Improved Vehicle Based Multibeam Bathymetry using

《Improved Vehicle Based Multibeam Bathymetry using》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、ImprovedVehicleBasedMultibeamBathymetryusingSub-mapsandSLAMChrisRomanHanumantSinghMIT/WHOIJointPrograminAppliedDepartmentofAppliedOceanPhysicsandEngineeringOceanScienceandEngineeringWoodsHoleOceanographicInstitutionMassachusettsInstituteofTechnologyWoodsHole,MA,USA

2、Cambridge,MA,USAhanu@whoi.educroman@whoi.eduAbstract—Thispaperpresentsanalgorithmtoimprovesub-groundreferencedpositionestimateswithboundederror.Un-seaacousticmultibeambottommappingbasedonthesimultane-fortunately,theactualerrorstatisticsarehighlycoupledtotheousmappi

3、ngandlocalization(SLAM)methodology.Multibeamenvironmentandaredifficulttocharacterize[3].LBLsystemsbathymetryfromunderwaterwatervehiclescanyieldvaluablealsorequirethedeploymentofadditionalinfrastructurethatlargescaleterrainmapsoftheseafloor,buttheoverallaccuracyofthes

4、emapsistypicallylimitedbytheaccuracyofthevehiclemakesquicklysurveyinganareadifficult.positionestimates.Thesolutionpresentedhereusessmallbathy-Subseavehiclesalsoemploydeadreckoning(DR)navi-metricpatchescreatedovershorttimescalesinasub-mappinggationusingacousticmeasur

5、ementsofthevehicle’sgroundcontext.Thesepatchesareregisteredwithrespecttooneanotherrelativevelocityandattitude,obtainedfromaninertialnav-andassembledinasinglecoordinateframetoproduceamoreigationsystem(INS).Thismethodofnavigationissubjectaccurateterrainestimateandpro

6、videimprovedrenavigationofthevehicletrajectory.Themappingisimplementedusingatounboundedpositionerrorgrowthoverlongtimescales.delayedstateextendedKalmanfilter(EKF)andresultsareshownHowever,thiserrorgrowthcanbeslowedbyusinghighqualityforarealworldmultibeamdatasetcolle

7、ctedattheMid-AtlanticINSsensorstoproduceaccurateshorttermnavigation[4].TheridgeusingtheJASONROV.primarynavigationsensorforDRnavigationunderwateristheIndexTerms—bathymetry,acousticmapping,SLAM,multi-acousticDopplervelocitylog(DVL).beamInrecentyearsthesimultaneousmap

8、pingandlocalizationcommunityhasfocusedonthecoupledproblemofmappingI.INTRODUCTIONanareawitharobotwhileconcurrentlyderivingimprovedBathymetricterra