Research library

IoT Embedded Smart Monitoring System with Edge Machine Learning for Beehive Management

2024 open-access IoT/TinyML beehive monitoring proof-of-concept using sensors, hive-noise feature extraction, edge inference, and low-power microcontroller deployment for automated hive-management support.

Publication details

Authors
Mihai Doinea, Ioana Trandafir, Cristian-Valeriu Toma, Marius Popa, Alin Zamfiroiu
Organizations
🇷🇴 Bucharest University of Economic Studies🇷🇴 National Institute for Research & Development in Informatics
Year
2024
Type
Journal

PDF

Abstract

The need of an automated support system that helps beekeepers maintain and improve beehive population was always a very stressing aspect of their work considering the importance of a healthy bee population. This paper presents a proof of concept based on Internet of Things technology, proposing a smart monitoring system using machine-learning processes and edge computing for communication and control. IoT sensors collect data and extract features from hive noises, while a TinyML network performs inference on low-power microcontroller devices for decision support. By moving inference to the edge, the system improves the autonomy of beekeeping solutions and supports healthier hive maintenance without relying on high-power cloud processing.

Relevancy to Gratheon

This paper is directly relevant to Gratheon's edge-device roadmap because it combines the same product constraints Gratheon faces in remote apiaries: low-power sensing, local audio/noise processing, and actionable hive-state inference without continuous cloud dependence. Its TinyML framing is useful for deciding which colony-health signals should run on-device and which should be uploaded to the monitoring dashboard. The proof-of-concept also provides a concrete benchmark for packaging sensor acquisition, feature extraction, and beekeeper decision support into one deployable smart-hive node.