Synchronously recorded gait kinematic data with Inertial Measurement Units and a photogrammetry system for a validation assessment

Description A gait database of 32 healthy adult subjects was built , volunteers were between 20 and 63 y.o. (33.64 ± 12.44) and 71.88% were females. Every individual underwent a barefoot walking test recorded simultaneously with Inertial Measurement Units (IMUs) and the photogrammetry system Vicon. The dataset contains the kinematic gait information of the hip, knee, and ankle joints in the three planes of motion: sagittal, frontal, and transversal. The signals recorded by the IMUs are referred to as I(t) and were captured with a sampling frequency of 50 Hz, and those recorded by the photogrammetry system are called V(t) and were captured with a sampling frequency of 100 Hz. To perform a comparative study of both systems, the V(t) signals must be resampled to 50 Hz. Then, the delay between the two signals must be corrected to align them. Finally, gait cycles can be extracted for each pair of trials following the data information provided, obtaining a pair of waveforms for each gait cycle [I(t), V(t)]. A total of 268 synchronous gait cycles [I(t), V(t)] can be recovered and analyzed in the three planes of motion per limb. Data information raw_data: folder containing the 32 subjects raw kinematic signals recorded with IMUs (sampling frequency 50 Hz) and photogrammetry system (sampling frequency 100 Hz) synchronously. For IMUs records: Z: sagittal plane. X: frontal plane. Y: transversal plane. For photogrammetry system records: X: sagittal plane. Y: frontal plane. Z: transversal plane. captures_information.xlsx: table containing the delay correction and the samples corresponding to the events of the gait cycles. The delay correction is the number of samples for which each photogrammetry signal V(t), after being resampled to 50 Hz, must be moved to be completely aligned with its synchronous IMUs signal couple I(t). If the delay is positive (+) the V(t) signal must be delayed by adding zeros at the beginning. If the delay is negative (-) the V(t) signal must be moved forward by removing zeros at the beginning.

本数据集构建了包含32名健康成年受试者的步态数据库，受试者年龄介于20至63岁之间，平均年龄为33.64±12.44岁，其中女性占比71.88%。所有受试者均接受赤足行走测试，测试数据由惯性测量单元（Inertial Measurement Unit，IMU）与摄影测量系统Vicon同步采集。本数据集涵盖髋关节、膝关节与踝关节在矢状面、冠状面、横断面三个运动平面上的运动学步态信息。其中，惯性测量单元采集的信号记为I(t)，采样频率为50 Hz；摄影测量系统采集的信号记为V(t)，采样频率为100 Hz。为开展两类采集系统的对比研究，需先将V(t)信号重采样至50 Hz，随后校正两路信号间的时延以实现对齐。后续可根据配套数据信息从每一组测试中提取步态周期，得到每个步态周期的一对波形[I(t), V(t)]。最终可恢复并分析每侧肢体在三个运动平面下共计268组同步步态周期[I(t), V(t)]。原始数据文件夹（raw_data）：存储32名受试者的原始运动学信号，这些信号由IMU（采样频率50 Hz）与摄影测量系统（采样频率100 Hz）同步采集得到。IMU记录的轴对应关系如下：Z轴对应矢状面，X轴对应冠状面，Y轴对应横断面；摄影测量系统的记录轴对应关系为：X轴对应矢状面，Y轴对应冠状面，Z轴对应横断面。captures_information.xlsx：该表格包含时延校正参数与步态周期事件对应的样本信息。其中时延校正参数指：将V(t)信号重采样至50 Hz后，为使其与同步配对的IMU信号I(t)完全对齐，所需移动的样本数。若时延为正（+），则需在V(t)信号开头补零以实现延迟对齐；若时延为负（-），则需移除V(t)信号开头的零值以实现前移对齐。