大语言模型辅助混合遥感星群管控框架

doi:10.16182/j.issn1004731x.joss.26-0038

摘要/Abstract

摘要：

针对混合遥感星群管控调度无法处理用户原始自然语言观测需求、未实现从需求到调度方案全流程自动化的问题，提出一种LLM辅助的混合遥感星群管控框架。利用LLM增强领域知识图谱构建、需求理解及任务表征，并实现调度方案智能生成与调度模型优化求解。设计基于“规划与执行”双Agent架构的调度方案生成方法，结合星群调度工具集，可有效拓展LLM的能力边界。仿真实验验证了该框架在混合遥感星群管控中的应用可行性，具备较高的自动化与智能化水平。梳理了LLM在星群管控中的关键技术，探讨了其在需求理解、方案生成等方面的应用潜力，为未来研究提供了参考。

关键词: 遥感卫星, 星群调度, 集中式管控, 大语言模型, 任务规划

Abstract:

A LLM-assisted control framework for mixed remote sensing satellite constellations is proposed to address the problem that the control and scheduling of mixed remote sensing satellite constellations cannot process users' raw natural language observation requirements and realize full-process automation from requirement input to scheduling schemes. LLMs are utilized to enhance the construction of domain knowledge graphs, requirement understanding and task characterization, and realize intelligent generation of scheduling schemes and optimal solution of scheduling models. A scheduling scheme generation method based on the "planning and execution" dual-Agent architecture is designed. Combined with the constellation scheduling toolset, the capability boundary of LLMs is effectively extended. Simulation experiments verify the application feasibility of this framework in the control of mixed remote sensing satellite constellations, which exhibits a high level of automation and intelligence. The key technologies of LLMs in satellite constellation control are systematically sorted out, and their application potential in aspects such as requirement understanding and scheme generation is discussed to provide reference for future research.

Key words: remote sensing satellite, satellite constellation scheduling, centralized control, large language model, task planning

中图分类号:

TP391.9

伍国华,游子琰,李雷等 . 大语言模型辅助混合遥感星群管控框架[J]. 系统仿真学报, 2026, 38(10): 3098-3112.

Wu Guohua,You Ziyan,Li Lei,et al . Large Language Model-assisted Control Framework for Mixed Remote Sensing Satellite Constellations[J]. Journal of System Simulation, 2026, 38(10): 3098-3112.

图/表 17

图1

图2

图3

图4

图5

图6

图7

图8

图9

图10

图11

表1

表2

表3

表4

表5

图12

参考文献 27

[1]	马文臻, 邹自明, 黎建辉, 等. 空间科学卫星数据处理框架研究与系统实现[J]. 系统仿真学报, 2022, 34(7): 1605-1618.
	Ma Wenzhen, Zou Ziming, Li Jianhui, et al. Space Science Satellite Data Processing Framework Research and System Implementation[J]. Journal of System Simulation, 2022, 34(7): 1605-1618.
[2]	余茂涛. 面向卫星集群的观测任务分配方法研究与实现[D]. 成都: 电子科技大学, 2024.
	Yu Maotao. Research and Implementation of Allocation Methods for Observation Tasks of Satellite Clusters[D]. Chengdu: University of Electronic Science and Technology of China, 2024.
[3]	李庥甜, 王凌, 陈英武, 等. 基于自适应大邻域搜索的多场景多卫星任务规划方法[J]. 系统仿真学报, 2025, 37(7): 1836-1847.
	Li Xiutian, Wang Ling, Chen Yingwu, et al. Multi-scenario Multi-satellite Mission Planning Method Based on Adaptive Large Neighborhood Search[J]. Journal of System Simulation, 2025, 37(7): 1836-1847.
[4]	陈佳威, 陈盈果, 吕济民, 等. 大语言模型在规划与调度问题上的应用[J]. 指挥与控制学报, 2024, 10(6): 740-751.
	Chen Jiawei, Chen Yingguo, Jimin Lü, et al. The Application of Large Language Models in Planning and Scheduling Problems[J]. Journal of Command and Control, 2024, 10(6): 740-751.
[5]	Petroni F, Rocktäschel Tim, Riedel S, et al. Language Models as Knowledge Bases?[C]//Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). Stroudsburg: ACL, 2019: 2463-2473.
[6]	Qiu Zhaopeng, Wu Xian, Gao Jingyue, et al. U-BERT: Pre-training User Representations for Improved Recommendation[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2021, 35(5): 4320-4327.
[7]	Wei Bingyang. Requirements Are All You Need: From Requirements to Code with LLMs[C]//2024 IEEE 32nd International Requirements Engineering Conference (RE). Piscataway: IEEE, 2024: 416-422.
[8]	Mu Fangwen, Shi Lin, Wang Song, et al. ClarifyGPT: A Framework for Enhancing LLM-based Code Generation via Requirements Clarification[J]. Proceedings of the ACM on Software Engineering, 2024, 1: 2332-2354.
[9]	谷学强, 罗俊仁, 周棪忠, 等. 智能博弈决策大模型智能体技术综述[J]. 系统仿真学报, 2025, 37(5): 1142-1157.
	Gu Xueqiang, Luo Junren, Zhou Yanzhong, et al. Survey on Large Language Agent Technologies for Intelligent Game Theoretic Decision-making[J]. Journal of System Simulation, 2025, 37(5): 1142-1157.
[10]	Schick Timo, Dwivedi-Yu Jane, Dessí Roberto, et al. Toolformer: Language Models Can Teach Themselves to Use Tools[C]//Proceedings of the 37th International Conference on Neural Information Processing Systems. Red Hook: Curran Associates Inc., 2023: 68539-68551.
[11]	Patil S G, Zhang Tianjun, Wang Xin, et al. Gorilla: Large Language Model Connected with Massive APIs[C]//Proceedings of the 38th International Conference on Neural Information Processing Systems. Red Hook: Curran Associates Inc., 2024: 126544-126565.
[12]	Liu Weiwen, Huang Xu, Zeng Xingshan, et al. ToolACE: Winning the Points of LLM Function Calling[EB/OL]. (2024-09-02) [2025-04-22]. .
[13]	Yang Chengrun, Wang Xuezhi, Lu Yifeng, et al. Large Language Models as Optimizers[C]//ICLR 2024 Conference. New York: ICLR, 2024: 12028-12068.
[14]	Kim Minsoo, Bursztyn Victor, Koh Eunyee, et al. RaDA: Retrieval-augmented Web Agent Planning with LLMs[C]//Findings of the Association for Computational Linguistics: ACL 2024. Stroudsburg: ACL, 2024: 13511-13525.
[15]	AhmadiTeshnizi A, Gao Wenzhi, Brunborg H, et al. OptiMUS-0.3: Using Large Language Models to Model and Solve Optimization Problems at Scale[EB/OL]. (2024-07-29) [2025-04-26]. .
[16]	Romera-Paredes B, Barekatain M, Novikov A, et al. Mathematical Discoveries from Program Search with Large Language Models[J]. Nature, 2024, 625(7995): 468-475.
[17]	Jing Heng, Sun Qinbo, Dang Zhaohui, et al. Intention Recognition of Space Noncooperative Targets Using Large Language Models[J]. Space: Science & Technology, 2025, 5: 0271.
[18]	丛诣钦, 刘天喜, 魏承, 等. 基于自然语言的星座人智协同管控方法[J]. 宇航学报, 2025, 46(6): 1061-1072.
	Cong Yiqin, Liu Tianxi, Wei Cheng, et al. Natural Language-based Human-AI Collaborative Management and Control Method for Satellite Constellations[J]. Journal of Astronautics, 2025, 46(6): 1061-1072.
[19]	Chen Jiawei, Chen Yingguo, Pham D T, et al. A Large Language Model-based Multi-agent Framework to Autonomously Design Algorithms for Earth Observation Satellite Scheduling Problem[J/OL]. Engineering. (2025-11-06) [2025-11-18]. .
[20]	程昊翔, 龚治兴, 王浩铭, 等. 基于大语言模型的知识图谱逻辑规则挖掘框架及应用[J]. 指挥与控制学报, 2025, 11(2): 202-209.
	Cheng Haoxiang, Gong Zhixing, Wang Haoming, et al. A Logical Rule Mining Framework Based on Large Language Models and Its Applications[J]. Journal of Command and Control, 2025, 11(2): 202-209.
[21]	陈逸, 邱思航, 朱正秋, 等. 基于启发式的人-大模型协作寻源方法[J]. 系统仿真学报, 2025, 37(12): 3112-3127.
	Chen Yi, Qiu Sihang, Zhu Zhengqiu, et al. A Method of Heuristic Human-LLM Collaborative Source Search[J]. Journal of System Simulation, 2025, 37(12): 3112-3127.
[22]	Cobbe Karl, Kosaraju Vineet, Bavarian Mohammad, et al. Training Verifiers to Solve Math Word Problems[EB/OL]. (2021-11-18) [2025-04-28]. .
[23]	余快, 张鹏, 陈斌, 等. 基于卫星数字并行系统的星座成像调度方法[J]. 空间控制技术与应用, 2023, 49(2): 68-75.
	Yu Kuai, Zhang Peng, Chen Bin, et al. Imaging Scheduling of Satellite Constellation Based on Digital Parallel System[J]. Aerospace Control and Application, 2023, 49(2): 68-75.
[24]	周棪忠, 罗俊仁, 谷学强, 等. 大语言模型视角下的智能规划方法综述[J]. 系统仿真学报, 2025, 37(4): 823-844.
	Zhou Yanzhong, Luo Junren, Gu Xueqiang, et al. Survey on Intelligent Planning Methods from Large Language Models Perspective[J]. Journal of System Simulation, 2025, 37(4): 823-844.
[25]	秦龙, 武万森, 刘丹, 等. 基于大语言模型的复杂任务自主规划处理框架[J]. 自动化学报, 2024, 50(4): 862-872.
	Qin Long, Wu Wansen, Liu Dan, et al. Autonomous Planning and Processing Framework for Complex Tasks Based on Large Language Models[J]. Acta Automatica Sinica, 2024, 50(4): 862-872.
[26]	Gu Yi, Han Chao, Chen Yuhan, et al. Large Region Targets Observation Scheduling by Multiple Satellites Using Resampling Particle Swarm Optimization[J]. IEEE Transactions on Aerospace and Electronic Systems, 2023, 59(2): 1800-1815.
[27]	Wu Guohua, Xiang Zhiqing, Wang Yalin, et al. Improved Adaptive Large Neighborhood Search Algorithm Based on the Two-stage Framework for Scheduling Multiple Super-agile Satellites[J]. IEEE Transactions on Aerospace and Electronic Systems, 2024, 60(5): 7185-7200.

模型	任务完成率	任务成功率
Qwen3-32B	0	0
Qwen3-32B-KGT	100	82.5
Qwen3-14B-KGT	100	40
Qwen3-8B-KGT	30	30
Qwen3-32B-KG	67.5	0
Qwen3-32B-T	37.5	25

任务属性	理解结果
优先级	高优先级
分辨率/m	10
观测频次	每天1次
中心经纬度坐标/(˚)	{北纬31.233 3, 东经121.483 3}
区域边界坐标/(˚)	{北纬31.0, 东经121.0}, {北纬31.0, 东经121.8}, {北纬31.5, 东经121.8}, {北纬31.5, 东经 121.0}
观测时长/s	5
卫星载荷类型	可见光遥感载荷
成像截止时间	2025-10-01T23:59:59
最小太阳高度角/(˚)	0

任务属性	理解结果
优先级	1
分辨率	10
观测频次	1次
中心经纬度坐标/(˚)	[31.230 4, 121.473 7]
区域边界坐标/(˚)	[31.200 0, 121.400 0] [31.200 0, 121.550 0] [31.260 0, 121.550 0] [31.260 0, 121.400 0]
观测时长	白天
卫星载荷类型	光学
成像截止时间	2025-10-01T18:00:00
最小太阳高度角/(˚)	0

参数	值
场景开始时间	2025-10-01T00:00:00
场景结束时间	2025-10-02T00:00:00
卫星1半长轴/km	6 878.14
卫星1真近点角/(°)	359.996
卫星1偏心率	3.646 83×10^-16
卫星1近地点辐角	0
卫星1轨道倾角/(°)	45.143 7
卫星1升交点赤经/(°)	89.672 2

任务信息	输出结果
网格编码	0000010111111010110001110000111111111111111111111111111111111111
可见窗口1	2025-10-01T00:12:55—2025-10-01T00:14:37
可见窗口2	2025-10-01T06:52:21—2025-10-01T06:54:45
可见窗口3	2025-10-01T23:50:25—2025-10-01T23:51:55