Collaborative Robot Mapping via Frequent Encounter-Based Fusion

Exploration strategies are essential for a robot team to explore and navigate efficiently in unknown environments. In this study, an exploration strategy designed to coordinate a robot team that experiences frequent encounters and has limited communication is introduced. The strategy takes advantage of these frequent encounters to exchange information and generate a global map by fusing the partial maps of each robot. The locality of the point clouds in the encounters is used to calculate the transformation between robots using the ICP algorithm. Additionally, a method for frontier detection that treats the maps generated by robots as images is proposed, thus facilitating the identification of areas to be explored. Finally, results obtained from both simulations and experiments in real environments are presented.