摘要
随着各领域对卫星遥感数据需求的日益深入,用户不再满足于卫星对地面目标单次观测所获得的遥感数据,而是希望组网卫星能够对目标进行周期性持续观测,以实现目标态势定期刷新.这对卫星任务规划研究提出了更高的要求,传统的多星多目标任务规划方法均假设目标一旦被观测即任务完成,难以适应周期性持续观测任务规划场景.本文分析了组网卫星周期性持续观测任务规划问题,建立了约束满足问题模型.基于分解的多目标进化算法框架,提出了组网卫星周期性持续观测任务规划方法,从观测周期超时程度和卫星能量消耗等两个维度进行多目标优化求解.最后,通过仿真实验,验证了该方法的可行性和有效性.
With the increasing demands of satellite remote sensing data in various fields,users are no longer satisfied with the remote sensing data of ground targets obtained from single observation. In order to achieve the goal of target situation periodic refreshment,a new observation request which need satellite cluster to do observations periodically to the same target emerges. This brings about a new challenge to satellite task scheduling. The traditional multi-satellite multi-target scheduling method assumes that if one target is observed,then the task is completed. There is no need for satellite observing it anymore. So it is difficult to apply traditional method to our problem. This paper analyzes the satellite periodic continuous observing task problem and establishes constraint satisfaction problem model with two objective functions which are the degree of timeout and energy consumption. Based on the MOEA/D algorithm framework,a method of satellite periodic continuous observation task scheduling is proposed. Finally,some experiments have been conducted to validate the correctness and practicability of our scheduling algorithm.
作者
王凌峰
陈兆荣
陈浩
陈宏盛
WANG Ling-feng;CHEN Zhao-rong;CHEN Hao;CHEN Hong-sheng(College of Electronic Science and Engineering,National University of Defense Technology, Changsha 410073, China;Unit 95874 of PLA,Nanjing 210022 ,China)
出处
《小型微型计算机系统》
CSCD
北大核心
2018年第6期1366-1371,共6页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(61101184
61174159)资助
关键词
组网卫星
成像任务规划
周期性持续观测
多目标优化
进化算法
satellite cluster
observing task scheduling
periodic continuous observation
multi-objective optimization
evolutionary algorithms
