利用加速規與陀螺儀進行樓梯、斜坡與平地行走之動作判讀

Abstract

目的:目前市面上的活動計量計僅能以加速度大小和步數計算使用者的身體活動量，對於步態相似但是消耗能量不同的動作 (如：上下樓梯、斜坡、平地行走），則可能產生估算的誤差，因此，本研究希望能透過加速規與陀螺儀感測器進行樓梯、斜坡與平地行走的動作判讀，未來協助提升活動計量計的準確度。方法:本研究共招募30位健康受試者，將感測器固定於左腳腳踝外側鞋面，於上下樓梯(兩種階高)、上下斜坡，以及平地行走各收取三筆資料，所有資料進行10 Hz低通濾波，步態分期後擷取著地期和離地期特徵點的角速度與加速度值繪製散佈圖，再以單因子變異數分析探討七種動作之間的差異性，找出不同動作間判讀的參數依據以及進行動作判讀時最有效率的軸向。結果:著地期的陀螺儀y軸（腳踝蹠屈/背屈角速度）在七種動作間的訊號分佈差異較為明顯，僅有下坡與平地走、上坡與上樓梯，以及上樓梯（低階）與上樓梯（高階）之間無顯著差異，其他組別間的差異性均達顯著。著地期踝關節特徵點的蹠屈/背屈角速度在下樓梯時，兩種階高之間有顯著差異，但上樓梯時差異則未達顯著，另外，不同階高的下樓梯踝關節角速度都和上樓梯有顯著差異。結論:腳踝的蹠屈/背屈角速度在判讀不同地形步行活動時，最能夠作為判讀依據。Purpose: Nowadays, we used digital sensors to calculate energy consumption only by the value of acceleration and the numbers of step. These methods might not be accurate with similar movements. Therefore, this study was designed to use accelerometer and gyro to recognize ascending and descending stairs as well as uphill, downhill, and level walking. Method: This study included 30 healthy subjects. The sensor was stabilized outside the shoe at left ankle. Subjects were asked to perform 7 movements including ascending/descending stairs (two different stairs), uphill, downhill, and level walking. All the accelerometer and gyro data were smoothed by a 10 Hz low-pass filter. One-way ANOVA was used to determine the difference among movements. Result: The angular velocity of ankle dorsi/plantar flexion at heel contact showed significant difference among the 7 movements, only downhill and level walking, uphill and upstairs, as well as upstairs (low) and upstairs (high) showed no significant difference. Therefore, the angular velocity of ankle dorsi/plantar flexion should be the indicator to determine activities on different landforms.