Wearable Motion Analysis in Sports: Advancing Biomechanical Insights Beyond the Laboratory

This study investigates the feasibility of estimating Ground Reaction Forces (GRFs) during various exercises using a wearable GPS sensor vest, combined with Artificial Neural Network (ANN) modelling.Traditional biomechanical assessments often rely on cumbersome optical motion capture systems and force platforms that limit real-world applicability.In contrast, this research leverages lightweight technology to gather kinematic data in natural training environments.A total of 6191 observations for the left leg and 5981 for the right leg were recorded, with the ANN trained to correlate GPS sensor data with GRFs.The results demonstrate predictive solid performance, evidenced by high R scores and low Root Mean Squared Error (RMSE) in both the development and hold-out datasets.While the left leg's predictions were slightly less accurate than those for the right leg, the model proved effective overall in estimating GRFs.These findings suggest that wearable technology, when paired with ANN techniques, can provide reliable biomechanical insights outside controlled laboratory settings, paving the way for further research involving larger participant groups and diverse exercises.This study investigates the validity of using a wearable GPS sensor vest to estimate Ground Reaction Forces (GRFs) during various exercises, employing an Artificial Neural Network (ANN) model.Traditional methods for measuring GRFs, such as force platforms, are limited to controlled laboratory settings, hindering real-world applications in sports.Our research aims to bridge this gap by leveraging kinematic data obtained from the GPS vest, allowing for real-time monitoring in dynamic environments.A comprehensive experimental setup involving a professional athlete performing a series of exercises was established, utilizing both force platforms and wearable sensors.The ANN was trained to correlate the sensor data with GRFs, yielding promising results.Validation metrics, including R and RMSE, demonstrated satisfactory accuracy of the model, particularly in predicting GRFs across different axes.The findings indicate that the proposed method can effectively estimate GRFs, providing valuable insights for performance optimization and injury prevention in sports.Future studies are recommended to expand the model's applicability across a broader range of athletes and exercise types.