Abstract: The application of earth observation satellite (EOS) is shifting from static to dynamic mission scenario, leading to an increasing demand for real-time observing capabilities. Consequently, an autonomous scheduling method is urgently needed to enable real-time mission responses and overcome the window constraints imposed by satellite telemetry command control (TT&C) systems. An event-triggered distributed autonomous scheduling method is proposed, which enables the autonomous closed-loop of target discovery, evaluation, and imaging. Firstly, an event-triggered distributed multi-satellite task negotiation framework is established utilizing the double layer contract network protocol (DLCNP). Secondly, a dynamic scheduling algorithm considering task priority based on the minimum conflict set is proposed to achieve real-time task assertion and conflict resolving. This algorithm provides an online solution for multi-satellite task negotiation. Finally, an iterative density cluster method is introduced to conduct clustering of high-value point targets. This method ensures the clustered point targets can be covered by a single imaging satellite, which effectively reduced observations required. The superiority of the event-triggered distributed autonomous scheduling method is verified by comparing its task completion rate with global optimization algorithm and communication load with blackboard inter-satellite structure. The pattern of task merging ratio is revealed by varying the size of the spread area and the number of targets. The proposed method can enhance the autonomy and emergency response capability of EOS swarm effectively.

Key words: swarm autonomous negotiation, double layer contract network protocol, dynamic scheduling, targets clustering, planning algorithm