Development and Sensitivity Analysis of a Fuzzy Logic Controller for Predicting ACL Injury Risk Values: a Study Using a Ski-Equipped Leg Surrogate

This research was conducted during an internship at the Sports Engineering & Sports Material laboratory of the Technical University of Munich (TUM) and falls within the broad scope of research aimed at developing a new mechatronic ski binding. It is a safety system designed to modulate the force/torque required for releasing the ski boot based on various knee injury risk scenarios that the skier may encounter. One major challenge towards a mechatronic ski binding is the development of an appropriate fuzzy logic controller to regulate the system according to the computed output ACL injury risk values. This thesis work was intended to enrich the knowledge base regarding how such a fuzzy logic controller might work. Moreover, this research aimed to provide useful starting points, results or pipelines for future developments of a fuzzy logic controller for a mechatronic ski binding. To do so, a smaller version of a fuzzy controller predicting ACL injury risk values has been developed and studied. An extensive literature study on the knee joint and its injury mechanism in alpine skiing was performed. Studies targeting the influence of parameters such as knee flexion angle, tibial torque and varus/valgus loading on ACL stress have been analysed. Furthermore, the theory of fuzzy logic controller was addressed. The knowledge base acquired in these preliminary steps was used to develop a ‘smaller’ version of a fuzzy controller in Python. The proposed algorithm predicts ACL injury risk values considering as input parameters the knee flexion angle, the internal/external torque, and the valgus/varus moment. Experimental tests using a ski-equipped leg surrogate and the correspondent multy-body-simulation model allowed the collection of ACL force data in different situations of knee flexion and external loads. These data were used to draw a comparison with the predicted risk levels in different situations to evaluate the behaviour and the performances of the newly developed controller. Lastly, a study on the possibilities of performing a sensitivity analysis on the fuzzy controller was performed and reported. The initial approach to such an analysis provided in this work aimed to contribute to the advancement of the understanding of the system and to give useful insights and methods for a future sensitivity analysis which might be performed on a ‘complete’ version of the control system for the mechatronic ski binding.