摘要： 分析了空间低轨目标群的运行特点，提出了基于时序向量相似性的空间目标群匹配算法，提高了对低轨巨型星座的识别管理能力。首先，介绍了时序向量的降维方法，将目标群高维观测时序向量简化为空间构型序列；而后，提出了基于动态时间规整（Dynamic Time Warping，DTW）的目标群空间构型序列相似性判别算法；最后，利用星链卫星目标群仿真和实测数据对算法的匹配能力进行验证。结果表明该算法可实现空间目标群监测数据快速匹配，仿真数据匹配过程中，在群内目标缺失30%的条件下匹配成功率可达100%，在低缺失条件下（缺失率5%以内）群内目标识别成功率平均超过75%；实测数据匹配成功率可达100%。
Abstract： This article analyzes the motion characteristics of loworbit space target groups and proposes a space target group matching algorithm based on time series vector similarity, which improves the recognition and management ability of low orbit satellite constellations. Firstly, this article introduces the dimensionality reduction method of time series vectors, which simplifies the highdimensional observation time series vectors of the target group into spatial configuration sequences. Secondly, an algorithm based on Dynamic Time Warping (DTW) is proposed to identify the phase sequence similarity of observed data of target group. Finally, the matching ability of the algorithm is verified by the simulation and measured data of Starlink satellite target group. The results show that this algorithm can match the observed data of the space target group rapidly. During the simulation data matching process, the success rate of matching can reach 100% under the condition of 30% random missing targets within the group, and the average success rate of target recognition within the group exceeds 75% under low missing conditions (with a missing rate of less than 5%). The success rate of matching real observed data can also reach 100%.
Key words : low orbit space target group; time series vector; dynamic time warping; similarity identify
