基于灰色系统理论的表面肌电分类

为提高假肢分类的准确率和速度 ,提出采用灰色系统理论中的灰关联法进行动作辨识。首先用小波变换方法对表面肌电信号进行分析 ,通过对小波系数奇异值分解提取信号特征 ,根据待分类动作与各标准动作模式间特征矢量的灰关联系数做出判断。从掌长肌和肱桡肌采集的两道表面肌电信号中识别四种运动模式 ,准确率达87.5 %。与神经网络等识别方法相比 ,此方法不需大量训练样本数量 ,运算量小 ,在识别率相近的情况下 ,辨识速度大大提高。

Surface Electromyography Signal Classification Using Gray System Theory

A new method based on gray correla ti on was introduced to improve the identification rate in artificial limb. The ele ctromyography (EMG) signal was first transformed into time-frequency domain by wavelet transform. Singular value decomposition (SVD) was then used to extract feature vector from the wavelet coefficient for pattern recognition. The decisio n was made according to the maximum gray correlation coefficient. Compared with neural network recognition, this robust method has an almost equivalent recognit ion rate but much lower computation costs and less training samples.