Simultaneous Gas Exploration and Network Localization with Robotic Swarms
Fabio Broghammer, Thomas Wiedemann, Siwei Zhang, Benjamin Noack · 2024
A common issue in state-of-the-art robotic gas distribution mapping and gas source localization is the fact that the positions of the robots are assumed to be perfectly known and the position uncertainties are not considered in movement strategies. However, for cooperative mobile sensor networks with relative localization systems, the geometry of the network plays an important role for localization accuracy. In this paper, we introduce the problem of Simultaneous Exploration and Localization (SEAL). Our approach incorporates position uncertainties of gas concentration measurements into the estimation of environment features, such as a source position or wind direction, and exploits the position uncertainties to design an exploration strategy for sampling gas distributions with mobile robots. More precisely, the paper presents a swarm control algorithm that improves the estimation by adapting the swarm formation for better localization. In simulations, we analyze how position uncertainties affect estimation and exploration performance and we compare the presented control algorithm with a method unaware of position uncertainties.