運用多尺度熵與迭代高斯濾波器於人體平衡狀態分析之研究

Abstract

近年來在醫學領域與大腦研究方面，有相當多關於人體平衡分析的文獻，也發展出各種的不同的技術與方法，但主要的量測方法取決於檢測儀器與感測器，像是測力板或加速規。所以在本文中提出一個有效的分析方法，可從人體晃動的狀況來分析不同平衡狀態的差異。分析的方法主要包含三個步驟：首先會透過三軸加速規來量測人體晃動的加速度訊號，接著再使用迭代高斯濾波器來濾除訊號的雜訊部分，此濾波器不會有相位誤差且較能分離出相異的訊號。之後再使用多尺度熵來量化濾波後的訊號，此方法較常用於訊號複雜度的分析。從實驗數據來看，我們認為多尺度熵的曲線可比較出各種不同狀態的差異，當人體處於較不平衡的狀況時，多尺度熵的曲線也會隨著降低。在實驗設計方面則分為兩個部分，首先是找出視覺對人體平衡的影響，例如張開雙眼與閉上雙眼；接著是針對人體在大腦思考時是否會對平衡造成影響，例如數學運算與英文閱讀。最後由實驗結果可獲得兩點結論：(1)當人體喪失視覺的時候會使平衡感降低；(2)人體思考的時候會占用大腦的資源，進而降低其平衡能力。另外由這些實驗的結果顯示本文所提出的分析方法，在人體平衡量測方面的效果相當顯著也具有可行性。

The analysis of human equilibrium, also known as postural stability, is a topic of great interest for the brain research and medicine community. A wide range of techniques and methodologies has been developed, but the choice of instrumentations and sensors depends on the requirement of the specific application. In this paper, we propose a methodology to analyze the human sway dynamics under several different stability conditions. The proposed methodology consists of three major steps: Firstly, the human postural sway acceleration signals were collected by using some 3-axis MEMS accelerometers. Secondly, we use a zero phase filter, named Iterative Gaussian Filter (IGF), to remove the noise from the collected signals. Thirdly, a popular complexity measure, named multi-scale entropy (MSE), is used to quantify the complexity of the filtered acceleration signal. We found that the MSE curves can be used to quantify different human stability conditions. The MSE curves derived from weak-stability-conditions lies below that derived from strong-stability conditions. Several experiments were designed to study the effect of the vision (eyes closed, eyes open) and brain resources (calculation, reading) on the human postural stability. Through the experimental results, we give two hypotheses: (1) loss of vision will decrease the human postural stability; (2) decrease of the brain resources will decrease the human postural stability. These experimental results demonstrated the feasibility and effectiveness of the proposed methodology on the analysis of human postural sway dynamics.