论文部分内容阅读
为了解决秩亏RCF(robust Capon filter-bank)谱估计方法的估计性能不稳健问题,提出一种基于奇异协方差矩阵的谱估计方法。该方法根据空间约束和导向矢量误差约束获得相应的窄带滤波器,并利用其对Capon滤波矢量进行滤波得到谱估计。仿真结果表明:在采样协方差矩阵奇异的情况下,该谱估计方法获得的谱估计在精度及分辨率上均明显优于秩亏RCF谱估计方法。该方法是一种可以利用少量快拍进行谱估计的有效方法。
In order to solve the problem of unsteady performance estimation of robust Capon filter-bank (RSF) algorithm, a spectral estimation method based on singular covariance matrix is proposed. The method obtains the corresponding narrow-band filter according to space constraint and steering vector error constraint, and uses it to filter the Capon filter vector to get the spectral estimation. The simulation results show that the spectral estimation obtained by the spectral estimation method is significantly superior to the rank-loss RCF spectral estimation method in terms of accuracy and resolution when the sampling covariance matrix is singular. This method is an effective method for spectral estimation with a small amount of snapshots.