摘要
针对现有空间目标威胁评估方法存在的指标选取单一、可信度较差的问题,通过分析空间目标威胁特点和特征指标的依赖关系,提出一种基于贝叶斯网络的威胁评估模型,该模型融合形态异常特征、轨道异常特征、历史行为等多源信息实现对空间目标威胁程度的综合评价。模型采用模糊隶属度函数对相对速度、相对距离等连续变量进行离散化隶属度表示,利用专家知识给出贝叶斯网络节点间的条件概率表。对抵近绕飞和碰撞两种典型空间场景进行仿真验证,结果表明该模型可以较为准确地刻画空间目标的威胁程度,具有一定实际意义。
Aiming at the single selection of indicators and poor credibility of existing space target threat assessment methods,a threat assessment model based on the Bayesian network is proposed by analyzing the dependency of space target threat characteristics and feature indicators,which integrates multi-source information such as morphological anomaly characteristics,orbit anomaly characteristics and historical behavior to achieve a comprehensive evaluation of space target threat level.A fuzzy affiliation function is used in this model to discretize the afiliation representation of continuous variables such as relative velocity and relative distance,and expert knowledge is token advantage to give a conditional probability table among Bayesian network nodes.Simulations are conducted for two typical spatial scenarios known as approaching fly-around and collision,and the results show that the threat level of spatial targets can be more accurately portrayed by using this model which has certain practical value.
作者
曲宗华
魏春岭
Qu Zonghua;Wei Chunling(Beijing Institute of Control Engineering,Beijing 100094,China;Science and Technology on Space Intelligent Control Laboratory,Beijing 100094,China)
出处
《航天控制》
CSCD
北大核心
2023年第4期67-76,共10页
Aerospace Control
基金
国家自然科学基金(U21B6001)。
关键词
空间安全
威胁评估
贝叶斯网络
模糊集合
Space security
Threat assessment
Bayesian networks
Fuzzy ensemble
