摘要
提出了一种基于时域和小波能量的步态分类算法。运用三轴加速度传感器采集走路、上楼、下楼三种步态下的上臂和胯部加速度信号,提取平均值、标准差、百分位数、平均绝对偏差等时域特征和2-6层小波能量特征,构建贝叶斯分类器,对三种步态进行分类。分类结果显示,时域与小波能量结合的分类方法的精度高于仅使用时域特征和仅使用小波能量特征的分类精度。
A novel gait classification algorithm analysis and proposed. S based on time domain wavelet energy analysis is ubjects are asked to walk on the level ground, walk upstairs and walk downstairs. The acceleration signals are collected from the 3D accelerometers worn around upper arnls and crotches. Time domain features of mean, standard deviation, percentiles, mean variation and features of wavelet energy of floor 2 to 6 are calculated And Bayes classifiers are produced The classification accuracy is improved compared with gait just classified by time domain features and wavelet energy features.
出处
《传感器世界》
2013年第4期10-13,共4页
Sensor World
关键词
步态
加速度传感器
贝叶斯
时域
小波能量
gait pattern
accelerometer
Bayes
time domain
wavelet energy