AGENT MEDIATED INTERFACE BETWEEN C4I AND SIMULATION

1
AGENT MEDIATED INTERFACE BETWEEN C4I AND
SIMULATION
Zach Furness The MITRE Corporation 7515 Colshire
Drive McLean, VA 22102 [phone] Fax (703)
883-1370 zfurness_at_mitre.org
Ray Emami Global InfoTek, Inc. 1920 Association
Drive, Reston, VA 20191 [phone] Fax
[phone] gemami_at_globalinfotek.com

• Ranjeev Mittu
• Naval Research Laboratory
• 4555 Overlook Avenue
• Washington, DC 20375
• [phone]
• Fax [phone]
• mittu_at_ait.nrl.navy.mil

2
INTRODUCTION

• The ability to interface Command and Control
  systems, with simulations, represents a powerful
  approach in the ability to analyze military
  plans.
• The simulations provide a context of the real
  world in which the plan can be exercised,
  what-ifs can be performed, and intelligent
  courses of action can be generated.
• To add more intelligence to our COA, we must be
  able to decompose the plans, so that critical
  events and actions/consequences can be
  understood.
• Through this understanding, we can begin to
  effectively monitor how the simulated plan may be
  deviating from the original plan (particularly
  the critical actions/events).

3
SIMULATION-AIDED EXECUTION MONITORING

• It is through Plan understanding that we will be
  able to generate intelligent what-ifs and
  courses of action.
• Plan Understanding will allow us to decompose and
  comprehend the relationships in the plan, and how
  they affect the outcome.
• Most operational analyses using simulation
  (planning, course-of-action analysis) are run in
  non-real time
• However, using simulations for execution
  monitoring, in real-time, to alert commanders to
  potential changes in plans could have significant
  benefits
• Once the plan is understood, Plan Monitoring will
  allow us to compare how the simulation of the
  plan relates to the original plan.
• Plan monitoring can reveal deviations.
• These can be measured and alerts can be issued
• These alerts may warrant what-ifs and exploration
  of COA.

4
POTENTIAL BENEFITS OF AGENTS FOR SIMULATION BASED
EXECUTION MONITORING

• The ability to understand monitor crucial
  events in the simulation and compare them to the
  original plan is a tedious and time-consuming
  task, which will require acute human attention.
• With the complexity of military plans, even the
  most attentive user will be prone to making
  mistakes.
• Software agents can be programmed to work
  autonomously, coordinating with each other
  through agent communication languages to
  accomplish the overall objective.
• The primary operational benefits of using
  software agents
• A reduction in the need for constant
  man-in-the-loop involvement between C4I systems
  and simulations
• A more cost effective simulation based training.

5
WHAT ARE SOFTWARE AGENTS?

• Agent-aided information retrieval and decision
  support has attracted the attention of the agent
  research community for several years.
• The concept of large ensembles of
  semi-autonomous, intelligent agents working
  together is emerging as an important model for
  building the next generation of sophisticated
  software applications.
• This model is especially appropriate for
  effectively exploiting the increasing
  availability of diverse, heterogeneous, and
  distributed on-line information sources, and as a
  framework for building large, complex, and robust
  distributed information processing systems.
• Software agents have been deployed in many
  domains, ranging from the commercial, academic to
  the military domains.

6
DEMONSTRATION CONCEPT

• Simulation based execution monitoring will draw
  upon a multidisciplinary set of technologies.
• This concept will be demonstrated using existing
  interfaces (with planned modifications) between
  GCCS and ITEM, and the CoABS agent grid.
• GCCS-NSS is an application, developed by the
  Defense Modeling and Simulation Office (DMSO) and
  NRL that allows information resident in the GCCS
  Track Database Manager to automatically populate
  the Integrated Theater Engagement Model (ITEM)
• The Control of Agent Based Systems (CoABS) Grid
  is a DARPA-developed application that provides a
  means for integration of different end-user
  applications using software agents.
• We will also leverage and build upon the software
  agent infrastructure.

7
DEMONSTRATION CONCEPT
8
DEMONSTRATION CONCEPT(GCCS/ITEM/NTMF)
9
GLOBAL COMMAND AND CONTROL SYSTEM MARITIME
(GCCS-M)

• GCCS incorporates the force planning and
  readiness assessment applications required by
  battlefield commanders to effectively plan and
  execute military operations.
• Its Common Operational Picture correlates and
  fuses data from multiple sensors and intelligence
  sources to provide warfighters the situational
  awareness needed to be able to act and react
  decisively.
• It also provides an extensive suite of integrated
  office automation, messaging, and collaborative
  applications.
• The GCCS HLA Ambassador is developed under DMSO
  funding
• It interfaces with the TDBM database to publish
  tracks to the HLA RTI

10
GCCS AMBASSADOR

• The Ambassador provides a direct linkage between
  the Track Database Manager (TDBM) and the HLA
  Runtime Infrastructure
• Developed by NRL as part of a DMSO HLA experiment
  on Simulation-C4I interoperability
• Relies upon standard simulation and C4ISR
  architecture standards, including HLA and DII COE
• Has been using to populate the TDBM with
  simulation data..
• JTLS-GCCS
• NSS-GCCS
• Pegasus-GCCS
• .And to initialize simulations based on data
  already resident in the TDBM
• GCCS-NSS
• GCCS-ITEM

GCCS Workstation
GCCS COP (application)
TDBM API
socket

TDBM libraries
GCCS Ambassador
RTI API
socket
TDBM
RTI 1.3 libraries
DII COE Kernel
TCP/IP
RTI-based FOM data exchanges over LAN or WAN
11
GCCS COA INITIALIZATION USING ITEM
Purpose - Provide automated initialization of
Course of Action (COA) simulations from
GCCS Partners - USFK, SPAWAR, NWC (Users) -
MITRE (Integration and Test) - NRL (GCCS
Ambassador) - SAIC (ITEM) Highlights - Reduces
initialization time for COA simulations from over
1 hour to a few minutes - Reduces data entry
errors due to manual input - Technical solution
based on DoD standard SW and Simulation
architectures (HLA and DII COE) Status -
Successfully demonstrated with NSS at Global 01
exercise (NWC) - Planned to be used in exercises
RSOI (Korea) and UFL (Korea) during spring and
summer 2002
OLD WAY Manual Initialization of Course of
Action (COA) Simulations
Simulation Wargame
COA Simulations (ITEM)
C4I System (GCCS)
ALSP
MDST
TACSIM
CSSTSS
RESA
CBS
HLA RTI
COE
NEW WAY Automated Initialization
COA Simulations (ITEM)
12
THE NAVAL TRAINING META-FOM (NTMF)

• The Naval Training Meta-FOM (NTMF) is being
  developed by Naval Modeling and Simulation (MS)
  trainer experts in response to the DoD MS vision
  of interoperability and consistency.
• The NTMF is focused on providing the mechanism to
  deliver a synthetic battlespace representation
  that is interoperable and consistent for use by
  Navy and Marine Force trainers.
• The primary goal of the NTMF is to provide a
  means to facilitate meaningful interoperability
  of Naval simulation/stimulation/C4I training
  systems in a consistent manner that supports
  stated training requirements and objectives.
• The NTMF will support many aspects associated
  with an OPLAN (including OPTASK messages and
  ATOs, etc).

13
DEMONSTRATION CONCEPT (CMDR AND GRID)
14
CRITICAL MISSION DATA OVER RTI (CMDR)

• CMDR is a tool for developing HLA RTI compliant
  applications.
• The software is a java library designed to enable
  developers to quickly federate HLA compliant
  simulation systems. It provides a general
  framework for interacting with the RTI.
• Some of CMDRs capability include
• Maintaining a database or internal representation
  of remotely simulated objects and their current
  states
• Managing the transmission of attribute updates
  for locally simulated objects
• Converting between raw data formats and actual
  objects
• Supports concept of Agile-FOM

15
CMDR PLUG-IN ARCHITECTURE

• Software plug-ins extend basic capability (via
  java introspection and reflection) through the
  network
• Plug-in process
• Application queries grid to locate services
• User queries the service to learn more, including
  model data, etc
• Application can locate and download plug-in to
  interoperate with service/models

16
CMDR MODEL REGISTRATION

• Dynamically discover advertised network services
• Advertisement consists of SOM and other meta-data
• Process includes
• Meta-data is created and registered on the grid.
• Plug-ins are downloaded to interact with service

17
DARPA CONTROL OF AGENT BASED SYSTEMS PROGRAM

• One of DARPAs largest program dealing with
  software agents approximately 2 dozen
  universities and companies
• The CoABS agent grid provides the foundation for
  integrating multiple, heterogeneous software
  agents with end-user applications. It is built
  on top of the Jini technology.
• The agent grid provides a means for software
  agents to advertise their capabilities, look up
  services and subscribe to these services. It
  also supports message logging and security
  services to provide authentication and message
  encryption.
• The grid will allow the software agents to
  discover simulations on the HLA RTI, as well as
  provide a mechanism for the software agents to
  subscribe to the FOM updates/interactions for
  monitoring purposes via CMDR

18
DEMONSTRATION CONCEPT(AGENT INFRASTRUCTURE)
19
AGENT INFRASTRUCTURE

• The complex nature of the problem is seen in the
  tables
• We are initially attempting to work with fairly
  structured information and well known ontological
  descriptions
• The challenge lies in working with ontology that
  are dynamic and planning information that is
  unstructured or semi-structured.
• Ontology negotiation
• Free text recognition

increasing complexity as we move along the arrow
20
PLAN UNDERSTANDING AGENTS

• Develop and build upon ontology efforts including
  the use of Darpa Agent Markup Language (DAML),
  Resource Description Framework (RDF) and/or
  extensible Markup Language (XML) for capturing
  problem domain associated with structured plan
  data.
• Task Ontology
• Spatial Event ontology
• Relationship ontology
• Temporal ontology
• DAML is built on top of the W3C Extensible Markup
  Language (XML), Resource Description Framework
  (RDF) and RDF Schema.
• Plan-understanding agents will use the concepts
  defined in these ontologies to look for
  relationships in spatial/temporal events
  associated with the tasking.

21
PLAN MONITORING AGENTS

• Uses the results from the interpretation of the
  planning information and subsequent relationship
  evaluation (via plan understanding agents),
• The individual events that "relate" the many
  plans together can be monitored in simulation for
  potential deviations.
• Allows cause-effect relationships to be
  identified in the simulation.
• More efficient Courses of action generation
  because a plan repair action can be compared
  against whether relationships identified by
  plan-understanding agents are violated.

22
AGENT BENEFITS

• Agents can understand multiple ontological
  descriptions
• Better distributed computing solution
  dynamically integrate new agent functionality.
• Agents can understand multiple inter-related
  descriptions
• Ontology negotiation to on-the-fly understand
  new meanings
• Early promising results in academia
• Agent teamwork models and theories
• One can imagine within this architecture how
  teams of agents with varying capabilities are
  able to decompose various aspects of the plans,
  monitor those plans, and perhaps even aid in
  repairing the plans based on the outcome of the
  simulated results.
• These teams of agents may converse with other
  teams of agents (or even teams of users) with
  differing ontological knowledge and negotiate
  meanings of information in reaching their goals
  in the process of performing courses of action
  analysis.
• As with ontology negotiation techniques, there
  are many research topics to be addressed in
  making this a reality, but the basic research is
  being conducted in this area and is being pushed
  to solve practical problems.

23
POTENTIAL FUTURE WORK

• Additional topics for investigation
• Federating additional simulations
• NETWARS is an HLA compliant simulation that
  provides the capability to analyze communications
  effects on the battlefield.
• An effort to link NETWARS and ITEM in FY03 for
  purposes of conducting synchronous planning is
  scheduled to be sponsored by DMSO.
• This will allow the effects of the communication
  infrastructure to be taken into consideration
  during development and refinement of an OPLAN
  that is being generated using ITEM.
• Integrating additional agent-based products
  emanating from the CoABS program
• Agent teamwork theories and models are of
  particular interest in order to support the
  capability of teams of agents (perhaps with
  different ontological representations) working
  together to decompose and monitor plans, and
  propose COA solutions based on individual and
  team goals.
• Expanding the capabilities of the
  plan-understanding and monitoring agents.
• Interact with systems that have a more formal
  specification for plans

24
REFERENCES

• 1 Web Address http//coabs.globalinfotek.com
• 2 Furness et al. Real-time Initialization of
  Planning and Analysis Simulations based on C4ISR
  System Data. In Proceedings of the 2002 Command
  and Control Research and Technology Symposium
  (CCRTS), Monterey, CA 11-13 June 2002.
• 3 Nielsen, J., Salisbury, M., Challenges in
  Developing the JTLS-GCCS-NC3A Federation, in
  proceedings of the 1998 Fall Simulation
  Interoperability Workshop (SIW), Orlando,
  Florida, 14-18 September 1998.
• 4 Lutz, R., Salisbury, M., Bidwell, G., A
  Demonstration of C2I System-to-Simulation
  Interoperability The NSS/GCCS-M Federation, in
  proceedings of the 1999 Fall Simulation
  Interoperability Workshop (SIW), Orlando,
  Florida, 12-17 September 1999.
• 5 Ogren, J., Command and Staff Training and
  the Practical use of the HLA, in proceedings of
  the 2000 Fall Simulation Interoperability
  Workshop (SIW), Orlando, Florida, 17-22 September
  2000.
• 6 Prochnow, D., King, R., Harrington, J.,
  Regala, B., Sonnenshein, J., Daly, J., Womble,
  J., The NSS-GCCS Federation Course of Action
  Analysis (COAA) Using a C4I-Simulation
  Interface, in proceedings of the 2001 Fall
  Simulation Interoperability Workshop (SIW),
  Orlando, Florida, 10-14 September 2001

25
REFERENCES

• 7 Prochnow, D., Harrington, J., Walter, D.,
  Womble, J., Automate Initialization of an
  Analysis Simulation With GCCS Track Data, in
  proceedings of the 2002 Fall Simulation
  Interoperability Workshop (SIW), Orlando,
  Florida, 9-13 September 2002.
• 8 Layman, G., Daly, J., Furness, Z., Womble,
  J., "C4I-Simulation Interoperability Using the
  HLA and DII COE", in proceedings of the 2001
  Command Control Research and Technology Symposium
  (CCTRS), Annapolis, MD, 19-21 June 2001.
• 9 Mittu et al., A Case Study for the Naval
  Training Meta FOM Analyzing the Requirements
  from MAGTF FOM. 7th International Command and
  Control Research and Technology Symposium, Quebec
  City, Canada. 9-12 Sept. 2002.
• 10 Web Address http//www.daml.org
• 11 Web Address http//www.w3c.org
• 12 Tamma et al, An Ontology for Automated
  Negotiation. In Proceedings of the International
  Workshop on Ontologies in Agent Systems. AAMAS
  02 Conference, Bologna, Italy
• 13 Web Address http//www.disa.mil/tis/netwars.