Classification and visualization of skateboard tricks using wearable sensors

The application of wearables and customized signal processing methods offers new opportunities for motion analysis and visualization in skateboarding. In this work, we propose an automatic trick analysis and visualization application based on inertial-magnetic data. Skateboard tricks are detected and classified in real-time and visualized by means of an animated 3D-graphic. We achieved a trick detection recall of 96.4%, a classification accuracy of 89.1% (considering correctly performed tricks) and an error of the board orientation visualization of 2.2°  ±  1.9°. The system is extendable in its application and can be incorporated as support for skateboard training and competitions.