A Family of Optimized LMS-Based Algorithms for System Identification

被引：0

作者：

Ciochina, Silviu ^{[1
]}

Paleologu, Constantin ^{[1
]}

Benesty, Jacob ^{[2
]}

Grant, Steven L. ^{[3
]}

Anghel, Andrei ^{[1
]}

机构：

[1] Univ Politehn Bucuresti, Bucharest, Romania

[2] Univ Quebec, INRS EMT, Montreal, PQ, Canada

[3] Missouri Univ Sci & Technol, Rolla, MO USA

来源：

2016 24TH EUROPEAN SIGNAL PROCESSING CONFERENCE (EUSIPCO) | 2016年

关键词：

STEP-SIZE NLMS;

D O I：

暂无

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

The performance of the least-mean-square (LMS) algorithm is governed by its step-size parameter. In this paper, we present a family of optimized LMS-based algorithms (in terms of the step-size control), in the context of system identification. A time-variant system model is considered and the optimization criterion is based on the minimization of the system misalignment. Simulations performed in the context of acoustic echo cancellation indicate that these algorithms achieve a proper compromise in terms of fast convergence/tracking and low misadjustment.

引用

页码：1803 / 1807

页数：5