精準運動科學：慣性感測器量測體操單槓動作表現之指標建構

Determining performance indicators for gymnasts’horizontal bar athletic performance using IMUs

緒論：穿戴式慣性感測器（inertial measurement unit, IMU）能精準監控運動員動作加速度與旋轉角速度的數據，提供動作技術及訓練調整的依據。本研究旨在探討利用IMU評估體操單槓動作表現的可行性，建立貓跳及特卡切夫動作的測量方法，並找出能反映動作表現的指標參數。方法：研究對象為一位具備貓跳及特卡切夫高實施分的世界級菁英體操選手作為標準動作，以及四位大專甲組體操選手為對照組。兩位大專甲組選手分別以中實施分（n = 1）和低實施分（n = 1）完成貓跳動作演練，另外兩位則分別以中實施分（n = 1）和低實施分（n = 1）完成特卡切夫動作演練。本研究建立貓跳及特卡切夫標準測試流程，並使用IMU測量上背、下背及右腳踝動作的加速度、角速度等貓跳7項、特卡切夫8項訊號特徵。評估訊號特徵的再測變異係數（三次演練），並比較不同動作實施分選手之間的差異（計算效果量），找出能反映動作表現的關鍵指標。結果：貓跳動作三次測驗之間，7項參數中的6項參數變異係數皆小於10%。特卡切夫三次測驗動作間，8項參數變異係數皆小於10%。貓跳及特卡切夫動作中，高實施分動作與中、低實施分動作相比，幾乎所有參數皆具有大效果量（Cohen´s d > 0.8）差異。其中貓跳的擺腳下背加速度峰值與壓槓上背加速度峰值，以及特卡切夫的向前擺腳角速度峰值、向後擺腳角速度峰值以及騰空上背角速度峰值等，這些參數高實施分皆具有較高的表現，可作為關鍵指標參數。結論：本研究建立了具再測穩定性的IMU量測貓跳與特卡切夫之標準測驗流程與方法，並且證實所建立的參數可做為動作監測之關鍵指標，可用於監控選手訓練品質、技術改善和評估訓練效果之參考。

Introduction: Wearable inertial measurement units (IMUs) can accurately monitor athletes’movement data, including acceleration and angular velocity, and thus provide a basis for technique improvement and training adjustments. This study explored the feasibility of using inertial sensors to (1) evaluate the athletic performance of gymnasts in horizontal bar routines, (2) establish measurement methods for the Yamawaki and Tkachev skills, and (3) identify the key performance indicators. Methods: A world-class elite gymnast with high execution scores for the Yamawaki and the Tkachev was recruited to provide the standard for the two techniques, and four university-level gymnasts were recruited as the control group for the study. Two of the university-level gymnasts performed the Yamawaki, one with a medium and one with a low execution score; the other two performed the Tkachev, one with a medium and one with a low execution score. The study established standard testing procedures for the Yamawaki and the Tkachev. Three IMUs were used to measure acceleration and angular velocity at the upper back, lower back, and right ankle, resulting in 16 signal features for the Yamawaki and 7 for the Tkachev. The study assessed the coefficients of variation (3 trials) for these signal features, and compared the differences between the gymnasts with high, medium, and low execution scores to identify the key performance indicators. Results: In the Yamawaki, six out of the seven parameters exhibited coefficients of variation below 10% across the three trials. With the Tkachev, all eight parameters showed coefficients of variation below 10% across the three trials. When comparing the performances of the gymnast with high execution scores and those with medium and low scores, almost all the parameters demonstrated large effect size differences (Cohen’s d > 0.8). Notably, in the Yamawaki, the peak acceleration of the lower back during the leg swing, peak acceleration of the upper back during the bar push, and peak acceleration of the lower back during the bar push appeared to be the key performance indicators. In the Tkachev, the peak angular velocity of the forward leg swing, peak angular velocity of the backward leg swing, and peak angular velocity of the upper back during the flight phase appeared to be the key performance indicators. Conclusion: This study established a reliable standard testing method and procedure for the Yamawaki and Tkachev skills using IMUs. The established parameters can be used as key indicators for monitoring movement, and to help assess training quality, improve technique, and evaluate the effectiveness of training.