Quantifying Sprint Determinants: Development of Motion Capture Analysis tools to support elite coaching.

Sports science is a field that has successfully leveraged the potential of Motion Capture technologies in the past years. Detailed real-time data allows us to carry precise analysis over the captured movements, potentially leading to new findings, insights or corrections. By observing mocap technologies at work, it was noticed that there is a gap between the captured data and the results. The Vicon Nexus software, paired with reflective markers and force plates, allows to render a skeleton and visualize its gait and the forces expressed on the ground. These, however, are lacking both in terms of clarity and proper analysis. The purpose of this work is twofold. First, to define a set of parameters and elements that characterize performance in sprinting. This was done by: referring to the prevailing literature, studying the parameters chosen by leading analysts in the field, and by discussing with experienced athletics coaches. 22 able bodied athletes of different levels participated in this study, performing block starts and sprinting trials. An analysis framework was created to extract relevant information about such efforts. The collected dataset on kinematic and kinetic parameters was investigated, and the latter were found to be more related to performance. Further analysis is required to better understand the weight of specific parameters in the sprinting performance. Second, to program an automatic tool capable of extracting raw data according to the framework and analyze it. Such software tool, developed in Matlab, provides a detailed report about the performance of the athlete, bridging the gap between the running effort and its objective details. This pipeline with automated analysis reduces processing time from hours to minutes and allows us to perform more running sessions and gather more data, with potential applications to Paralympic athletes. The report structure and content was based on the analysis conducted on the full dataset and on feedback by experts. Its graphs and organized tables strongly enhance the limited visualization provided by the tools previously available.