摘要
为了有效识别肌电信号EM G(E lectrom yography)的运动模式,利用小波分析的方法对采集的肌电信号进行消噪处理,最大限度地清除混杂在肌电信号中的噪声;然后提取各尺度小波系数最大值作为P i-S igm a神经网络分类器的输入,完成基于EM G信号多运动模式的识别.与此同时,利用EM G信号的能量特性,对各模式的起始和终止时刻进行界定,配合模式分类器的识别结果控制电动假手完成相应的动作.实验表明,基于小波分析的二次消噪方法能很好地消除混杂在EM G信号中的噪声,在正确的运动模式识别情况下,依据提取的运动模式时间信息,能够方便地实现假手的实时控制.
In order to recognize the movement patterns of electromyography （EMG） signals effectively, the acquired EMG signals were disposed first by the method of wavelet analysis to eliminate the noise to the maximum extent. Then the maximum of wavelet coefficients under each scale was taken as the input of Pi-Sigma neural network to recognize the movement patterns of EMG signals. At the same time, based on the energy characteristic of EMG signals, the points of motion start and stop were obtained to control the movement of powered prosthetic hands with the results of pattern recognition. The experimental results indicate that the method of the second wavelet de-noising can eliminate the noise of EMG effectively. Under the condition of correct movement pattern recognition, the real-time control of powered prosthetic hands can be realized expediently based on the time information of pattern recognition.
出处
《测试技术学报》
2006年第4期344-348,共5页
Journal of Test and Measurement Technology
基金
国家自然科学基金资助项目(60474054)
教育部新世纪优秀人才支持项目(NCET-04-0558)
关键词
肌电信号
运动模式
小波分析
模式分类器
电动假手
electromyography signal
movement Pattern
wavelet analysis, pattern classifier
powered prosthetic hands
