Home // SENSORCOMM 2021, The Fifteenth International Conference on Sensor Technologies and Applications // View article

Tensor-Based Recursive Least-Squares Algorithm with Low Computational Complexity

Authors:
Ionut-Dorinel Ficiu
Cristian-Lucian Stanciu
Laura-Maria Dogariu
Camelia Elisei-Iliescu
Constantin Paleologu

Keywords: adaptive filter; multilinear forms; recursive least-squares (RLS); dichotomous coordinate descent (DCD); tensor decomposition

Abstract:
Many system identification problems can be addressed based on tensor decomposition methods. In this framework, the conventional Recursive Least-Squares (RLS) algorithm requires a prohibitive amount of arithmetic resources and is sometimes prone to numerical stability issues. This paper presents a low-complexity RLS-based algorithm for multiple-input/single-output system identification, which results as a combination between the exponentially weighted RLS algorithm and the Dichotomous Coordinate Descent (DCD) iterations.

Pages: 32 to 33

Copyright: Copyright (c) IARIA, 2021

Publication date: November 14, 2021

Published in: conference

ISSN: 2308-4405

ISBN: 978-1-61208-917-1

Location: Athens, Greece

Dates: from November 14, 2021 to November 18, 2021