University of Twente Student Theses
Machine learning-based prediction of running-induced fatigue, during outdoor recreational running using IMUs, heart rate, and smartwatch data
Stojanac, Milica (2024) Machine learning-based prediction of running-induced fatigue, during outdoor recreational running using IMUs, heart rate, and smartwatch data.
|
PDF
2MB |
| Abstract: | Running has significant health benefits but comes with a high risk of running-related injuries (RRIs), often due to overuse. Traditional fatigue monitoring methods like CPET and lactate measurement are effective but impractical for outdoor use. Wearable sensors combined with machine learning (ML) offer a promising solution for real-time fatigue monitoring in realistic conditions. This study involved 19 recreational runners—14 in the first phase and five in the second—who completed three types of outdoor runs: endurance, interval, and a 5 km run. Participants wore seven IMUs on both tibias, thighs, pelvis, sternum, and wrist, plus a heart rate monitor and smartwatch to collect data. Fatigue was assessed using the Borg Rating of Perceived Exertion (RPE) scale (0-10) in the second phase, with a Random Forest regression model predicting RPE at 1-second intervals. The model, developed using nested Leave-One-Subject-Out (LOSO) cross-validation and RandomizedSearchCV, was tested with various sensor combinations. The wrist-only sensor configuration had the best RPE prediction performance (MSE 1.89), outperforming setups with more sensors. Results were strongest in endurance runs, with less accurate predictions in 5 km runs. Future work should increase sample sizes, integrate biometric data, and compare results with gold-standard fatigue assessments like EMG and VO2 max. |
| Item Type: | Essay (Master) |
| Faculty: | TNW: Science and Technology |
| Subject: | 44 medicine, 54 computer science |
| Programme: | Biomedical Engineering MSc (66226) |
| Link to this item: | https://purl.utwente.nl/essays/104549 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page