一种新的空间飞行器机动规避策略  

A New Maneuvering and Evading Strategy for Space Vehicle

作　　者：刘晓宇 高晓颖[1] 吕春红 赵建军 陈哲一 LIU Xiaoyu;GAO Xiaoying;LYU Chunhong;ZHAO Jianjun;CHEN Zheyi(Beijing Aerospace Automatic Control Institute,Beijing 100854,China)

机构地区：[1]北京航天自动控制研究所,北京100854

出　　处：《航天控制》2024年第3期23-28,共6页Aerospace Control

摘　　要：针对空间飞行器和具有潜在威胁运动体碰撞交会的场景,提出了一种基于遗传算法优化,并结合BP神经网络预测的空间飞行器机动规避策略。该算法区别于传统碰撞交会末段的机动策略,在完成预警探测之后即进行机动策略选择,扩大了时间和空间范围。其以双方位置和速度为输入参数,机动时刻、机动方向和机动时长为输出参数,脱靶量和燃料消耗为性能指标。数值仿真结果表明,相比于瞬时机动策略,该策略有效地降低了燃料消耗,增大了脱靶量,提高了机动规避成功的可能性,为空间飞行器遭受威胁时的安全防护技术提供了新思路。Aiming at the collision rendezvous scene between the space vehicle and the potentially threatening object,a maneuvering and evading strategy of the space vehicle is proposed,which is based on genetic algorithm optimization and BP neural network prediction.By presenting the difference against the traditional maneuver strategy in the last phase of collision rendezvous,the evading strategy is selected after the warning and detection by the proposed strategy algorithm that expands the scope of time and space and takes the position and speed at both sides as input parameters,the time,direction and duration of maneuver as output parameters,miss distance and fuel consumption as performance indexes.The numerical simulation shows that the proposed strategy effectively reduces fuel consumption,increases miss distance and improving the possibility of successful evading,which is compared with the instantaneous maneuver strategy and provides a new idea of safety protection technology for space vehicle under security threat.

关 键 词：空间飞行器 碰撞交会 机动规避 遗传算法 神经网络 

分 类 号：V448[航空宇航科学与技术—飞行器设计]