Conclusion

In this paper, we propose a SEC-enabled framework that jointly optimizes observation scheduling, routing and computation node selection for wide-area and real-time EOM (SECO). SECO incorporates a distributed observation scheduling strategy based on game theory. This approach enables fine-grained boundary-side scheduling and multi-satellite cooperative observations of a large RoI area. To enhance efficiency, shard splitting is introduced to facilitate parallel transmission and computation. Regarding routing and computation node selection, SECO considers queuing delay within the LEO satellite networks during the optimization process. Further, we propose a theoretically guaranteed system-wide greedy-based strategy to minimize the time cost while considering queuing delay at a system level for simultaneous shard capturing.

SECO 采用了一种基于博弈论的分布式观测调度策略。该策略能够实现对大范围感兴趣区域（RoI）的精细化边界调度与多卫星协同观测。为提升效率，我们引入了分片分割 (shard splitting) 技术以促进并行传输与计算。

在路由与计算节点选择方面，SECO 在优化过程中充分考虑了 LEO 卫星网络内部的排队延迟。此外，针对并发分片捕获的场景，我们提出了一种具有理论保障的全系统贪心策略，旨在系统层面考虑排队延迟的同时，最小化总时间成本。