Architectural Design of an IoT-Enabled Real-Time Athlete Monitoring System: A Multi-Sensor Integration Framework

Abstract

This study presents a comprehensive architectural framework for real-time monitoring in sports training, leveraging Internet of Things (IoT) technology. The system integrates wearable multi-sensor nodes—including 9-axis IMUs, PPG, and sEMG sensors—within a structured network architecture that employs hybrid BLE-Wi-Fi communication for low-latency data transmission to a cloud-based platform. The backend, built on a microservices architecture using AWS cloud services, incorporates sensor fusion algorithms, real-time feature extraction, and an XGBoost-based machine learning module for fatigue classification. A React.js dashboard enables intuitive visualization and interactive feedback. The proposed architecture was validated through controlled experiments involving 25 basketball players, demonstrating high accuracy in kinematic measurement (MAE of 1.52° for knee joint angles) and fatigue detection (94.2% classification accuracy). The study highlights the architectural scalability, interoperability, and real-time processing capabilities of the system, offering a robust and transparent model for IoT-driven applications in sports science and human performance monitoring.