COOPERATION WITHOUT COMMUNICATION - MULTIAGENT SCHEMA-BASED ROBOT NAVIGATION

Multiagent robotics affords the opportunity to solve problems more efficiently and effectively than any single agent could achieve. Nonetheless, communication bottle-necks between agents pose potentially serious drawbacks in coordinated behavior. In this research, we demonstrate the efficiency of multiagent schema-based navigation for object retrieval. Primitive motor behaviors are specified for each of the individual robotic agents, which produce safe task-achieving action in an unstructured environment. When implemented over several identical units, retrieval is facilitated in the absence of interagent communication as evidenced by spontaneous recruitment of several agents to accomplish a task. Simulation results are provided to demonstrate these effects. Extensions to other task-achieving behaviors are also feasible.