简介:Sequentialmeasurementprocessingisofbenefittobothestimationaccuracyandcomputationalefficiency.Whenthenoisesarecorrelatedacrossthemeasurementcomponents,decorrelationbasedoncovariancematrixfactorizationisrequiredinthepreviousmethodsinordertoperformsequentialupdatesproperly.Anewsequentialprocessingmethod,whichcarriesoutthesequentialupdatesdirectlyusingthecorrelatedmeasurementcomponents,isproposed.Andatypicalsequentialprocessingexampleisinvestigated,wheretheconvertedpositionmeasurementsareusedtoestimatetargetstatesbystandardKalmanfilteringequationsandtheconvertedDopplermeasurementsarethenincorporatedintoaminimummeansquarederror(MMSE)estimatorwiththeupdatedcross-covarianceinvolvedtoaccountforthecorrelatederrors.Numericalsimulationsdemonstratethesuperiorityoftheproposednewsequentialprocessingintermsofbetteraccuracyandconsistencythantheconventionalsequentialfilterbasedonmeasurementdecorrelation.