根据肌电信号产生机理 ,本文对双通道前臂肌电信号建立单输入多输出 FIR系统模型 ,由于模型输入未知且不可测 ,采用了盲信号处理方法对模型参数进行辨识 .通过提取模型冲激响应作为信号特征 ,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 ,该方法仅需建立较低阶数的模型即可达到较好的分类目的 ,性能要优于传统的 AR模型方法 According to the mechanism of EMG signal generation, this paper establishes a single-input multiple-output FIR system model of dual-channel forearm electromyography signal, and blind signal processing method is used to identify model parameters because the model input is unknown and unpredictable.By extracting the model impulse response As a signal feature, it can recognize four kinds of forearm movements such as fist, flexing, forearm pronation and forearm pronation.Experiments show that this method can achieve better classification purpose only by establishing a lower order model, and performance Better than the traditional AR model method