2013 IEEE International Conference on Robotics and Automation (ICRA)Karlsruhe, Germany, May 6-10, 2013
TheNextBestTouchforModel-BasedLocalization
PaulHebert,ThomasHoward,NicolasHudson,JeremyMa
JetPropulsionLaboratory,CaliforniaInstituteofTechnology
Pasadena,CA,91109,USA
JoelW.Burdick
CaliforniaInstituteofTechnologyPasadena,CA,91124,USA
Abstract—Thispaperintroducesatactileorcontactmethodwherebyanautonomousrobotequippedwithsuitablesensorscanchoosethenextsensingactioninvolvingtouchinordertoaccuratelylocalizeanobjectinitsenvironment.Themethodusesaninformationgainmetricbasedontheuncertaintyoftheobject’sposetodeterminethenextbesttouchingaction.Intu-itively,theoptimalactionistheonethatisthemostinformative.Theactionisthencarriedoutandthestateoftheobject’sposeisupdatedusinganestimator.Themethodisfurtherextendedtochoosethemostinformativeactiontosimultaneouslylocalizeandestimatetheobject’smodelparameterormodelclass.ResultsarepresentedbothinsimulationandinexperimentontheDARPAAutonomousRoboticManipulationSoftware(ARM-S)robot.
I.INTRODUCTION
Wewouldlikeautonomousroboticstoperformdexteroustaskssuchasgraspingscrewdriversandhouseholdobjectssuchascups,insertingcablesintopoweroutlets,andusingkeystoopenandunlockdoors[1]–[3].Allofthesetasksrequirepreciselocalizationofkeyobjectsintherobot’sframe,withtaskssuchaskeyinsertionrequiringlocalizationontheorderofamillimeter.Whilemanyofthesetaskscouldbedonewithhighlypreciseroboticsystems,thereisabroadinterestinperformingthesetaskswithinexpensivesystems,whichmaylacktherequiredprecision.Eveninknownenvironments,withprecisevisualsensors,theprecisionofaclosedchainroboticmechanismmaybeinsuf cienttoexecutemanyofthesetasksopenloop[4].Moregenerally,errorstypicallyintroducedbyvisualposeestimationand/orinternalkinematicerrorsareoftentoolargeforprecisegraspingandmanipulation.Instead,estimationalgorithmscanre nethelocalposeestimateduringtaskexecution,whenphysicalinteractionwiththeobjectoccurs[5]–[7].Thispaperbuildsupontheseestimationalgorithms,andprovidesanautomatedmethodtoselectroboticmanipulationactionswhichwillprovideincreasedknowledgeaboutthesystemstate.
IntherecentDARPAAutonomousRoboticManipulationSoftware(ARM-S)competition,teamswererequiredtocreateautonomousmanipulationsystemstoexecutevarioustasks.Inparticular,teamswererequiredtounlockandopenadoorautonomously.Aninterestingobservationisthatalmostallteams,regardlessoftheunderlyingapproach,createdactionsthat rstre-localizedthedoorhandle(byeithertouchingthedoororthehandle)beforeproceedingwithkeyinsertion.Inthispaper,theactionselectionalgorithmautomaticallydecideswhichprobingmotionstomake,

as
Fig.1:Exampledoorhandletouchingactions.opposedtousingprede nedprobes,motionsorpolicies.Therearesituationsinwhichtherobotmaybepresentedwithanovelversionofaknownobjectclass(e.g.imaginearobottasktopickupapriorinewscrewdriver.Therobot’sperceptionsystemmaybeabletoeasilyrecognizethescrewdriverclass,buttheprecisedimensionswillnotbeknown.).Asaconsequence,itwouldbeadvantageoustoselectinteractingactionsthatwouldsimulataneouslylocalizetheobjectandestimatecharacteristicparametersofthemodelthatmayimpactlocalization.Similarly,arobotmayberequiredtomanipulateanobjectofwhichitisunabletodiscerntheobject’sclassorcategory.Thismayoccurfromvisualocclusionorviewpointdependency.Thispaperpresentsamethodtoselectactionsthataimstobothlocalizeanddeterminetheobjectmodel.
Here,anactionisde nedasthecombinationofatrajec-tory,andasetofendconditions.Thetrajectoryencodestheplannedopenloopexecutionoftherobotmotion,andtheend-conditionsspecifywhenanactionterminates(e.g.uponacertainforcethreshold,orwhencontacthasoccurred).Ouractionselectiontheorychoosesactionswhichareexpectedtominimizetheentropyoftherelativerobot-environmentbelief.Accuratelocalizationrequiresminimaluncertaintyaboutthepose.Inordertominimizeuncertainty,theentropyofthestatedistributionshouldbeminimized,asdistributionswithlowentropyarepeakedandcertain.Equivalently,theinformationgaincanbeviewedasthechangeinrelativeentropycausedbyundergoinganactionandupdatingthebeliefbyacquiringanewmeasurement.
VIRE:Active RFID-based Localization Using Virtual ...The paper will focus on the first model in ...automatically determining the best communication path...
-Based and Range-Free Localization Schemes for Sensor Networks Localization ?...Radio model ? Upper & lower bounds on signal strength ? Beyond the UB, ...
The Continuum Analytic... 49人阅读 7页 2.00Adiabatic shear loca...
Source localization problems are often approached using non-model-based techniques. B隐藏>> Model-based Solution Techniques for the Source Localization Problem...
To self-con?gure and adapt the localization in ...Factor (WAF) based signal propagation model. The...for best results with their centralized convex positio...
A Process Meta- Model Based Approach for the Development of Collaborative Applications Built on_专业资料。JunlfmmuiainadCoue 21)1013oraoCo nct n mptro8(0 2...
A fractal-based model for the microstructure evolution of silicon bronze wires fabricated by die_专业资料。IenainlounaoMieasMealryantrtaJrlfnrl,otlgundMaeil tras...
A GCM-Based Forecasting Model for the Landfall of Tropical Cyclones in China_专业资料。ADVANCESIATM0S N PHERISENCES,VOL.2C CI8,N0.5,01109052,14—15 A...
y 1 Model-based recognition and localization are fundamental tasks in ...set of candidate objects fO1; O2; : : :g best explains the input ...
暂无评价 10页 免费 Model-based localization... 暂无评价 6页 免费喜欢...From the experiments we conclude that the second strategy is the best and ...
热门文档