面向即时响应的卫星在轨分布式协商智能任务规划
doi: 10.11728/cjss2024.01.2022-0074 cstr: 32142.14.cjss2024.01.2022-0074
-
李英玉：女, 1976年11月出生于河北省迁安市. 现为中国科学院国家空间科学中心副研究员, 硕士生导师, 主要研究方向为空间信息组织、复杂系统仿真等. E-mail: liyingyu@nssc.ac.cn
作者简介:
通讯作者:
On-orbit Distributed Negotiation Intelligent Mission Planning for Instant Response
-
摘要: 低轨遥感星座任务规划是一个复杂的多目标优化问题, 目前基于深度强化学习的卫星任务规划研究存在试验数据星座规模小、优化目标单一、任务重复安排或模型适应性差等问题. 针对上述问题, 提出CON_DQN(Contract network and Deep Q Network)算法, 采用主从星在轨分布式协商机制, 从星基于规划决策, 主星基于深度强化学习算法决策, 从任务优先级、资源代价和负载均衡等方面进行多目标优化, 实现面向即时响应的卫星在轨分布式协商智能任务规划. 针对用户需求高频动态到达重点观测区域的场景, 进行百星级星座不同规模任务集的仿真实验, 结果表明本文所提算法的响应速度较快且能达到较高的任务收益.Abstract: The mission planning of LEO remote sensing constellation is a complex multi-objective optimization problem. At present, there are some problems in satellite mission planning research based on deep reinforcement learning, such as small scale of test data constellation, single optimization objective, repeated task arrangement and poor model adaptability. To solve the above problems, the CON_DQN (Contract network and Deep Q Network, CON_DQN) algorithm is proposed in this paper, which adopts the master-slave on-orbit distributed negotiation mechanism, the slave satellite makes decisions based on the planning, and the master satellite makes multi-objective optimization decisions from the aspects of priority, resource cost and load balancing based on the deep reinforcement learning algorithm, and processes on-orbit distributed negotiation intelligent mission planning for instant response. Aiming at the scene where the user demand reaches the key observation area dynamically at high frequency, the simulation experiment of different scale task sets of 100-star constellation is carried out. The results show that the proposed algorithm has a fast response speed and can achieve higher task benefits.
-
图 1 遥感星座任务规划模型
Figure 1. Remote sensing sconstellation scmission scmplanning scmpmodel
图 6 CON_DQN与其他算法各项指标对比
Figure 6. Comparison of indexes between CON_DQN and other algorithms
图 7 各算法针对各优先级任务的完成情况
Figure 7. Each algorithm aims at the completion of each priority task
图 8 不同任务规模下各算法完成情况
Figure 8. Completion of different algorithms in different task scales
-
[1] 王俊, 夏维, 胡笑旋, 等. 基于多Agent的遥感星座自主协同任务规划[J]. 指挥与控制学报, 2021, 7(3): 287-294 doi: 10.3969/j.issn.2096-0204.2021.03.0287WANG Jun, XIA Wei, HU Xiaoxuan, et al. Autonomous cooperative mission planning for remote sensing constellation based on multi-Agent[J]. Journal of Command and Control, 2021, 7(3): 287-294 doi: 10.3969/j.issn.2096-0204.2021.03.0287 [2] XIANG M S, DENG Q C, DUAN L S, et al. Dynamic monitoring and analysis of the earthquake worst-hit area based on remote sensing[J]. Alexandria Engineering Journal, 2022, 61(11): 8691-8702 doi: 10.1016/j.aej.2022.02.001 [3] 章罗娜, 马忠成, 饶建兵, 等. 低轨卫星互联网发展趋势及市场展望[J]. 国际太空, 2020(11): 28-31 doi: 10.3969/j.issn.1009-2366.2020.11.006ZHANG Luona, MA Zhongcheng, RAO Jianbing, et al. Development trend and market prospect of low-orbit satellite Internet[J]. International Space, 2020(11): 28-31 doi: 10.3969/j.issn.1009-2366.2020.11.006 [4] SINHA P K, DUTTA A. Multi-satellite task allocation algorithm for earth observation[C]//2016 IEEE Region 10 Conference. Singapore: IEEE, 2016: 403-408 [5] 王冲, 景宁, 李军, 等. 一种基于多Agent强化学习的多星协同任务规划算法[J]. 国防科技大学学报, 2011, 33(1): 53-58 doi: 10.3969/j.issn.1001-2486.2011.01.012WANG Chong, JING Ning, LI Jun, et al. An algorithm of cooperative multiple satellites mission planning based on multi-agent reinforcement learning[J]. Journal of National University of Defense Technology, 2011, 33(1): 53-58 doi: 10.3969/j.issn.1001-2486.2011.01.012 [6] HUANG H, SUN C Y, HU J X, et al. Optimization design of response satellite deployment for regional target emergency observation[C]//Proceedings of 2020 International Conference on Guidance on Advances in Guidance, Navigation and Control. Tianjin: Springer, 2022: 579-591 [7] SUTTON R S, BARTO A G. Reinforcement Learning: An Introduction[M]. Cambridge: MIT Press, 1998 [8] 马一凡, 赵凡宇, 王鑫, 等. 密集观测场景下的敏捷成像卫星任务规划方法[J]. 浙江大学学报(工学版), 2021, 55(6): 1215-1224 doi: 10.3785/j.issn.1008-973X.2021.06.023MA Yifan, ZHAO Fanyu, WANG Xin, et al. Agile imaging satellite task planning method for intensive observation[J]. Journal of Zhejiang University (Engineering Science), 2021, 55(6): 1215-1224 doi: 10.3785/j.issn.1008-973X.2021.06.023 [9] 周碧莹, 王爱平, 费长江, 等. 基于强化学习的卫星网络资源调度机制[J]. 计算机工程与科学, 2019, 41(12): 2134-2142 doi: 10.3969/j.issn.1007-130X.2019.12.006ZHOU Biying, WANG Aiping, FEI Changjiang, et al. A satellite network resource scheduling mechanism based on reinforcement learning[J]. Computer Engineering and Science, 2019, 41(12): 2134-2142 doi: 10.3969/j.issn.1007-130X.2019.12.006 [10] 彭双, 伍江江, 陈浩, 等. 基于注意力神经网络的对地观测卫星星上自主任务规划方法[J]. 计算机科学, 2022, 49(7): 242-247 doi: 10.11896/jsjkx.210500093PENG Shuang, WU Jiangjiang, CHEN Hao, et al. Satellite onboard observation task planning based on attention neural network[J]. Computer Science, 2022, 49(7): 242-247 doi: 10.11896/jsjkx.210500093 [11] 王海蛟. 基于强化学习的卫星规模化在线调度方法研究[D]. 北京: 中国科学院大学(中国科学院国家空间科学中心), 2018WANG Haijiao. Massive Scheduling Method Under Online Situation for Satellites Based on Reinforcement Learning[D]. Beijing: University of Chinese Academy of Sciences (National Space Science Center, Chinese Academy of Sciences), 2018 [12] 李大林. 天文观测卫星任务规划模型与方法研究[D]. 哈尔滨: 哈尔滨工业大学, 2021LI Dalin. Research on Observation Scheduling Model and Method of Astronomy Satellite[D]. Harbin: Harbin Institute of Technology, 2021 [13] LIU Y C, CHEN Q F, LI C Y, et al. Mission planning for earth observation satellite with competitive learning strategy[J]. Aerospace Science and Technology, 2021, 118: 107047 doi: 10.1016/j.ast.2021.107047 [14] 吴白轩. 基于合同网协议的多星多任务规划方法研究[D]. 哈尔滨: 哈尔滨工程大学, 2020WU Baixuan. Research on Multi-star and Multitask Planning Based on Contract Network Protocol[D]. Harbin: Harbin Engineering University, 2020 [15] 姜维, 庞秀丽. 组网成像卫星协同任务规划方法[M]. 哈尔滨: 哈尔滨工业大学出版社, 2016JIANG Wei, PANG Xiuli. Collaborative Mission Planning Method for Networking Imaging Satellite[M]. Harbin: Harbin Institute of Technology Press, 2016 [16] MNIH V, KAVUKCUOGLU K, SILVER D, et al. Playing Atari with deep reinforcement learning[OL]. arXiv preprint arXiv: 1312.5602, 2013 -
-
下载:
图(9)
计量
- 文章访问数: 1170
- HTML全文浏览量: 289
- PDF下载量: 138
-
被引次数:
0(来源:Crossref)
0(来源:其他)
