Simultaneous multi-robot mapping and localization

Deploying autonomous robots in an unknown or partially destroyed environment is a major challenge for our societies: sensitive or dangerous sites for humans are a good illustration. This article focuses on a central issue of this challenge, namely the ability of mobile robots to autonomously explore and construct a map of an unknown environment. The article is organized in three parts. First, we present the SLAM function (Simultaneous Localization And Mapping – Cartographie et localisation simultanées), which generates a map from the fusion of data perceived by the robot(s). We then introduce the multi-robot context by considering, on the one hand, the communication and processing organization prerequisites within a fleet and, on the other, the deployment strategy aspects, which ensure efficient and coordinated coverage of the environment. Based on these elements, we then present the various techniques for implementing a multi-robot SLAM. The article concludes with a set of illustrations of multi-robot SLAM applications in industrial and academic contexts.