ScholarLens Back to search
2002•Unpublished venueRequires access

Performance analysis of a new implementation for the frequency-domain LMS adaptive filter

A.O. Ogunfunmi

Open publisher page 2 citations

Abstract

The author analyzes the performance of a novel implementation of the frequency-domain LMS (least mean square) adaptive filter. This implementation is based on using the LMS adaptive algorithm to implement the DFT (discrete Fourier transform) part of the frequency-domain LMS adaptive filter. Numerical issues are also discussed. This is significant in this implementation because the performance of the frequency-domain LMS adaptive filter is dependent on the accuracy of the DFT computation obtained by using the LMS algorithm. The effects of inaccurate transform (DFT) computation of the performance of the transform-domain LMS adaptive filter are analyzed. The author also used a linear filter model to explain some of the observed nonideal effects. This indicated that, for large filter lengths, the accuracy of the DFTs computed diminishes.>

About this research paper

What this paper is about

The author analyzes the performance of a novel implementation of the frequency-domain LMS (least mean square) adaptive filter. This implementation is based on using the LMS adaptive algorithm to implement the DFT (discrete Fourier transform) part of the frequency-domain LMS adaptive filter. Numerical issues are also discussed. This is significant in this implementation because the performance of the frequency-domain LMS adaptive filter is dependent on the accuracy of the DFT computation obtained by using the LMS algorithm. The effects of inaccurate transform (DFT) computation of the performance of the transform-domain LMS adaptive filter are analyzed. The author also used a linear filter model to explain some of the observed nonideal effects. This indicated that, for large filter lengths, the accuracy of the DFTs computed diminishes.>

Why it matters

OpenAlex reports 2 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

The author analyzes the performance of a novel implementation of the frequency-domain LMS (least mean square) adaptive filter. This implementation is based on using the LMS adaptive algorithm to implement the DFT (discrete Fourier transform) part of the frequency-domain LMS adaptive filter. Numerical issues are also discussed. This is significant in this implementation because the performance of the frequency-domain LMS adaptive filter is dependent on the accuracy of the DFT computation obtained by using the LMS algorithm. The effects of inaccurate transform (DFT) computation of the performance of the transform-domain LMS adaptive filter are analyzed. The author also used a linear filter model to explain some of the observed nonideal effects. This indicated that, for large filter lengths, the accuracy of the DFTs computed diminishes.>

Key concepts: Multidelay block frequency domain adaptive filter, Adaptive filter, Least mean squares filter, Frequency domain, Filter (signal processing), Discrete Fourier transform (general), Computer science, Kernel adaptive filter

Related papers

Related paper 1Related paper 2Related paper 3
Back to paper searchBrowse research topicsOriginal source
Performance analysis of a new implementation for the frequency-domain LMS adaptive filter — Research Paper | ScholarLens