有关电信方面的论文.docx

有关电信方面的论文.docx

《有关电信方面的论文.docx》由会员分享，可在线阅读，更多相关《有关电信方面的论文.docx（10页珍藏版）》请在冰点文库上搜索。

有关电信方面的论文

TheKalmanfilterbasedresearchofT-wavealternansdetectionalgorithm

Abstract:

Asthesecondlargestcauseofdeathinhumanbeingsonlybehindcancer,suddencardiacdeath（SCD-SuddenCardiacDeath）isfeaturedassuddenonsetanddifficulttorescue.Therefore,forthesepatients,earlydiagnosis,earlyinterventionisthemosteffectivetreatment.TWA（Twavealternans）asthedetectionofSCD-likeillnessisanimportantindicatorofhow（non-invasive）toobtainaccuratedatainaTWAresearchwhichisfocusedoninrecentyearsintheindustry.

ThispaperpresentsamodifiedKalmanfilter-based（non-steadystate）TWAdetectionalgorithm:

Firstofall,preprocessing,filteringnoiseandbaselinedriftofECG;methodusingwavelettransformmodulusextremumofQRSandTwaves,seevehiclefeatureslocationofthepoint;furtheralignmentandextractstheTwaveTwavematrix;accordingtothestrategy（inaccordancewiththeoddandeven）ontheTwavematrixre-grouping;KalmanfilterwereusedontwogroupsofTwavematrix;calculatethedifferencebetweentwogroupsofTwavematrix,thenthevectormodulusmaximatoidentifythevectorthroughvectormaximamodules,andmakesthemeanvectorasthetestingstandards.Throughsimulationresults，thisalgorithmforthecorrelationcoefficientofTWAdetectionvaluesandrealvaluesreaches97%,andnotonlycanittesttherateofTwavealternans,butalsocanidentifyshort-termTwavealternans.

[keyword]:

TWAKalmanextremumofwavelettransformmodulus

SuddenCardiacDeath

[Introduction]:

Suddencardiacdeath（SCD-SuddenCardiacDeath）isadeathcausedbyheartattack.Accordingtotheagency'sstatistics,suddencardiacdeathhasbecomethesecondlargestcauseofdeathaftercancer.Thereareabout45.000casesinUnitedStates,andover1.8millioncasesperyearinChinaAdditionally,theincreasinglyseriousenvironmentalpollution,thepressurecontinuedtoincrease,thisfigureisconstantlyincreasingbecauseoftheincreasinglyseriousenvironmentalpollutionandthecontinuedpressure.Astheincidencesuddencardiacdeathhavethecharacterofsuddenonsetanddifficulttorescueoncethediseasehappens,thereislittletimeleftformedicinepersonnel,thusthesurvivalrateisverylow.Therefore,themaintreatmentforsuddencardiacdeathistheearlypredictionandintervention-based.Clinically,therearetwomaintreatmentsforthepotentialriskforpeoplewithSCD:

invasiveandnon-invasivemethod.Invasivemethodistousetheinstigationofelectrophysiologicalmethods（Electrophysiology,EP）todetectthesizeofSCDoccurrence,wherebytheuseofdrugsanddefibrillatorsfortreatment.Non-invasiveisamethod,whichincludesQTdispersion,averagedelectrocardiogramandotherpredictionmethods.Invasivemethod,hasdamageonthehumanbodytosomeextent,highcost,whileECGQTdispersionandtheaccuracyoftheaverageforecastisnothigh.ThespecificphenomenonofTWA（TWaveAltenate）Twavealternantis:

repeatedappearanceofTwavemorphologyandamplitude.Alargenumberofclinicalstudiesfoundthat:

thesumofmalignantarrhythmiaSCDTWAoftenoccurredsimultaneously.Therefore,Twavealternanshasbecamearesearchhotspotforitsnon-invasive,themostvaluableandefficientpredictionofelectrophysiologicalindicatorsofSCD.

Currently,therearetwomainclassificationsaboutTWAdetectionalgorithm:

oneisthetimedomainandfrequencydomainbasedmethodandtheotheristheevolutionofalgorithm-basedclassification.Timedomainincludes:

complexdemodulation（CD）,modifiedmovingaveragemethod（MMA）,correlationanalysismethod（CM）.Frequency-domainalgorithmsinclude:

Adamproposedbytheenergyspectrumanalysis（ESM.EnergySpectralMethod）,Smithmadeonthisbasis,thespectralanalysismethod（SM.SpectralMethod）,LagunafurtherproposedKL+spectralanalysis,time-frequencydomainanalysis,wavelettransform,etc.Therearethreeclassificationsthroughtheevolutionalgorithm:

BasedontheshorttimeFouriertransform,basedonnon-linear,basedonsymbolictransformation.ShorttimeFouriertransformbasedalgorithms,including:

complexdemodulation（CD）,cyclemethod（PT）,statisticaltestingmethod（ST）,andextensiveclinicalapplicationofspectralanalysis.AlgorithmbasedonnonlinearTwavealternansusinganonlinearcharacteristic,usingthenonlinearitytotestparameterTwavealternans.Thereareseveralmethodsasfollows:

modifiedmovingaverage（MMA）,theLaplacelikelihoodratiomethod（LLR）,Poincareplotmethod（PM）andsoon.Transformalgorithmbasedonsymbolictime-domainsignalthroughthesigntransformTwavealternansofthecriteriainclude:

correlationanalysismethod（CM）,Karhunen-loevetransform（KLT）,andstatisticaldetectionofRayleighdetection.

Althoughthespectralanalysismethodwiththeinputdatamaintainsthecharactersofrequirementsforlownoiseandrespiratorymodulationofanti-interferenceabilityoftheadvantages,itrequirestheinsurancethattheheartratetoremainatafixedfrequencyforsometimewhenitworks,anditissensitivefortheheartrate.Atthesametimeitdoesnothavethetimeresolution,thedatacannotdetectnon-steady-statephenomenon.Thepremiseofcomplexdemodulationmethodistwofold:

First,theperiodis2timestheTWAheartrate,andsecond,TWAamplitudeandphasechangeswithtimecosinecurve.ComplexdemodulationmethodcantrackthedynamicinthetimedomainwaveformTWA,butitlacksofaccuratesensitivitytonoise,,andhasalargeamountofcomputation.CorrelationanalysismethodquantifiesthemagnitudeofACMTWAbycalculatingtheaverageTwaveandTwaveofthealternatingsingleindex（AlternatesCorrelationIndex,ACI）andalternatingthetwoindicators.ItcandetectnotonlytheamplitudechangesofT,butalsothelocationofTWAontheECGandduration.Howeverithashighrequirementsontheinputdata,buthaspoorperformanceagainstbaselinedrift.

Kalmanfilter，asoneofthecontemporarythreemostoptimizedtheories,employsthereal-timerecursivealgorithmtooptimallyestimatethesignalswhichincludeobservationnoiseandinterference,whichiswidelyusedinnon-stationaryandmulti-dimensionalrandomprocess.NotethatthemeaningandmethodofKalmanfilteriscompletelydifferentfromtheconventionalfilters,andtheoptimalestimateisitsessence.Thisarticle,takesthestrategyofmodulusmaximatoobtaintheamplitudeofTwavealternansafterdoingthefollowingactivities:

accesstheTwavematrixwhichwasdividedintotwogroupsaccordingtothepolicy,useKalmanfiltertocorrecttheimpactofnoiseanddriftlimits,updatetheTwavematrix,andreducetheimpactofthedegreeofalignmentofTwave,irregularshapeofTwave.Then,Twavealternanscanbeobtainedthroughtheapplicationofmaximumorderwindowedstrategies,meanwhile,itcaneffectivelydetectthephenomenonofshort-termTwaves.Also,frequencydomainmethodsorstatisticalmethodscanbeavailableindetectingTWA.

TWAdetectionandanalysisalgorithm

1.Kalmanfiltertheory

KalmanfilteringwasadvocatedbyKalman（R.E.Kalman）in1960.itisakindoffilteringalgorithmtoestimatetherequiredsignalfromtherelatedobservationswhichwereextractedfromthesignal.theconceptofstatespacewasintroducedtotherandomestimationtheory.Kalmanfilteralgorithmisformedbyemployingthestatisticalpropertiesofthesystemstateequation,theobservationequation,thesystemnoiseandtheobservationnoise.Kalmanfilterisatime-domainfiltering,usingthestatespaceapproachtodescribethesystem.Itsalgorithmistherecursiveform,anditisofasmalldatastorage,soitcannotonlydealwithstationarysignalsbutalsocanhandlemulti-dimensionalandnon-stationaryrandomprocess.Moreover,Kalmanfilteringisdifferentfromtheconventionalfilters,itsessenceisthemostoptimalestimationmethodanditisoneofthethreemostoptimizedtheories.Ishasbeenwidelyusedininertialnavigation,targettracking,communicationsandsignalprocess.ThesuccessofU.S.ApolloprogramisapracticeofthesuperiorityofKalmanfiltertheory.JustbecauseoftheseadvantagesoftheKalmanfilter,itcanfurtherfilteroutnoiseanddriftlimits,reducetheinfluenceontheTwavealignmentandirregularshapeinsertontheTWAdetection.Kalman滤波理论的公式：

Kalmanfiltertheoryformula:

Stateestimation:

Statestepprediction：

Sequenceofnewinformation：

Filtergainmatrix:

Steppredictionerrorvariancematrix：

Estimatingthevariancematrix:

Intheabove,

isthen-dimensionalpredictivevalue,

isthemeasuredvalue,

isthetransitionmatrixofn*n-dimensionalstate,Histhem*n-dimensionalobservationmatrix.

2.Theprocedureofcalculation

​​2.1ECGsignalpretreatment

Thefrequencyrangeofthenoisessuchasfrequencyinterference,EMGinterference,baselinedrift,motionartifactinECGsignalcanbedetermined,thecentralizedbandandtheemergencelocationoftheenergyinvariouspartsoftheECGalsohaveacertainregularity.Therefore,waveletheursurecanbeusedinthresholdingnoisereductionandthemedianfilteringmethodinbaselinedrift.

2.2ThedetectionofQRSWaveandTWave

Employathresholdstrategyofwaveletmodulusmaximaalgorithmtocalculatethepeakpoint,thestartingpointandtheendingpointoftheTwave.

2.3TwaveTwavematrixalignmentandaccesstoTwave

AstheconventionalalignmentmethodisinaccordancewiththeTwavestart,endalignment,buttheTwaveshapesvariably,isdifficulttoaccuratelydetectthefeaturepoints,thiswilldefinitelyaffecttheTwavealignment,therefore,inthisarticlethefollowingstrategyisused:

First,