Chain based path formation in swarms of robots

Abstract

In this paper we analyse a previously introduced swarm intelligence control mechanism used for solving problems of robot path formation. We determine the impact of two probabilistic control parameters. In particular, the problem we consider consists in forming a path between two objects which an individual robot cannot perceive simultaneously. Our experiments were conducted in simulation. We compare four different robot group sizes with up to 20 robots, and vary the difficulty of the task by considering five different distances between the objects which have to be connected by a path. Our results show that the two investigated parameters have a strong impact on the behaviour of the overall system and that the optimal set of parameters is a function of group size and task difficulty. Additionally, we show that our system scales well with the number of robots. © Springer-Verlag Berlin Heidelberg 2006.