A swarm robot methodology for collaborative manipulation of non-identical objects

In this paper we investigate an algorithm that improves the task completion rate of a swarm of simple robots implementing a leaf-curling task. In this biologically inspired task, robots collaborate to find a suitable place to bend a leaf, which allows them to successfully fold it up. To complete the task simple robots were developed that are not equipped with any direct communication devices. They communicate via sematectonic stigmergy, which means robots can only exchange information via changes they make to their working environment. This type of communication has proved beneficial in helping swarm robots monitor the performance of other swarm members without direct contact, team mate localization or recognition. However, in earlier experiments, implementing the leaf-curling task, information perceived by every robot has not been effectively used to create meaningful collaboration. This disadvantage becomes evident via the low task completion rate. If robots explore their environment, this will improve the outcome by increasing the probability of finding the most suitable part of the leaf to work on. In this paper, an algorithm enabling swarm robots to effectively explore the environment and find the most effective place to perform the leaf-curling task is described in detail. The improvement of completion rate, achieved by this exploring rule, is verified by both simulation and physical experiments with a group of W-AntBots.