Utilization of neural networks within sports to define movement frequency for product development

Engineering of Sport 15 - Proceedings from the 15th International Conference on the Engineering of Sport (ISEA 2024)

The application of technology in sports is rapidly expanding, with the latest innovations including the use of neural networks (NNs)to classify motions. Initial applications of NNs have been made across sports with activity-specific motions, such as volleyball, table tennis and tennis. The use of Inertial Measurement Units (IMUs) is preferred due to the versatility of application to the body, testing location, and sampling rates. Class imbalance of data appears in sports involving ‘standard’ motions such as running and sprinting, alongside occasional execution movements, such as basketball shooting, requiring data augmentation. The aim of this work is to show the efficacy of these applied methods, with metrics of how well motions can be classified. This applies NNs for sporting movement classification in women’s tennis, to improve knowledge of frequency and magnitude of movements experienced on the female torso. This method can identify highload and frequent movements, to better inform sports bra design to improve comfort.