基于边缘化高斯厄密特卡尔曼滤波的有偏量测下机动目标跟踪方法  

Maneuvering target tracking algorithm based on marginalized Gauss-Hermite Kalman filter with sensor measurement biases

作　　者：余润华 孙少帅[1] 邓洪高[1] YU Runhua;SUN Shaoshuai;DENG Honggao(School of Information and Communication,Guilin University of Electronic Technology,Guilin 541004,China)

机构地区：[1]桂林电子科技大学信息与通信学院,广西桂林541004

出　　处：《桂林电子科技大学学报》2024年第2期162-166,共5页Journal of Guilin University of Electronic Technology

基　　金：国家自然科学基金(62061010);广西重点研发计划(桂科AB23026150)。

摘　　要：针对传感器存在测量偏差情况下的机动目标跟踪问题,提出一种基于边缘化高斯厄密特卡尔曼滤波的有偏量测下机动目标跟踪方法。首先,通过相邻时刻的量测差分构建差分量测方程,从而消除测量偏差,将相邻时刻的目标状态扩维形成增广状态向量以匹配差分量测方程,实现实时滤波估计,同时对增广状态向量进行边缘化处理以降低高斯厄密特滤波的采样维度,减小滤波负担;然后,将交互式多模型方法融入边缘化高斯厄密特卡尔曼滤波以解决机动目标运动模型不确定的问题,在此基础上利用差分量测方程和增广状态向量推导出相应的滤波方程。实验结果表明,本方法能有效消除传感器测量偏差,目标跟踪精度远高于传统的交互式多模型高斯厄密特卡尔曼滤波方法。In order to address the issue of maneuvering target tracking with sensor measurement biases,a new method based on marginalised Gauss-Hermite Kalman filter(MGHKF)is proposed.Firstly,the differential measurement equation is constructed by differencing consecutive measurements to eliminate measurement biases,and the target state of adjacent moments is expanded to form an augmented state vector to match the differential measurement equation,enabling real-time filtering estimation.Meanwhile,the augmented state vector is marginalized to reduce the sampling dimension of Gaussian Hermite filtering,thereby reducing the filtering burden.Then,the interactive multiple model method is integrated into the marginalized Gaussian Hermite Kalman filtering to handle the uncertainty of motion models of maneuvering target.Based on the differential measurement equation and augmented state vector,the corresponding filtering equations are derived.Experimental results demonstrate that the proposed method can effectively eliminate sensor measurement biases and achieves significantly higher tracking accuracy compared to the traditional interactive multiple model Gaussian Hermite Kalman filtering method.

关 键 词：测量偏差 边缘化高斯厄密特卡尔曼滤波(MGHKF) 状态扩维 差分量测方程 交互式多模型 

分 类 号：TN953[电子电信—信号与信息处理]