A K-Nearest Neighbours model deployed on a PYNQ-Z2 health monitoring system achieved 95.55% overall accuracy in classifying health conditions using heart rate, SpO2, and temperature data.
An FPGA-based health monitoring system utilizing a KNN model achieved 95.55% accuracy in classifying health conditions based on heart rate, SpO2, and body temperature.
Absolute Event Rate: 0% vs 0%
Introduction: The integration of Information and Communication Technology (ICT) into healthcare presents significant opportunities to optimize medical resource utilization and deliver effective, reliable care to elderly patients, individuals with chronic illnesses, and those with physical impairments. This paper proposes an advanced health monitoring system featuring remote, self-powered monitoring, real-time data collection, and control capabilities. Methodology: Leveraging Personal Health Devices (PHDs) and modern mobile technologies, the system enables continuous health monitoring through sensor-based data acquisition, ensuring accessibility and affordability for users. The proposed framework incorporates microcontroller- based sensors and uses the MQTT protocol for efficient, lightweight messaging. An intelligent health prediction module is designed to classify health conditions based on key physiological parameters such as heart rate, SpO₂, and body temperature. Machine learning algorithms, including Support Vector Machine (SVM), K-Nearest Neighbours (KNN), and Decision Tree (DT), were implemented to train and test the system for accurate health classification. Real-time data collected from multiple individuals under various scenarios served as the training dataset. The system integrates temperature, heart rate, and oximeter sensors with an ESP32-based MQTT client and a Raspberry Pi acting as the MQTT broker, allowing for seamless access to health data over the internet. Result: The machine learning models are deployed on a PYNQ-Z2 development board powered by the Xilinx Zynq-7000 System-on-Chip (SoC), where the KNN model achieves an overall accuracy of 95.55% with faster processing time than other models. Discussion: This FPGA-based framework ensures high-speed, low-power, and cost-effective operation, making it a suitable solution for patients requiring continuous or intermittent health monitoring. Conclusion: The proposed system demonstrates scalability, efficiency, and precision, making a significant contribution to the evolving domain of connected e-health.
Roy et al. (Mon,) reported a other. A K-Nearest Neighbours model deployed on a PYNQ-Z2 health monitoring system achieved 95.55% overall accuracy in classifying health conditions using heart rate, SpO2, and temperature data.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: