一种改进滤波算法及其在目标跟踪中的应用研究  

An Improved Filtering Algorithm and Its Application in Target Tracking

作　　者：邱照原 倪龙强[1] 姚桐 杨蕴萌 江腾耀 耿晓虎 QIU Zhaoyuan;NI Longqiang;YAO Tong;YANG Yunmeng;JIANG Tengyao;GENG Xiaohu(Northwest Institute of Mechanical&Electrical Engineering,Xianyang 712099,Shaanxi,China)

机构地区：[1]西北机电工程研究所,陕西咸阳712099

出　　处：《火炮发射与控制学报》2024年第1期68-73,81,共7页Journal of Gun Launch & Control

摘　　要：针对无迹卡尔曼滤波算法在滤波迭代过程中可能存在状态协方差矩阵非正定情况,导致跟踪误差较大甚至滤波易发散等问题,对无迹卡尔曼滤波算法进行了改进设计。在滤波迭代过程中对状态协方差矩阵进行QR分解和Cholesky分解,通过分解后的协方差平方根矩阵来进行滤波迭代,从而保证状态协方差矩阵的正定性,并利用加权最小二乘法进行量测数据同步融合。为验证改进算法的有效性,设计了雷达红外联合跟踪系统数据融合仿真测试实验,并与传统无迹卡尔曼滤波算法进行了比较。仿真实验结果表明:改进算法能够有效抑制跟踪误差、提升目标跟踪系统的鲁棒性,可应用于防空武器系统多传感器航迹信息融合系统设计。Aiming at the problem that the unscented Kalman filter algorithm may have a non positive covariance matrix in the process of filtering iteration,which leads to large tracking error and even easy divergence of filtering,the unscented Kalman filter algorithm is improved and designed.In the process of filtering iteration,QR decomposition and Cholesky decomposition are carried out on the state cova-riance matrix,and the filtering iteration is carried out through the decomposed covariance square root matrix,so as to ensure the positive definiteness of the state covariance matrix,and the weighted least square method is used for synchronous fusion of measurement data.In order to verify the effectiveness of the improved algorithm,the data fusion simulation test experiment of radar infrared joint tracking system is designed and compared with the traditional unscented Kalman filter algorithm.The simulation results show that the improved algorithm can effectively suppress the tracking error and improve the robustness of the target tracking system,which can be applied to the design of multi-sensor track information fusion system of air defense weapon system.

关 键 词：数据融合 目标跟踪 时间配准 平方根无迹卡尔曼滤波 

分 类 号：TN953[电子电信—信号与信息处理]