Home // ICWMC 2018, The Fourteenth International Conference on Wireless and Mobile Communications // View article

Diffusion Recursive Least Square Adaptive Networks with Neighbor-Selection

Authors:
Wael Bazzi
Vahid Vahidpour
Amir Rastegarnia
Azam Khalili

Keywords: Adaptive network, diffusion, neighbor selection, re-cursive least-squares

Abstract:
Constrained communication resources and limited communication bandwidth are key issues for any task involv-ing wireless sensor networks. This phenomenon motivated the authors to examine diffusion networks where only a frac-tion of neighbors participle in the communication process. In this context, we modify the Diffusion Recursive Least Square (DRLS)algorithm by allowing each node to receive intermediate estimates from a subset of its neighbors, called neighbor-selection DRLS. This results in significant reduction in communication overhead at the cost of some possible deterioration in the network performance. We derive a theoretical expression for the steady state Mean Square Deviation (MSD). Both numerical simulations and theoretical findings are used to validate the effectiveness of the proposed algorithm in providing a trade off between communication burden and estimation performance.

Pages: 22 to 25

Copyright: Copyright (c) IARIA, 2018

Publication date: June 24, 2018

Published in: conference

ISSN: 2308-4219

ISBN: 978-1-61208-642-2

Location: Venice, Italy

Dates: from June 24, 2018 to June 28, 2018