Multi-robot Informative and Adaptive Planning for Persistent Environmental Monitoring

Kai-Chieh Ma, Zhibei Ma, Lantao Liu, Gaurav S. Sukhatme

Abstract: To gain a better understandingMa, Kai-Chieh of environmentalMa, Zhibei processesLiu, Lantao we are interestedSukhatme, Gaurav S. in the problem of deploying multi-robot systems for efficient collection of environmental data. For long-term autonomy, enabling persistent monitoring, it is important to consider the spatio-temporal variations of environmental phenomena. We develop a multi-robot persistent path planning method that reduces uncertainty in the environmental model. Our framework contains two components: the first component computes potential observation points that minimize model prediction uncertainty, and the second component uses this for online planning of multi-robot paths, while also taking into account the efficiency of information collection. We validated our method via simulations, and the results show that it produces multi-robot routing paths that are conflict-free, informative, and adaptive to the environmental dynamics.