09-07-2012, 09:49 AM
A Feasibility Study using Intelligent Software Agent Technology for Weapon-Target Pairing
[attachment=27201]
Introduction: Intelligent Software as a Paradigm Shift.
A very powerful new type of information systems technology is rapidly emerging, driven by
government and commercial needs for expert decision-support and knowledge management. One
very apparent result of this technology is increasingly intelligent software systems. Computer
programs with collaborative agents that are capable of automatically reasoning about data and the
dynamic changes in data that occur in real world decision-making situations are already in use by
the military and are now transitioning to the commercial world.
It can be argued that our human view of computer software has been shortsighted in respect to two
popular notions: first, that data and information are essentially synonymous terms; and, second,
that computer intelligence is largely a misnomer because computers are machines. Neither of these
notions is accurate. While we human beings are able to convert data (i.e., numbers and words
without relationships) automatically into information due to the experience (i.e., context) that is held
in our cognitive system, computers do not have the equivalent of a human cognitive system and
therefore store data simply as the numbers and words that are entered into the computer. For a
computer to interpret data it requires an information structure that provides at least some level of
context. This can be accomplished utilizing an ontology of objects with characteristics and a rich set
of relationships to create a virtual version of real world situations and provide the context within
which agent logic can automatically operate.
Purpose and Structure of Report
The purpose of this report is to explore the software needs of unmanned vehicles and investigate
whether intelligent software agent technology incorporating artificial intelligence methodologies
would be able to satisfy some or all of these needs.
Section 2 briefly defines the different types of unmanned vehicles, describes their principal military
application areas, and summarizes the existing functional limitations that have been identified in
recent US Department of Defense studies. The authors believe that a strong case is made at the end
of this Section for the highly beneficial application of intelligent agent technology to unmanned
vehicles. Sections 3, 4, 5, and 6 provide an explanation of information-centric, service-oriented
architecture concepts, and more detailed descriptions of the principles of intelligent agents, web
services, and interoperability bridges, respectively. Section 7 describes a demonstration of
automated object interpretation as an example of just one application of intelligent agents to
unmanned vehicles.
Technical Challenges Identified by DoD
The most critical near-term technical challenge facing unmanned vehicles is the development of an
autonomous capability to detect, assess and respond to near-field objects in their path of travel. For
a UAV the near-field can extend to several nautical miles in all directions, while for a UGV or
UUV the distance at which an object must be detected is likely to be less than 100 yards.
Existing Intelligent Software in Military Domains
Ontology-based software applications with rule-based agents have been successfully applied in both
tactical command and control (C2) and logistical planning and execution since the late 1990s.
Examples include the Integrated Marine Multi-Agent Command and Control System (IMMACCS)
developed under sponsorship of the Marine Corps Warfighting Laboratory, which was successfully
tested in 1999 as the C2 system of record in the Urban Warrior Advanced Warfighting Exercise
(Pohl et al. 1999). Subsequently, in 2002, IMMACCS was evaluated by two independent teams
during a Limited Technical Assessment (LTA) to determine its ability to maintain a common
tactical picture in a wireless communication environment operating within extremely limited
bandwidth conditions (SPAWAR 2002). It was found that an information-centric4 software system,
such as IMMACCS, requires less bandwidth than a data-centric system in which there is no internal
understanding of the meaning of the data being exchanged and processed.
[attachment=27201]
Introduction: Intelligent Software as a Paradigm Shift.
A very powerful new type of information systems technology is rapidly emerging, driven by
government and commercial needs for expert decision-support and knowledge management. One
very apparent result of this technology is increasingly intelligent software systems. Computer
programs with collaborative agents that are capable of automatically reasoning about data and the
dynamic changes in data that occur in real world decision-making situations are already in use by
the military and are now transitioning to the commercial world.
It can be argued that our human view of computer software has been shortsighted in respect to two
popular notions: first, that data and information are essentially synonymous terms; and, second,
that computer intelligence is largely a misnomer because computers are machines. Neither of these
notions is accurate. While we human beings are able to convert data (i.e., numbers and words
without relationships) automatically into information due to the experience (i.e., context) that is held
in our cognitive system, computers do not have the equivalent of a human cognitive system and
therefore store data simply as the numbers and words that are entered into the computer. For a
computer to interpret data it requires an information structure that provides at least some level of
context. This can be accomplished utilizing an ontology of objects with characteristics and a rich set
of relationships to create a virtual version of real world situations and provide the context within
which agent logic can automatically operate.
Purpose and Structure of Report
The purpose of this report is to explore the software needs of unmanned vehicles and investigate
whether intelligent software agent technology incorporating artificial intelligence methodologies
would be able to satisfy some or all of these needs.
Section 2 briefly defines the different types of unmanned vehicles, describes their principal military
application areas, and summarizes the existing functional limitations that have been identified in
recent US Department of Defense studies. The authors believe that a strong case is made at the end
of this Section for the highly beneficial application of intelligent agent technology to unmanned
vehicles. Sections 3, 4, 5, and 6 provide an explanation of information-centric, service-oriented
architecture concepts, and more detailed descriptions of the principles of intelligent agents, web
services, and interoperability bridges, respectively. Section 7 describes a demonstration of
automated object interpretation as an example of just one application of intelligent agents to
unmanned vehicles.
Technical Challenges Identified by DoD
The most critical near-term technical challenge facing unmanned vehicles is the development of an
autonomous capability to detect, assess and respond to near-field objects in their path of travel. For
a UAV the near-field can extend to several nautical miles in all directions, while for a UGV or
UUV the distance at which an object must be detected is likely to be less than 100 yards.
Existing Intelligent Software in Military Domains
Ontology-based software applications with rule-based agents have been successfully applied in both
tactical command and control (C2) and logistical planning and execution since the late 1990s.
Examples include the Integrated Marine Multi-Agent Command and Control System (IMMACCS)
developed under sponsorship of the Marine Corps Warfighting Laboratory, which was successfully
tested in 1999 as the C2 system of record in the Urban Warrior Advanced Warfighting Exercise
(Pohl et al. 1999). Subsequently, in 2002, IMMACCS was evaluated by two independent teams
during a Limited Technical Assessment (LTA) to determine its ability to maintain a common
tactical picture in a wireless communication environment operating within extremely limited
bandwidth conditions (SPAWAR 2002). It was found that an information-centric4 software system,
such as IMMACCS, requires less bandwidth than a data-centric system in which there is no internal
understanding of the meaning of the data being exchanged and processed.