基于δ-广义标签多贝努利的群分裂算法  

Group Target Splitting Algorithm Based on δ-GLMB

作　　者：马艳琴[1] 甘林海 王刚[3] MA Yanqin;GAN Linhai;WANG Gang(Nanjing Research Institute of Electronics Technology,Nanjing 210039,China;School of Graduate,Air Force Engineering University,Xi'an 710051,China;Air and Missle Defense College,Air Force Engineering University,Xi'an 710051,China)

机构地区：[1]南京电子技术研究所,南京210039 [2]空军工程大学研究生院,西安710051 [3]空军工程大学防空反导学院,西安710051

出　　处：《现代雷达》2018年第12期46-51,共6页Modern Radar

摘　　要：群分裂是群目标的重要运动形式,基本的广义标签多贝努利(GLMB)滤波器计算复杂,且只对存活和新生目标建模,不能有效描述群分裂运动。针对此问题,在δ-GLMB滤波算法的基础上建立群分裂模型,通过最小化Kullback-Leibler散度(KL-DIV)的方法近似目标真实密度分布,推导了一种考虑群分裂的GGIW-δ-GLMB滤波算法。仿真实验表明,该算法能够较好地跟踪群分裂运动。Splitting is an motion manner of group target. The original generalized labeled multi-Bernoulli( GLMB) filter,which is complex in calculation,only the surviving and new born targets are modeled in the filter,and the group spawning could not be handled.Aimed at this,one group spawning model is introduced in δ-GLMB filter,and the Gamma Gaussian Inverse Wishart δ-GLMB( GGIW-δ-GLMB) filter with group spawning considered is derived with the true density of the group approximated by Kullback-Leibler divergence( KL-DIV). Simulation shows,the proposed algorithm does improve the performance for group spawning events.

关 键 词：群目标 分裂 δ-广义标签多贝努利算法 Kullback—Leibler散度 伽马高斯逆威夏特 

分 类 号：TN957.52[电子电信—信号与信息处理]