Automatic Identification of Electrical Appliances in Domestic Settings

Authors

  • Antonio D. T. Amurim Universidade Federal do Ceará
  • Elvis M. G. Stancanelli Universidade Federal do Ceará (UFC)

DOI:

https://doi.org/10.5335/rbca.v17i2.16305

Keywords:

Machine learning, Classification, Load signature, Internet of things

Abstract

The Internet of Things (IoT) is pivotal for automation, monitoring, and control in modern environments. However, effectively managing numerous electrical devices within these settings presents significant challenges. A critical gap in understanding the operational status and power consumption of these devices often results in energy waste and can shorten their lifespan. This paper focuses on the automatic identification of electrical devices in IoT environments, leveraging features extracted from the frequency domain of electrical current signals. We collected power consumption measurements from household appliances and fed them to machine learning algorithms. Our results highlight the efficacy of the k-Nearest Neighbors classifier, which achieved an impressive F1-score of 99% in device identification.

Downloads

Download data is not yet available.

Published

2025-08-23

Issue

Section

Original Paper

How to Cite

[1]
2025. Automatic Identification of Electrical Appliances in Domestic Settings. Brazilian Journal of Applied Computing. 17, 2 (Aug. 2025), 87–102. DOI:https://doi.org/10.5335/rbca.v17i2.16305.