基于RBF網(wǎng)絡的傳感器校正和補償系統(tǒng)研究與實現(xiàn).pdf

1、南京航空航天大學碩士學位論文基于RBF網(wǎng)絡的傳感器校正和補償系統(tǒng)研究與實現(xiàn)姓名：劉月清申請學位級別：碩士專業(yè)：計算機應用技術指導教師：章勇20060301基于RBF網(wǎng)絡的傳感器校正和補償系統(tǒng)研究與實現(xiàn)iiAbstractSensingtechnologyisthefoundationofinfmationscience.Senstechnologyisoneoftheimptantfootstonesofthemoderninfmati

2、ontechnologywhichhasalmostpermeatedthrougheverycnerofsciencetechnologynationaleconomy.Senswkisameasurementcontrolsystemlinkedbymountsofsensnodescommunicatedthroughcertainwiredwirelessprotocols.Indertoimprovetheprecisiono

3、fthesensstheintelligentintensityinthewirelesssenswksystemthispaperappliesRBFneuralwktothecalibrationtemperaturecompensationofpiezesistancesenstoimplementtwofunctions.Thefirstoneistoobtainthenonlinearacteristics.Thesecond

4、oneontheotherhistocalibratethenonlinearacteristicsofthepiezesistancesensfthepurposeofgreatlyreducingthesenswirelesssenswkontheinfluenceoftheambienttemperaturethepowerfluctuation.OnthebasisofstudyingthedeterminedmethodofR

5、BFcenterfirstthispaperconcentratesonanalyzingthedynamicnearestneighbalgithmsummarizesitsdeficiencyprovidesanimproveddynamicadaptivenearestneighbalgithm.BasedonsuchalgithmIDARBFneuralwkisconstructed.Seconditsgeneralizatio

6、nabilitythroughtheconcretefunctionapproximationexampleisanalyzedamethodfobtainingbettergeneralizationabilitythroughdynamicGausswidthsearchingispresented.Thesimulatedresultsindicatethattheimprovednearestneighbalgithmhasbe

7、tterapproachingabilitycanobtainbetterwkperfmance.ThirdthisIDARBFneuralwkisappliedtopiezesistancesens.Thecalibrationtemperaturecompensationfunctionareimplementedthroughtwoneuralwksmodeling.Thedirectmodelingtechnologyenabl

8、esanestimateofthenonlinearsensacteristicswhereastheinversemodelingtechnologyestimatestheappliedpressure.FourthaftercomparingthepressurefusionperfmancewiththeiginaladaptedneuralwkSLFMneuralwkthesimulatedresultsshowtheIDAR