Analysis of Satellite Observation Task Clustering Based on the Improved Clique Partition Algorithm.

Xiangyu Long,Shufan Wu,Benjie Cui,Zhongcheng Mu,Yixin Huang,Sunhao Chu

DOI: https://doi.org/10.1109/cec.2019.8790243

2019-01-01

Abstract:Satellite task clustering can execute more tasks at a lower cost within a certain time, thus reducing energy consumption and improving the efficiency of satellite planning and scheduling. In this paper, the clustering algorithm design for the tasks of the observation satellite has three parts: Firstly, selecting valid tasks in this current orbit according to the time window and slewing angle of the tasks. Secondly, according to the time window constraint, observation duration, maximum startup time, slewing angle, pitch angle constraints and the sensor, resolution and other task constraints, a graph model that meets the clustering conditions is established. Finally, the problem of the clustering graph model becomes the minimal clique partition problem. Based on the classical clique partition algorithm, this paper improves the minimal clique partition algorithm. Considering the task priority and the cluster tasks minimum slewing angle to achieve the tasks clustering. The simulation results show that the task-clustering algorithm has the advantages of high clustering efficiency and short running time and it is an effective algorithm for clustering observation targets. Improved clique partition algorithm can effectively avoid the problems of the previous algorithm and relatively simple and more suitable for practical problems.