Faculty of Engineering and IT, University of Technology Sydney, 81 Broadway, Ultimo, NSW 2007, Australia.
School of Sport, Exercise and Rehabilitation, Faculty of Health, University of Technology Sydney, Moore Park, NSW 2007, Australia.
Sensors (Basel). 2022 Jun 25;22(13):4812. doi: 10.3390/s22134812.
Given the popularity of running-based sports and the rapid development of Micro-electromechanical systems (MEMS), portable wireless sensors can provide in-field monitoring and analysis of running gait parameters during exercise. This paper proposed an intelligent analysis system from wireless micro-Inertial Measurement Unit (IMU) data to estimate contact time (CT) and flight time (FT) during running based on gyroscope and accelerometer sensors in a single location (ankle). Furthermore, a pre-processing system that detected the running period was introduced to analyse and enhance CT and FT detection accuracy and reduce noise. Results showed pre-processing successfully detected the designated running periods to remove noise of non-running periods. Furthermore, accelerometer and gyroscope algorithms showed good consistency within 95% confidence interval, and average absolute error of 31.53 ms and 24.77 ms, respectively. In turn, the combined system obtained a consistency of 84-100% agreement within tolerance values of 50 ms and 30 ms, respectively. Interestingly, both accuracy and consistency showed a decreasing trend as speed increased (36% at high-speed fore-foot strike). Successful CT and FT detection and output validation with consistency checking algorithms make in-field measurement of running gait possible using ankle-worn IMU sensors. Accordingly, accurate IMU-based gait analysis from gyroscope and accelerometer information can inform future research on in-field gait analysis.
鉴于跑步类运动的普及和微机电系统（MEMS）的快速发展，便携式无线传感器可以在运动过程中提供现场监测和分析跑步步态参数。本文提出了一种基于无线微惯性测量单元（IMU）数据的智能分析系统，该系统可以基于单个位置（脚踝）的陀螺仪和加速度计传感器来估计跑步时的触地时间（CT）和腾空时间（FT）。此外，还引入了一种预处理系统来检测跑步周期，以分析和提高 CT 和 FT 的检测精度，并减少噪声。结果表明，预处理系统成功地检测到指定的跑步周期，从而去除了非跑步周期的噪声。此外，加速度计和陀螺仪算法在 95%置信区间内具有很好的一致性，平均绝对误差分别为 31.53 毫秒和 24.77 毫秒。反过来，组合系统在分别为 50 毫秒和 30 毫秒的容差范围内获得了 84-100%的一致性。有趣的是，准确性和一致性都随着速度的增加而呈下降趋势（高速前足着地时下降 36%）。成功的 CT 和 FT 检测以及使用一致性检查算法进行输出验证使得使用脚踝佩戴的 IMU 传感器进行现场跑步步态测量成为可能。因此，基于陀螺仪和加速度计信息的精确 IMU 步态分析可以为未来的现场步态分析研究提供信息。
