简介：Analgorithmofhighlymaneuveringtargettrackingisproposedtosolvetheproblemoflargetrackingerrorcausedbystrongmaneuver.Inthisalgorithm,anewestimator,namedasmulti-parameterfusionSinger(MF-Singer)modelisderivedbasedontheSingermodelandthefuzzyreasoningmethodbyusingradialaccelerationandvelocityofthetarget,andappliedtotheproblemofmaneuveringtargettrackinginstrongmaneuveringenvironmentandoperatingenvironment.ThetrackingperformanceoftheMF-Singermodelisevaluatedandcomparedwithothermanueveringtrackingmodels.ItisshownthattheMF-Singermodeloutperformsthesealgorithmsinseveralexamples.

简介：Theoptimalestimationperformanceoftargetparametersisstudied.First,thegeneralformofCramer-Raobound(CRB)forjointestimationoftargetlocationandvelocityisderivedforcoherentmultipleinputmultipleoutput(MIMO)radars.TogainsomeinsightintothebehavioroftheCRB,theCRBwithasetofgivenorthogonalwaveformsisstudiedasaspecificcase.Second,amaximumlikelihood(ML)estimationalgorithmisproposed.Themeansquareerror(MSE)oftheMLestimationoftargetlocationandvelocityisobtainedbyMonteCarlosimulationanditapproachesCRBinthehighsignal-to-noiseratio(SNR)region.

简介：AmultisensordistributedextendedKalmanfilteringalgorithmispresentedfornonlinearsystem,inwhichthedynamicequationofthesystemandtheequationsofsensor’smeasurementsarelinearizedintheglobalestimateandglobalpredictionrespectivelyandthesuboptimalglobalestimatebasedonallavailableinformationcanbereconstructedfromtheestimatescomputedbylocalsensorsbasedsolelyontheirownlocalinformationandtransmittedtothedatafusioncenter.Ananalysisofthepropertiesofthealgorithmpresentedhereshowsthattheglobalestimatehashigherprecisionthanthelocaloneandsmallerlinearizationerrorthantheexistingmethod.Finally,anapplicationofthealgorithmtoradar/IRtrackingofamaneuveringtargetisillustrated.Simulationresultsshowtheeffectivenessofthealgorithm.

简介：Themeanshifttrackerhasdifficultyintrackingfastmovingtargetsandsuffersfromtrackingerroraccumulationproblem.Toovercomethelimitationsofthemeanshiftmethod,anewapproachisproposedbyintegratingthemeanshiftalgorithmandframe-differencemethods.Theroughpositionofthemovingtargetisfirstlocatedbythedirectframe-differencealgorithmandthree-frame-differencealgorithmfortheimmobilecamerascenesandmobilecamerascenes,respectively.Then,themeanshiftalgorithmisusedtoachieveprecisetrackingofthetarget.Severaltrackingexperimentsshowthattheproposedmethodcaneffectivelytrackfirstmovingtargetsandovercomethetrackingerroraccumulationproblem.

简介：ThispaperproposesaPCAandKPCAself-fusionbasedMSTARSARautomatictargetrecognitionalgorithm.Thisalgorithmcombinesthelinearfeatureextractedfromprincipalcomponentanalysis(PCA)andnonlinearfeatureextractedfromkernelprincipalcomponentanalysis(KPCA)respectively,andthenutilizestheadaptivefeaturefusionalgorithmwhichisbasedontheweightedmaximummargincriterion(WMMC)tofusethefeaturesinordertoachievebetterperformance.Thelinearregressionclassifierisusedintheexperiments.Theexperimentalresultsindicatethattheproposedself-fusionalgorithmachieveshigherrecognitionratecomparedwiththetraditionalPCAandKPCAfeaturefusionalgorithms.

简介：Thispaperconstructsasimulationsystemofnear-fieldlaserimagingfor3Dgridmodeloftarget,providessomemethodsforthekeyproblems,suchasthemodelingoftargetandlasertransceiver,thecalculationoflaserechopower,theimagingalgorithmsandsoon.Atargetimagelibraryisestablishedbyanewimagingmethodinanyrendezvousconditions.Thefourreal-timerecognitionalgorithmswhichareefficientandsuitableforhardwareimplementationarepresentedattheconditionsoftheimageincompleteness,intensivenoiseandarbitraryattitudeoftarget.Theexperimentalresultsshowthatallthefouralgorithmscanindependentlyrecognizethetargeteffectivelyandabetterrecognitioneffectisobtainedbytheintegrationoffouralgorithms.

简介：ThispaperproposedarobustmethodbasedonthedefinitionofMahalanobisdistancetotrackgroundmovingtarget.Thefeatureandthegeometryofairbornegroundmovingtargettrackingsystemsarestudiedatfirst.Basedonthisfeature,theassignmentrelationoftime-nearbytargetiscalculatedviaMahalanobisdistance,andthenthecorrespondingtransformationformulaisdeduced.Thesimulationresultsshowthecorrectnessandeffectivenessoftheproposedmethod.

简介：定位一个目标的一个有效解决方案被建议，由使用到达(TDOA)的时差，大小它面对随机的传感器放错误增加评价的精确性。在二阶段的加权的最少的广场(TSWLS)的位置评价错误的原因方法被分析为改进本地化表演开发一个简单、有效的方法。明确地，参考传感器再被选择，坐标系统被使用TSWLS方法，和目标的最后的位置评价根据初步的估计的目标位置旋转被使用加权的最少的广场(WLS)获得。建议途径展览一靠近形式并且象TSWLS一样有效方法。模拟结果证明建议途径与增加传感器位置错误产出低评价偏爱和改进坚韧性并且能容易因此容易完成Cramer-Rao更低的界限(CRLB)并且有效地改进本地化精确性。

简介：Single-cameramobile-visioncoordinatemeasurementisoneoftheprimarymethodsof3D-coordinatevisionmeasurement,andcodedtargetplaysanimportantroleinthissystem.Amultifunctionalcodedtargetanditsrecognitionalgorithmisdeveloped,whichcanrealizeautomaticmatchoffeaturepoints,calculationofcamerainitialexteriororientationandspacescalefactorconstraintinmeasurementsystem.Theuniquenessandscalabilityofcodingareguaranteedbytherationalarrangementofcodebits.Therecognitionofcodedtargetsisrealizedbycross-ratioinvariancerestriction,spacecoordinatestransformoffeaturepointsbasedonspacialposeestimationalgorithm,recognitionofcodebitsandcomputationofcodingvalues.Theexperimentresultsdemonstratetheuniquenessofthecodingformandthereliabilityofrecognition.

简介：Thepaperfirstdiscussesshortcomingsofclassicaladjacent-framedifference.Secondly,basedontheimageenergyandhighorderstatistic(HOS)theory,backgroundreconstructionconstraintsaresetup.Underthehelpofblock-processingtechnology,backgroundisreconstructedquickly.Finally,backgrounddifferenceisusedtodetectmotionregionsinsteadofadjacentframedifference.TheDSPbasedplatformtestsindicatethebackgroundcanberecoveredlosslesslyinaboutonesecond,andmovingregionsarenotinfluencedbymovingtargetspeeds.Thealgorithmhasimportantusagebothintheoryandapplications.