IoT-Based Isolation Ward Monitoring System Prototype.

Area of Science:

• Biomedical Engineering
• Internet of Things (IoT)
• Artificial Intelligence (AI)

Background:

• The COVID-19 pandemic highlighted risks to healthcare workers (HCWs) due to direct patient contact, with HCWs facing significantly higher infection rates.
• Over 7000 HCW deaths globally by mid-2020 underscore the urgent need for solutions to minimize physical exposure.

Purpose of the Study:

• To design and validate a prototype IoT-based remote patient monitoring system.
• To enable continuous, real-time physiological surveillance while minimizing patient-to-HCW physical contact.

Main Methods:

• Integrated multi-sensor hardware (ECG, pulse oximeter, thermistor, CO2 sensor) with Arduino UNO for data acquisition.
• Utilized Raspberry Pi 3B for edge computing and a Convolutional Neural Network (CNN) for heartbeat classification.
• Employed SMOTE resampling for model training, achieving low on-device inference latency (<200 ms).
• Data transmission to Firebase Realtime Database via REST API, with front-end interfaces in LabVIEW and Flutter.

Main Results:

• Laboratory validation confirmed system functionality with round-trip cloud latencies of 300-800 ms.
• The system demonstrated error-free threshold alert generation and sub-second latency for its chat utility.
• The AI model achieved efficient heartbeat classification on-device.

Conclusions:

• The developed IoT system offers a unified, low-cost platform for remote patient monitoring.
• It effectively combines sensing, AI-driven analysis, cloud synchronization, and bidirectional communication.
• The prototype provides a viable solution for reducing HCW exposure in healthcare settings, particularly during infectious disease outbreaks.