SWT及其在分数阶滤波器设计中的应用

SWT and its application in design of fractional filter

为解决受噪声干扰的非平稳信号的滤波问题,最大化地保留有用信号,提出了由Stockwell-Wigener变换(SWT)辅助设计的分数阶滤波器方法。提出了SWT,它是结合了S变换多分辨率和Wigner-Ville分布高清晰度优点的一种新的信号时频方法,并将SWT用于辅助设计分数阶滤波器,从而进一步最小化被滤波截止线包围的含噪信号时频表示面积,提高滤波后的信噪比。理论分析和仿真实验结果表明,该方法有效且可行,可使恢复信号的归一化均方差得到进一步减小。

To deal with filtering problems of nonstationary signal interfered by noise and reserve the original signal part after filtering as much as possible, the fractional filter designed by the interfered signal＇ s Stockwell-Wigner Transform （SWT） is presented. Firstly, the novel SWT is derived, combining the advantages of ST＇ s multi-resolution and WVD＇ s high clarity, and then it is applied to help design the fractional filter, minimizing the area circled by cutofflines in the time-frequency domain. Theoretically analysis and numerical implementation indicate that the proposed method is effective and could decrease the NMSE of the reconstructed signal.