A low-complexity RLS-DCD algorithm for Volterra system identification

Raffaello Claser; Vitor Nascimento; Yuriy Zakharov

doi:10.1109/EUSIPCO.2016.7760199

A low-complexity RLS-DCD algorithm for Volterra system identification

Raffaello Claser, Vitor Nascimento, Yuriy Zakharov

Electronic Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Adaptive filters for Volterra system identification must deal with two difficulties: large filter length M (resulting in high computational complexity and low convergence rate) and high correlation in the input sequence. The second problem is minimized by using the recursive least-squares algorithm (RLS), however, its large computation complexity (O(M2)) might be prohibitive in some applications. We propose here a low-complexity RLS algorithm, based on the dichotomous coordinate descent algorithm (DCD), showing that in some situations the computational complexity is reduced to O(M). The new algorithm is compared to the standard RLS, normalized least-mean squares (NLMS) and affine projections (AP) algorithms.

Original language	English
Title of host publication	24th European Signal Processing Conference (EUSIPCO)
Place of Publication	Budapest, Hungary
Number of pages	5
ISBN (Electronic)	978-0-9928-6265-7
DOIs	https://doi.org/10.1109/EUSIPCO.2016.7760199
Publication status	Published - 29 Aug 2016

Access to Document

10.1109/EUSIPCO.2016.7760199

Cite this

@inproceedings{7de3767b727a4a9a952ef7553c15990b,

title = "A low-complexity RLS-DCD algorithm for Volterra system identification",

abstract = "Adaptive filters for Volterra system identification must deal with two difficulties: large filter length M (resulting in high computational complexity and low convergence rate) and high correlation in the input sequence. The second problem is minimized by using the recursive least-squares algorithm (RLS), however, its large computation complexity (O(M2)) might be prohibitive in some applications. We propose here a low-complexity RLS algorithm, based on the dichotomous coordinate descent algorithm (DCD), showing that in some situations the computational complexity is reduced to O(M). The new algorithm is compared to the standard RLS, normalized least-mean squares (NLMS) and affine projections (AP) algorithms.",

author = "Raffaello Claser and Vitor Nascimento and Yuriy Zakharov",

year = "2016",

month = aug,

day = "29",

doi = "10.1109/EUSIPCO.2016.7760199",

language = "English",

booktitle = "24th European Signal Processing Conference (EUSIPCO)",

}

TY - GEN

T1 - A low-complexity RLS-DCD algorithm for Volterra system identification

AU - Claser, Raffaello

AU - Nascimento, Vitor

AU - Zakharov, Yuriy

PY - 2016/8/29

Y1 - 2016/8/29

N2 - Adaptive filters for Volterra system identification must deal with two difficulties: large filter length M (resulting in high computational complexity and low convergence rate) and high correlation in the input sequence. The second problem is minimized by using the recursive least-squares algorithm (RLS), however, its large computation complexity (O(M2)) might be prohibitive in some applications. We propose here a low-complexity RLS algorithm, based on the dichotomous coordinate descent algorithm (DCD), showing that in some situations the computational complexity is reduced to O(M). The new algorithm is compared to the standard RLS, normalized least-mean squares (NLMS) and affine projections (AP) algorithms.

AB - Adaptive filters for Volterra system identification must deal with two difficulties: large filter length M (resulting in high computational complexity and low convergence rate) and high correlation in the input sequence. The second problem is minimized by using the recursive least-squares algorithm (RLS), however, its large computation complexity (O(M2)) might be prohibitive in some applications. We propose here a low-complexity RLS algorithm, based on the dichotomous coordinate descent algorithm (DCD), showing that in some situations the computational complexity is reduced to O(M). The new algorithm is compared to the standard RLS, normalized least-mean squares (NLMS) and affine projections (AP) algorithms.

U2 - 10.1109/EUSIPCO.2016.7760199

DO - 10.1109/EUSIPCO.2016.7760199

M3 - Conference contribution

BT - 24th European Signal Processing Conference (EUSIPCO)

CY - Budapest, Hungary

ER -