Knowledge-Rich Agents for Modeling Naval C3

Period of Performance: 05/29/2008 - 03/29/2009

$70K

Phase 1 SBIR

Recipient Firm

Soar Technology, Inc.
3600 Green Court Array
Ann Arbor, MI 48105
Principal Investigator

Abstract

Current traditional warfare assessment simulations have been criticized for their simplistic modeling of command and control behavior, often resulting in inappropriate reactions to complex tactical situations, which can produce erroneous simulation results. Soar Technology, in partnership with Rolands & Associates, proposes to address this problem in the specific context of the Naval Simulation System (NSS). We will identify a framework to integrate agent software with NSS to allow different kinds of intelligent agents to enhance command and control behavior. We will also prototype a knowledge-rich intelligent agent based on Soar, a cognitive architecture for human behavior modeling that has been used successfully to build agents in military simulations over the past 10-15 years. Such knowledge-rich agents have the potential to bring human-level decision-making capabilities to simulations like NSS, resulting in more accurate modeling and assessment of new tactics.