The integration of multi-modal wearable smart devices is revolutionizing sports science, facilitating a paradigm shift from subjective coaching to a quantitatively-driven, personalized athletic development model. This study designs, implements, and validates a sophisticated wearable system for monitoring and optimizing the training of university-level track and field athletes. We present a custom-designed system architecture integrating high-frequency Inertial Measurement Units (IMUs) and electrocardiography (ECG) sensors, synchronized via a proprietary protocol. The software system employs edge computing for low-latency biomechanical feature extraction using an Extended Kalman Filter and a gradient-boosting regression model on the cloud backend for fatigue prediction. A 12-week randomized controlled trial was conducted with 50 athletes. Results demonstrated that the experimental group achieved statistically significant improvements in running economy (a 6.5% reduction in VO₂ at 16 km/h), a 10.3% enhancement in a novel Movement Quality Index, and a 45% reduction in time-loss injuries. The study concludes that a deeply integrated, multi-modal wearable system, underpinned by robust data processing pipelines and machine learning, can profoundly enhance training precision, performance outcomes, and athlete health.
XIAO et al. (Thu,) studied this question.