Projects

Current projects

The following are the current main projects under the Cougaar source tree. See the downloads page for access to the source tree.

CougaarSE

Cougaar is a java-based architecture for the construction of large-scale distributed agent-based applications.

CoreProject

The core of Cougaar, including Agent, Node definitions and initialization, and the MTS.

ServiceDiscovery

Agent discovery & dynamic relationships using the yellow pages.

Older Projects

CSMART

CSMART is an integrated toolset for building, running, monitoring, and analyzing all Cougaar societies, and for performing experiments on those societies by systematically varying their properties and comparing the resulting behaviors. CSMART users will likely include Cougaar component/Plugin developers as well as researchers interested in understanding Cougaar societies and mechanisms. See the CSMART_UserGuide_10.0.

CougaarIDE

An Integrated Development Environment for Cougaar. It was developed by Cougaar Software Inc., to support developers interested in using Cougaar. CougaarIDE is comprised of a number of Plugins for Eclipse, you will need to install the Eclipse SDK before attempting to run CougaarIDE. CougaarIDE was tested on Windows, MacOS X, and Red Hat Linux. Originally written to work with Eclipse Platform 2.1.x and 3.0, but the published version is for use with Eclipse 3.4 Ganymede only.

  1. Build or Download Binary Release
  2. Install Cougaar IDE
  3. Use
  4. See Cougaar Tutorials with Cougaar IDE
  5. CougaarIDERepo

Downloads
– In order to install, unzip the release file into the dropins folder of your eclipse ganymede installation (usually something like C:\eclipse\dropins).

CougaarMicro

The Cougaar MicroEdition project developed a JAVA based distributed agent software suite derived from the Cougaar architecture that executes on very small wireless computing devices integrated with sensor and robotic packages for sensor-web and cooperative robotic applications. This project provides the necessary programming environment and distributed application tools to facilitate the development of sophisticated, self-organizing networks of wireless connected agent systems for military and space applications. The project culminated in the demonstration of the utility of such systems by connecting a simple robotic sensor-web with a fully enabled Cougaar society, providing distributed data and information management down to the simple sensor packages.
The project is focused on the extension to the Cougaar infrastructure to support operation on resource-limited embedded devices. These sensing and acting agents operate as a wireless web to perform data fusion operations that are not possible with single-sensor systems. They are interoperable with Cougaar agents, which can perform higher-level aggregation, additional data fusion and problem solving using the Cougaar cognitive model.

This distribution of CougaarME from 2002 is available.

Objective: Extend the Cougaar distributed agent technology to support embedded devices in a way that is:

  • Randomly-distributed
  • Embedded agents need spatial awareness.
  • Many must sense their position, surroundings

Approach:

  • Miniaturize the Cougaar core, make it less than 1 MB memory for VM and require less than 100 MHz processor
  • Develop sensor interfaces and algorithms
  • Develop actuator interfaces and algorithms
  • Develop platform coordination software

Architecture Concept: The concept is that one CougaarSE (Standard Edition) would manage several CougaarME modules. These modules provide access to sensors and actuators in the real world, effectively serving as the eyes and hands for the agent in the real world.

SupportProject

The core of Cougaar, including Agent, Node definitions and initialization, and the MTS.