跳转至

文章背景与核心概要

大型低轨（LEO）星座的迅猛发展彻底革新了对地观测（EO）领域。然而，如何在例行任务计划执行过程中处理突发的“应急”观测请求，依然是一项巨大的技术挑战。本文正式提出了动态应急观测调度问题（DEOSP），并创新性地提出了一种任务驱动的三层分布式调度（T3L-DS）方法来解决这一难题。

该方法通过利用基于地理网格的表征方式以及多层协同机制，将紧急任务高效地融入现有调度方案中，同时将对原有计划的干扰降至最低。实验表明，T3L-DS 的性能超越了传统的分布式方法，并逼近了集中式解决方案的效率，为未来卫星星座的智能化自主运行提供了强有力的技术支撑。

Task-Driven Three-Layer Distributed Scheduling for Emergency Earth Observation in Large Low-Earth-Orbit Constellations

Authors: Qian Yin, Xinwei Wang, Guohua Wu
Date: August 14, 2026
Subject: Artificial Intelligence (cs.AI)
DOI: 10.48550/arXiv.2608.14789

摘要

大型低轨（LEO）星座的迅猛发展彻底革新了对地观测（EO）领域。然而，管理在例行任务计划已在执行后才到达的“应急”观测请求，构成了巨大的挑战。本文介绍了动态应急观测调度问题（DEOSP），并提出了一种新颖的任务驱动三层分布式调度（T3L-DS）方法来解决该问题。

通过利用基于地理网格的表征和多层协调机制，T3L-DS 以最小的干扰将紧急任务有效地整合到现有调度中，其性能超越了传统的分布式方法，并接近集中式解决方案的效率。

Summary

The rapid growth of large Low-Earth-Orbit (LEO) constellations has revolutionized Earth observation (EO). However, managing "emergency" observation requests—which arrive after routine mission plans are already in execution—poses a significant challenge. This paper introduces the Dynamic Emergency Observation Scheduling Problem (DEOSP) and proposes a novel Task-Driven Three-Layer Distributed Scheduling (T3L-DS) method to solve it.

By utilizing a geographic grid-based representation and a multi-layered coordination mechanism, T3L-DS effectively integrates urgent tasks into existing schedules with minimal disruption, outperforming traditional distributed methods and approaching the efficiency of centralized solutions.

T3L-DS 的核心特征

  • 地理网格映射： 在统一的网格上表示任务需求和传感器覆盖范围，以促进高效的空间协同。
  • 动态聚类： 根据实时观测能力和星间链路形成临时集群。
  • 多层协同：
    • 集群内： 采用星载双计划竞标和联合边际评估来优化本地任务分配。
    • 集群间： 提供强大的机制来处理无法由单个集群满足的未解决需求。

Key Features of T3L-DS

  • Geographic Grid Mapping: Represents task demands and sensor footprints on a unified grid to facilitate efficient spatial coordination.
  • Dynamic Clustering: Forms temporary clusters based on real-time observation capabilities and inter-satellite links.
  • Multi-Layered Coordination:
    • Intra-cluster: Employs onboard dual-plan bidding and joint marginal evaluation to optimize local task allocation.
    • Inter-cluster: Provides a robust mechanism to handle unresolved demands that cannot be satisfied by a single cluster.

性能亮点

大量的计算实验表明，T3L-DS 是一种高效的分布式方法： * 卓越的覆盖率： 在分布式方法中实现了最高的应急覆盖率，比 A-SeTVBRP 提高 2.8%，比传统的合同网协议（CNP）提高 17.1%。 * 高效率： 相对于集中式模拟退火（SA）算法，性能差距仅为 7.1% 左右。 * 最小干扰： 在高冲突负载下，T3L-DS 显著维护了例行计划的完整性，将例行覆盖率损失分别降低了 57.9%（对比 A-SeTVBRP）和 87.7%（对比 CNP）。

Performance Highlights

Extensive computational experiments demonstrate that T3L-DS is a highly effective distributed approach: * Superior Coverage: Achieved the highest emergency coverage among distributed methods, with a 2.8% improvement over A-SeTVBRP and a 17.1% improvement over the conventional Contract-Net Protocol (CNP). * Efficiency: The performance gap relative to centralized Simulated Annealing (SA) is only approximately 7.1%. * Minimal Disruption: Under high-conflict loads, T3L-DS significantly preserved routine-plan integrity, reducing routine-coverage loss by 57.9% (vs. A-SeTVBRP) and 87.7% (vs. CNP).

访问与资源

Access & Resources

license icon