航空学报 > 2024, Vol. 45 Issue (21): 330149-330149   doi: 10.7527/S1000-6893.2024.30149

一种多尺度级联地磁匹配相似性度量准则MP-S

尤逸轩1,2, 纪新春1,3(), 魏东岩1, 陆一1,2, 袁洪1   

  1. 1.中国科学院 空天信息创新研究院,北京 100094
    2.中国科学院大学 电子电气与通信工程学院,北京 100049
    3.西北工业大学 电子信息学院,西安 710072
  • 收稿日期:2024-01-15 修回日期:2024-03-21 接受日期:2024-04-08 出版日期:2024-04-12 发布日期:2024-04-10
  • 通讯作者: 纪新春 E-mail:jixc@aircas.ac.cn
  • 基金资助:
    国家自然科学基金(42204048)

Criterion MP-S for multi-scale cascaded geomagnetic matching similarity measurement

Yixuan YOU1,2, Xinchun JI1,3(), Dongyan WEI1, Yi LU1,2, Hong YUAN1   

  1. 1.Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China
    2.School of Electronic,Electrical and Communication Engineering,University of Chinese Academy of Sciences,Beijing 100049,China
    3.School of Electronics and Information,Northwestern Polytechnical University,Xi’an 710072,China
  • Received:2024-01-15 Revised:2024-03-21 Accepted:2024-04-08 Online:2024-04-12 Published:2024-04-10
  • Contact: Xinchun JI E-mail:jixc@aircas.ac.cn
  • Supported by:
    National Natural Science Foundation of China(42204048)

摘要:

地磁匹配定位具有无源、隐蔽、无累积误差等优点,在军事领域展现出广阔的应用前景。地磁匹配定位的核心是通过度量磁场实测序列与地磁基准图中待匹配序列的相似性,来确定载体当前的位置。针对现有地磁匹配相似性度量准则未能兼顾地磁序列全局特征和细节特征、易受磁场扰动影响等缺点,提出一种基于矩阵轮廓(MP)算法的空间矩阵轮廓准则(MP-S)来实现多尺度级联地磁匹配相似性度量。通过级联细节特征和全局空间约束构造空间最近邻(NN-S)函数,能够有效筛选地磁序列细节特征的错误匹配,提高定位精度和容错性能;同时,针对MP算法计算繁杂的问题,设计了一种二维磁图轻量化(2D-MLight)搜索与计算策略,通过避免相邻待匹配序列相似性重复计算,显著降低相似性度量算法的运行时间。某飞行器实测地磁数据的处理结果表明,使用MP-S准则进行相似性度量的水平定位误差为48.6 m,匹配概率为93.99%,优于目前常用的均方差(MSD)准则,并且2D-MLight轻量化搜索与计算过程能够满足实时定位需求,对于提高特征较为平坦、存在局部扰动等复杂场景的地磁匹配定位性能具有良好的工程应用意义。

关键词: 地磁匹配定位, 相似性度量, 多尺度级联, Matrix Profile算法, 轻量化

Abstract:

Geomagnetic matching positioning has the advantages of being passive, covert, and free from cumulative errors, showing broad prospects of application in the military field. The core of geomagnetic matching positioning is to determine the current position of the carrier by measuring the similarity between the measured geomagnetic field sequence and the target sequence in the geomagnetic reference map. However, existing criteria for geomagnetic matching similarity measurement do not fully consider the global and detailed features of the magnetic sequence, and are easily affected by magnetic field disturbances. To address these limitations, this paper proposes a Matrix Profile-Space (MP-S) criterion based on the Matrix Profile (MP) algorithm to achieve measurement of multi-scale cascading geomagnetic matching similarity. A Nearest Neighbor-Space (NN-S) function is constructed based on cascade detailed features and global space constraints, effectively filtering out incorrectly matched detailed features of magnetic sequences, thereby improving positioning accuracy and fault tolerance. Additionally, we design a lightweight 2D-Magnetic Light (2D-MLight) search and calculation strategy to address the complex computation issues of the MP algorithm, significantly reducing the runtime of the similarity measurement algorithm by avoiding redundant similarity calculations of adjacent target sequences. The processing results of geomagnetic data from a certain aircraft show that using the MP-S criterion for similarity measurement results in a horizontal positioning error of 48.6 m and a matching probability of 93.99%, outperforming the commonly used Mean Square-Difference (MSD) criterion. Furthermore, the lightweight search and calculation process of 2D-MLight can meet real-time positioning requirements, demonstrating good engineering application significance for improving the performance of magnetic matching positioning in complex scenarios with flat features and local disturbances.

Key words: geomagnetic matching positioning, similarity measurement, multi scale cascading, Matrix Profile algorithm, lightweight

中图分类号: