融合表面肌电和加速度信号的下肢运动模式识别研究

发布时间：2018-04-21 21:43

本文选题：下肢运动模式识别 + 表面肌电信号　；参考：《电子学报》2017年11期

【摘要】：为了提高下肢运动模式识别率,本文设计了一种融合表面肌电和加速度信号的下肢运动模式识别方法.首先,用局部均值分解将表面肌电信号分解为多个乘积函数(Product Functions,PFs),再计算PF成分的多尺度排序熵.然后,通过拉普拉斯权重(Laplacian score,LS)特征选择算法选定每路肌电信号的一个尺度排序熵为特征,并把该特征和加速度信号的排序熵组成特征向量.最后,根据类内欧氏距离和类间样本分布,设计了改进的二叉树支持向量机,把特征向量输入该支持向量机进行下肢运动模式分类.实验结果表明所提方法对七个日常动作的平均识别率达到98.62%,相较于其他方法有较高的识别率.
[Abstract]:In order to improve the recognition rate of lower extremity motion pattern, a method of lower limb motion pattern recognition combining surface electromyography and acceleration signals is designed in this paper. Firstly, the surface EMG signal is decomposed into several product functions, Product FunctionsPFS, and then the multi-scale ordering entropy of PF components is calculated by means of local mean decomposition. Then, a scale sorting entropy of each EMG signal is selected by Laplacian scorene-LS-based feature selection algorithm, and the feature and the sorting entropy of acceleration signal are composed of eigenvector. Finally, an improved binary tree support vector machine is designed according to the Euclidean distance within the class and the distribution of samples between classes. The feature vector is input into the support vector machine for lower limb motion pattern classification. The experimental results show that the average recognition rate of the proposed method for the seven daily movements is 98.62, which is higher than that of other methods.
【作者单位】：杭州电子科技大学智能控制与机器人研究所;
【基金】：浙江省自然科学基金(No.LY17F030021) 国家自然科学基金(No.61671197)
【分类号】：R318;TN911.7

【相似文献】