空间平滑和盖氏半径变换的宽带相干信源数估计

Estimation of wideband coherent source number based on spatial smoothing and transformed Gerschgorin radii

采用空间平滑处理和不同信源数估计准则进行宽带相干信源数估计。该方法针对宽带相干信号,不需要聚焦处理,直接对宽带信号阵列采样输出分段并进行快速傅立叶变换(fast Fourier transform,FFT),得到不同频点处的窄带采样协方差矩阵,然后采用空间平滑处理技术进行解相干处理,接着利用信源估计准则进行信源数估计,对每个频点处的结果进行加权处理得到宽带信源数目。仿真结果表明,通过对信息论准则、基于盖氏半径的似然准则、盖氏圆估计准则的性能的对比,进而得到了一种在色噪声下估计宽带相干信源数的方法。

In this paper, spatial smoothing processing and several different sources number estimation criterions are adopted to estimate wideband coherent sources number. The method aims to wideband coherent sources, in which the focusing process is not needed. Firstly, the array sample outputs are divided into some subsections and are transformed into frequency domain data by FFT, then narrowband sampling covarianee matrix on each frequency point can be obtained. Further, spatial smoothing technique is adopted to de-correlate coherent sources, and different criterions are adopted to estimate the sources number, then the number of wideband coherent sources can be determined by weighted the processing results of each frequency point. Finally, the simulation results show that, the performance of the information theory criterion, Gerschgorin likelihood estimators （GLE） criterion, Gerschgorin disk estimator （GDE） criterion are contrasted through computer simulations, and a method is got to estimate wideband coherence source number under the condition of color noise environment.