当前位置: X-MOL 学术Med. Eng. Phys. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Wearable sensor validation of sports-related movements for the lower extremity and trunk.
Medical Engineering & Physics ( IF 1.7 ) Pub Date : 2020-08-05 , DOI: 10.1016/j.medengphy.2020.08.001
Kimi D Dahl 1 , Kristin M Dunford 1 , Sarah A Wilson 1 , Travis Lee Turnbull 1 , Scott Tashman 1
Affiliation  

Inertial Measurement Units (IMUs), an alternative to 3D optical motion capture, are growing in popularity to assess sports-related movements. This study validated an IMU system against a “gold-standard” optical motion capture system during common sports movements. Forty-nine healthy adults performed six movements common to a variety of sports applications (cutting, running, jumping, single leg squats, and cross-over twist) while simultaneously outfitted with standard, retroreflective markers and a wireless IMU system. Bias, RMSE, precision, and maximum absolute error (MAE) were calculated to compare the two systems at the lower extremity joints and the trunk in all planes of movement and for all activities. The MAE difference between fast and slow activities for the sagittal, transverse, and frontal planes were 11.62°, 7.41°, and 5.82°, respectively. For bias, the IMU system tended to report larger angles than the optical motion capture system in the transverse and frontal planes and smaller angles in the sagittal plane. Average intraclass correlation coefficients for the sagittal, transverse, and frontal planes were 0.81±0.17, 0.38±0.19, and 0.22±0.37, respectively. When calculating a global bias across all three planes, the IMU system reported nearly identical angles (< 3.5° difference) to the optical motion capture system. The global precision across all planes was 2–6.5°, and the global RMSE was 7–10.5°. However, the global MAE was 11–23°. Overall, and with suggestions for methodological improvement to further reduce measurement errors, these results support current applications and also indicate the need for continued validation and improvement of IMU systems.



中文翻译:

下肢和躯干运动相关运动的可穿戴传感器验证。

惯性测量单元 (IMU) 是 3D 光学运动捕捉的替代方案,在评估与运动相关的运动方面越来越受欢迎。这项研究验证了 IMU 系统与常见运动运动中的“黄金标准”光学运动捕捉系统的对比。49 名健康成年人进行了各种运动应用中常见的六种运动(切割、跑步、跳跃、单腿深蹲和交叉扭转),同时配备了标准的逆反射标记和无线 IMU 系统。计算偏差、RMSE、精度和最大绝对误差 (MAE) 以在所有运动平面和所有活动中比较下肢关节和躯干的两个系统。矢状面、横断面和额面的快慢活动之间的 MAE 差异为 11.62°、7.41° 和 5.82°,分别。对于偏差,IMU 系统倾向于在横向和正面报告比光学运动捕捉系统更大的角度,而在矢状面报告更小的角度。矢状面、横向和额状面的平均组内相关系数分别为 0.81±0.17、0.38±0.19 和 0.22±0.37。在计算所有三个平面的全局偏差时,IMU 系统向光学运动捕捉系统报告几乎相同的角度(< 3.5° 差异)。所有平面的全局精度为 2-6.5°,全局 RMSE 为 7-10.5°。然而,全球 MAE 为 11-23°。总体而言,并提出了方法改进以进一步减少测量误差的建议,这些结果支持当前的应用,也表明需要继续验证和改进 IMU 系统。与光学运动捕捉系统相比,IMU 系统倾向于在横向和正面报告更大的角度,而在矢状面报告更小的角度。矢状面、横向和额状面的平均组内相关系数分别为 0.81±0.17、0.38±0.19 和 0.22±0.37。在计算所有三个平面的全局偏差时,IMU 系统向光学运动捕捉系统报告几乎相同的角度(< 3.5° 差异)。所有平面的全局精度为 2-6.5°,全局 RMSE 为 7-10.5°。然而,全球 MAE 为 11-23°。总体而言,并提出了方法改进以进一步减少测量误差的建议,这些结果支持当前的应用,也表明需要继续验证和改进 IMU 系统。与光学运动捕捉系统相比,IMU 系统倾向于在横向和正面报告更大的角度,而在矢状面报告更小的角度。矢状面、横向和额状面的平均组内相关系数分别为 0.81±0.17、0.38±0.19 和 0.22±0.37。在计算所有三个平面的全局偏差时,IMU 系统向光学运动捕捉系统报告几乎相同的角度(< 3.5° 差异)。所有平面的全局精度为 2-6.5°,全局 RMSE 为 7-10.5°。然而,全球 MAE 为 11-23°。总体而言,并提出了方法改进以进一步减少测量误差的建议,这些结果支持当前的应用,也表明需要继续验证和改进 IMU 系统。

更新日期:2020-08-29
down
wechat
bug