手臂運動之肌電訊號特性分析與應用

Abstract

基於對人類運動程式與運動策略的研究，我們利用量測到的肌電訊號，以及因應的人手運動，進一步探討人類肌電訊號（Electromyograms）用於手臂控制之特徵；而在本論文研究中，將利用ICA演算法(Independent Component Analysis)來分析當人類手部前臂在做快慢彎曲動作時的手臂肌電訊號，建立系統架構，同時藉由分析找出適用於運動辨識之特徵，並完成一可分辨手臂不同彎曲速度的分類器，以應用至機器手臂或義肢之控制。

To investigate the command pattern and corresponding strategy in human movement, we study the relationship between the arm movement and measured electromyography(EMG).The independent component analysis technique is employed analyze the EMG signals exhibited during human forearm bending with different speeds. Based on the analysis, proper features are identified are used to arm movement study. A motion classifier is then developed to recognize arm movements of different speeds. The research can serve as a basis for the control of the robotic arm and prosthesis in the future.