Khatri-Rao积变换下的离格信号DOA估计

当存在离格信号时,基于稀疏表示的波达角（DOA）估计算法性能损失严重。为解决这个问题,在对接收数据协方差矩阵进行Khatri-Rao积变换的基础上,推导了离格信号网格偏离量与紧邻信号原子系数之间的关系,提出了一种单一离格信号DOA估计方法。为提高对邻近离格信号DOA的估计性能,利用矩阵的广义逆性质提出了基于多原子系数的联合估计方法。仿真实验表明,单一离格信号DOA估计方法在低信噪比下有较好的性能,联合估计方法在高信噪比条件下对邻近离格信号DOA有较高的估计精度,同时所提算法估计性能几乎不受网格划分间距的影响,可以通过增大网格间距降低算法运算量。相关研究对阵列天线DOA估计具有一定的参考价值。...

Full description

Saved in:
Bibliographic Details
Published in电讯技术 Vol. 57; no. 2; pp. 203 - 209
Main Author 李根 梁玉英
Format Journal Article
LanguageChinese
Published 军械工程学院 电子与光学工程系,石家庄,050003 2017
Subjects
DOA估计
Khatri-Rao积变换
压缩感知
离格信号
稀疏表示
离格信号
DOA estimation
压缩感知
off-grid signal
DOA估计
Khatri-Rao积变换
Khatri-Rao product transform
com―pressive sensing
稀疏表示
sparse representation
Online AccessGet full text
ISSN1001-893X
DOI10.3969/j.issn.1001-893x.2017.02.014

Cover

More Information
Summary:当存在离格信号时,基于稀疏表示的波达角（DOA）估计算法性能损失严重。为解决这个问题,在对接收数据协方差矩阵进行Khatri-Rao积变换的基础上,推导了离格信号网格偏离量与紧邻信号原子系数之间的关系,提出了一种单一离格信号DOA估计方法。为提高对邻近离格信号DOA的估计性能,利用矩阵的广义逆性质提出了基于多原子系数的联合估计方法。仿真实验表明,单一离格信号DOA估计方法在低信噪比下有较好的性能,联合估计方法在高信噪比条件下对邻近离格信号DOA有较高的估计精度,同时所提算法估计性能几乎不受网格划分间距的影响,可以通过增大网格间距降低算法运算量。相关研究对阵列天线DOA估计具有一定的参考价值。
Bibliography:When the off-grid signals appear, the performance of direction-of-arrival （DOA） estimation algorithms based on sparse representation suffer serious loss. In order to solve this problem, the Khatri-Rao product transform is performed for covariance matrix of received data, the theoretical relationship between grid deviation and atomic coefficients close to the off-grid signal is derived, and then a method of single off -grid signal DOA estimation is proposed. For improving the DOA estimation performance of off-grid signals with approaching angles, a joint estimation method based on the multiple atomic coefficients is proposed by utilizing the generalized inverses properties. The simulation experiments show that the single off- grid signal DOA estimation method has excellent performance under low signal-to-noise ratio（SNR） and the joint estimation method has better performance facing the adjacent off-grid signal under high SNR. At the same time, the grid spacing has few influences on the estimation performance o
ISSN:1001-893X
DOI:10.3969/j.issn.1001-893x.2017.02.014