Success in figure skating depends on executing complex, rapidly rotating jumps. Training involves repeated practice of specific jump types, requiring a balance between performance gains and the risk of overuse injuries. Tracking jump counts by type helps manage training load, but manual counting is difficult - especially at the sub-elite level with limited resources. Prior work suggests that body-worn inertial measurement units (IMUs) can classify jump types. We investigate how IMU placement and quantity affect classification accuracy, addressing the broader challenge of applying SportsHCI research in real-world settings where practical considerations, like sensor location and count, significantly impact system usability. Our results show that ankle-mounted sensors yield the highest classification accuracy, while wrist-mounted sensors perform the worst. Future work should explore how sensor placement and skater skill level influence jump detection.
© Copyright 2025 SPORTSHCI'25: Proceedings of the First Annual Conference on Human-Computer Interaction and Sports. Veröffentlicht von Association for Computing Machinery. Alle Rechte vorbehalten.