关键词: 磁性目标跟踪, 磁偶极子, 多初值模型, 最大似然选择, 自适应跟踪算法

Abstract: In order to solve the divergence of the magnetic target filter tracking method when the initial prior information is missing, this paper proposes a solution framework of multiple initial value models. Taking the square-root central difference Kalman filter(SRCDKF) as an example, the SRCDKF adaptive magnetic target tracking algorithm is obtained by combining multiple initial value models. Firstly, based on the magnetic field equivalence of long-distance magnetic dipoles, a multi-initial filter tracking model is established. Then based on the maximum likelihood selection theory, the method of how to choose the best result from multiple models is derived.Finally,using SRCDKF filter as the filtering unit,an adaptive magnetic target tracking algorithm based on SRCDKF is obtained. The simulation experiments show that: (1) the validity of the multi-initial model establishment and selection method; (2) the adaptive magnetic target tracking algorithm based on SRCDKF can effectively complete the tracking of magnetic targets in the absence of initial position information; (3) the tracking results of the adaptive tracking algorithm with different filters as the filtering unit show that the solution framework of the multi-initial value model can provide a method to solve the tracking problem under the initial value unknown.

Key words: magnetic target tracking, magnetic dipole, multiple initial value model, maximum likelihood selection, adaptive tracking algorithm