SoC Design for Mobile Real-time Badminton Stroke Classification Design
Wen-Hsin Tsai, Kuei-Ann Wen · 2022 19th International SoC Design Conference (ISOCC) · 2022
This paper provides an innovative mobile real-time badminton stroke classification platform. Badminton is one of the fastest racket-based sport. Since it's difficult for eye-detecting the fast movements, camera system as Vicon are used for stroke detection but with high cost and limited space. We proposed a real-time badminton stroke classification system with a single IMU (Inertial Measurement Unit) mounted on the racket. In consideration of low power design for specific SoC to embedded on the racket, the computation complexity is highly reduced without using deep learning approaches. The accuracy rate can reach to 90.3%. Test chip have been implemented in TSMC 0.18 μm process. Die size is 760.6 μm by 740.0 μm and power dissipation is 0.0097mW.