跺脚-走路-摔倒步态时小腿角加速度信号的希尔伯特谱分析

Hilbert spectrum analysis of calf angular acceleration signals in stomping-walking-falling gait

目的:解决运动信号因非平稳性而难以提取特征的难题。方法:以人体小腿角加速度信号为例,用希尔伯特-黄变换对其进行时频分析,即先对信号进行经验模态分解得到多个特征模态函数,再对全部或多个特征模态函数进行希尔伯特变换得到该信号的希尔伯特谱。结果:在跺脚、行走、摔倒的行为模式下,当试验脚着地时,能量分别集中在频率4、3、2 Hz处,时频分辨力较高,且这些行为的信息主要集中于高频分量的特征模态函数中。结论:该方法可成功提取人体小腿角加速度信号的时频特征,也适用于对其他下肢运动学参数信号的提取,以得到整个下肢的运动特征。

Objective To solve the problem in motion signal extraction due to nonstationarity. Methods The calf angular acceleration signal was taken as an example, and the time-frequency analysis was carried out by using Hilbert-Huang transform(HHT). Then, the empirical mode decomposition(EMD) of the signal was executed to obtain some intrinsic mode functions(IMFs), and then all or some intrinsic mode functions were subjected to a Hilbert transform to obtain Hilbert spectrum of the signal. Results In the mode of stomping, walking and falling, when the test-feet touched the ground, the energy was concentrated at the frequencies of 4 Hz, 3 Hz and 2 Hz, and the time-frequency resolution was higher, and the information of these behaviors was mainly focused on the high-frequency components of the intrinsic mode functions.Conclusion The time-frequency characteristics of the human body calf angular acceleration signal can be applied to other kinematic parameters of the lower limbs in order to obtain the movement characteristics of the lower limbs.