基于支持向量機(jī)的傳感器非線性校正及應(yīng)用.pdf

1、上海大學(xué)碩士學(xué)位論文基于支持向量機(jī)的傳感器非線性校正及應(yīng)用姓名：李靜申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：檢測(cè)技術(shù)與自動(dòng)化裝置指導(dǎo)教師：付敬奇王靈20090201上海大學(xué)碩士學(xué)位論文ABSTRACTSensortechnologyistheSOUIC七ofinformationtechnologyandoneofthethreebasictechniques，howeverlimitedbytheperformanceofthesensoritsel

2、ftheoutputofsaisoralwaysbenonlinearAndinmostcases，becauseoftheimpactionofelectromagneticwaves，noiseandoutsideenvironment，thenonlinearityofSenSOrSbecomemorecomplex，whichseriouslyimpactonmeasurementaccuracyandcontrolaccura

3、cyTherefore，thestudyofnon—linearcalibrationofsensorshasaveryimportantapplicationvalueandpracticalsignificanceInthispaperSensorsnonlinearcorrectionbasedonSupportVectorMachine(SVM)werestudied，whichtakethedatafromCapacitive

4、differentialpressuretransmitterthermocouplethermalresistanceandinclinationsensorasresearchsamplesThemaintaskWasreflectedasfollows：1IntroducedtheSVMtheory，includingclassificationandregression；DiscussedseveralSVMlearningal

5、gorithm，especiallySequentialMinimalOptimization(SMO)；andintroducedtheoptimizationofSVMparameters一一P硎cleSwarmOptimization(PSO)detailed2DevelopedtheSupportVectorRegressionmachineSensorsnonlinearcorrectionmodel，whichtakethe

6、sinsensitiveErrorFunctionasthelosefunctionIntroducedthestepsdetailed，includeddatapreprocessingsamplessetpartitionandkernelfunctionchosen；Duetothistypeofgeneralizationperformanceiseasytobeimpactedbyparametersselection，Par

7、ticleSwarmOptimizationarithmeticWasusedtoselectparameters3￡SupportVectorRegressionmachineSensorsnonlinearcorrectionmodelWasappliedincapacitivedifferentialpressuretransmitterthermocouple，thermalresistance，andinclinationse