Variation of Kinematic Metrics With Perceived Fatigue in Ice Skating Measured Using Wearable Sensors

Khandan, A, Fathian, R, Carey, JP, and Rouhani, H. Variation of kinematic metrics with perceived fatigue in ice skating measured using wearable sensors. J Strength Cond Res 39(10): e1178-e1187, 2025-Enabling to obtain ice skaters' kinematics, wearable technology can track skaters' performance and thus detect performance fatigue in real-world settings. Therefore, this study aimed to investigate the potential of wearable inertial measurement units (IMU) to track skaters' performance, predict perceived fatigue, and detect severe fatigue onset before serious fatigue-related sequelae. In a multistage aerobic experiment, 19 subjects, 2 groups of high- and low-caliber skaters clustered by a novel algorithm, were asked to skate at a self-selected speed around an ice rink. During the experiments, subjects skated with 2 IMUs on their dominant leg's shank and thigh and 4 IMUs on their skates, pelvis, and trunk. These IMU outputs were used to develop 22 kinematic metrics whose variations were monitored with self-reported perceived fatigue by a linear mixed model, considering the effect of caliber. Finally, a machine learning algorithm was implemented to predict severe fatigue onset using the proposed kinematic metrics. The variations of intersegment correlation, joint angle fluctuations, and trunk angle were considerable (6-17% variation) during this intermittent skating experiment. In addition, a gradient-boosting model could predict severe fatigue onset with average precision, sensitivity, accuracy, and F1 score of 75, 81, 74, and 78%, respectively, in 196 skating stages captured from the subjects. The proposed kinematic metrics, as performance indicators, could also indicate perceived fatigue during an aerobic ice skating experiment and predict severe fatigue onset. The kinematic metrics introduced in this study equip coaches with quantitative tools to monitor performance and assess perceived fatigue in ice skating.