A Case of Identity for NLMS and Kaczmarz Algorithms

Authors

Muhammad Yasir Siddique Anjum
Assistant Professor, Department of Electrical Engineering, National University of Modern Languages, Rawalpindi Campus, Pakistan
DOI: https://doi.org/10.12928/jti.v9i2.
Download PDF
Published: 2021-06-24 • Issue: Vol. 9 No. 2 (2021) • Section: Articles

Abstract

In this paper, we will demonstrate, by employing basic linear algebra, that two seemingly disconnected algorithms, the Kaczmarz algorithm, familiar to mathematics community as an iterative solver of linear systems, and the Normalized LMS (NLMS) algorithm, known to the signal processing community as a self-learning adaptive filter, are identical.

Author Biographies

Muhammad Yasir Siddique Anjum
Muhammad Yasir Siddique Anjum is an Assistant Professor of Electrical Engineering at National University of Modern Languages, Rawalpindi Campus, PAKISTAN. His research focuses on reduction of transceiver complexity in massive MIMO systems. He is a professional engineer and member of Pakistan Engineering Council.