IoT System for Monitoring and Electrical Automation using Edge Architecture with AI for Energy Savings in Local Networks

Internet of Things (IoT), Edge Computing, local Artificial Intelligence, Offline-First home automation, energy efficiency, MQTT, FastAPI, Flutter.

Sergio Suarez Barajas
Unidades Tecnológicas de Santander,
Colombia, ssuarez@correo.uts.edu.co
https://orcid.org/0000-0002-5074-5997
1

Leydi Johana Polo Amador
Unidades Tecnológicas de Santander,
Colombia, lpolo@correo.uts.edu.co
https://orcid.org/0000-0002-5335-2814
2

Javier Mendoza Paredes
Unidades Tecnológicas de Santander,
Colombia, jmendoza@correo.uts.edu.co
https://orcid.org/0000-0001-6883-398X
3

Alejandro Bianchá Hernández
Unidades Tecnológicas de Santander,
Colombia, abiancha@correo.uts.edu.co
https://orcid.org/0000-0003-3248-1198
4

*Correspondence to
Sergio Suarez Barajas
Unidades Tecnológicas de Santander,
Colombia, ssuarez@correo.uts.edu.co
https://orcid.org/0000-0002-5074-5997

Received-14-05-2026
Revised-18-06-2026
Accepted-23-06-2026

European Journal of Prosthodontics and Restorative Dentistry (2026) 34(4S), 475–483

IoT System for Monitoring
and Electrical Automation
using Edge Architecture with
AI for Energy Savings in
Local Networks
Abstract — Commercial home automation systems present a critical
dependency on internet connectivity and cloud infrastructure, generating
operational vulnerabilities and user data privacy risks. This paper proposes
the design and development of EcoMind IoT, a residential energy
management prototype based on an Edge Computing architecture with
event-driven processing, integrating strictly local artificial intelligence. The
system captures real-time electrical telemetry via Shelly 1PM Mini smart
switches through the MQTT protocol (Mosquitto), processes data on a
central node implemented with FastAPI and Python, stores consumption
histories in PostgreSQL, and employs a lightweight local language model
(local LLM) to analyze electricity bill tariff information and make
autonomous actuator control decisions. The user interface was developed in
Flutter under an Offline-First approach, ensuring operational continuity
without cloud dependency. Results conceptually validate the proposed
architecture, establish the complete system data flow, and demonstrate the
technical feasibility of integrating local AI in residential IoT solutions
oriented toward energy savings with full data sovereignty.
I. INTRODUCTION
Today, home automation and the Internet of Things (IoT) have profoundly
transformed the way users interact with their residential and commercial
environments. According to recent estimates, the number of connected IoT
devices worldwide will exceed 30 billion by 2030 (International Data
Corporation, 2023), which is evidence of the massification of these
technologies. However, the vast majority of available business solutions
share a fundamental structural problem: their near-total reliance on cloud
servers to run automation logic.
This centralized cloud architecture creates at least two critical
vulnerabilities. The first is operational disruption: when the internet
connection fails, the home automation system immediately loses its ability
to automate, leaving the user without intelligent control over their devices
(Shi et al., 2016). The second is privacy risk: data on consumption habits,
hours of presence, and household routines are transmitted and processed on
third-party servers, exposing them to security breaches and information
monetization models (Atlam & Wills, 2019).
In parallel, energy efficiency is emerging as a global priority. The
residential sector accounts for approximately 27% of global energy
consumption (International Energy Agency, 2022), and smart energy
management systems have the potential to reduce this consumption by 15%
to 30% (Moreno et al., 2014). However, current commercial solutions such
as Sonoff, Tuya Smart or Sense are mostly limited to remote switching
on/off of devices without offering real optimization strategies based on the
variable cost of electricity or the user's historical consumption patterns.
In response to this problem, this article proposes EcoMind IoT, a system
that integrates Edge Computing, event-oriented processing and local
artificial intelligence to offer autonomous, private and resilient energy
management.
••••••••••••••••••••••••••••••••
ejprd.org- Published by Riset Publishing Services LLC.

EJPRD

Copyright © 2026 by Riset Publishing Services LLC