基于EMD的時頻分析與濾波研究.pdf

1、電子科技大學(xué)碩士學(xué)位論文基于EMD的時頻分析與濾波研究姓名：羊初發(fā)申請學(xué)位級別：碩士專業(yè)：通信與信息系統(tǒng)指導(dǎo)教師：黃煒20090501ABSTRACTThctimefrequencyanalysismethods，whichareappliedinnonlinearandnonstationarysignalprocessingandanalysisquicklyandefficientlyhavebecomeahottopicofmo

2、dernsignalprocessingmethodsstudyfieldItgraduallyshowsthattherearemanyadvantagesofEMD(EmpiricalModeDecomposition)inthefieldofnon—linearandnonstationarysignalprocessingEMDdecomposesacomplexsignalintoafinitenumberofIMF(Intr

3、insicModeFunctions)，andEMDisahi曲frequencyresolutionandadaptivemethodInthispapertheentiretheoreticalsystemsofempiricalmodedecompositionhavebeenresearcheddeeplyfocusedonthefilteringmethodsbasedonEMI)andEMDapplicationoftime

4、frequencyanalysisFurthermoreanewfilteringmethod。_’HHTspectrumfilteringmethodbasedonwaveletthresholdfilteringpreprocessingisproposedThemaincontentsaleillustratedasfollows：(1)Theconceptofinstantaneousfrequencyisintroducedf

5、irstlyThenthebasicconceptsofIMFandprinciplesofEMDdecompositionalgorithmalediscussed，andthecharacteristicsoftheEMDalgorithmareanalyzedCombinedwithMF，itisgiventhatthephysicalmeaningoftheHilbertmarginalspectrumandtheHilbert

6、spectrumareformulated，andthetheoriesofthewholetimefrequencyanalysisalediscussedindetailBesides，theproblemofEMDtimefrequencymethodisalsoanalyzedincludingmodesmixingendeffectsiftingstopconditionandSOon(2)Comparedtotraditio

7、naltimefrequencyanalysismethodsandusedtoanalyzetimefrequencycharacteristicsofnumericalresultsofclassicalnonlinearequationsystemandsignalsingularitydetection，ThetheoreticanalysisandsimulationexperimentresultsshowthatEMDsp

8、ectrumhasperfecttime—frequencyconcentrationandhightimefrequencyresolutioncharacteristics(3)Thethreefilteringmethods：theEMDfiltering，waveletthresholdfilteringbasedonEMDandtheHilbertspectrumfilteringarediscussedfortheiradv