Experiences in Developing An Intelligent Ground Vehicle (IGV) Ontology In Protégé Craig Schlenoff, NIST Randy Washington, DCS Corporation Tony Barbera, NIST July 8, 2004
Agenda • Background: – What is an Intelligent Ground Vehicle (IGV)? – NIST 4D/RCS Methodology and Architecture • Ontology Development: – 4D/RCS to Ontology Mapping – Interchange Formats and Upper Ontologies – IGV Military Equipment – IGV Behaviors – IGV Conditions • Current Status • Issues and Lessons Learned
Intelligent Ground Vehicle? What is an
4D/RCS Mapping to IGV Commander & Ontology Service Vehicle Agent Instances Instances C onduc tR oadM a rchTo A s sem blyA rea ConductRoadMarchToAnAssemblyArea(AA) (ConductRightFlankRecon) (Executing) L ight C avalry Troop KNOWLEDGE VALUE P re pare ForR oadM a rch P re pare ForR oadM a rch C on ductR outeR ec on Follow P latoo nToA s se m bly A re a DATABASE M o veIntoForm ation M o veIntoForm ation S ec ureA ss em blyA rea JUDGMENT C onduc tA rea R e con E ecu x teR oad M arch E ecu x teR oad M arch Conduct FollowPlatoon PrepareFor PrepareDetailed Secure Organize MoveInto E x ecu teH alt E x ecu teH alt O rganizeA sse m b ly A re a RouteRecon ToAssemblyArea RoadMarch MovementPlans AssemblyArea AssemblyArea MarchFormation O cc upyA ss em bly A rea O cc upyA ss em bly A rea C om m and S cou t Q uartering Scout G roup P lato on Platoon P arty WORLD BEHAVIOR SENSORY DOT Driving Manuals (FSO ) 1st 3rd PROCESSING MODEL GENERATION and STATUS COMMANDED DeployTo Conduct Conduct Conduct Locate&Secure Conduct MoveTo Conduct D eployToS tartPo int DeployToStartPoint TASK (GOAL) ARMY Field Manuals D eploy ToS tartP oint StartPoint LeftFlank RightFlank MainRoute ObstacleBypass Obstacle ControlPoint DominateTerrain C onductR ightFla nkR ec on ConductMainRouteRecon C onduc tLe ftFla nkR ec on RouteRecon RouteRecon Recon Recon Recon C onductD o m inan tTerrainR e con Evaluate&ClassifyObstruction . Lo cate& S ecureO bs ta cleB ypas s C onductA m bus hS iteR ec on MoveToControlPoint M ov eToC ontrolPoint (Platoon “B” Section M ov eToC ontrolP oint + Agent Control SENSORY B S ectio n C S ection BEHAVIOR A S ection INPUT Module) DriveOnTwoLaneRd NegotiateLaneConstrict (Executing) (ReconToRoute) GENERATION PassVehInFront DriveOnTwoLaneRd PassVehInFront Loc ateW ate rB ypas s MoveTo Locate Perform Ford Secure MoveTo Overwatch P erform Ford R con e ControlPoint WaterBypass Ford Water Area Cover/Concealed Section PassVehInFront . FordW a terO bstac le STATE-TABLES Recon Obstacle Position S ec ureA re a O verw atchS ec tion S quad S quad Squ ad H M M W V H M M W V H M M W V H M M W V H M M W V H M M W V H M M W V #2 H M M W V #3 H M M W V #7 #5 #6 #8 #4 PS G #9 #10 NEXT A s se ssFo rdTerrain C ross Ford Assess ScanArea Cross MoveTo S e ndR e port S anA c rea ForE nem y STATUS M ov eToP os ition SUBGOAL FordTerrain ForEnemy Ford Position S urveillanc Lethality C om m s M obility (R ST A ) (G unner) Domain Experts M o veToW ater S hiftTo4W hee lL o S c anP ath M o veToO ppositeB ank Select S hiftTo4W hee lH i Scan MoveTo ShiftTo MoveTo ShiftTo Dry D ry B ra kes Path Water 4WhLo Opposite 4WhHi Brakes Bank “ConductWaterObsRecon” FLIR , L aser P an /T ilt G aze Prim itive/ T argeting W eapon R adar, Co ntrol T rajectory C on trol Sensors C ontrol Plan State-Table C am era Task Decomposition Tree Hierarchical Organization of ConductRightFlankRecon (Route Reconnaissance Example) Agent Control Modules Selection Conditions Selected Plan NormalRouteReconSituation ConductRightFlankRecon WaterObstacleDetected WaterObstacleDetected ConductWaterObstacleRecon MinefieldDetected AssessMinefieldRecon MostlyTrees,SomeBushes DefileDetected ConductDefileRecon Saturated Plant ExtensiveWaterSurfaceVisible LateralRouteDetected ConductLateralRouteRecon SlowMovingWaterCoveredLand SwampDetected Ground Height PLAN SELECTION TABLE 3 cm 6-18" LongLeafGrasses - very flat, long green SignificantTractionSlip leaves-purple/rose/yellow flowers. MajorGroundDeformation 6-48" Sedges - triangular tan/green stem plants, StagnantWater 15 cm papyrus, narrow green to tan grass- OrganicMaterialOnWaterSurface like leaves, spikelets of inconspicuous IndeterminantGroundLevel BogDetected tan-to-yellow-to-white flowers. Mosses,Evergreens,andShrubs SignificantTractionSlip 1-6' Reeds - tall, woody, thin, round, hollow 2.4 cm .9 to 2.7m jointed (tan-to-green) stem plants, ConductWaterObstacleRecon long narrow green blade leaves, WaterCoveredLand SomeWaterVisible S1 SetupControlMeasures large feathery panicles (elongated NewPlan sq5_ReconToRouteFlank ExtensiveGrassyVegetation clusters of tan/white/purple flowers sq6_ConductTravelingOverwatch MarshDetected MajorGroundDeformation LegalToPass along main stem). sq8_VisuallyClearObstacle-FarSide WaterDepthToSixFeet TractionSlip 3-9' Bulrushes - tall tan-to-green stems, with S2 SetupReport&NewControlMeasures sq5_ReconToFarFlank dark brown cylindrical seed heads S1 sq5_NoBypassOnRouteSide sq6_ConductTravelingOverwatch that explode into white down, long FlowingWater sq8_VisuallyClearObstacle-FarSide flat green sword shaped leaves, Cattails NarrowWidth,IndeterminantLength StreamDetected cattails. NonVegetatedWaterInMiddle S0 SetupReport&NewControlMeasures sq5_MoveTacticallyToControlPoint ErodedEarthEmbankments Float S2 sq5_NoBypassTowardBoundary Water Lilies - very large floating green sq6_ConductTravelingOverwatch sq8_MoveTacticallyToControlPoint Six Feet leaves with white flowers. LargeSurfaceAreaOfStillWater of Water S4 SetupReport&NewControlMeasures Object Groupings and OrganicMaterialMayBeOnWaterSurface Pond/LakeDetected LargeAreaWithoutGrasses,Trees,Shrubs S1 sq5_PossibleFordDetected sq5_ConductFordRecon sq6_ConductNearSideOverwatch BoundedBySwamp,Marsh Classifications sq8_VisuallyClearObstacle-FarSide FlowingWater S5 SetupReport&NewControlMeasures SignificantWidth,IndeterminantLength Objects and Maps S4 sq5_FordLooksPassable sq5_MoveToCover/ConcealPosition RiverDetected NonVegetatedWaterInMiddle sq6_ConductNearSideOverwatch ErodedEarthEmbankments sq8_VisuallyClearObstacle-FarSide Features and Attributes S1 SetupControlMeasures WaterSheenOnGroundSurface sq5_ReconToRouteFlank Process S4 sq5_FordNotPassable RuttedWithStandingWater sq6_ConductTravelingOverwatch Segmented Groupings sq8_VisuallyClearObstacle-FarSide LittleToNoVegetation MudDetected SignificantTractionSlip S7 SetupReport&ControlMeasures SignificantGroundDeformation Model sq5_ConductObstacleFarSideRecon S1 sq5_LateralBypassFound sq6_ConductTravelingOverwatch Objects & Attributes World States Situations sq8_VisuallyClearToRoute Instances Input Conditions Output Commands ColorCameras LADAR PLAN STATE-TABLE Radar Stereo FLIR Nav SENSORY WORLD MODEL BEHAVIOR PROCESSING KNOWLEDGE GENERATION Environment Condition Instances Instances
Interchange Formats and Upper Ontologies • OWL – Neutral (W3C) interchange format – XML base enables use XSLT transforms – Provides access to emerging semantic web technologies • OWL-S – Rich semantics for describing complex processes (without being too complicated) – Well suited to agent architectures • Pieces of SUMO (Suggested Upper Merged Ontology) – Class structure and properties provide a good starting point for developing domain specific ontology – Native KIF format too complex for target community and not necessary for requirements capture • Namespaces – Used quite a bit to make ontology more manageable
IGV Conceptual Model External Service Request by a process Troop Commander Service Agent Platoon Platoon Platoon Leader Leader Leader Service Service Service Agent Agent Agent Section Section Section Lead Lead Lead Service Service Service Agent Agent Agent Vehicle Vehicle Vehicle Commander Commander Service Commander Service Service Agent Agent Agent Mobility Surveillance Survivability Lethality Support System System System System Service System Service Service Service Service Agent Agent Agent Agent Agent Automotive Auxiliary Propulsion Localization Navigation Subsystem Subsystem Subsystem Subsystem Subsystem Service Service Service Service Service Agent Agent Agent Agent Agent Brake Engine Transmission Track-Drive Component Component Component Component
Representing an IGV (cont.)
Recommend
More recommend
