Optimal Linear Fusion of Smoothed State Estimation

《Optimal Linear Fusion of Smoothed State Estimation》由会员上传分享，免费在线阅读，更多相关内容在行业资料-天天文库。

1、I.INTRODUCTIONEstimationfusionistheproblemofhowtobestutilizeusefulinformationcontainedinmultiplesetsOptimalLinearFusionofofdataforthepurposeofestimatingaquantity[1].Therearetwobasicestimationfusionarchitectures:SmoothedStateEstimatescentralizedanddistributed,dependingo

2、nwhetherrawdataareuseddirectlyforfusionornot.Centralizedfusion,inwhichallsenorssendtheirrawdatatothefusioncenter,canproducegloballyoptimalYONGXINGAOestimatesbutneedsheavycomputationanddataXi’anJiaotongUniversitytransmission.Indistributedfusion,eachlocalsensorX.RONGLI,F

3、ellow,IEEEsendsitsprocesseddata,whichmayormaynotbeUniversityofNewOrleansstateestimates,tothefusioncentertoobtainglobalestimates.Whentheprocesseddataarestateestimates,thefusionisreferredtoasstandarddistributedfusion[1].Inthispaper,wefocusonstandarddistributedTheworkpres

4、entedheredealswithestimationfusionoffusionforaparticularcase,whereunliketheexistingsmoothedstateestimates.Twoproblemsoffusionforsmoothingwork,localestimatesaresmoothedestimates.areconsidered:fixedpointandfixedinterval.OptimallinearSmoothingistheestimationofapastvalue.I

5、tfusionrulesinthesenseoftheoptimalweightedleastsquaresusesmeasurementsthroughtimektoestimatethestateofasystemattimen,wheren

6、enientforlag,andfixedinterval.Fixed-pointsmoothingistoimplementation.Wealsoproposeamorepracticalmethodforestimatethevalueatafixedpasttimen=pbasedonreal-timesmoothing,whichinessenceisfusingsmoothedanddatathroughtimek;fixed-lagsmoothingistoestimatefilteredestimates.Illus

7、trativenumericalresultsareprovidedtothevalueLstepsbeforebasedondatathroughtimeverifytheperformanceandcredibilityofthefusionrules.k(kvariesandListheconstantlag);fixed-intervalsmoothingistoestimateallpastvaluesbasedonafixedsetofdatathroughk=N(Nisfixed)[3,4].Sincesmoothin

8、gleadstoanestimationimprovementbyusing“future”measurements,itiswidelyusedinstateestimation(see,e.g.,[5]—[7])andtarget