Presentation Information
[B-15-12]Estimation of Sprint Acceleration State Using a Foot-mounted IMU with Camera-derived Ground Truth
◎△Shotaro Hirano1, Jotaro Kai1, Kenshin Tomizawa2, Akihito Kohiga1, Takahiro Koita2 (1. Doshisha Univ., 2. Graduate School of Doshisha Univ.)
Keywords:
Foot-mounted IMU,Sprint running,Acceleration state estimation
Recently, motion analysis using cameras and inertial measurement units (IMUs) has attracted attention in sports. In sprint running, the acceleration phase over the first several tens of meters strongly affects performance, and understanding speed changes at each step is important. Camera-based measurement can provide reliable ground truth for speed changes, but it requires installation and calibration, making it difficult to use routinely in school club activities. This study investigates a method for estimating per-step sprint acceleration using a foot-mounted IMU, with camera-derived speed changes used as ground truth. First, the position of the waist or trunk is obtained from video, and the average speed for each step is calculated. The difference between consecutive step speeds is then used as the target speed change. In parallel, acceleration and angular velocity data from the foot-mounted IMU are divided step by step, and features such as contact time, acceleration peaks, and angular velocity peaks are extracted. By relating these IMU features to camera-derived speed changes, this study aims to build a model that can estimate sprint acceleration using only an IMU. Ultimately, the proposed approach is intended to support athletes in objectively reviewing their acceleration state and improving their running motion.