Active Contour Model-based control of multi-agent robots

Abstract

Applications for multi-agent robot systems have been extensively studied because of their flexibility under a wide variety of situations and there are various proposals for control methods. Our research focuses on the control of robot positions for a generalized task: the transportation of some object of unknown shape to some goal destination. We propose a method that places each robot at a position along the contour of the object based on the Active Counter Model, a method often utilized in image processing. With our ACM-inspired algorithm, robots can surround the target object without knowing its shape in advance. If object transportation by a multi-agent system which was less expensive and utilized exchangeable robots were realized it would have properties that would lend itself to applications in real world environments.