Investigations on Sparse System Identification with l0-LMS, Zero-Attracting LMS and Linearized Bregman Iterations

Andreas Gebhard; Michael Lunglmayr; Mario Huemer

doi:10.1007/978-3-319-74727-9_19

Investigations on Sparse System Identification with l0-LMS, Zero-Attracting LMS and Linearized Bregman Iterations

Andreas Gebhard
, Michael Lunglmayr
, Mario Huemer

Institute of Signal Processing

Research output: Chapter in Book/Report/Conference proceeding › Conference proceedings › peer-review

Abstract

Identifying a sparse system impulse response is often performed with the l0-LMS-, or the zero-attracting LMS algorithm. Recently, a linearized Bregman (LB) iteration based sparse LMS algorithm has been proposed for this task. In this contribution, the mentioned algorithms are compared with respect to their parameter dependency, convergence speed, mean-squared error (MSE), and sparsity of the estimate. The performance of the LB iteration based sparse LMS algorithm only slightly depends on its parameters. In our opinion it is the favorable choice in terms of achieving sparse impulse response estimates and low MSE. Especially when using an extension called micro-kicking the LB based algorithms converge much faster than the l0-LMS.

Original language	English
Title of host publication	Computer Aided Systems Theory - EUROCAST 2017
Editors	Franz Pichler, Roberto Moreno-Diaz, Alexis Quesada-Arencibia
Place of Publication	Cham
Publisher	Springer International Publishing
Pages	161-169
Number of pages	9
Volume	10672
ISBN (Print)	978-3-319-74727-9
DOIs	https://doi.org/10.1007/978-3-319-74727-9_19
Publication status	Published - Jan 2018

Publication series

Name	Lecture Notes in Computer Science (LNCS)

Fields of science

202 Electrical Engineering, Electronics, Information Engineering
202022 Information technology
202037 Signal processing

JKU Focus areas

Computation in Informatics and Mathematics
Mechatronics and Information Processing

Access to Document

10.1007/978-3-319-74727-9_19

Cite this

Gebhard, A., Lunglmayr, M., & Huemer, M. (2018). Investigations on Sparse System Identification with l0-LMS, Zero-Attracting LMS and Linearized Bregman Iterations. In F. Pichler, R. Moreno-Diaz, & A. Quesada-Arencibia (Eds.), Computer Aided Systems Theory - EUROCAST 2017 (Vol. 10672, pp. 161-169). (Lecture Notes in Computer Science (LNCS)). Springer International Publishing. https://doi.org/10.1007/978-3-319-74727-9_19

Gebhard, Andreas ; Lunglmayr, Michael ; Huemer, Mario. / Investigations on Sparse System Identification with l0-LMS, Zero-Attracting LMS and Linearized Bregman Iterations. Computer Aided Systems Theory - EUROCAST 2017. editor / Franz Pichler ; Roberto Moreno-Diaz ; Alexis Quesada-Arencibia. Vol. 10672 Cham : Springer International Publishing, 2018. pp. 161-169 (Lecture Notes in Computer Science (LNCS)).

@inproceedings{2a2e3ba72f924c21ac39716a8916f7c7,

title = "Investigations on Sparse System Identification with l0-LMS, Zero-Attracting LMS and Linearized Bregman Iterations",

abstract = "Identifying a sparse system impulse response is often performed with the l0-LMS-, or the zero-attracting LMS algorithm. Recently, a linearized Bregman (LB) iteration based sparse LMS algorithm has been proposed for this task. In this contribution, the mentioned algorithms are compared with respect to their parameter dependency, convergence speed, mean-squared error (MSE), and sparsity of the estimate. The performance of the LB iteration based sparse LMS algorithm only slightly depends on its parameters. In our opinion it is the favorable choice in terms of achieving sparse impulse response estimates and low MSE. Especially when using an extension called micro-kicking the LB based algorithms converge much faster than the l0-LMS.",

author = "Andreas Gebhard and Michael Lunglmayr and Mario Huemer",

year = "2018",

month = jan,

doi = "10.1007/978-3-319-74727-9\_19",

language = "English",

isbn = "978-3-319-74727-9",

volume = "10672",

series = "Lecture Notes in Computer Science (LNCS)",

publisher = "Springer International Publishing",

pages = "161--169",

editor = "Franz Pichler and Roberto Moreno-Diaz and Alexis Quesada-Arencibia",

booktitle = "Computer Aided Systems Theory - EUROCAST 2017",

}

Gebhard, A, Lunglmayr, M & Huemer, M 2018, Investigations on Sparse System Identification with l0-LMS, Zero-Attracting LMS and Linearized Bregman Iterations. in F Pichler, R Moreno-Diaz & A Quesada-Arencibia (eds), Computer Aided Systems Theory - EUROCAST 2017. vol. 10672, Lecture Notes in Computer Science (LNCS), Springer International Publishing, Cham, pp. 161-169. https://doi.org/10.1007/978-3-319-74727-9_19

Investigations on Sparse System Identification with l0-LMS, Zero-Attracting LMS and Linearized Bregman Iterations. / Gebhard, Andreas; Lunglmayr, Michael ; Huemer, Mario.
Computer Aided Systems Theory - EUROCAST 2017. ed. / Franz Pichler; Roberto Moreno-Diaz; Alexis Quesada-Arencibia. Vol. 10672 Cham: Springer International Publishing, 2018. p. 161-169 (Lecture Notes in Computer Science (LNCS)).

Research output: Chapter in Book/Report/Conference proceeding › Conference proceedings › peer-review

TY - GEN

T1 - Investigations on Sparse System Identification with l0-LMS, Zero-Attracting LMS and Linearized Bregman Iterations

AU - Gebhard, Andreas

AU - Lunglmayr, Michael

AU - Huemer, Mario

PY - 2018/1

Y1 - 2018/1

N2 - Identifying a sparse system impulse response is often performed with the l0-LMS-, or the zero-attracting LMS algorithm. Recently, a linearized Bregman (LB) iteration based sparse LMS algorithm has been proposed for this task. In this contribution, the mentioned algorithms are compared with respect to their parameter dependency, convergence speed, mean-squared error (MSE), and sparsity of the estimate. The performance of the LB iteration based sparse LMS algorithm only slightly depends on its parameters. In our opinion it is the favorable choice in terms of achieving sparse impulse response estimates and low MSE. Especially when using an extension called micro-kicking the LB based algorithms converge much faster than the l0-LMS.

AB - Identifying a sparse system impulse response is often performed with the l0-LMS-, or the zero-attracting LMS algorithm. Recently, a linearized Bregman (LB) iteration based sparse LMS algorithm has been proposed for this task. In this contribution, the mentioned algorithms are compared with respect to their parameter dependency, convergence speed, mean-squared error (MSE), and sparsity of the estimate. The performance of the LB iteration based sparse LMS algorithm only slightly depends on its parameters. In our opinion it is the favorable choice in terms of achieving sparse impulse response estimates and low MSE. Especially when using an extension called micro-kicking the LB based algorithms converge much faster than the l0-LMS.

UR - http://eurocast2017.fulp.ulpgc.es/

U2 - 10.1007/978-3-319-74727-9_19

DO - 10.1007/978-3-319-74727-9_19

M3 - Conference proceedings

SN - 978-3-319-74727-9

VL - 10672

T3 - Lecture Notes in Computer Science (LNCS)

SP - 161

EP - 169

BT - Computer Aided Systems Theory - EUROCAST 2017

A2 - Pichler, Franz

A2 - Moreno-Diaz, Roberto

A2 - Quesada-Arencibia, Alexis

PB - Springer International Publishing

CY - Cham

ER -

Gebhard A, Lunglmayr M , Huemer M. Investigations on Sparse System Identification with l0-LMS, Zero-Attracting LMS and Linearized Bregman Iterations. In Pichler F, Moreno-Diaz R, Quesada-Arencibia A, editors, Computer Aided Systems Theory - EUROCAST 2017. Vol. 10672. Cham: Springer International Publishing. 2018. p. 161-169. (Lecture Notes in Computer Science (LNCS)). doi: 10.1007/978-3-319-74727-9_19

Investigations on Sparse System Identification with l0-LMS, Zero-Attracting LMS and Linearized Bregman Iterations

Abstract

Publication series

Fields of science

JKU Focus areas

Access to Document

Other files and links

Christian Doppler Laboratory for Digitally Assisted RF Transceivers for Future Mobile Communications

Cite this

Investigations on Sparse System Identification with l0-LMS, Zero-Attracting LMS and Linearized Bregman Iterations

Abstract

Publication series

Fields of science

JKU Focus areas

Access to Document

Other files and links

Projects

Christian Doppler Laboratory for Digitally Assisted RF Transceivers for Future Mobile Communications

Cite this