Convergence Performance Comparison of Transform Domain LMS Adaptive Filters for Correlated Signal

Author

Murmu, Govind ; Nath, Ravinder

Author_Institution

Dept. of Electron. Eng., Indian Sch. of Mines, Dhanbad, India

fYear

2011

fDate

24-25 Feb. 2011

Firstpage

1

Lastpage

5

Abstract

Adaptive filters are very useful in the present world. It is a well known fact that the performance of adaptive filters is highly dependent upon the type of input sequence. Further the uncorrelated input shows a better convergence speed than the correlated input. But most of the real world signals are correlated. Various methods have been proposed to preprocess these inputs by transforming these inputs to different domain and then filtering. These include DFT-LMS, DCT-LMS, DSTLMS and DWT-LMS among some of the popular algorithms.In this paper we analyzed and compared convergence properties of these different transform domain algorithms via computer simulations. The results showed that DWT-LMS perform better than others under certain criterion.

Keywords

adaptive filters; convergence; correlation methods; discrete cosine transforms; discrete wavelet transforms; least mean squares methods; DCT-LMS; DFT-LMS; DST- LMS; DWT-LMS; computer simulation; convergence performance comparison; convergence property; convergence speed; correlated signal; discrete cosine transform; discrete sine transform; discrete wavelet transform; least mean square method; transform domain LMS adaptive filters; transform domain algorithm; Convergence; Correlation; Discrete wavelet transforms; Eigenvalues and eigenfunctions; Least squares approximation; Matrix decomposition;

fLanguage

English

Publisher

ieee

Conference_Titel

Devices and Communications (ICDeCom), 2011 International Conference on

Conference_Location

Mesra

Print_ISBN

978-1-4244-9189-6

Type

conf

DOI

10.1109/ICDECOM.2011.5738534

Filename

5738534