基于灰关联的分类信息辅助多目标跟踪  

Classification-Aided Tracking Multiple Targets Based On Gray Correlation Analysis

作　　者：李斌[1] 姚康泽[2] 王岩[2] 慈林林[2] 万建伟[1] 

机构地区：[1]国防科技大学电子科学与工程学院,长沙410073 [2]北京信息高技术研究所,北京100085

出　　处：《信号处理》2009年第3期398-402,共5页Journal of Signal Processing

基　　金：武器装备项目预研基金项目(9140A07020806JB3302)

摘　　要：密集多回波环境下对弹道多目标跟踪时,通过灰关联方法融合了目标运动状态空间似然值和HRRP特征分类信息,提出了分类信息辅助多目标跟踪的CAT算法;结合目标电磁仿真数据,建立了多目标跟踪和识别交互仿真场景,并与JPDA和NN跟踪算法进行比较,仿真结果表明,CAT克服了轨迹合并和误跟的现象,较好的实现了多目标跟踪。Tracking ballistic muhi-targets in densely multi-return environment,HRRP classification and state-sapce feature of targets is fused through their gray correlation coefficients. A novel CAT （ Classification-Aided Tracking） algorithm is investigated. With electromagnetic simulation data of targets, a simulative scene of combining tracking and classification is constituted. By comparing with conditional JPDA and NN algorithm, the result shows that the CAT algorithm overcomes the phenomena of track-coalition and errortrack, and has nice performance of tracking multiple targets.

关 键 词：多目标跟踪 HRRP 分类辅助跟踪 灰关联 

分 类 号：TN957.56[电子电信—信号与信息处理]