基于盲辨识理论的双通道肌电信号建模与分类

Modeling and Classification of Two-Channel Electromyography Signals Based on Blind Channel Identification Theory

基于肌电信号产生机理 ,对双通道前臂肌电信号建立单输入多输出 IIR系统模型 ,由于模型输入未知且不可测 ,采用了盲信道辨识方法对模型传递函数进行辨识 .通过提取模型参数作为信号特征 ,能够对握拳、展拳、前臂内旋和前臂外旋四类前臂动作进行识别 .实验表明 ,该方法运算量小 ,适合在线实现 ,性能要优于传统的

According to the physiology of myoelectric singals, a two channel single input multiple output (SIMO) IIR model was proposed to the two channel upper limb surface electromyography (EMG) signals. As the input of the model is unknown and unaccessible, a neural network based blind channel identification technique was employed to identify the model's transfer function. By extracting the model parameters as signal features, four types of forearm motions: hand grasp, hand extension, forearm supination and forearm pronation are classified. The experimental results demonstrate that this method has better classification accuracy than the classical AR parameters based method. This paper shows a promising application of blind signal processing method to the analysis of physiological signals.