UNCLASSIFIEDUNCLASSIFIED
1
Kevin [email protected]
Modeling & Simulation Information Management BranchSignal and Information Sciences LaboratoryApplied Research LaboratoriesThe University of Texas at Austin
Applications of Ontology OWL to:• Geospatial Feature Data Dictionaries• Rapid Data Generation: Order of Battle and
Entity Type Data Management
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN2
My involvement• Participation
– M&S COI Data Management Working Group– ASW COI Data Management Working Group– NATO M&S Group (MSG) 085 – C2 &
Simulation Interoperability– Simulation Interoperability Standards
Organization (SISO)• Standards Activity Committee• Military Scenario Definition Language (MSDL)• Coalition Battle Management Language (C-
BML)• Simulation Conceptual Modeling (SCM)• Architecture-Neutral Data Exchange Model
(ANDEM)
• Projects: – M&S Coordination Office– US Army Simulation to C4I
Interoperability (SIMCI) OIPT– Joint Staff J7 Joint Coalition
Warfighting (formerly JFCOM)
• Coordinated with– US Army Operational Test Command– AMSAA– Global Force Management Data
Initiative (GFM DI)– US Army PD Tactical Network
Initialization
UNCLASSIFIEDUNCLASSIFIED
3
10F-SIW-068Mapping Data Models and Data Dictionaries – Removing the Ambiguity
Kevin [email protected] 512-835-3679
Eric [email protected]
Roy [email protected] 512-835-3857
Bruce [email protected] 512-835-3120
Modeling & Simulation Information Management BranchSignal and Information Sciences LaboratoryApplied Research LaboratoriesThe University of Texas at Austin
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN4
Overview• Background
– Data dictionaries must be mapped to enable translation and reuse of datasets and tools based on one data dictionary or another.
• Problems– Current mapping processes use English language and
spreadsheets to capture the mappings.– Too much room for interpretation.– Difficult to evaluate or compare mapping results.– No clear path to using mappings in data mediation software.
• Our objective– Explore and demonstrate the benefits of an ontology-base
approach to data dictionary mapping.
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN5
We focused on mapping of EDCS and NFDD ...
• EDCS – SEDRIS Environmental Data Coding Specification
• NFDD – National System for Geospatial-Intelligence (NSG) Feature Data Dictionary
• Both are dictionaries of geospatial feature concepts• Both contain concepts as:
• Features / Classifications• Attributes• Enumerations
• Both provide definitions for each Concept, but little or no taxonomy or relationships
• Both are available as MS Access Databases
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN6
...But recognized that there are others.
Sub-schemes and implementation schemes• Some schemes are “based” on a common data dictionary, but semantics have
drifted and diverged for various reasons.• Some schemes are not based on any common data dictionary.
Environment-related thesauri:• GEMET – GEneral Multilingual Environmental Thesaurus• AGROVOC – a thesaurus of agriculture, forestry, fisheries, and other domains• NALT – National Agriculture Library Thesaurus
General use knowledge bases:• DBPedia – a structured extraction of the Wikipedia body of knowledge• OpenCyc – Open source Cycorp general knowledge base• WordNet – Lexical database of the English Language
Unlike NFDD and EDCS, these are actual thesauri with broader / narrower relations, preferred and alternate names, definitions, etc.
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN7
AGC Mapping Relations AGC Relation
(only 7 of 18 shown here) AGC Example Set Theory Relations
Concepts are completely disjoint (EDCS:Terrain Plain, NFDD:Slope Region)(EDCS:Complex Outline, NFDD:Facility)
Concepts overlap completely (EDCS:Parcel, NFDD:Parcel)(EDCS:Marine Port, NFDD:Port)
Concepts overlap well (EDCS:Glacier, NFDD:Glacier)(EDCS:Opera House, NFDD:Building)
A \ B and B \ A are “small”
Concepts overlap somewhat (EDCS:Sports Arena, NFDD:Sports Stadium)
Concept A is a generalization of Concept B
(EDCS:Harbour, NFDD:Harbour)(EDCS:Route, NFDD:Ice Route)
Concept A is a slight generalization of Concept B
(EDCS:Traffic Light, NFDD:Traffic Light)(EDCS:Astronomical Station, NFDD:Astronomical
Observatory) A \ B is “small”
Concept A is an aggregate of Concept B
(EDCS:Airfield, NFDD:Runway)(NFDD:Tent, EDCS:Camp)
A “has part” BB “is part of” A
BA
BA
ABBABA
\\
ABBABA
\\
BA
BA
Set Theory shows duplicate relationships with ambiguous differences.
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN8
AGC Relation(another 3 of 18 shown here) Class Diagram Visualization
Concept A is a generalization of Concept B
Concept A is a slight generalization of Concept B
Concept A is an aggregate of Concept B
AGC Mapping Relationships
AGC Relation(only 4 of 18 shown here) Class Diagram Visualization
Concepts are completely disjoint
Concepts overlap completely
Concepts overlap well
Concepts overlap somewhat
BA
A B
BA
A B
BAA \ B
“small”B \ A
“small”
A B
BAA \ B B \ A
A
B
A
BA \ B“small”
has partA B
Set Theory shows duplicate relationships with ambiguous differences.
OWL is built upon set theory, where OWL classes are sets.
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN9
Unqualified Qualified
Clearly all A are in B, but we don’t know if b in B is in A.
We can map A to B but not B to A.
All A are in B, and we know what subset of B equals A.
We can map A and B bidirectionally (lossy).
Combined qualification examples:
B
BQ1=AB \ A
B
A=B?B \ A
B
BQ1=A1 BQ2=A2 BQ3=A3
A B
22 QQ BBAA AQ1=A \ B BQ1=B \ A
“Qualified” Relationships• A “qualified” relationship is one that hold under some known condition or
criteria.• Described as one or more attributes having certain values.• In the examples below, a qualification Q1 on concept B forms
a subconcept BQ1.
Both AGC and SEDRIS team schemes capture qualified relationships.
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN10
Example of “Qualified” Relationship
EDCS: WellQ1: well type = ‘Fountain’ NFDD: Fountain
equivalent
AGC Relation: “EDCS (Well) and NFDD (Fountain) concepts overlap completely (qualified)”
EDCS
NFDDEDCS: Well
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN11
AGC Relation: “NFDD (Railway) is an aggregate of EDCS (Railway track)”
EDCS
NFDD
GEMET
AGROVOC
WordNet
Integration and Linking of Dictionaries• Potential outcome: Integration of data dictionary concepts
– More than just mapping• Semantic alignment across multiple data dictionaries• Example: NFDD “railway” and EDCS “railway track”
Railway trackRailway
Underground railway
High-speed railway Railroad
Aggregate of
Equivalent
Railway network
Infrastructure Track
Railway
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN12
Using Mappings in Data Translation
Relation Class Diagram Visualization Mapping A to B? Mapping B to A?Concepts A and B overlap somewhatUnqualified
NO
Some elements map, but we don’t know which ones!
NO
Some elements map, but we don’t know which ones!
Concepts A and B overlap somewhatQualified
YES, WHEN APPROPRIATE
The qualification tells us which elements map.
YES, WHEN APPROPRIATE
The qualification tells us which elements map.
Concept A is a generalization of Concept BUnqualified
NO
Some elements map, but we don’t know which ones!
YES, ALWAYS
Subset relationship implies membership in superset A.
Concept A is a generalization of Concept BQualified
YES, WHEN APPROPRIATE
The qualification tells us which elements map.
YES, ALWAYS
Subset relationship implies membership in superset A.
Concept A is an aggregate of Concept B
A might imply existence of B B might imply existence of A
A B
BAA \ B B \ A
A
B
has partA B
A B
22 QQ BBAA AQ1=A \ B BQ1=B \ A
A
AQ1=BA \ B
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN13
Roadblocks: The same old problemsGarbage In Garbage Out
• With poor mappings, we get wrong data faster.• Weak semantics in data dictionaries beget poor
mappings.• Both EDCS and NFDD Concepts have:
– Short definitions.– No scoping or context statement.– No relationships to other Concepts (internal or external) to
capture the intended “world view”.• Perhaps NFDD and EDCS should be mapped onto
themselves first?– EDCS includes a partial taxonomy in its definitions, but can be
more precise.
Weak semantics in EDCS and NFDD perpetuate ambiguity.
UNCLASSIFIEDUNCLASSIFIED
14
Kevin [email protected]
Modeling & Simulation Information Management BranchSignal and Information Sciences LaboratoryApplied Research LaboratoriesThe University of Texas at Austin
Rapid Data Generation (RDG)
Rapid Data Generation
RDG
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN15
RDG Background• RDG is a High Level Task (HLT) selected by the DoD M&S Steering
Committee (M&S SC) for funding through the M&S Coordination Office PE to address M&S Enterprise Data issues
• Mr. Tom Irwin, Joint Staff (J7), and Dr. Amy Henninger, Army, are the M&S SC co-leads for governance of RDG
• Government PM was Mike Willoughby, JTIEC; replacement TBA• Performers are JS J7 JCW (MITRE & GDIT), University of Texas
Applied Research Laboratory, Oak Ridge National Laboratory and others
• Objective: Reduce the resources required to integrate and initiate data, eliminate or reduce duplicative efforts, and promote data commonality for M&S activities across the DoD.
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN16
RDG Summary• RDG implements the DoD Net-Centric Data Strategy (NCDS) by
making data – Visible – search via SOA services or a user interface– Accessible – access via SOA services– Understandable / Interoperable – described by structural metadata– Trusted – controlled access to data integrated from authoritative data
sources• RDG implements the DoD Net-Centric Services Strategy (NCSS) by
– making information and functional capabilities available as SOA services• RDG implements the DoD M&S Enterprise Data Strategy by
– Implementing the NCDS and NCSS for M&S data– Using the M&S Community of Interest (COI) Data Management Working
Group to gain stakeholder input
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN17
Rapid Data Generation
FY09/10HLT-IC2
Capability
HLT IC2 GFM JTDS
5 Year M&S Data Enterprise Investment Strategy
Enterprise ApproachSC Oversight
Metrics Immediate ProgressRequirements DrivenD P
D P
D P
Geospatial, Atmosphere, Space, Ocean
Logistics
Command & Control
Red & BlueOrder of Battle
L IFE CYCLE
MANAGEMENT
SC Governance, Community Participation, Cross-Doman Interoperability
OOB Mid Term
ExamOOB Final
Exam= SC Decision Points
D P
D P D P
D P = Development Planning
FY 11 FY 12 FY 13 FY 14 FY 15FY 10
Other Capability On/Off-ramps
Common Data Production Environment
Year of Funding
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN18
RDG M&S CDPE OOB Data Services Conceptual Overview(Draft Pre-decisional)
“Non-US” Force OOB Data Provider
RDGM&S CDPE CDPE
Discovery Metadata
Catalog
M&S Catalog
OperationalOOB DataProviders
(i.e. GFM DI, JPES/APEX, etc.)
Joint Training Data Services
(JTDS) OBS
USAF Scenario Generation Server (SGS)
US Special Operations Command
(USSOCOM)
CDPE Portal
Authentication/ Authorization ServiceOOB Discovery Service
OOB Subscription ServiceOOB Edit/Build Service
OSD/CAPE Joint Data Support
(JDS)
Discovery Metadata Update ServiceData Retrieval Service
Integrated Gaming System
(IGS)
USN Common Distributed
Mission Training Station (CDMTS)
Other M&S OOB Data Provider, Integrator, or
Consumer Systems
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN19
DATA ISSUESRapid Data Generation
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN20
Discovery, Retrieval, and Understanding
Discovery• Need to tag data products with
“discovery metadata” to enable visibility through search services.
• Specifically, need to tag data products containing Unit, Task Organization, and related data so they are discoverable based on– Unit identifiers and names– Unit types and capabilities– Major end-item equipment types– Mission, Scenario, garrison and
other contexts
Retrieval and Understanding• Need to support exchange
of data in multiple data formats, including incompatible ones.
• Need to define and align the semantics of the formats.
• Promote convergence of formats (or schema fragments) for Order of Battle-related data.
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN21
Metadata TypesDoD Directive 8320.02, “Data Sharing in a Net-Centric Department of Defense”
Discovery Metadata[Information about a data asset] that allows data assets to be found using enterprise search capabilities.
Structural MetadataInformation provided about a data asset that describes the internal structure or representation of a data asset (e.g., database field name, schemas, web service tags).
Semantic MetadataInformation about a data asset that describes or identifies characteristics about that asset that convey meaning or context (e.g., descriptions, vocabularies, taxonomies).
Descriptions about the content and context of the asset, including author, title, pedigree, source, media type, and more.
Schemas, grammars, and structures that data assets conform to.
The definitions, references, and models that define the meaning of data assets to capture intent and preclude misinterpretation. Typically tightly related to the Structural Metadata.
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN22
Relationship of Discovery Metadata to OOB Data
OOB Data Asset
Format / XML Schema
“Metacard” for Data Asset
• Stored in a metadata repository• Shared to catalogs for search and discovery• Conforms to either
• DDMS• MSC-DMS
• Augmented with• Ucore/C2 Core content for discovery• RDG extensions for OOB discovery
• Stored in a data repository• Tagged with a metacard• Conforms to some structure
metadata (format or structure).
• Stored in the DoD Metadata Registry (MDR)
• Tagged with a metacard• Conforms to some structure metadata
(format or structure).
Discovery Metadata
Structural Metadata
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN23
What is meant by “Order of Battle?”UNITS / ORGANIZATIONS‘SIDES’ Nations
CoalitionCivilianOPFOR
Perspective:• Authorized• On-Hand• Planned• Anticipated• Reported• Scenario• Organic / garrison• Task Organized (OPORD / FRAGO)
Scope / resolution:• Operational vs. Systems Architecture• Aggregated vs. entity-level• Contains network?• Contains readiness and holdings?• Contains locations?• Contains plans and orders?
Validated for purpose:• Acquisition• Analysis• Experimentation• Intelligence• Planning• Training• Test & evaluation
Verified for system needs:• C4I system initialization• C4I network initialization• Simulation and instrumentation initialization
OrganicAssignedAttachedOPCONTACONDirect supportReinforcingGeneral support-reinforcingGeneral Support
FriendlyHostileNeutral
PLATFORMS & LIFE FORMSLocationsC2 Network
Logistics Plans, orders, control graphics
Entity (unit, platform, and life form) type definitions
Platform / weapon / sensor composition
Application-specific detailsSymbols, icons,
3D models
Agent/Behavior models
Characteristics and Performance
P(hit), P(kill), P(detect), P(classify)
System environment
System data format
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN24
Making OOB Searchable
OOB Data Asset
“Metacard” for Data Asset
Discovery Metadata
Based on eitherDDMS
orMSC-DMS
Annotated with UCore content to support
IC/DoD CDR OpenSearch
orRDG OOB discovery
metadata extensionsUnits• Name• UIC & FMID• UTC• Symbol code• Echelon• Capabilities• Force relationships
Equipment Types• NSN• LIN• FMID
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN25
Order of Battle FormatsFormatGFM DI XML Joint Staff J8 GFM DIUCore DoD CIO / DISA, ODNI, DOJ, DHSC2 Core C2 DSSC / Joint Staff and DoD CIOMSDL Simulation Interoperability Standards
Organization (SISO)OBS XML Joint Staff J7 JCWArmy LDIF address books PEO C3TSIMCI/PD TNI XML PD TNI and Joint Staff J7 JCW
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN26
Global Force Management Problem Statement
We need Global Force Management DataCurrent Unit Locations“Event” data Operational AvailabilityTotal US InventoryHistorical archiveTimely, reliable, and
authoritative
What forces do I have?Where are the forces today?
What residual capability exists?How do I manage forces, manpower, & equipment from acquisition to end of
service?What happens if…?
GFM DI is the Department-wide enterprise solution that:1. Enables visibility/accessibility/sharing of entire DoD force structure2. Allows integration of data across domains and systems
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN27
Data from Org Servers exposed to the enterprise via NCES messaging
Standardized,
Authoritative
DataFeeder systems document authorizations in without enterprise-wide standards
6 Org Servers on NIPRmirrored and augmentedin 7 Org Servers on SIPR(Defense Intel only on SIPR)
GFM DI Task 1: Organization Servers
Feeder Systems
ARM
YAI
R FO
RCE
OrgServerAI
R FO
RCE
OSDN
AVY
MAR
INE
COPR
S
JOIN
T ST
AFF
Inte
l Com
mun
ityO
SD
Force Structure
DOD
USAFUSA USMCUSN
States
JOIN
T ST
AFF
ANG ARNGN
AVY
MAR
INE
CORP
SAR
MY
Raw Data
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN28
GFM DI: Document “Authorized” Force Structure as the Basis for “On-Hand” and “Execution”
What are you authorized?
Authorization dataAuthorized by Law and
organized by the Components
What do you actually have?
“On-Hand” dataProperty Books & Personnel Systems
What do you have to operate with and where is
it?
Execution dataReadiness, Logistics &
Personnel Systems
Org Servers ITAPDB, MCTFS, MilPDS, etc.
DRRS, JOPES etc.
GFM DI Task 1 GFM DI Next Steps Task 2 -- Service/User Systems
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN29Person
SSG Smith
Organizations & Authorizations
Gunner
Loader
Driver
M1 TK 1
A
E-6SSG19K3OASI: K4
M1A2
TK 4 M1
Tank Cdr
GFM DI Next Steps: Using OUIDs as Reference for Real Equipment, People, other IDs and Reorganizations
OUID
EDIPI
Military Force Tracking
URN, UIC, ...OUID
OUID
Real PropertyRPUID
OE: Organization ElementOUID: Organization Unique Identifier UII: Unique Item IdentifierRPUID: Real Property Unique IdentifierEDIPI: Electronic Data Interchange
Personal Identifier URN: Unit Reference NumberUIC: Unit Identification Code
OE
OE
OE Equipment
C-2
UII
Fort Hood
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN30
Example Format Utilization
JTDS OBS
Simulation systems
JDLM WARSIM / WIMSIMPLE OneSAF IGSJCATS
AR
MY
OSDNAV
YM
AR
INE
CO
PRS
JOIN
T ST
AFFAIR
FO
RC
E
Inte
l Com
mun
ityO
SD
AIR
FO
RC
E
Force Structure
DOD
USAFUSA USMCUSN
States
JOIN
T ST
AFF
ANG ARNG
NAV
YM
AR
INE
CO
RPS
AR
MY
GFM Org Servers
Other consumers and data integrators
OBS XML
Other sources
Other formats or unstructured
Army PD TNI DPDE
GFM DI XML
SIMCI / PD TNI XML
ABCS
Army LDIF, etc.
Army CADIE
MSDL
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN31
ENTITY TYPE DEFINITION ANDPARAMETRIC DATA
has-parts
has-BOIP
. . .
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN32
What are Entity Type Compositions (ETCs)?
• The “real world” / battle space (C2/Log) objects that must be accurately and consistently modeled across different simulations of a federation.
• Entity types are “compositions” of a base platform or person with associated
– weapon systems, – sensors, and – other (simulation-relevant) equipment.
• Examples of ETC names:
– M1A2 Tank– M1A2 with mine plow– M998 Cargo HMMWV– M1114 Armored HMMWV with Mk-19– Scout HMMWV with 50 CAL MG and LRAS3– Airborne Soldier with M4 rifle– Infantry Soldier with SAW
• Could include organization and facility types too.
M1114 HMMWV Up-Armored Armament Carrier
FBCB2/BFT
LRAS3M2 .50 CAL MG
M&S ETC Name : SCOUT HMMWV Armored 50 CALDIS Enum: 1-1-225-6-1-21-0
• Some ETC enumeration schemes:– SISO DIS enumerations– National Stock Numbers (NSNs)– Line Item Numbers (LINs)– US Army Standard Nomenclature– JLCCTC MRF enumerations– OneSAF enumerations
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN33
Every• LVC M&S federation, • individual simulation, • local M&S federation site
can have different definitions and namesfor the same “real world” ETC.
ETCs in Practice
ETCs relate to other data
ETC
PD TNIEvery Simulation Site
Service-level Force Management
Weapon / sensor effects evaluators
SISO EWG
Object Models
3D model Repositories
ETCs are managed in multiple places
ETC
Force Management
Platform properties Weapon
stations
Behavior Models
Characteristics and Performance Data
Federation Object Models3D Models
Scenario / Order of Battle
Logistics / Readiness
instance data reference data Logistics Databases
AMSAA
JTDS OBS
GFM ORG Servers
Foreign / intel databases OTC JLVC
JLCCTCARCIC
TRADOC
C2 / logistics simulation users
If ETC definitions are not aligned across the C2 and M&S enterprise,
• OOB data is not interoperable or reusable• C&P and PH/PK parameter data cannot be
published or consumed…without human-analyst intervention.
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN34
• Realization: ETCs are Classes.– Not just simple enumerations– ETCs are “sets of like things”,
corresponding to classes in the Web Ontology Language (OWL)
• OWL has class semantics “built in”– Subclass, restricted class, identifying
properties and relationships
• Use existing rules and tools– To avoid OWL is to redefine the same
semantics and software that is available today.
• Easier alignment of “enumerations” to other data standards:– MSDL– C-BML– JC3IEDM– C2 Core– RPR FOM– TENA LROMs
• We can now use existing OWL tools for basic editing of ETC knowledge-bases.
ETCs as OWL Classes
HMMWV M1114 w/ .50 CAL
Vehicle
Equipment
Aircraft
F/A-18M2A3
Bradley IFV
AC-130E
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN35
JC3IEDM and OWLOrganizing ETCs using JC3IEDM-based object-type taxonomy.
But JC3IEDM has three problems that had to be resolved first:1. Dual taxonomies for Object-Item and
Object-Type.– In OWL, they can be combined.
2. Flattening of class hierarchy using “category codes” to reduce table count.
– Not a problem in OWL, so we fleshed out the full class hierarchies.
3. Only supports single-inheritance– Many of JC3IEDM’s conflicts can now be cleaned up by
reconnecting the multiple inheritances.– e.g., Fox M93A1 – is it a Vehicle or a CBRN equipment?
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN36
RDG PLAN FOR OOB MODELS & FORMATS
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN37
RDG Concept for OOB Formats1. Support what exists today: Enable exchange of any existing or future
data format.(in accordance with IC/DoD Content Discovery and Retrieval (CDR) Retrieve specifications)
2. Define a common, extensible OOB logical data model (LDM) and format to be a managed union of existing data requirements.
a. Start with GFM DI XML as a “common core” and extend; align with UCore and C2 Core efforts
b. Require data providers to support the common OOB format (in addition to any legacy format, optionally)
3. Leverage Entity Type management efforts 4. Align M&S to C2 and logistics representations and data sharing solutions.
Work to converge solutions, where appropriate.
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN38
XML
Principles of OOB LDM• Goal is to support GFM DI XML,
OBS XML, MSDL, Army LDIF, etc. content completely.
• Enable dynamic extensibility to support future data exchange requirements without imposing schema changes to established CDPE producers or consumers.
• Recognize that there are more than one valid way of viewing and modeling the world: structures, resolution, dimensions.
• Define foundation for aligning semantics for disparate formats, schemas, and data requirements.
• Enable more automated data translation, and quantify lossiness.
• Stop inventing ambiguous, unnecessary M&S corollaries to real-world concepts.– Align to operational semantics:
architectures, data models, doctrine, vocabularies, taxonomies, etc.
– Coordinate activities with GFM DI, UCore, etc.
XML
OWL
XML
XMLXMLXML
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN39
Creation of OOB LDM for RDG1. Reverse engineer grammars / XML formats into OWL.
2. Construct modular composed ontologies
XSDOWL
XSDOWL
XSDOWL
GFM DI XML
SIMCI / PD TNI XML
OBS XML
XSDOWLOther formats
GFM DI
PD TNIMSDL
OBS
Other models
• UCore / C2 Core• DIS Enums• Logistics sources• C-BML• NFDD / EDCS
……
Drafts complete
In progress…XSD
OWLMSDL
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN40
OOB LDM Elements
OBJECT-TYPE
FACILITY-TYPE
MATERIEL-TYPE
ORGANISATION-TYPE
PERSON-TYPE
OBJECT-ITEM
FACILITY
MATERIEL
ORGANISATION
PERSON
ADDRESS
ELECTRONIC-ADDRESS
PHYSICAL-ADDRESS
…
……
…
OBJECT-ITEM-ADDRESS
OI-ALIAS
ALIAS-TYPEOBJECT-ITEM-ASSC
OBJECT-ITEM-TYPE
OT-ESTABLISHMENT
DETAIL ESTAB-ALIAS
…
Started with GFMIEDM v3.5 …
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN41
OOB LDM Elements
OBJECT-TYPE
FACILITY-TYPE
MATERIEL-TYPE
ORGANISATION-TYPE
PERSON-TYPE
OBJECT-ITEM
FACILITY
MATERIEL
ORGANISATION
PERSON
ADDRESS
ELECTRONIC-ADDRESS
PHYSICAL-ADDRESS
…
……
…
OBJECT-ITEM-ADDRESS
OI-ALIAS
ALIAS-TYPE
OBJECT-ITEM-ASSC
OBJECT-ITEM-TYPE
OT-ESTABLISHMENT
DETAILESTAB-ALIAS
…
Extended to also support OBS v3 …
LOCATION
LINE
OBJECT-ITEM-LOCATION
POINT…
SIDE
SCENARIO
PLATFORM
NETWORK-MEMBER
ABCS-COMPONENT
FACTIONDIS-CODE
OWNING-FEDERATE
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN42
Other models to fold in…• OGRE/JACOB• MIDB• TRAC - Paul Works, Lee Lacy and Dean Hartley
are developing an ontology for irregular warfare
• Army PD Tactical Network Initialization• Coalition Battle Management Language
UNCLASSIFIED
APPLIED RESEARCH LABORATORIES
THE UNIVERSITY OF TEXAS AT AUSTIN43
QUESTIONS?
19 APRIL 2012
Kevin [email protected]
Modeling & Simulation Information Management BranchSignal and Information Sciences LaboratoryApplied Research LaboratoriesThe University of Texas at Austin