采用加权L1范数稀疏模型构造DOA估计的方法  

DOA estimation using weighted L1 norm sparse model

作　　者：刘楠[1] 宋文龙[1] 董光辉[1] 冷欣[1] 

机构地区：[1]东北林业大学机电工程学院,黑龙江哈尔滨150040

出　　处：《哈尔滨工程大学学报》2016年第4期603-607,共5页Journal of Harbin Engineering University

基　　金：国家自然科学基金项目(31270757);中央高校基本科研业务费专项资金项目(2572014EB03;DL13BB16);高等学校博士学科点专项科研基金项目(20130062120005)

摘　　要：在以GPS辐射源发出信号作为探测信号的无源雷达系统中,针对回波通道内强干扰低信噪比情况下的多目标波达方向（DOA）估计问题,根据GPS辐射信号以及无源雷达系统特点,提出一种采用改进加权L1范数的约束模型构造DOA估计的方法。先采用批处理抵消算法（extensive cancellation algorithm,ECA）估计目标信号的时延和多普勒频移等参数,消除直达波和多径干扰,然后利用改进权值的L1范数作为约束条件,建立稀疏模型进行DOA估计,在低信噪比环境中无需估计干扰参数,以较低的计算复杂度进行准确DOA估计。仿真结果表明：该方法减少了计算复杂度,在相同配置下运行时间比MUSIC-like方法降低了1.18 s;同时也提高了准确性,其均方误差较MUSIC-like方法和Candes方法降低了0.5°～3.7°,低信噪比环境下分辨概率较MUSIC-like方法和Candes方法提高了0.4～0.6。For a passive radar system based on a GPS illuminator as the detection signal,estimating direction of arrival（ DOA） of multiple targets in an echo channel with strong interference and low signal-to-noise（ SNR） ratio is difficult. Based on the characteristics of GPS-based passive radar system,a DOA estimation method using an improved weighted L1 norm constraint model is proposed. After using the extensive cancellation algorithm（ ECA） to estimate time delay and Doppler frequency shift for removing direct- and multi-path interference,a sparse model for DOA estimation is constructed with the L1-norm as a constraint without parameter estimation; DOA estimation is accurately conducted with reduced computational complexity. Simulation results show that the proposed method can perform well with lower computational complexity,and the execution time reduces by 1.18 s compared to the MUSIC-like method with the same configuration. The mean square error is 0.5° ～ 3.7° lower than that of the MUSIC-like and Candes methods,and resolution probability with low SNR is 0.4 ～ 0.6 higher than MUSIC-like and Candes method.

关 键 词：GPS辐射源 无源雷达 波达方向 L1范数 批处理抵消算法 稀疏模型 均方误差 分辨概率 

分 类 号：TN957.51[电子电信—信号与信息处理]