基于分数低阶统计模型的自适应滤波器组谱估计方法

Some Adaptive Filterbank Spectral Estimation Methods Based on FLOS

本文研究了一些自适应滤波器组谱估计方法(AFSE)在对称alpha-stable噪声(SαS)环境下的性能.提出了三种新的基于分数低阶统计模型(FLOS)的自适应滤波器组谱估计方法(AFSE-FLOS).这些新方法是基于分数低阶协方差(FLOC),因此在对称alpha-stable噪声环境下比基于二阶和一阶统计模型的方法估计性能好.计算机仿真结果表明这些基于分数低阶统计模型的新方法在对称alpha-stable噪声环境下的性能比对应的基于二阶和一阶统计模型的方法性能好,在高斯噪声环境下,与其对应的方法性能相似.

The performance of some adaptive filterbank spectral estimation（APSE） methods in the symmetric alpha-stable noise environment is studied, Three new APSE methods based on the fractional lower order statistics（FLOS） （ AFSE_ FLOS） are presented. These new methods are based on the fractional lower order covariance（FLOC） which are better than the methods based on second order statistics（SOS） and first order statistics（FOS） in the alpha-stable noise. The simulation shows the new methods based on FLOS outperforms their respective methods based on SOS and FOS in the alpha-shable noise while performs similar to their respective methods in the Gaussian noise.