The Game of War:
Military Simulation and Game Development
Jennifer Sandercock
Michael Papasimeon
Air Operations Division, DSTO
• Defence Science and Technology Organisation
Department of Defence
• Conduct research into air operations
to help the Australian Defence Force
(usually the Royal Australian Air
Force)
• Make smart acquisition decisions on
new systems
• Make best use of existing systems
(tactically, operationally etc.)
• Military Simulation is one of our
primary tools.
Scope and Purpose
• Raise awareness of Australian
military simulation activities amongst
Australian game developers.
• Raise awareness of Defence
simulation activity in Melbourne.
• Explore the potential for future
collaboration.
• Focus on particular areas, will not be
exhaustive.
Games and Military Simulations
• Similarities
• Technology
• Software Design
• Construction
(Software Engineering)
• Some common skill sets.
• Differences
•
•
•
•
Purpose
Emphasis
Funding
Social
Purpose of Military Simulation
MILITARY
SIMULATION
ION
USIT
ACQ
DE
IN L PLOY
M
IFE
SUP ENT
POR
T
DESIGN
H
M
G R
IN O
N CT
AI FA
TR N
A
U
PS EN
O
N PM
CO LO
VE
E
D
T
S
Example Simulations
• Human In The Loop
• Flight Simulators (e.g. AOSC Dome)
• Training
• Human Factors
• Constructive and Wargaming
•
•
•
•
Operational Analysis for Acquisition,
Tactics Development,
Concepts of Operations
Experimentation
• Campaign
• Strategic
• Sortie Generation
• Force-Element Mix
Military Experimentation for AEW&C
Types of Simulation
103 -104 km2 , 10-1000+ entities, days- months
Strategic
CAMPAIGN
102 -103 km2 , 2-32+ entities, hours
Operational
MISSION
< 102 km2 , 2-4 entities, minutes
Tactical
ENGAGEMENT
1 entity
ENTITY
COMPONENT
(ENGINEERING)
A Flavour Military Simulation in DSTO
• F/A-18 Air Combat (“Top Gun”)
• SWARMM and XCombat
• AEW&C Acquisition
• Air Operations Simulation Centre (AOSC) Dome
• Synthetic Environment Research Facility (SERF)
Tactics for F/A-18 Hornets
F/A-18 Fighter Pilot AI
Example: Cognitive Modelling (AI)
• Ingelligent Agents
• Situational Awareness
• The BDI Reasoning
Model
• BDI Languages
• dMARS
• JACK
• Graphical
Representation
• The Melbourne Link
BDI Agent Mechanisation
Plan Library
Observe
External
Information
Plan
selection
Orient
Current
beliefs
Plan A
Plan
BC
Plan
Plan ..
How to perceive
act
communicate
tactics etc..
Plan x
Plan
yn
Plan
Plan ..
‘To do’ list
Decide
Prioritisation
Act
Plan execution
External
Action
Pilot Agent Architecture and Cognitive Model
Pre-briefed (mission) data
Data from the environment
Situation Awareness
Situation Assessment
Computational
representation
of the OODA Loop
Tactics Selection
Standard Operating
Procedures
Actions
Agent Architecture (detail)
Present / Reactive
Balancing & Management
Future / Proactive
Situation
Awareness
Situation Memory
Situation
Prediction
Situation
Assessment
Situation
Management
Future
Assessment
Tactics
Nomination
(reaction)
Tactics
Management
Future Tactics
Nomination
(Planning)
Implement Tactic
Distributed Simulation
• Distributed Mission
Training
• Example: The Virtual Air
Environment
• DIS/HLA Standards
• Interoperability between
simulators that were not
originally designed to work
together.
Some things military simulations have…
•
•
•
•
•
•
•
•
Artificial Intelligence and Agent Technology
Simulation Architectures and Design Patterns
Complexity
Sophisticated Distribution
Credibility and IV&V
Standardisation
Complex Analysis and Scenario Setup Tools
Commercial heavy duty software engineering
Some things games have…
• Cool and easy to use user-interfaces
Some things games have…
• Cutting Edge Graphics and Visualisation
Some things games have…
• Cutting Edge Sound
Chess: Real, Virtual and Mixed
Human Chess Players
Human Playing Against Chess Computer
Human Playing Against
Robotic Chess Simulator
Chess Playing Agent Versus another
simulated Chess Playing Agent
AP-3C Maritime Patrol: Real, Virtual and Mixed
Real Crew in Real AP-3C
Real Crew in AP-3C
Tactical Simulator
Real Crew in AP-3C Simulator
Virtual Crew in OR Simulator
Classifying Games
• Methods:
•
•
•
•
•
•
•
•
Genre
Complexity
Number of Entities
Number of Interactions
Time scale (minutes vs. hours vs. days)
Levels
AI
Graphics
Complexity?
Civilization
Sim City
CAMPAIGN
The Sims
MISSION
Unreal Tournament 2004
ENGAGEMENT
Pinball
ENTITY
Grand Turismo
COMPONENT
(ENGINEERING)
Getting Games to Play with
Military Simulations
• Should be able to:
•
•
•
•
•
•
Be validated
Plug & play for models
Control game/simulation time
Run for extended periods of simulation time
Have access to API documentation
Modify look & feel of game
Current Active and Relevant Research
• Representing teams and team tactics
• Mixing teams of virtual and real military operators
• Representing virtual environments for agents and
other A.I
• Emotion Moderators in Agents
• Strategy and Tempo of War
• Speech Recognition for Controlling Agents
Existing Cooperation
• United States
• Full Spectrum Warrior & Full Spectrum Command (ICT,
US Army, Pandemic, Quicksilver)
• America’s Army (MOVES Institute, US Navy)
• Australia
• Turing Tests of FPS games (DSTO)
• Modifying FPS for Army soldiers (ADFA)
Opportunities for Interactions
• Conferences
• Local: AGDC, SimTecT,
• International: GDC, IJCAI, AAMAS, OzCogSci
• Books and Magazines
• Game Developer, Gems Series, AI Wisdom, AI Magazine,
IEEE, ACM
• Associations
• IGDA (Melbourne Chapter), Agents-VIC
Groups to Interact With
• Academia
• Agents – The University of Melbourne, RMIT
• Government and Defence
• Defence Science and Technology Organisation (DSTO)
• Australian Defence Force Academy (ADFA)
• Local Technology Companies
• Agent Oriented Software (AOS)
• Agentis
Questions?
Descargar

Document