Distributed Mobile Robotics by the Method of Dynamic Teams

Distributed teams of agents (robots and workstations) hold great promise for solving complex tasks efficiently, reliably, and automatically. But automatic coordination of the actions of the autonomous agents within a team remains a difficult problem. We suggest that, in order to engender cooperation and to avoid interference among the agents in a team, the general organization of the team reflect the structure of the task. Moreover, the team itself may be a dynamic, fluid entity whose logical structure is maintained by a run-time system, even in the face of attrition, substitution of team members, and explicit recruitment of new members. The Method of Dynamic Teams is based 011 this principle of fluidity.