Unlocking the potential of GPS data to further improve load prediction models

This project aims to improve the prediction of mechanical load during running by making use of GPS data. Instead of relying solely on running speed, it is based on the premise that mechanical load is influenced by multiple factors that drive injury risk. By extracting additional features such as elevation changes, accelerations, decelerations, and surface type from GPS time series, the load signal can be modeled more accurately. Both a general model trained across all runners and individualized models capturing personal biomechanical characteristics are explored. Model performance is systematically benchmarked against a speed-only baseline. Ultimately, the project aims to support safer training progression…

‎Read more