东北大学学报(自然科学版) ›› 2013, Vol. 34 ›› Issue (5): 654-657.DOI: -

• 信息与控制 • 上一篇    下一篇

基于加权L1范数的CS-DOA算法

刘福来1,彭泸2,汪晋宽1,杜瑞燕1   

  1. (1东北大学秦皇岛分校,河北秦皇岛066004;2东北大学信息科学与工程学院,辽宁沈阳110819)
  • 收稿日期:2014-08-19 修回日期:2014-08-19 出版日期:2013-05-15 发布日期:2013-07-09
  • 通讯作者: 刘福来
  • 作者简介:刘福来(1975-), 男,河北唐山人,东北大学教授,博士生导师;汪晋宽(1957-),男,辽宁沈阳人,东北大学教授,博士生导师.
  • 基金资助:
    国家自然科学基金资助项目(60904035);中央高校基本科研业务费专项资金资助项目(N120423002);河北省科技厅资助项目(11213502D);辽宁省自然科学基金资助项目(20102064);河北省教育厅资助项目(Z2009105);辽宁省高等学校优秀人才支持计划项目(LJQ2012022).

CSDOA Algorithm Based on Weighted L1 Norm

LIU Fulai1, PENG Lu2, WANG Jinkuan1, DU Ruiyan1   

  1. 1. Northeastern University at Qinhuangdao, Qinhuangdao 066004, China; 2. School of Information Science & Engineering, Northeastern University, Shenyang 110819, China.
  • Received:2014-08-19 Revised:2014-08-19 Online:2013-05-15 Published:2013-07-09
  • Contact: LIU Fulai
  • About author:-
  • Supported by:
    -

摘要: 针对基于L1范数约束的压缩感知理论的恢复算法出现虚假目标,恶化DOA估计性能的问题,提出了一种基于加权L1范数的CS-DOA估计算法.该算法利用噪声子空间与信号子空间的正交性,构造了一个加权矩阵,然后对L1范数约束模型进行加权.通过此加权处理,该算法能够使恢复的系数向量具有更好的稀疏性,并能有效地抑制伪峰,从而获得更精确的DOA估计.仿真结果验证了算法的有效性.

关键词: 波达方向估计, 压缩感知, 奇异值分解, 加权矩阵, L1范数最小化

Abstract: The recovery algorithm of compressive sensing (CS) based on L1 norm constraint may lead to many false targets and deteriorate the performance of DOA estimation. To solve the above problem, a CSDOA algorithm based on weighted L1 norm was proposed. Using the orthogonality between noise subspace and signal subspace, a weighted matrix was constructed to penalize the L1 norm constrained model. By the weighted processing, the reconstructed coefficient vector with better sparsity could be achieved by using the presented algorithm. What’s more, the spurious peaks could also be effectively suppressed. Finally, more accurate DOA estimation could be obtained. Simulation results showed the efficiency of the proposed method.

Key words: DOA estimation, compressive sensing, SVD(singular value decomposition), weighted matrix, L1 norm minimization

中图分类号: