多通道肌电信号的FFT识别技术研究

Identifying technology of multi-channel EMG based on FFT

为了提高单个通道肌电信号识别效率,对多通道肌电信号进行融合处理.选择了具有kaiser窗的FFT变换,对多通道肌电信号进行频谱特性分析,并提取出能表征各通道肌电信号特性的奇异值.在奇异值的基础上,利用BP神经网络的数据融合技术,从多通道中实现了单个通道的信息识别,有效地识别各动作模式,为康复机器人的智能化控制提供了有效的多通道肌电信号识别技术.

In order to improve the identifying efficiency of one channel, we dispose the multi-channel EMG using Fusion technology. The FFT with kaiser window was used to analyze the spectral characteristics of multichannel EMG and the singular value that expressed the property of every EMG about each channel was taken out. On the basis of the singular value, using the Data Fusion technology of BP neural network, we identified one-channel information from multi-channel EMG and each movement mode effectively. It offers effective identifying technology to multi-channel EMG for the intelligent control of rehabilitating robot.