Modal analysis of the l0-LMS and l0-NLMS sparse adaptive algorithms

Abdullah Al-Shabili, Luis Weruaga, Shihab Jimaa

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

The l0-Least Mean Squares (l0-LMS) and l0- Normalized LMS (l0-NLMS) are arguably the best among gradient adaptive algorithms for sparse system identification. However, due to the non-linear and non-convex sparse penalty term in their cost functions, deriving analytical modals for the Mean Square Deviation (MSD) update equation is quiet challenging. In this paper, the significant and zero taps misalignment is studied separately, and then joined in a dynamical manner. Thus, we propose the MSD update equations for both l0-LMS and l0-NLMS, with reasonable assumptions for white input signal. Moreover, the steady state MSD of both algorithms is presented. Simulation results illustrate strong agreement between the derived analytical modals and the empirical simulation.

Original languageBritish English
Title of host publication2016 IEEE 59th International Midwest Symposium on Circuits and Systems, MWSCAS 2016
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781509009169
DOIs
StatePublished - 2 Mar 2017
Event59th IEEE International Midwest Symposium on Circuits and Systems, MWSCAS 2016 - Abu Dhabi, United Arab Emirates
Duration: 16 Oct 201619 Oct 2016

Publication series

NameMidwest Symposium on Circuits and Systems
ISSN (Print)1548-3746

Conference

Conference59th IEEE International Midwest Symposium on Circuits and Systems, MWSCAS 2016
Country/TerritoryUnited Arab Emirates
CityAbu Dhabi
Period16/10/1619/10/16

Fingerprint

Dive into the research topics of 'Modal analysis of the l0-LMS and l0-NLMS sparse adaptive algorithms'. Together they form a unique fingerprint.

Cite this