Autonomous IoT systems for real-time environmental monitoring using edge intelligence and deep learning

The integration of the Internet of Things (IoT), Edge Intelligence (EI), and Deep Learning (DL) has enabled a novel form of real-time environmental health monitoring. A wide gap in data collection characterizes traditional surveillance and, consequently, prevents it from identifying health-defining incidents at a glance (e.g., a sudden change in air quality or an increase in noise in health institutions). This study presents an autonomous IoT system that seeks to remove these restrictions by making smart decisions at the network edge. The system involves deploying distributed networks of IoT sensors to continuously collect environmental data (e.g., PM, VOCs, temperature, and humidity). Edge Intelligence, that is, small Deep Learning models, can enable real-time analysis and autonomous actuation on a decentralized basis, not just aggregate data, but also risk aversion. It considers the resource-constrained architecture of the Lightweight Environmental CNN (LE-CNN) within an edge-device-friendly framework to identify and categorize the tendencies of abnormal environmental signatures that directly affect patient health and wellbeing. The model will reduce backhaul traffic, enhance system stability, and increase the likelihood of an urgent response to defend vulnerable groups in medical facilities. The deployment and performance of the evaluation indicate that the system is more efficient than traditional cloud-based methodologies in terms of the accuracy of real-time anomaly detection and the likelihood of responding to anomalies with lower latency. A self-managed system such as this is a strong sign of better and more sustainable environmental health systems.

Перевод пока недоступен