HomeEducationDoctorate (PhD & EngD)For current candidatesPhD infoUpcoming Public DefencesPhD Defence Luca Marotta | Development of inertial sensor-based methods to assess physical fatigue in running applications

PhD Defence Luca Marotta | Development of inertial sensor-based methods to assess physical fatigue in running applications

Development of inertial sensor-based methods to assess physical fatigue in running applications

The PhD Defence of Luca Marotta will take place in the Waaier building of the University of Twente and can be followed by a live stream.
Live Stream

Luca Marotta is a PhD student in the department Biomedical Signals and Systems. (Co-)Supervisors are prof.dr. J.H. Buurke, dr.ir. B.J.F. van Beijnum and dr. J. Reenalda from the faculty of Electrical Engineering, Mathematics and Computer Science.

Monitoring physical fatigue has short-term and long-term benefits for runners, but quantitative identification of physical fatigue is lacking in literature and in practice. Continuous monitoring of physical fatigue using inertial sensors could provide meaningful quantitative information on the physical state of a runner on the short term, and could reduce the risk of running-related injuries on the long term. Aim of this thesis was to assess whether physical fatigue can be identified in running using inertial sensors, while contributing to the overarching goal of prevention of running-related injuries and providing building blocks in the framework of developing fatigue monitoring wearable devices.  Since low sampling frequency inertial sensors help reducing data streams and processing times, it was first assessed whether lower sampling frequency inertial sensors can track biomechanical changes over time similarly to a high- resolution inertial sensor across different anatomical locations. Then, the optimal combination of inertial measurement unit (IMU) locations at the lower limbs and trunk to detect fatigue levels was assessed in a fatiguing outdoor run Machine learning techniques allowed to learn from IMU-derived biomechanical and statistical parameters and detect fatigue in prolonged running activities with increasingly higher accuracy from a single IMU location up to eight locations. Tibias are the sensor locations that resulted in higher fatigue detection accuracy. The experimental findings of this thesis support the hypothesis that machine learning models can identify running-induced fatigue with reasonable accuracy regardless of running intensity, while minimal sensor setups should be applied with careful consideration depending on the type of running activity. An interdisciplinary effort should be made in future research to efficiently use fatigue information extracted from inertial sensor technology as a mean to provide feedback to the runner and ultimately improve training loads and decrease the risk of injuries.