中国空间科学技术

中国空间科学技术 ›› 2023, Vol. 43 ›› Issue (1): 29-43.doi: 10.16708/j.cnki.1000-758X.2023.0003

• 论文 • 上一篇    下一篇

面向卫星多目标重复观测任务的分层聚类规划

张耀元，杨洪伟，袁荣钢，梁奕瑾，李爽   

  1. 1 南京航空航天大学 航天学院，南京211106
    2 上海卫星工程研究所，上海201109
  • 发布日期:2023-01-13 出版日期:2023-02-25

Satellite multi-object repetitive observation mission planning based on hierarchical clustering

ZHANG Yaoyuan，YANG Hongwei，YUAN Ronggang，LIANG Yijin，LI Shuang   

  1. 1 College of Astronautics,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China
    2 Shanghai Institute of Satellite Engineering,Shanghai 201109,China
  • Online:2023-01-13 Published:2023-02-25

PDF (PC)

401

摘要： 针对敏捷卫星多目标重复观测任务规划面临的可行任务执行序列集合规模庞大困难，提出了一种任务执行序列时间解耦的分层聚类任务规划方法。该方法以规划过程中的可观测窗口和任务执行窗口为聚类对象，通过单次可观测窗口聚类和多次任务执行窗口聚类将任务集合按任务窗口属性分解为一系列时间解耦的小规模集合，在两次任务执行窗口聚类间使用基于贪婪优化的搜索算法对聚类生成的小规模任务集合分别进行集合内任务规划，最后将各集合的任务规划结果合并后得到所有任务的执行序列。仿真结果表明，该分层聚类方法可有效降低全局优化复杂度，消解不同优先级观测任务的冲突，提高任务规划质量，能够在不降低目标点观测完成率的前提下对有多个观测机会的目标点进行重复观测，且算法稳定性好，能在数秒内得出规划结果，适用于星上自主任务规划。

关键词: 敏捷卫星, 可观测窗口聚类, 任务规划, 对地观测, 重复观测任务

Abstract:  In order to solve the problem of large scale of feasible task execution sequences for agile satellite multi-target repetitive observation mission planning,we proposed a hierarchical clustering task planning method with time decoupling of task execution sequence.In this method,the observable windows and task execution windows in the planning process are taken as the clustering objects,and the task set is decomposed into a series of time-decoupled,small-scale sets according to the task attributes in the set by single clustering of the observable windows and multiple clustering of the task execution windows.Then,a search algorithm based on greedy optimization was used to plan the small-scale task sets generated by clustering between two task execution windows.Finally,the execution sequence of all tasks was obtained by combining the task planning results of each set.The simulation results indicate that the hierarchical clustering method can effectively reduce the global optimization complexity,eliminate the conflicts between different types of observation tasks,improve the quality of mission planning,and can perform repeated observation of multiple observation opportunities without reducing the coverage rate of the target points.The algorithm has good stability and can get planned results in a few seconds,which is suitable for satellite onboard autonomous mission planning.

Key words: agile satellites, observable window clustering, mission planning, earth observation, repeated observation missions