Open AccessIterative Wiener filter
Author(s) -
Xi Bin,
Liu Yuehong
Publication year - 2013
Publication title -
electronics letters
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.375
H-Index - 146
eISSN - 1350-911X
pISSN - 0013-5194
DOI - 10.1049/el.2013.0009
Subject(s) - wiener filter , covariance matrix , mathematics , adaptive filter , covariance , convergence (economics) , filter (signal processing) , residual , control theory (sociology) , matrix (chemical analysis) , algorithm , kernel adaptive filter , stability (learning theory) , iterative method , mathematical optimization , computer science , filter design , statistics , artificial intelligence , materials science , control (management) , machine learning , economics , composite material , computer vision , economic growth
A new adaptive filter algorithm, the iterative Wiener filter (IWF), is proposed to overcome the drawback of slow convergence speed for most LMS‐type algorithms. The adaptive filter is posed as a problem of finding the solution of a linear matrix equation, equivalent to the Wiener equation. Then the step size is optimised, which is time variant in terms of the residual error in each step. This property gives the IWF the ability of fast convergent speed. The stability of the algorithm can be secured when the estimation of covariance and cross‐covariance statistics become stationary. Only the product of the matrix and vector is needed for the implementation in each iteration. Numerical results demonstrate the superior performance of the IWF over some other LMS‐type algorithms.