A collective moving algorithm in modular robotics: contribution of communication capacities

The goal of self-reconfigurable robotics is to replace a complex robot by a swarm of more simple robots called modules which can change the way they are connected, creating thus a robot which can adapt his shape to the environment in which it is evolving. The goal of this approach is to obtain a large robustness against breakdowns. In order to further increase this robustness, we can use autonomous modules which can move by their own and which can make/break connections with other modules as desired. We present a behaviour model for autonomous modules to permit the swarm to get over simple difficulties: climbing on straight walls, getting over ditches and moving down straight walls. Our model is an identical finite state machine implanted in all modules which is dedicated to the movement of autonomous modules swarm. We also present three simulations in which we study three different levels of communication, thus estimating influence of these capacities on the modules swarm speed to get over three consecutive difficulties.