中国空间科学技术

中国空间科学技术 ›› 2023, Vol. 43 ›› Issue (3): 123-133.doi: 10.16708/j.cnki.1000-758X.2023.0045

• 巨型星座/低轨大规模星座专栏 • 上一篇    下一篇

基于A2C算法的低轨星座动态波束资源调度研究

刘伟，郑润泽，张磊，高梓贺，陶滢，崔楷欣   

  1. 1 国家航天局卫星通信系统创新中心，北京100094
    2 中国空间技术研究院 通信与导航卫星总体部，北京100094
    3 西北工业大学，西安710072
    4 北京理工大学，北京100081
  • 出版日期:2023-06-25 发布日期:2023-05-23

Research of dynamic beam resource scheduling of LEO constellation based on A2C algorithm

LIU Wei，ZHENG Runze，ZHANG Lei，GAO Zihe，TAO Ying，CUI Kaixin   

  1. 1 Innovation Center of Satellite Communication System，CNSA，Beijing 100094，China
    2 Institute of Telecommunication and Navigation Satellites，China Academy of Space Technology，Beijing 100094，China
    3 Northwestern Polytechnical University，Xi′an 710072，China
    4 Beijing Institute of Technology，Beijing 100081，China
  • Published:2023-06-25 Online:2023-05-23

PDF (PC)

257

摘要： 巨型低轨星座为载人飞船、空间站、遥感卫星等用户航天器提供低时延、大容量的通信通道存在波束资源分配优化的难题。针对采用离散时间的深度强化学习A2C（advanced actor critic）的智能优化框架进行了研究，结合遗传算法中个体和基因概念、形成了可有效满足多用户、动态、并发接入需求的波束资源调度算法。基于仿真分析，提出的算法可在多种典型场景下具有适用性，支持在20s内完成超过3000个任务的有效规划，任务成功率不低于91%。通过算法优化实现复杂度的降低，相对传统遗传算法可节约时间45%以上。同时对传统A2C算法框架中的收敛问题进行了优化，解决了传统全连接A2C算法无法收敛的难题，同时相比DQN（deep q-network）算法框架收敛速度提升38%以上。

关键词: 低轨星座, 波束调度, 任务规划, 深度强化学习, A2C算法

Abstract: The giant low-orbit constellation provided low-latency，largecapacity communication channels for user spacecraft such as manned spacecraft，space stations and remote sensing satellites，and there was a resource allocation optimizing problem of satellite beams.The intelligent optimization framework of A2C（advanced actor-critic）using discrete-time deep reinforcement learning was studied，and the beam resource scheduling algorithm that could effectively meet the needs of multi-users，dynamic and concurrent access was formed by combining the concepts of individuals and genes in genetic algorithms.Based on simulation and analysis，the proposed algorithm could be applicable in a variety of typical scenarios.The method could provide effective scheduling results for more than 3000 tasks in 20s，and the task success rate was not less than 91%.The complexity was reduced by algorithm optimization，which could save more than 45% of the time compared with traditional genetic algorithms.At the same time，the convergence problem in the traditional A2C algorithm framework was optimized，which solved the non-convergence problem in the traditional fully connected A2C algorithm.Meanwhile，the convergence speed was increased by more than 38% compared with the DQN（deep q-network）algorithm.

Key words: LEO constellation, beam scheduling, task planning, DRL, A2C algorithm