存在幅相误差时谱峰快拍实现频率和到达角估计

Frequency and DOA estimation using snapshots nearby frequency spectrum peaks in the presence of unknown sensor gains and phases error

针对多通道测频测向系统中存在的幅相误差,将频域谱峰3 dB内的数据取作快拍,提出对谱峰存在单/多信源时分别采用改进单次快拍模型和加权子空间投影实现到达角估计的方法.该方法无需协方差矩阵运算和频率到达角的配对操作,对独立或相干源均有效,且并行传输数据量压缩比为3 dB信号带宽与采样率之比,适用于工程实时性要求高的场合.等距线阵实测数据处理验证了该方法对相干源的有效性和对幅相误差的鲁棒性;平面阵仿真数据处理证明,存在幅相误差情况下,该方法比MUSIC法具有更高的分辨力,能够分辨出到达角小于1/2波束宽度以内的两信号.

A new method for frequency and direction of arrival（DOA） estimation using snapshots nearby frequency spectrum peaks is proposed for application in the multichannel system in the presence of unknown sensor gains and phase.A modified single snapshot mode for only one source contributing to a frequency spectrum peak and a weighted subspace projection for multiple sources are proposed,respectively.Without the calculation of the spatial covariance matrix and pairing operation of frequency and direction of arrival,the method is effective for both coherent and incoherent sources scenarios and may be applied in the multichannel system due to the high compress ratio of the 3dB signal bandwidth to the sampling rate.The measured data based on a uniform linear array illustrate the effectiveness for coherent sources and the robustness to gains and phases mismatch and the simulated data based on a plane array verify the higher angle resolution in comparison with the MUSIC algorithm.