Xiong Dezhen, Zhang Daohui, Zhao Xingang, Zhao Yiwen
Annu Int Conf IEEE Eng Med Biol Soc. 2020 Jul;2020:3094-3097. doi: 10.1109/EMBC44109.2020.9175238.
Gait can reflect human biological status during walking, which can be used for disease detect, identity verification or robot control, etc. Traditionally, gait analysis only classifies a gait cycle into a few discrete stages. In this paper, human gait will be decoded continuously using surface electromography (sEMG). The angle of knee joint and ankle joint during walking at different speed will be estimated at the same time by the proposed scheme. Four time domain features combined together will be used for the task. Six estimation methods will be compared and the best performance reaches the RMSE of 6.64° for knee and 3.89° for ankle. The proposed method shows great potential for the gait tracking problem.
步态能够反映人类行走过程中的生理状态，可用于疾病检测、身份验证或机器人控制等。传统上，步态分析仅将一个步态周期划分为几个离散阶段。本文将使用表面肌电图（sEMG）对人类步态进行连续解码。所提出的方案将同时估计不同速度行走时膝关节和踝关节的角度。四个时域特征将结合起来用于该任务。将比较六种估计方法，最佳性能在膝关节处达到均方根误差（RMSE）为6.64°，在踝关节处为3.89°。所提出的方法在步态跟踪问题上显示出巨大潜力。
