摘要
目的提出一种基于多通道肌电信号小波变换提取人手多种运动模式的方法。方法对人手上肢肌肉记录的表面肌电信号进行多尺度分解,利用肌电信号小波系数的方差构造特征空间。然后从这个特征空间选择指定动作对应的两块肌肉的小波系数方差,求得两者间的比值。结果和结论构造了新的特征空间,这个空间中包含了可识别各种人手动作的特征值。这种方法可以用于人手动作的识别,为研究多自由度假手表面肌电信号控制方法提供了新途径。
Objective： To present an identification method of human hand motion using the ratio of variance in each frequency section of multi-channel EMG signal. Method： The first step was to analyze the surface EMG signal of human upper alto by using multi-resolution of wavelet transform, then to establish a feature space consisting of the variance of wavelet coefficients. Secondly, a new feature space for a certain motion was constituted by selecting the variance of wavelet coefficients of corresponding two muscles and calculating the ratio of these two columns. The feature in the new space was used to identify the human hand motion. Result and Conclusion： This method is effective in identification of human hand motion. Thus, it provides an alternative novel approach to use the surface EMG in control the muhi-freedom prosthetic hand.
出处
《中国康复医学杂志》
CAS
CSCD
北大核心
2006年第1期22-24,共3页
Chinese Journal of Rehabilitation Medicine
基金
国家自然科学基金项目(50435040)
国家"八六三"计划资助项目(2001AA320601)
关键词
表面肌电信号
小波变换
假手
多自由度
surface electromyography
wavelet analysis
motion identification
prosthetic hand
