α稳定分布噪声下一种稳健加权滤波的统一框架

针对传统滤波方法在α稳定分布噪声环境下性能退化的问题,从加权Myriad滤波以及加权Merid滤波方法出发,以M估计理论为基础,推导得到稳健加权(robust weighted,RW)滤波方法的统一算法结构,并据此提出了基于RW滤波的新算法,即基于稳健加权滤波的统一框架,从而将加权Myriad、加权Merid以及基于广义柯西分布的加权滤波器统一起来。此外,针对线性调频(linear frequency modulation, LFM)信号采用基于RW的LVD(RW LVD)方法估计其参数,并根据估计性能对RW方法的抑噪效果进行分析。仿真结果表明,与基于加权Myriad滤波、加权Merid滤波以及基于广义柯西分布的加权滤波等多种方法相比,在强脉冲噪声下RW滤波方法能有效抑制脉冲噪声,并具有良好的稳健性。

Unified framework of robust weighted filtering in α stable noise

In order to solve the problem that the performance of traditional filtering methods degrades significantly inαstable noise environment, a unified structure of robust weighted (RW) filtering is derived based on the M estimation theory, the weighted Myriad and the weighted Merid filtering methods, and accordingly a RW filtering method, i.e., the unified framework, is proposed. The weighted Myriad, the weighted Merid and the generalized Cauchy distribution based weighted filtering methods are interpreted within the unifying framework. The parameters of noisy linear frequency modulation (LFM) signals can be estimated by the Lv’s distribution method based on the RW (RW-LVD) theory, and the estimation results are used to analyze the noise suppression performance. Simulation results show that compared with the weighted Myriad, the weighted Merid as well as the generalized Cauchy distribution based filtering methods, the proposed RW filtering method has better performance and it is robust to αstable noise.