摘要
针对传统的使用小波分解系数作为特征对走路、上楼、下楼进行分类时不能对具有相同强度加速度信号的步态样本进行分类的问题,提出了一种基于样本熵和小波能量相结合作为特征的分类算法。利用三轴加速度传感器采集走路、上楼、下楼3种步态下的上臂加速度信号,将信号进行小波分解,提取能量特征和样本熵特征,构建决策树分类器和贝叶斯分类器。决策树分类器和贝叶斯分类器的总体分类精度分别为75%和78.75%,使用样本熵与小波能量作为特征的分类精度比仅使用小波能量的分类精度提高了15.85%和19.17%。就步态分类精度而言,样本熵与小波能量相结合的方法优于仅使用传统小波能量方法。
It is unavailable for the classification of gait samples with a same-intensity acceleration signal by use of the traditional wavelet decomposition for three gait patterns（walking,up stair and down stair）,so a new method of sample entropy combined with wavelet energy was proposed.The three kinds of acceleration signals of one ＇ s upper limb were captured,then the characters of energy and sample entropy were measured for the construction of decision tree and Bayes classifier after wavelet decomposition of these signals were made.The general classification accuracy of decision tree and Bayes classifier were up to 75% and 78.75% which improved 15.85% and 19.17% compared with wavelet energy alone.As far as the precision of gait classification is concerned,the method based on sample entropy and wavelet energy was better than the wavelet energy alone.
出处
《传感技术学报》
CAS
CSCD
北大核心
2013年第4期545-549,共5页
Chinese Journal of Sensors and Actuators
关键词
步态分类
样本熵
小波能量
贝叶斯
决策树
计步器
加速度
gait pattern classification
sample entropy
wavelet energy
Bayes
decision tree
pedometer
acceleration
