基于微多普勒特征的空间锥体目标识别  

Micro-Doppler signature based cone-shaped target recognition

作　　者：关永胜[1,2] 左群声[3] 刘宏伟[1] 

机构地区：[1]西安电子科技大学雷达信号处理国防科技重点实验室,陕西西安710071 [2]西安电子科技大学通信工程学院,陕西西安710071 [3]中国电子科技集团公司,北京100846

出　　处：《电波科学学报》2011年第2期209-215,412,共8页Chinese Journal of Radio Science

基　　金：长江学者和创新团队发展计划(IRT0954);国家自然科学基金(编号:60772140和60901067);国防预研项目;国防预研基金

摘　　要：空间锥体目标在大气层外飞行过程中存在不同的微动,目标的微动为进动,诱饵的微动为摆动或自旋。根据真假目标的微动差异,提出了低分辨雷达的目标识别方法。在多散射中心信号模型下分析了真假目标的微多普勒特性,指出目标回波信号微多普勒谱与其进动参数密切相关,而诱饵回波信号微多普勒谱与其摆动\自旋频率密切相关。应用低分辨雷达多次回波信号抽取微多普勒域熵与波形标准差作为识别特征,并采用两种分类器进行分类识别。仿真结果表明:所提方法可以有效地实现真假目标的分类。When the cone-shaped target is flying in exo-atmosphere, the micro-mo tion of the targets belongs to precession and that of the decoy belongs to wobble or spin. In order to recognize the targets from its decoys, a method is proposed based on the low resolution radar signal. Based on the difference in the micro-Doppler, entropy and variance in frequency domain as recognition feature are extracted. Analyze the micro-Doppler signature of the target and the decoy in the multi-sections model, a conclusion is reached that the micro-Doppler spectrum of echo from target depends on the precession parameters and that of the decoy depends on the wobble／ spin parameter. Entropy and variance in frequency domain as recognition feature are extracted. The experiment shows that the method is efficient.

关 键 词：雷达目标识别 微动 微多普勒 进动 摆动 自旋 

分 类 号：TN957.51[电子电信—信号与信息处理]