Evolving Group Transport Strategies for e-Puck Robots: Moving Objects Towards a Target Area

This paper describes a set of experiments in which a homogeneous group of simulated e-puck robots is required to coordinate their actions in order to transport cuboid objects towards a target location. The objects are heavy enough to require the coordinated effort of all the members of the group to be transported. The agents’ controllers are dynamic neural networks synthesised through evolutionary computation techniques. The results of our experiments indicate that the most effective transport strategies generated by artificial evolution are those in which the robots exploit occlusion by pushing the objects across the portion of their surface, where they occlude the direct line of sight to the goal. The main contribution of this study is the analysis of the relationships between the characteristics of the object (i.e., mass and length), the morphology of the robots, and the group performance. We also test the scalability of the occlusion-based transport strategies to group larger than those used during the evolutionary design phase.