Multi-robot frontier-based exploration strategies for mapping unknown environments

With the recent developments in robotics, mobile robots are gaining momentum and are increasingly involved in our lives. Robot exploration and mapping are essential tasks for robot navigation in unknown environments. In the last three decades the challenge has grown and many researches have tried to accomplish these tasks through a team of robots. The use of multi-robot systems represents a big challenge because they require robots able to collaborate each other and, then, to create coordination techniques to make the multi robot system efficient. Multi-robot mapping introduces many advantages, providing a faster and efficient map building even in high dimensional unknown environments. In a multi-robot system, each robot has to be able to move autonomously in the map, avoiding obstacle and reaching desired goals to explore the environment. Furthermore, in order to explore optimally the environment, e.g. minimizing the exploration time, it is necessary to de ne an effcient technique to determine optimal target points to each robot, considering the multi-robot context, that makes the exploration fast. In this work ve algorithms for multi-robot exploration have been evaluated; four of these are de ned and used in other works and represent the state-of-art in this eld, whereas the fth strategy is a novel solution proposed in this work as a valid and effective multi-robot exploration strategy. Each of these algorithms is frontier-based, i.e., exploiting the frontier-detection approach, which is one of the most common and efficient exploration strategy. These algorithms are tested and analyzed, as well as compared with each other. Moreover, evaluation criteria are presented and discussed to efficiently evaluate the proposed frontier-based algorithm. Finally the outcome of the tests are provided with a discussion emphasizing the difference between different strategies and the improvements of the proposed technique.