資源描述：
《(已發(fā)表) Quasi-Stochastic Integration Filter for Nonlinear Estimation》由會(huì)員上傳分享，免費(fèi)在線閱讀，更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫。
1、HindawiPublishingCorporationMathematicalProblemsinEngineeringVolume2014,ArticleID967127,10pageshttp://dx.doi.org/10.1155/2014/967127ResearchArticleQuasi-StochasticIntegrationFilterforNonlinearEstimationYong-GangZhang,Yu-LongHuang,Zhe-MinWu,andNingLiCollegeofAutomation,HarbinEngi
2、neeringUniversity,No.145NantongStreet,NangangDistrict,Harbin150001,ChinaCorrespondenceshouldbeaddressedtoYu-LongHuang;heuedu@163.comReceived21October2013;Revised18May2014;Accepted24May2014;Published23June2014AcademicEditor:DanSimonCopyright?2014Yong-GangZhangetal.Thisisanopenacc
3、essarticledistributedundertheCreativeCommonsAttributionLicense,whichpermitsunrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycited.Inpracticalapplications,numericalinstabilityproblem,systematicerrorproblemcausedbynonlinearapproximation,andn
4、onlocalsamplingproblemforhigh-dimensionalapplications,existinunscentedKalmanfilter(UKF).Tosolvetheseproblems,aquasi-stochasticintegrationfilter(QSIF)fornonlinearestimationisproposedinthispaper.nonlocalsamplingproblemissolvedbasedontheunbiasedpropertyofstochasticsphericalintegrat
5、ionrule,whichcanalsoreducesystematicerrorandimprovefilteringaccuracy.Inaddition,numericalinstabilityproblemissolvedbyusingfixedradialintegrationrule.Simulationsofbearing-onlytrackingmodelandnonlinearfilteringproblemwithdifferentstatedimensionsshowthattheproposedQSIFhashigherfilt
6、eringaccuracyandgoodnumericalstabilityascomparedwithexistingmethods,anditcanalsosolvenonlocalsamplingproblemeffectively.1.IntroductionTheunscentedtransformation-(UT-)basedunscentedKalmanfilter(UKF)isatypicalGaussianapproximatefilterNonlinearfilteringhasbeenwidelyusedinmanyappli-
7、andhasbeenwidelyusedduetoitseaseofimplementation,cations.Generally,filteringproblemcanbeaddressedbymodestcomputationalcost,andappropriateperformanceusingBayesianestimationtheory,whichprovidesanoptimal[12,13].However,UKFsuffersfromthreemainproblemssolutionfordynamicstateestimatio
8、nproblembycomputinginitspracticalapplications:n
