablenoisetermswiththeir

7、estimates，therecursiveleastsquaresalgorithmfornonuniformlysampledARXsystems，therecursiveextendedleastsquaresalgorithmfornonuniformlysampledARMAXsystemsandtherecursivegeneralizedextendedleastsquaresalgorithmfornonuniforml

8、ysampledARARMAXsystemsarederivedSimulationresultsindicatethatthealgorithmsareeffectiveConvergencepropertiesoftheformertwoalgorithmsarestudjed3Fornonuniformlysampledoutputerrorsystems，twoidentificationmethodsarepresentedo

9、neisarecursive1eastsquaresidentificationalgorithmbasedonthebiascompensationtechniquefortheunbiasedestimatesandtheotherisanattxiliarymodelbasedrecursiveleastsquaresalgorithmusingtheauxiliarymodelideaConvergencerateandesti

10、mationaccuracyoftheproposedalgorithmsareverifiedbysinmlationexamplesandtheconvergencepropertiesoftheauxiliarymodelbasedrecursiveleastsquaresalgorithmisstudiedbyusingthemartingaleconvergencetheorem4FOrnon—uniformlysampled