Cooperative Autonomous Agents Testbed
Final rept. Feb 1989-May 1993
HUGHES RESEARCH LABS MALIBU CA
Pagination or Media Count:
The goal of the Cooperative Autonomous Agents Testbed CAAT program is to develop new techniques by which multiple autonomous agents can interact intelligently and effectively in both cooperative or competitive modes of operation. Among the most significant accomplishments of this program are the development of two techniques that improve the effectiveness of decision making and that simplify the acquisition of tactics knowledge for computer generated forces used in large-scale distributed interactive simulations. We have developed a concurrent control paradigm that vastly improves decision making in the presence of multiple competing objectives. We have also developed case-based learning techniques to simplify the acquisition of tactics through direct interaction with subject experts. These techniques have been applied to automate simulations of Ml tank platoons as well as F-14 fighter aircraft. SAFOR, SIMNET, DIS, Intelligent agents, Cased-based reasoning, SOAR, Behavior-based control, Reactive planning, Machine learning.
- Computer Programming and Software
- Navigation and Guidance