磁性目标跟踪的多模型自适应滤波方法

Multiple⁃model Adaptive Filtering Method for Magnetic Target Tracking

摘要:

针对非协作磁性目标定位中初始条件难以确定的问题，提出一种基于静态多模型滤波思想的自适应跟踪方法。 首先对磁场观测模型进行分析，设计了一种磁性目标初始参数估算方法，通过该方法得到不同假设条件下的磁性目标状态初值及其误差矩阵，在此基础上起始多个扩展卡尔曼滤波器(Extended Kalman Filter， EKF) 进行求解，根据各个滤波器的求解结果利用最大似然准则选取最佳解作为当前时刻估计结果。 通过仿真实验验证了方法的有效性，结果表明，方法可在目标源及位置等先验信息完全未知条件下准确估计出真实目标状态，对磁性目标实际应用具有参考价值。

Abstract:

With the problem of uncertain initial conditions in non?cooperative magnetic target localization， an adaptive tracking method based on the idea of static multiple?model filtering is proposed． Through analyzing the observation model of magnetic field， an estimation method for initial parameters of magnetic target is designed to obtain initializations of magnetic target state and error matrices in deferent assumed conditions， which subsequently are used for initializing multiple extended Kalman filters to solve， then a maximum likelihood principle is used to select the optimal solution as estimate result at current time according to solving results of each filter． The effectiveness of the proposed approach is tested by simulation experiment． The results show that the method can estimate the truth target state on the condition of that prior information of target resource and location are unknown completely， which has reference value for practical application of magnetic target．