COOPERATIVE ROBOTIC SYSTEM USING DISTRIBUTED DECISION MECHANISMS WITH DELIBERATIVE CENTRAL SUPERVISOR

Abstract

Cooperative multi-robot systems with distributed decision mechanisms and distributed sensing may be the source of decisional conflicts which can lead to severe performance deterioration. A deliberative central supervisor is a simple approach to correct any incoherent decisions in the system. Given an application, the supervisor can be an autonomous software agent or a human-machine interface. Using Hierarchical Decision Machines (HDM) as distributed decision mechanisms, the decision supervision can use simple matrix representations of decisional data. The resulting architecture has been tested on a fully autonomous team of soccer-playing robots and results indicate that it is well adapted for, but not restricted to, the specific needs of autonomous multi-robot systems with real-time distributed sensing and decision taking. 1.