基于改进量测划分策略的概率假设密度滤波器  

Probability Hypothesis Density Filter Based on Improved Measurement Partition Scheme

作　　者：孙志强[1] SUN Zhiqiang(Department of Mechanical and Electronic Engineering,Shangqiu Polytechnic,Shangqiu 476000)

机构地区：[1]商丘职业技术学院机电系,商丘476000

出　　处：《舰船电子工程》2021年第1期20-22,39,共4页Ship Electronic Engineering

基　　金：河南省科技攻关项目(编号:182102210116)资助。

摘　　要：针对密集杂波场景下标准概率假设密度(PHD)滤波器的计算负担大及滤波精度低等问题,论文提出一种基于改进量测划分策略的高斯混合概率假设密度(IMPS-GM-PHD)滤波器。通过将目标权重参数叠加融合到固定经验门阈值中,IMPS-GM-PHD滤波器改进了标准的量测门技术,一定程度上避免了由经验门阈值的量测门技术导致的量测漏选问题。实验结果表明,IMPS-GM-PHD滤波器不仅具有较高的目标状态估计精度而且其计算代价相对较低。For the problems of high computational burden and low filtering accuracy of the standard probability hypothesis den⁃sity(PHD)filter in the dense clutter scenes,a Gaussian mixture probability hypothesis density filter based on the improved mea⁃surement partition scheme,namely IMPS-GM-PHD,is proposed in this paper.By superimposing the target weight parameter into the fixed empirical gate threshold,the IMPS-GM-PHD filter improves the standard measurement gate technology,which avoids the problem of missing measurement caused by the empirical threshold-based measurement gate technology.Experiment results show that the IMPS-GM-PHD filter not only has high accuracy of target state estimation but also has a relatively low computation cost.

关 键 词：目标跟踪 概率假设密度 量测划分 状态估计 

分 类 号：TN953[电子电信—信号与信息处理]