Misalignment analysis and insights into the performance of clipped-input LMS with correlated Gaussian data

Mehdi Bekrani, Andy W.H. Khong

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

The three-level clipped input least-mean-square (CLMS) adaptive algorithm is known to have low complexity that is suitable for the identification of long finite impulse response of unknown systems. In this paper we analyze the performance of CLMS which allows one to gain insights into its convergence property and the amount of steady-state misalignment error for both time-invariant and time-varying systems perturbed by correlated Gaussian input. Arising from our analysis, we derive the optimal step-size for CLMS to achieve the minimum possible steady-state misalignment and compare its results with the performance of LMS adaptive algorithm. The accuracy of our derivations is evaluated with simulation results.

Original languageEnglish
Title of host publication2014 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2014
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages5929-5933
Number of pages5
ISBN (Print)9781479928927
DOIs
Publication statusPublished - 2014
Externally publishedYes
Event2014 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2014 - Florence, Italy
Duration: May 4 2014May 9 2014

Publication series

NameICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
ISSN (Print)1520-6149

Conference

Conference2014 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2014
Country/TerritoryItaly
CityFlorence
Period5/4/145/9/14

ASJC Scopus Subject Areas

  • Software
  • Signal Processing
  • Electrical and Electronic Engineering

Keywords

  • Adaptive filter
  • Clipped input LMS
  • Misalignment
  • Tracking

Fingerprint

Dive into the research topics of 'Misalignment analysis and insights into the performance of clipped-input LMS with correlated Gaussian data'. Together they form a unique fingerprint.

Cite this