Abstract: Martial arts, such as Taekwondo, require knowledge of one ’s own skill level for skill improvement. In this study, we propose a method of Taekwondo performance evaluation by attaching a smartphone to a performer’s waist and collecting acceleration data. When a martial performer stops moving, they instantly hold the entire body still. Acceleration data shows a peak at this moment of stillness, known as "focus". If the "focus" timing is identical across all three axes, the performer is considered to possess considerable skill. In this study, we use the timing differences between the three axes to calculate feature values and create decision trees from the acceleration data to evaluate movement based on these features as explanatory variables. We evaluate the effectiveness of the decision trees to confirm their accuracy.
0 Replies
Loading