基于两次傅里叶变换的时域MUSIC波达方向估计

针对基于频域多重信号分类（MUSIC）的波达方向（DOA）估计方法在有效快拍数较少情况下的不稳定问题，提出了一种基于两次傅里叶变换的时域MUSIC波达方向估计方法。首先，通过傅里叶变换将各阵元接收数据转换为频域数据，并按扫描角度对各阵元数据进行相位补偿；然后，再通过傅里叶变换将补偿后的频域共轭数据转换为时域复解析数据，在时域构建相移后的协方差矩阵；最后利用特征分解求取具有正交特性的噪声子空间，获得扫描方位空间谱，实现对波达方向估计。数值仿真及实测数据处理结果均表明，相比频域MUSIC方法，在一次有效快拍条件下，所提方法可稳定获得具有正交特性的噪声子空间，实现对波达方向估计；稳定性得到了5 dB的改善，背景噪声级和旁瓣级得到了3 dB以上的改善，因此该方法可明显提高目标检测和方位估计性能。

Time-domain MUSIC for DOA estimation based on twice Fourier transform

To solve the instability of Direction of Arrival(DOA) estimation of frequency-domain Multiple Signal Classification(MUSIC) in the case of few snapshots,a time-domain MUSIC method for DOA estimation based on twice Fourier Transform is proposed.Firstly,the frequency-domain form of received data of each element is obtained by the Fourier Transform whose phase is compensated in the search angle.Then,the time-domain complex analysis data is translated from the frequency-domain conjugate data after phase compensation by the Fourier Transform,and the covariance matrix is constructed in time-domain.Finally,the orthogonal noise subspace is solved by eigen-decomposition,the spatial spectrum of search angle is obtained,and the DOA is estimated.The processed results of numerical simulation and measured data show that,compared with the frequency-domain MUSIC method,under the case of one effective snapshot,the proposed method can statistically solve the orthogonal noise subspace and realize DOA estimation.The stability is improved 5 dB,and the background noise level and side-lobe level of spatial spectrum is improved more than 3 dB.So the proposed method has better performance of target detection and DOA estimation.