Extending device noise measurement capacity for OFDM-based PLC systems: Design, implementation, and on-field validation

Aymen Omri*, Javier Hernandez Fernandez, Roberto Di Pietro

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

1 Scopus citations

Abstract

Noise measurement in power line communication (PLC) systems is a common activity performed by grid operators for network tuning operations. Usually, these measurements are carried out with portable devices that have a fixed sensing and storage capacity. In this context, this paper presents a software-only solution for enhancing the performance of noise measurements in PLC systems. In detail: (i) we extend the measurement capacity in terms of the maximum number of samples that can be detected continuously, by using a machine learning (ML)-powered low complexity algorithm; and, (ii) we reduce the discontinuity period between successive measurements. This latter feature enables the possibility of collecting more continuous data. To show the viability of our proposal, we conducted a field measurements campaign to measure the scheme's accuracy and the measurement capacity extension ratio (MCER). The introduced approach is able to increase the MCER by up to 8 times, in the considered PLC environments, with an accuracy above 90%. While the proposed approach has a clear application – improving current devices’ noise measurements capability without requiring costly hardware upgrades –, the technique herein shown has a general applicability, and could hence pave the way for further applications in related fields.

Original languageEnglish (US)
Article number110038
JournalComputer Networks
Volume237
DOIs
StatePublished - Dec 2023

Keywords

  • Machine learning
  • PLC noise measurement
  • Smart grid applications

ASJC Scopus subject areas

  • Computer Networks and Communications

Fingerprint

Dive into the research topics of 'Extending device noise measurement capacity for OFDM-based PLC systems: Design, implementation, and on-field validation'. Together they form a unique fingerprint.

Cite this