基于噪声特征向量重构的地波雷达单次快拍超分辨算法  

Single snapshot superresolution algorithm for HFSWR based on noise eigenvector reconstruction

作　　者：位寅生[1] 童鹏[1] 郭晓江[1] 

机构地区：[1]哈尔滨工业大学电子与信息工程学院,黑龙江哈尔滨150001

出　　处：《系统工程与电子技术》2013年第3期493-498,共6页Systems Engineering and Electronics

基　　金：国家自然科学重点基金(61032011)资助课题

摘　　要：高频地波雷达在一个相参积累时间内通常只能得到频域一次快拍,利用其直接进行波达方向估计性能较差。针对这一问题,在分析时、频域协方差矩阵特征分解后差异的基础上,在频域采用降维方法估计协方差矩阵。根据频域目标信噪比相对较大的特点,利用最大特征值对应的信号特征向量构造原始的数据矩阵,并对该矩阵进行奇异值分解得到新的噪声子空间,进而构造出新的噪声特征向量,最后利用该噪声特征向量进行方位角估计。仿真和实测数据分析验证了算法的有效性,相比降维Toeplitz算法和前后向空间平滑算法有着更高的分辨力和估计精度。High frequency radar can only obtain one snapshot of targets in frequency domain during a coherent processing interval. It leads to a degraded estimation of direction of arrival （DOA）. By comparing the eigen decomposition of the covariance matrix estimation method in temporal with that in frequency domain, a dimension-reduction method is used to estimate the covariance matrix with a single snapshot in frequency domain, in which a relatively higher signal-to-noise ratio （SNR） is achieved. An original data matrix is then constructed by the signal eigenvector correspond to the biggest eigenvalue after the eigen decomposition of the covariance matrix. The noise subspace is obtained by the singular value decomposition （SVD） of the data matrix. After that, a noise eigenvector is reconstructed by combining all the noise eigenvectors. Finally, the new noise eigenvector is used to perform the DOA estimation. Simulation and experiment data analysis demonstrate that the proposed algorithm outperforms both dimensionreduction Toeplitz method and forward-backward smoothing method on precision and resolution.

关 键 词：超分辨 单次快拍 噪声特征向量 地波雷达 

分 类 号：TN958[电子电信—信号与信息处理]