摘要
为实现老年人的跌倒与日常行为动作的模式识别,提出了一种基于排列组合熵和加权核Fisher线性判别的表面肌电信号跌倒识别方法.以腓肠肌和股外侧肌2路肌电信号对应的排列组合熵为特征向量输入加权核Fisher线性分类器进行模式识别,对跌倒与坐下、蹲下和行走进行识别.实验结果表明,该方法的跌倒识别率为93.33%,特异度100%,优于其他分类方法.
A EMG fall recognition method based on permutation entropy and weighted kernel Fisher linear discriminant analysis（WKFDA）was proposed to achieve recognition of the elderly fall and activities of daily living.The permutation entropy of sEMG on the gastrocnemius and vastus lateralis muscle was inputted to a weighted kernel fisher linear discriminant analysis（WKFDA）proposed in this paper to complete the pattern recognition.This method can successfully identify fall from some activities of daily life like walking,squatting and sitting down.The experimental results show that the sensitivity is 93.33%,and the specificity is 100%.This method has better recognition than the other classification method.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2015年第11期1685-1689,1700,共6页
Journal of Shanghai Jiaotong University
基金
国家自然科学基金(61172134)
浙江省自然科学基金(LY13F030017)
浙江省科技计划(2014C33105)资助项目
关键词
表面肌电信号
跌到识别
排列组合熵
加权核Fisher线性判别
surface electromyography
fall detection
permutation entropy
weighted kernel fisher linear discriminant analysis（WKFDA）
