Synthetic AWACS Weapons

1
Synthetic AWACS Weapons Director
S. David Kwak Lockheed Martin 781-505-9545 kwak_at_la
ds.is.lmco.com
2
SAWD Demo

• Objective
• Training Support Man Power Reduction
• Operational Approach
• SAWD automates AWACS weapons director operations.
• Is fully adaptive to new situation
• Does not require scripted scenario
• Facilitates real time C4ISR injection
• SAWD controls/collaborates AirSF agents
• Anticipates the battle scenes (both BLUE and
  OPFOR)
• Guides the AirSF agents
• Monitors the engagement progress.

Synthetic AWACS Weapons Director
3
SAWD Demo Technical Approach

• Break the mold
• The known restrictions (high cost and poor higher
  level behaviors) of present CGF's are not really
  intrinsic to CGF.
• A new approach to CGF technology can provide
  command level behaviors and also interface to the
  existing CGF installed base .
• Prove concept of a new enabling technology
  COREBA
• COREBA means Cognition ORiented Emergent Behavior
  Architecture.
• COREBA facilitates construction of human-like
  behaviors through its unique hybrid paradigm.
• COREBA requires significantly less development
  efforts.
• Construct Synthetic AWACS Weapons Director (SAWD)
  demo
• SAWD controls/coordinates many Air Synthetic
  Forces (AirSF) agents.
• SAWD communicates with AirSF by exchanging radio
  messages.

4
SAWD Demo Configuration
Synthetic AWACS Weapons Director COREBA Agent
Rules
Human friendly
Rational Reasoning
Strategic

Policy, Doctrine, etc

Rules of thumb
S
Emergent Patterns
Human friendly
Tactical

Behavior
T

Decisions
Common Sense

Information Fusion

Collaboration
E
Sensors and Actuators
Interfacing to
Execution

Physical world
Reactions

Other systems

Other COREBA agents
5
l
Scenario for COREBA/Soar Tech Demo during Coyote
98
8
16
X
4
A
0
Red
Blue
88
4
Y
4
B
1
88
4
4
4
Z
C
2
75
75
50
50
125
16
6
6
0
2
6
10
20
45
36
14
23
16
38
12F15 blue Sweep
Counter Attack Launch 16 orange
ENDEX
16 red
Red Blue Battle
Green Border Cross
4 red
Kill ROE Change
Green/blue Border Cross
Blue Orange Battle
VID
Flush 12 blue
Launch 16 Green 16 Blue
6
What You Saw in the Demo

• SAWD makes many AirSF agents collaborate by
• Anticipating the battle scenes (both BLUE and
  OPFOR)
• Paring and Guiding the AirSF agents
• Monitoring the engagement progress.
• SAWD automates AWACS weapons director operations.
• Is fully adaptive to new situation
• Does not require scripted scenario
• Facilitates real time C4ISR injection
• Reduces man power requirement for AWACS operation
  in support of pilot training
• SAWD was developed in two staff month effort.
• The openness of COREBA facilitates integration
  with existing AirSF.
• COREBA technology significantly reduces
  development cost and time.

Automated AWACS Operations
7
SAWD in the Demo

• Uses perceived reality detected by AWACS radar
• Provides the basis for SAWD's cognitive
  operations
• Guides AirSF by sending and receiving radio
  messages
• Continually performs predictive assessment of
  paired engagement
• Anticipates the battle scene for pairing
• Uses both side tactics and policies
• Force ratio
• Considers aircraft/weapon capabilities and roles
• Includes known intention

8
SAWDs Anticipation
Reds engagement range
Anticipated interception point based on ordinary
linear extrapolation concept
Blues engagement range
9
SAWDs Anticipation
10
SAWDs Anticipation
11
SAWDs Anticipation
Interception point anticipated by SAWD
12
SAWDs Anticipation
Interception vector generated by SAWD
13
COREBA SAWD Agent Instantiation
COREBA SAWD Agent
Genetics Algorithm
NASA CLIPS Rules Rule based system
Learning Discovery
S
Santa Fe Institute Swarm Objects Complex Adaptive
System
T

• Other Systems
• Soar Agents
• JointSAF
• Ordinance Server

Objective C Objects Object Oriented System
E
14
How does SAWD guide Soar Agents?
COREBA SAWD Agent
Soar Agents (40 entities)
JointSAF (52 entities)
Meta rules (defrule set-force-ratio )
S
S
T
T
Send/Receive Messages
E
E
Inside of SAWD

• Radio Messages
• (Outgoing)
• Judy
• Bogey-dope
• State-message
• On Station
• Off Station (RTB)
• Enroute to station
• Confirmed bandit
• Confirmed friend

• Radio Messages
• (Outgoing)
• Vector and Kill
• What State
• Tactical Control
• Launch Alert
• VID
• Kill Bogey
• Broadcast control
• Close control

DIS 2.0.4 PDU Packets
Computing resource
100 BaseT Ethernet
15
Example Meta Rule and Force Parameters
16
Why does SAWD exhibit human-like behavior
characteristics?
Policies
Govern behavior generation
Conventional Discretized Behavior Approach
17
COREBA Behavioral Characteristics in Behavior
Space
Behavior Quality Improvement (Learning)
Global Behavior Trajectory
Locally Interacting Behavior Trajectory (Adaptiven
ess supported by emergence)
Rationality Boundaries
18
Thus, COREBA Architecture

• Is a Multiple Paradigm Behavior Architecture
• Rule based paradigm
• Complex adaptive paradigm
• Object oriented paradigm
• Evolutionary computing paradigm
• Facilitates implementation of
• Policy driven behaviors
• Cognitive behaviors
• Adaptive behaviors
• Human-like automated behaviors
• Man power reduction
• Is an Open/Nested Architecture
• Promotes reuse
• Reduces development cost

COREBA
Rational
Cognition Oriented
Learning
Adaptive
Emergent