Willow Garage May 2010 CRAM Team IAS brPage 3br CRAM CRAM Perception CRAM Knowledge Processing CRAM Planbased Control Demonstrations setting the table cleaning the table 2 simple meal preparation tasks strong emphasis on inferring right actions and ID: 44287
Download Pdf The PPT/PDF document "CRAM Cognitive Robot Abstract Machines ..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
CRAM—CognitiveRobotAbstractMachines MichaelBeetz,GordonCheng,andMatthiasKranz andIntelligentAutonomousSystemsGroup TechnischeUniversitätMünchenWillowGarage,May2010 CRAM CRAMPerceptionCRAMKnowledgeProcessingCRAMPlan-basedControl Cognition-enabledControlforEverydayManipulation *with compact controlprogramsperformingan open-endedset of everydaymanipulationtasks like I layingthetableand I cleaningup humanenvironments ina natural , general , skillful , exible and reliable manner. WillowGarage,May2010TeamIAS CRAM CRAMPerceptionCRAMKnowledgeProcessingCRAMPlan-basedControl Demonstrations I settingthetable,cleaningthetable I 2simplemealpreparationtasksstrongemphasison Iinferring rightactionsandactionparameterizations, Ianswering queriesaboutwhattherobotdidandwhyExamples:infer I missingitemsonthetable I whereobjectsbelong I howobjectsmustbehandled I wheretostand,howtoreach,whichgrasptype,wheretograsp,howmuchgraspforce,etc WillowGarage,May2010TeamIAS CRAM CRAMPerceptionCRAMKnowledgeProcessingCRAMPlan-basedControl CognitiveRobotAbstractMachine WillowGarage,May2010TeamIAS CRAM CRAMPerception CRAMKnowledgeProcessingCRAMPlan-basedControl CRAMPerceptionSystemCharacteristics I multimodal(tiltinglaser,tof,stereocamera) I taskable I basedonobjectdescriptions I reasonsaboutobjectidentities I beliefstatepassiveperception I storesatime-stampedmemoryofperceptions I symbolicscene \n \r WillowGarage,May2010TeamIAS CRAM CRAMPerception CRAMKnowledgeProcessingCRAMPlan-basedControl CRAMPerceptionSystemPCLContributions ! * WillowGarage,May2010TeamIAS CRAM CRAMPerception CRAMKnowledgeProcessingCRAMPlan-basedControl ScenePerception j KnowledgeBase j holds(on(o45,t2),t23)h312,248i),t23)object-data(o45, WillowGarage,May2010TeamIAS CRAM CRAMPerception CRAMKnowledgeProcessingCRAMPlan-basedControl PerceptionandKnowledgeProcessing I Providestherobotwithabstractsymbolicknowledgeabouttheperceivedscenes I Usesabstractsymbolicknowledgeforaccomplishingperceptiontasks I Answersqueriesthatrequireknowledgeprocessing and perception. WillowGarage,May2010TeamIAS CRAM CRAMPerception CRAMKnowledgeProcessingCRAMPlan-basedControl MissingObjectsQuery \n \r\n \n \r\n\n\r\n \n \r\n !\n \n \n \n" #\n $\n\r\n$\n \n \n" \n\n"\n" #\n !\n\r$\n \n \n"%\n\n"\n" \n \r #\r\n\n\r #%\n\n"\n" \r\n\n\r #$\n \n \n"! & ' \n \n \n \n "\n$ (\n )\n !* \n" +\n" \n \n \r \n \n \n\r \n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n \n \n \n\n\n\n\n\n\n\n\n\n\n\n\n perceivedObjectsOnPlane(Plane,Perceived):-setOf(Obj-Hyp,(on(Obj,Plane),category(Obj,Cat),objectInstace(Obj,KnownObj),Obj-Hyp=[Id,Obj,Cat,KnownObj]),Perceived). WillowGarage,May2010TeamIAS CRAMCRAMPerception CRAMKnowledgeProcessing CRAMPlan-basedControl CRAMKnowledgeProcessing I terminologicalknowledge I pragmatic:PROLOG I OWL:semanticmap,SemanticRobotDescriptionLanguage I graphicalinterface I statisticalrelationallearningand WillowGarage,May2010TeamIAS CRAMCRAMPerception CRAMKnowledgeProcessing CRAMPlan-basedControl TheKNOWOB \n \r\n \n \r\n \r \n \n \n \n ! "# "# $\r\n \n %\r & \r ' $ \r \r \n\n ( \r \r I ArchitectureoftheKnowRobframework. I Visualizedresultsofqueriesforhumanposes.Objectsthatareconsideredtobemissingforacompletetablesetup. I Positionsformanipulationactions. I methodsforprocessingandimportingnatural-languagetaskinstructionsfromwebpages. WillowGarage,May2010TeamIAS CRAMCRAMPerception CRAMKnowledgeProcessing CRAMPlan-basedControl UsingtheSemanticObjectMap WillowGarage,May2010TeamIAS CRAMCRAMPerception CRAMKnowledgeProcessing CRAMPlan-basedControl Computablepredicates I Usetheenvironmentas'virtualknowledgebase' I Computepredicatesondemandbasedonthebeliefstate I Inherentlygrounded,automaticallyupdated I Query-specicabstraction \n \r \r \r \r ! \n \r \n \n \n " \r # $ % & # \r \n\n\n\n \n \r \n \r WillowGarage,May2010TeamIAS CRAMCRAMPerception CRAMKnowledgeProcessing CRAMPlan-basedControl ExtendingRobotSimulation Simulationisusedasatoolfore.g.testingofrobot'sbehaviorinthepresenceofhumans. WillowGarage,May2010TeamIAS CRAMCRAMPerceptionCRAMKnowledgeProcessing CRAMPlan-basedControl CRAMPlanLanguage I CPL=high-levelprogramminglanguageforautonomousrobotcontrol I specieshowtherobotistorespondtopercepts,events,successandfailure I concurrentsensor-directedcontrolwithvarioussynchronization I rst-classobjects:objectdescriptions,failuredescriptions,(computational)controlprocesses WillowGarage,May2010TeamIAS CRAMCRAMPerceptionCRAMKnowledgeProcessing CRAMPlan-basedControl CRAMPlanLanguage I robotplan=robotcontrolprogramthatcannotonlybeexecutedbutalsoreasonedaboutandmanipulated I declarativegoalsforachievement,perception,prevention,maintenance I allowsforreasoning:teleologicalandcausalreasoning,prediction,transformationalplanning WillowGarage,May2010TeamIAS CRAMCRAMPerceptionCRAMKnowledgeProcessing CRAMPlan-basedControl Thanks WillowGarage,May2010TeamIAS