A fast and reliable technique for muscle activity detection from surface EMG signals
A fast and reliable technique for muscle activity detection from surface EMG signals
复制标题
DOI：
10.1109/tbme.2003.808829
复制
发表时间：
2003-03-01
影响因子：
4.6
通讯作者：
Merletti, R
中科院分区：
文献类型：
--
作者：
Merlo, A;Farina, D;Merletti, R
运动过程中人体骨骼肌开关时间的估计是表面肌电信号处理及相关临床应用中的一个重要问题。在本文中，提出了一种新的方法来解决这个问题。该方法是基于使用连续小波变换的表面肌电信号的单个运动单位动作电位的识别。一个表现变量被计算为在不同尺度上的匹配滤波器的输出的最大值。将阈值应用于表现变量以检测EMG活动。一个模型，基于肌肉的物理结构，用于测试所提出的技术的合成信号，与已知的功能。在信噪比低至2dB的加性高斯有色噪声的情况下，起始估计的所得偏差低于40ms，标准偏差低于30ms。与以前开发的方法进行了比较，并提出了具有代表性的应用实验信号。该方法被设计用于完整的实时实施，因此，可以应用于临床常规活动。
The estimation of on-off timing of human skeletal muscles during movement is an important issue in surface electromyography (EMG) signal processing with relevant clinical applications. In this paper, a novel approach to address this issue is proposed. The method is based on the identification of single motor unit action potentials from the surface EMG signal with the use of the continuous wavelet transform. A manifestation variable is computed as the maximum of the outputs of a bank of matched filters at different scales. A threshold is applied to the manifestation variable to detect EMG activity. A model, based on the physical structure of the muscle, is used to test the proposed technique on synthetic signals, with known features. The resultant bias of the onset estimate is lower than 40 ms and the standard deviation lower than 30 ms in case of additive colored Gaussian noise with signal-to-noise ratio as low as 2 dB. Comparison with previously developed methods was performed, and representative applications to experimental signals are presented. The method is designed for a complete real-time implementation and, thus, may be applied in clinical routine activity.