Predictive fatigue estimation for exercise monitoring

This thesis aims to develop predictive fatigue monitoring frameworks leveraging machine learning/deep learning techniques. Increasingly sophisticated approaches are presented in this thesis, which are able to: (1) perform accurate user-specific fatigue detection from wearable sensing, (2) perform accurate user-independent fatigue detection, based on musculo-skeletal simulations, and (3) perform short-term prediction of user-independent fatigue. The outcomes of this research suggest that the proposed fatigue recognition frameworks could be used in the future to monitor fatigue changes in mobility online utilising wearable sensors for athletes and patients. Additionally, a real-time application for feedback during at-home workouts becomes feasible.