A local interaction based multi-robot hunting approach with sensing and modest communication

Abstract

A local interaction based hunting approach for multi-robot system in unstructured environments is proposed in this paper. The hunting task is modeled as three modes: initial leader-fixed following&search mode, leader-changeable following&search mode and hunting mode. The conditions for modes switching are given. In order to reduce the dependence on communication, an event-trigger communication scheme based on the evader's observation state is designed. For individual robot, it integrates local information from vision system, sonar sensors and encoders in its local coordinate frame as well as modest communication data to acquire situation-suited task mode, and then makes decisions based on behaviors with appropriate local coordination rules. The experiments with physical mobile robots verify the effectiveness of the proposed approach. © 2009 Springer-Verlag Berlin Heidelberg.