基於體感網路的高爾夫揮桿分析

Abstract

在運動學習的過程中，打下良好的基礎動作是必須的，但若是自行摸索很容易造成基礎動作錯誤，未來得耗費更大的精神來矯正錯誤姿勢，因此市面上有許多的教學書籍、影帶以及訓練器材能夠提供民眾對各項運動的基礎認識。但是這些器材往往因為過於靜態或是受限於器材本身的攜帶性不足，造成民眾無法在現場即時察覺動作上的錯誤，或是某些細部動作的變化。因此我們提出一個結合體感網路以及雲端運算為基礎的自我運動訓練系統，讓使用者能隨時隨地進行基礎訓練。我們以高爾夫運動為主要對象，配合發展成熟的慣性感測元件偵測使用者的動作，並將資訊透過網路傳送至雲端系統，其中處理了感測器的偏差校正以及分散式感測器的時間同步問題。接著為了描繪出球桿的軌跡，對持續旋轉的感測器座標做處理讓資料表示在地球座標下。並提出了高爾夫揮桿動作的五個分解階段，並利用球桿上的感測資訊自動切割動作，最後整合各種資訊提供使用者關於擊球速度、揮桿節奏以及揮桿軌跡資訊，讓打者藉由這些資訊自行修正動作。

Having a good foundation for action is necessary during athletics learning, but it is easily lead to the flaw of basic action by self training. Nowadays, these are many teaching books, videos and training equipments can provide people with a basic knowledge of athletics. But these devices lack for portable, so that the public can not find out the error or different of the detail of action immediately at the scene. Therefore, we propose a golf self-training platform based on body sensor networks and cloud services, so that user could practice anytime, anywhere. We use inertial sensors which placed on golf club and human body to detect user's motion and transmit the data to cloud server through wireless connections. The server corrects the bias of sensors and handles time synchronization problems. Then, to depict the trajectory of golf swing, we calculate the rotation matrix which transforms readings from the sensor frame to the Earth frame. The swing is segmented into five actions based on the inertial data from the sensor attached to the golf club. In the end, we provide users information about the correctness of their swing such that user could correct their motion.