高速高機動目標IMM跟蹤算法研究.pdf

1、南京理工大學(xué)碩士學(xué)位論文高速高機動目標IMM跟蹤算法研究姓名：陸晶瑩申請學(xué)位級別：碩士專業(yè)：導(dǎo)航、制導(dǎo)與控制指導(dǎo)教師：李銀伢20100620碩_：論文高速高機動目標IMM跟蹤算法研究AbstractThetargettrackingproblemisofinterestinbothmilitaryandcivilianapplicationsWiththerapiddevelopmentofthemodemaerospace，theve

2、locityandmaneBverabilityisbecomingincreasinglyhighInthestationofhighmaneuveringtargets，thereliableandexacttrackingisthekeyproblemTheInteractingMultipleModel(IMM)estimatorhasbeenshowntobeoneofthemosteffectiveschemesTheref

3、ore，thisstudyisofgreattheoreticalsignificanceandapplicationvalueFirsttheresearchanddevelopmentofthemodelsandtrackingalgorithmsisbrieflyreviewed，andtheirfeaturesandperformancearepresentedConsiderableresearchisundertakenon

4、thebasicIMMestimatorThetheoreticalratiocinationmathematicalexpressionselectionofthemodelsandparametersanditsapplicabilityarediscussedThesimulationresultsshowthat，comparedwithsinglemodel，multiplemodelalgorithmhaswidertrac

5、kingbandwidth，andCanbegenerallyapplicabletomaneuveringandnonmaneuveringtargetsThen，forhighspeedandhighmaneuveringconditions，thecurrentstatisticalmodelandconstantturningmodelareintroducedintoIMMalgorithmtoimprovethemodels

6、etsTheIMMalgorithmsusingcurrentstatisticalmodelorturningmodelarecomparedwithbasicIMMestimatorbyusingMonteCarlosimulationTheinfluenceofmaneuveringfrequency，maximumaccelerationandturningratetotrackingaccuracyisanalyzedThes

7、imulationresultsshowthatTheIMMalgorithmsusingcurrentmodelismoreeffectiveformostconditions，SOitismostlystudiedhereFinallybasedonIMMalgorithmusingcurrentmodel，someimprovementisdoneaimedatitsdisadvantagesFirst，byusingweight

8、edsquareresidualdetectiontodetectthetargetmaneuveraselfadaptivefunctionformaneuveringfrequencyisproposedtoformaselfadaptivealgorithmSecond，theVariableStructureMultiplemodel(VSMM)algorithmisstrudied，andaModeGroupSwitching

9、(MGS)methodisintroducedteducethenumberofmodelsThird，introducethefuzzyalgorithmTheaccelerationestimationisinputintoafuzzymoduletogetmaximumacceleration，basedonweightedsquareresidualdetectionMGSstructureFinallytheimprovedA