Minimised moving distance algorithm for geometric deployment of robot swarms

Abstract

This paper addresses the deployment problem for a swarm of autonomous mobile robots initially randomly distributed in 2 dimensional space. A new local interaction rule which decreases the total distance moved by all robots in the swarm during deployment is proposed. This work allows a swarm of robots to configure themselves into a two-dimensional triangular tessellation. Each robot agent has a limited viewing range. No explicit communication or leader robot is used. Each robot interacts selectively with two neighbouring robots so that three robots converge onto each vertex of an equilateral triangle with fixed and pre-determined side lengths. The robot model, local interaction and neighbour selection algorithms are described. Through extensive simulations, comparisons are drawn between the performance of the algorithm from previous works and the proposed algorithm. Both the number of required activation steps and the total distance moved are investigated, it is shown that the proposed algorithm minimises the total distance moved by the swarm during deployment with a required number of activation steps similar to that of previous works.