IEEEJOURNALOFSELECTEDTOPICSINSIGNALPROCESSING,VOL.3,NO.4,AUGUST2009695ABayesianApproachtoMultipathMitigationinGNSSReceiversPauClosas,StudentMember,IEEE,CarlesFernández-Prades,Member,IEEE,andJuanA.Fernández-Rubio,SeniorMember,IEEEAbstract—Multipathisknowntobeoneofthemostdominantbetweenthereceiverandthesetofin-viewsatellites.Thesesourcesofaccuracydegradationinsatellite-basednavigationdistancesarecalculatedestimatingthepropagationtimethatsystems.Multipathmaycausebiasedpositionestimatesthatcouldtransmittedsignalstakefromeachsatellitetothereceiver.Atjeopardizehigh-precisionapplications.Thispaperconsiderstheproblemoftrackingthetime-variantsynchronizationparam-leastfoursatellitesareneededinordertocomputeuserposi-etersofboththeline-of-sightsignal(LOSS)anditsmultipathtionandtime[1]–[4].Eachsatelliteisuniquelyidentifiedbyreplicas.Inparticular,theproposedalgorithmtrackstime-delays,itsowndirect-sequencespread-spectrum(DS-SS)signal,whichamplitudes,phasesandproposesaproceduretoextractDoppleraretransmittedsynchronouslybyallsatellites.GNSSreceiversshiftsfromcomplexamplitudes.However,theinterestisfocusedareonlyinterestedinestimatingthedelaysofsignalswhichonLOSStime-delayestimates,sincethoseprovidethemeanstocomputeuser’sposition.TheundertakenBayesianapproachisarereceiveddirectlyfromthesatellites,hereafterreferredtoasimplementedbyaparticlefilter.Theselectionoftheimportanceline-of-sight-signal(LOSS),sincetheyaretheonesthatcarrydensityfunction,fromwhichparticlesaregenerated,isperformedinformationofdirectpropagationtime.Hence,reflectionsdis-usingaGaussianapproximationoftheposteriorfunction.Thistortthereceivedsignalinawaythatmaycauseabiasindelayselectionprovidesaparticlegeneratingfunctionclosetotheandcarrier-phaseestimations[5],[6].Multipathisprobablytheoptimal,whichyieldstoanefficientusageofparticles.Thecom-plex-linearpartofthemodel,i.e.,complexamplitudes,istackleddominantsourceoferrorinhigh-precisionapplicationssinceitbyaRao–Blackwellizationprocedurethatimplementsaco