多通道不完全量測下的估計算法研究.pdf

1、南京理工大學碩士學位論文多通道不完全量測下的估計算法研究姓名：高璽璟申請學位級別：碩士專業(yè)：控制理論與控制工程指導教師：戚國慶20100620Abstract碩士論文AbstractIntargettrackingarea，therewidelyexistsincompletemeasurementInpracticalapplications，measurementsareoftenmulti—channelTherefore，the

2、studyonestimationalgorithminthecaseofmultichannelincompletemeasurementisofgreatsignificanceForthenonlinearintargettrackingsystemandmulti—channelincompletemeasurement，thispaperfocusesontheestimationalgorithminthecaseofmul

3、ti—channelincompletemeasurement，whichcontainstwokindsofsituationsOnesituationiSthatthemulti—channelincompleteiScoupled，andtheothersituationisthatthemulti—channelincompleteisnotrelatedThecriticaldetectionprobabilityisalso

4、studiedinthispaperThemainresearchresultsareasfollows：First，themulti—channelincompletemeasurementtrackingsystemismodeledBasedonthepreviousresearchandaccordingtothepracticalapplications，wedesignedtwokindsofmodelsOnekindisa

5、multichannelincompletemeasurementwithcoupleddetection(MIMCD)；theotherisamulti—channelincompletemeasurementwithdecoupleddetection(MIMDD)Next，westudytheestimationalgorithmbasedonMIMCDUsingtheprojectiontheoremwederivedUCMKF

6、basedonMIMCDBysolvingmodifiedRiccatidifferenceequationwegettherobustfiltergainwhichmakesthefilterresultsuniformstabilityBysimulationmethodthesetwomethodsarecomparedandanalyzed，thesimulationresultsshowthattherobustfilterg

7、ethigherprecisionandUCMKFgetbetterreal—timeThen，wealsostudytheestimationalgorithmbasedonMIMDDWiththehelpoftheprojectiontheoremwederivedtheEKFalgorithmbasedonMIMDDInthisbasis，AccordingtothestandardKalmanfilterwedesignedth

8、erobustfilterbasedonMIMDDWedesimaedtherobustfilteringalgorithmbasedonMIMDDSimulationresultsshowthattherobustfilterhasbetterrobustnessandhigherprecisionEKFmaybeunstableinthelowprobabilityofdetection，andtheelTorvarianceisl

9、argeButEKFrunmuchfasterthanrobustfilterFinallycriticaldetectionprobabilityisstudiedThispaperprovedtheexistenceofthecriticaldetectionprobabilityanditalsogivesalgorithmstocalculatethecriticaldetectionprobabilityKeywords：Mu