Potential Game Based Distributed Task Scheduling for Multi-Earth Observation Satellites

In this paper, we study the coordinated earth observation task scheduling problem of multi-earth observation satellites (EOSs). We aim to provide an optimal observation sequence to maximize the target observation ratio with the least repeated observation targets for all EOSs taking into consideration its motion constraints. Firstly, we establish an individual task graph model to formulate feasible observation sequences for each EOS. Then, by introducing a global utility function of all EOSs and a local utility function of each EOS as a metric, we show that the formulated problem is an exact potential game. Based on this conclusion, we propose a distributed task scheduling algorithm for EOSs, by which an optimal observation sequence can be obtained for each EOS. Finally, we provide a numerical simulation to verify the effectiveness of our proposed algorithm.