不同位置慣性感測元件對捷泳划手技術分析之可行性

目的：透過不同位置慣性感測元件所收取的捷泳划手參數尋找較適合收取划手資料位置。方法：招募10位年齡18至25歲男性長距離捷式選手實施200公尺捷泳，速度為最佳成績80%。使用慣性感測元件六顆貼於雙側肩峰、雙側手腕、胸骨及薦骨，擷取頻率120Hz；共使用二台攝影機拍攝，收取頻率為120Hz。於泳池側邊跟拍及固定於側邊池壁拍攝，收取划幅、划頻參數及矢狀面活動，判斷划手分期。慣性感測元件所收取資料分析判斷划頻、划幅、身體滾轉角度及划手分期等參數；受試者各項基本資料描述性統計呈現。不同位置慣性感測元件所收取之划頻、划幅、身體滾轉角度及划手分期參數與影片數據以皮爾森積差相關分析，並計算誤差值及誤差百分比。手部慣性感測元件資料與身體各部位資料做皮爾森積差相關分析，顯著水準均為p <.05。結果：趟次辨認上平均誤差值約為0.82~1.14秒，其平均誤差百分比為0.62%~0.85%。划手次數判斷準確率100%。在635個划手週期中各部位切分的資料點上具顯著高度相關(皆為r=0.968~989, p<.05)，週期長度皆未具顯著差異。身體旋轉角度上時間點具高度相關(r=0.999, p<.05)、週期長度未有顯著差異(p=.797, t=-.261)且軀幹上不同位置慣性感測元件收取之數據具顯著高度相關性(r=0.993~0.998, p<.05)。結論：除位於手部之慣性感測元件無法收取身體旋轉角度資料外，位於軀幹上的慣性感測元件在模擬練習的情況下皆能判斷並分析划幅、划頻、身體滾轉角度、IdC等數據，若應用於實際練習監控，可根據選手較舒適的位置選擇配戴位置。

關鍵字

穿戴式裝置；游泳動作分析；划手分期；身體滾轉

並列摘要

Purpose: Optimizing placement of inertial measurement unit, in order to collect the data of front crawl stroke better. Method: 10 long-distance major swimmers, whose age between 18 to 25, were recruited and asked to swim 200 meters front crawl with 80% of personal best record. 6 inertial measurement unit were placed on both side of the shoulders and wrists, chest and waist, the capture frequency was 120 Hz. Two cameras were used to collecting parameters of stroke rate (SR) and stroke length (SL) and stroke phase, the capture frequency was 120 Hz. One was shooting along the pool, another was set on the wall of the pool. Parameters were analyzed to obtain SR, SL, body roll angle, and stroke phase. The information of the participants was showed in descriptive statistics. The parameters from different IMU and video were analyzed by Pearson’s correlation(p<.05). Result: the average difference of lap time between video and IMU is 0.82 -1.14 second, the percentage of average difference is 0.62%-0.85%. The number of stroke times is 100% correct. The stroke recognition from every IMU in total 635 strokes and the body roll angle recognition between Vicon system and IMU have high significant correlation ( r=0.968~0.989, p<. 05) and there are no significant differences (p=.797, t=-.261) between the stroke length. And the data between different IMU also has high significant correlation(r=0.993~0.998, p<.05). Conclusion: IMU in every placement under the practice situation can obtain all the parameters successfully except the wrist placement can’t obtain the data of body roll angle. The best placement of IMU for application can be chosen by the feeling of the swimmers.