Large Scale Computation for Human Agent Team Coordination in Dynamic Environments

Michael Lewis and Katia Sycara

December 2006-November 2009

Air Force Office of Scientific Research

In this work we are developing techniques that allow coordination of large groups of operational assets in the execution of complex missions. To achieve this goal a variety of key issues are being addressed. Specific issues include the development of tractable automated coordination algorithms, techniques for representing responsibility and authority, efficient techniques for finding optimal coalitions to perform specific subtasks, high level intent inference and techniques for allowing effective and flexible control of such groups by small numbers of operators. By achieving these goals we will close the loop on control and coordination and demonstrate the overall approach on simulation test beds. Our work at the University of Pittsburgh focuses on human experimentation and developing effective command strategies and interfaces for controlling cooperating teams of platforms.