Abstract:

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.