Upload
others
View
12
Download
0
Embed Size (px)
Citation preview
IEEE: Robotics and automation Society (RAS)
Ontologies for Robotics and Automation
Study group
Craig SchlenoffNational Institute of Standards and Technology
Need
A robot can only achieve tasks and perform missions based on what it knows, which is primarily captured within the robot’s internal knowledge representation. This representation is usually very specialized to the individual robot and often very loosely defined. With the growing complexity of behaviors that robots are expected to perform as well as the need for multi-robot and human-robot collaboration, the need for a standard and well-defined knowledge representation is becoming more evident.
IEEE Ontologies for Robotics and Automation Goal
To develop a methodology for knowledge representation and reasoning in robotics and automation, together with the representation of concepts in an initial set of application domains, to allow for unambiguous knowledge transfer among any group of humans, robots, and other artificial systems.
Study Group Participants (78 subscribed to mailing list)
GroundAirUnderwater/SurfaceSpace
Scope (Robots)
Different Kinds of Robotic Knowledge
Cost-based Models
LADAR and Color Camera Images
Layered Terrain Maps
S3 TentativePlans_Done RoadMarchOrganizationInPlace
S4 sp3_ReadyToConductRouteRecon
S5 qp_ReadyToOrganizeAA
NewCommand
S1 MarchOrganizationDetermined
S2
CONDUCT TACTICAL ROAD MARCH TO ASSEMBLY AREA
S6 qp_ClearOfStartPoint
S7 mb_tp_ReconToStartPoint_Done
1
2
14
13
12
11
10
9
8
7
6
5
4
3
15
16
S8 sp3_RouteRecon_Done
S9 qp_AtReleasePoint
S10 qp_AreaReconOfAADone
17
S11 qp_Status_AssemblyAreaSuitable
S14 mb_Unit_AtReleasePoint qp_AssemblyAreaReady
S16 mb_Unit_AtReleasePoint
S12 mb_Unit_AtStartPoint
S12 mb_LastUnit_AtStartPoint
S13 tp_TrailPartyAtStartPoint
S14 HaltCriteriaMet
NewSituationReport
NewHigherLevelInformation
S15 HaltEnded_ReadyToContinueMarch
DisabledVehicleUnscheduledHalt
S16 tp_AtReleasePoint
S17 mb_tp_OccupyingAssemblyArea
S18 AllUnitsSecureInAssemblyArea
S4 sp3_PrepareForRouteReconnaissance qp_PrepareToOrganizeAssemblyArea
S5 sp3_ConductRouteReconnaissance
S6 qp_FollowReconPlatoonToAssemblyArea
S1 DetermineMarchColumnOrganization:
S2 all_FormTacticalRoadMarchOrganization
S3 MakeTentativePlan: Determine_Route, _FireSupport, _MovementFactors, _AA
S7 mb_tp_PrepareForRoadMarch
S8 PrepareDetailedMovementPlans:
S9 sp3_EstablishAssemblyAreaSecurity
S10 qp_ConductAreaReconnaissanceOfAA
S11 qp_OrganizeAssemblyArea
S12 1st_mb_Unit_MoveIntoRoadMarchFormation
S16 1st_mb_Unit_OccupyAssemblyArea
S16 mb_Unit_OccupyAssemblyArea
S12 mb_Unit_ExecuteTacticalRoadMarch next_mb_unit_MoveIntoRoadMarchFormation
S13 mb_Unit_ExecuteTacticalRoadMarchtp_MoveIntoRoadMarchFormation
S14 tp_SupportMarchColumnMovement
S15 mb_tp_ExecuteScheduledHalt
UpdateDetailedMovementPlans
UpdateDetailedMovementPlans
S14 mb_tp_ResumeExecutionOfTacticalRoadMarch
S15 mb_tp_ExecuteUnscheduledHalt
S17 tp_OccupyAssemblyArea
S18 all_FormStandardTroopOrganization
S0 TacticalRoadMarchToAssemblyArea_Done
25
24
23
22
21
20
19
18
State Tables
Ground TruthPrediction Equations
Databases
Per
cept
ion
Use Cases
Moving
Machine Ontology
Autonomous Vehicle Ontologies
Simulatio
n
State
Planning
Cost Based
Planning
Ontology
Judgment
Value Prediction
Object
Terms
General Logic
Thesauri
formalTaxonomies
Frames(OKBC)
Data Models(UML, STEP)
Description Logics
(DAML+OIL)
Principled, informal
hierarchies
ad hoc Hierarchies
(Yahoo!)structured Glossaries
XML DTDs
Data Dictionaries
(EDI)
‘ordinary’Glossaries
XML Schema
DB Schema
Glossaries & Data Dictionaries
MetaData,XML Schemas, & Data Models
Formal Ontologies & Inference
Thesauri, Taxonomies
Formalities of Knowledge Representations
What is an Ontology?
“a specification of a conceptualization”*Ontologies explicitly represent key concepts, their properties, their relationships, and their rules and constraints.Ontologies often focus more heavily on the meaning of concepts as opposed to terms that are used to represent them
Vocabulary + Structure = TaxonomyTaxonomy + (Relationships and Constraints) = Ontology
*Tom Gruber, Stanford Univ.
Ontology Application Scenarios
Common Access to InformationInformation required by multiple agents
Provides a single source of information for multiple applications
Ontology used as agreed standard
Benefits: knowledge reuse, maintainability, long term knowledge retention
Ontology
Application 1 Application 2 Application 3
referencesreferences
OA
references
• Ontology as Specification – build ontology for required domain– produce software consistent with ontology – manual or partially automated– Benefits: documentation, maintenance,
reliability, knowledge (re)use
Ontology Application Scenarios (cont.)
Ontology as an Exchange LanguageProvides an interlingua among disparate applications
Ensures semantic-based interoperability
Solves the point-to-point integration problem
Benefits: systems integration, semantic interoperability
• Ontologies for Reasoning– Allows one to run queries through
a reasoning engine– Helps to identify information that is
not explicitly represented– Benefits: knowledge inference Application
Ontology
Reasoning Engine
AD
OAaccesses
Interfaces with
Ontologies for Robotics and Automation Approach
Top downDevelop/identify an upper ontology to serve as the overarching structure that information can “hang from”
Develop a methodology to add new information to the ontology
Bottom UpDevelop detailed ontologies for a small set of domains
● Service Robots
● Autonomous Robots
Domains are intentionally broad to allow for overlapping concepts
Tying it all togetherIncorporate the domain ontologies into the upper ontology using the defined methodology
Reconcile any discrepancies that exist among concepts
Define a framework for building ontologies that allows robot designers to build domain-specific ontologies in a controllable wayDefine linguistic framework so that expressions using the ontologies can be communicated and so that it is possible to translate between different ontological realms (e.g. human and robot)Define top-level categories as a foundation for further extension
Upper Ontology and Framework SubgroupGoals
Service Robot Ontology SubgroupTypes of Service Robots
“A Service Robot is a robot which operates semi- or fully autonomously to perform services useful to the well-being of humans and equipment, excluding manufacturing operations” - International Federation of Robotics (IFR)
Industrial robots (e.g., radiographic inspection of welds)
Defense robots (e.g., autonomous scout vehicles)
Healthcare robots (e.g., surgical manipulation, wheel chairs)
Prosthetic robots (e.g., prosthetic arms, legs, etc.)
Scientific robots (e.g., gene sequencers)
Domestic robots (e.g., floor cleaners, lawn mowers)
Diffused robots (e.g., parallel park assist systems)
Military and law enforcement robots (e.g., drones, UAVs)
Service Robot Ontology SubgroupApproach
Phase 1: Identify and summarize the different definitions and glossaries used in the different Service Robots - bibliographical research
Phase 2: Requirement analysis to extend /modify the glossaries to meet the needs of industries and scientific communities
Phase 3: Aims at describing the robots in terms of well-defined ‘taxonomy’ by its components and operating environment
Phase 4: Defines and elaborates on the different knowledge layers - for high level Human-Robot/Robot-Robot Interaction
Autonomous robots are robots that can perform desired tasks in unstructured environments without continuous human guidance.
aerial photography using flying robots
customs and border security
electricity companies, who can inspect power lines, nuclear power plants, wind turbines, and other facilities using robots
gas and oil supply companies, who can use robots to inspect, maintain, and guard pipelines
local civic authorities
meteorological services, who can use UAVs to carry weather stations
river authorities and water boards
landmine detection and destruction.
Autonomous Robots Ontology SubgroupOverview
Description of robot hardware, softwareDescription of the activities that need to be performedDescription of the environment in which the robot needs to workUnderstand of cause and effect of performing actionsRelationship among other robots and/or people…
Autonomous Robots Ontology SubgroupKnowledge Representation
Investigate standards from other SDOs (ISO, ALFUS)Investigate existing efforts and formalisms and list the advantages/disadvantagesCollect practice in academia/industry (interview business management)Collect academia/industry requirements for standardizing autonomous robots ontologyHow to represent the terminology? Dictionaries, Glossaries, XML schemas, Databases, Logic-Based approaches, Others?Define ontology/glossary for autonomous robots
Define glossaries specific to autonomous robots
Define ontology for autonomous robots (air, ground, water, space)
Define ontology for sub-domains, sensor, perception, fusion, actuator, control, guidance, navigation, motion planning, mission planning, decision making (or software, hardware, electrical, mechanical, performance, …)
Define ontology for human machine interactions
Use of ontology to *describe* autonomous robots.
Set up a long-term roadmap for autonomous robots ontology
Autonomous Robots Ontology SubgroupApproach
PAR Procedures
PAR (Project Authorization Request) is the documentation necessary to start the process of turning a Study Group into a IEEE-Recognized Working GroupSeptember 16th - PAR drafted and submitted through IEEE Robotics and Automation Society/Standing Committee for Standards Activities (RAS-SCSA) mid-October - NesCOM (New Standards Committee) will review the PAR and make recommendations to the IEEE-SA Standards Board regarding approval.
Responsible for ensuring that proposed standards projects are within the scope and purpose of IEEE, assigned to the proper Society or other organizational body, and interested parties are appropriately represented in the development of IEEE standards.
How To Get Involved
Come to our meeting tomorrow (Sept 26th)!Union Square Rooms 23&24 (4th floor), 9am-12:30pm
Contact me:Email: [email protected]
Phone: 301-975-3456
Join our Google group
IEEE RAS Ontologies for Robotics and Automationhttps://groups.google.com/forum/#!forum/ieeeraswgVisit the Web site set up for Autonomous Robots Group (http://aro.svn.sourceforge.net/)