论文部分内容阅读
针对下肢肌电信号(EMG)的多运动模式分类问题,提出了一种基于小波支持向量机(WSVM)的多类识别方法.在小波框架理论和SVM核方法的基础上,构造基于二叉树结构的WSVM多类分类器,采用多尺度分析对下肢EMG进行消噪处理和特征提取,将特征向量输入WSVM多类分类器.以水平行走为例对支撑前期、支撑中期、支撑末期、摆动前期和摆动末期等5个细分运动模式进行分类,并与传统的神经网络和高斯核SVM分类器进行比较.实验结果验证了所提方法的有效性.
Aiming at the multi-motion classification of lower limb electromyography (EMG), a multi-class identification method based on wavelet support vector machine (WSVM) is proposed.Based on wavelet frame theory and SVM kernel method, WSVM multi-class classifier, multi-scale analysis of the lower extremity EMG denoising and feature extraction, the feature vector input WSVM multi-class classifier to horizontal walking as an example of early support, mid-support, support the end of the pre-swing and swing The last five subdivision movement patterns are classified and compared with the traditional neural network and Gaussian kernel SVM classifier.The experimental results verify the effectiveness of the proposed method.