Back

Low cost IoT for heritage monitoring

This project presents the design and evaluation of a modular, low cost Internet of Things (IoT) system for monitoring environmental conditions and visitor behaviour in cultural heritage contexts. It addresses the limitations of existing commercial monitoring solutions, which are often expensive, proprietary, and difficult to adapt to diverse heritage settings.

The system combines ESP32-based sensor nodes, edge computing on a Raspberry Pi, and cloud-based data storage and visualisation to monitor variables such as temperature, humidity, sound, proximity, and visitor presence. A key component is the integration of computer vision techniques for behavioural analysis, where camera nodes and deep learning models are used to identify potentially harmful visitor actions in near real time.

Modularity is a central design principle. Sensor nodes and system behaviour are configured using lightweight JSON files, allowing sensors to be added or removed without firmware changes. This enables the same framework to be adapted to different heritage environments, from museum displays to archaeological sites, while keeping hardware costs below those of most commercial alternatives.

The system was validated in a simulated museum environment, demonstrating reliable data collection, low power consumption, and effective edge-based image analysis. Fine-tuned Vision Transformer models achieved high accuracy in classifying visitor behaviour, while the overall architecture supported reproducible data collection and analysis workflows.

Further technical details, system architecture, and validation results are described in the accompanying preprint: https://arxiv.org/abs/2508.00849