一种改进的粒子滤波检测前跟踪算法  

An improved particle filter track-before-detect algorithm

作　　者：高广顺 陈霄[1] Gao Guangshun;Chen Xiao(National Defense Key Discipline Laboratory of Communication and Information Transmission and Fusion Technology,Hangzhou Dianzi University,Hangzhou 310018,China)

机构地区：[1]杭州电子科技大学通信信息传输与融合技术国防重点学科实验室,浙江杭州310018

出　　处：《电子技术应用》2020年第4期61-65,共5页Application of Electronic Technique

基　　金：国防科技重点实验室基金项目(6142804180407)。

摘　　要：针对多个信噪比相差较大时容易发生的目标漏检问题,提出了一种改进的多目标双层粒子滤波检测前跟踪算法(IM-PF-TBD)。算法采用双层粒子滤波结构,在目标检测层中,采用锦标赛选择方法对检测粒子群进行重采样,选取多个权重相差较大的粒子,通过粒子聚类同时检测多个目标,提高了检测初期较弱目标的存在概率。此外,算法提出了粒子群融合方法用于新发现目标的验证,便于目标检测后虚假目标的剔除。仿真结果表明,所提算法能有效改善信噪比较小的目标的检测概率并降低目标RMSE。Aiming at the problem of missed targets that easily occur when there are large differences in signal-to-noise ratio,this paper proposes an improved multi-target dual-layer particle filter track-before-detect algorithm(IM-PF-TBD).The algorithm uses a two-layer particle filter structure.In the target detection layer,the method of tournament selection is used to resample the detected particle group,select multiple particles with large weight differences,and simultaneously detect multiple targets through clustering.The existence probability of weak target in early detection is improved.In addition,the algorithm proposes a particle swarm fusion method for verification of newly discovered targets,which facilitates the removal of false targets after target detection.Simulation results show that the proposed algorithm can effectively improve the detection probability of targets with small signal-to-noise ratio and reduce the target RMSE.

关 键 词：粒子滤波 检测前跟踪 多目标 锦标赛选择 粒子群融合 

分 类 号：TN95[电子电信—信号与信息处理]