DYNAMIC MULTI-ROBOT COORDINATION

Communication among a group of robots should in principle improve the overall performance of the team of robots, as robots may share their world views and may negotiate task assignments. However, in practice, effectively handling in real-time multi-robot merge of information and coordination is a challenging task. In this paper, we present the approach that we have successfully developed for a team of communicating soccer robots acting in a highly dynamic environment. Our approach involves creating shared potential functions based on shared positions of relevant obstacles in the world. The biases introduced in the potential functions are general and they could in principle be provided by external sources, such as a human or robot coach. We provide controlled experiments to analyze the impact of our approach in the overall performance of a robot team.