Autonomous self-localization and mapping agents

Autonomous systems are increasingly seen as enablers for future applications in commercial and military domains. This paper is concerned with platform localization (key element of autonomy) using teams of robots, to support an ongoing research project investigating multi-robot exploration. To our knowledge, the work reported here represents the first instance of decentralized SLAM running onboard platforms. The algorithms operate using natural features, and thus support a fully autonomous capability. This paper describes the theory of decentralized slam, how it has been implemented on COTS platforms and provides results of experiments to assess correctness and performance of the current implementation.