IMU to Motion Capture Based Nonlinear Limb Joint Variability Based Kinematics Comparison
Advisor Information
Dr. Brian Knarr
Location
University of Nebraska at Omaha, Omaha,NE
Presentation Type
Poster
Start Date
1-3-2019 2:00 PM
End Date
1-3-2019 3:15 PM
Abstract
The aim of this study is to compare an IMU-based system with a motion capture system in relation to nonlinear analysis of kinematics. Ten older adult subjects participated in the study. The Xsens IMU system was synchronized with a Vicon motion capture system to record simultaneously. Subjects were asked to walk on an instrumented treadmill for 3 minutes at their self-selected walking speed. Both systems’ data were filtered using a 6Hz lowpass Butterworth filter. Lyapunov Exponent (LyE) was calculated bilaterally for sagittal plane ankle and knee joint angle for each system. Peak ankle and knee flexion were also calculated using Visual 3D software. Paired T-tests were performed on each joint angle (JA) comparing LyE and peak JA calculated form the motion capture system and Xsens systems. Linear regression was used to determine the relationship of the two systems with significance set to p<0.05. There are no differences between either motion capture of IMU systems for ankle or knee JA variability. There is also no difference between motion capture and IMU for peak ankle plantarflexion, however there is a significant difference between the two systems for peak knee flexion. There is a positive relationship between the Vicon motion capture and Xsens IMU systems for the knee JA variability, peak ankle and knee flexion demonstrating good agreeability between systems. The motion capture and IMU systems are not related for the ankle JA variability. For knee joint LyE, the two systems may be comparable but initial results suggest not for ankle joint LyE.
ACKNOWLEDGEMENTS Funding: NIH P20 GM191090, R15 HD094194, U54 GM115458
REFERENCES
[1] Al-Amri M et al. (2018) Sensors. 18,719: 1-29.
[2] Blair S et al. (2018). J BIOMECH. 73: 24-32
[3] Barak Y et al. (2006). Physical Therapy. 86:150
IMU to Motion Capture Based Nonlinear Limb Joint Variability Based Kinematics Comparison
University of Nebraska at Omaha, Omaha,NE
The aim of this study is to compare an IMU-based system with a motion capture system in relation to nonlinear analysis of kinematics. Ten older adult subjects participated in the study. The Xsens IMU system was synchronized with a Vicon motion capture system to record simultaneously. Subjects were asked to walk on an instrumented treadmill for 3 minutes at their self-selected walking speed. Both systems’ data were filtered using a 6Hz lowpass Butterworth filter. Lyapunov Exponent (LyE) was calculated bilaterally for sagittal plane ankle and knee joint angle for each system. Peak ankle and knee flexion were also calculated using Visual 3D software. Paired T-tests were performed on each joint angle (JA) comparing LyE and peak JA calculated form the motion capture system and Xsens systems. Linear regression was used to determine the relationship of the two systems with significance set to p<0.05. There are no differences between either motion capture of IMU systems for ankle or knee JA variability. There is also no difference between motion capture and IMU for peak ankle plantarflexion, however there is a significant difference between the two systems for peak knee flexion. There is a positive relationship between the Vicon motion capture and Xsens IMU systems for the knee JA variability, peak ankle and knee flexion demonstrating good agreeability between systems. The motion capture and IMU systems are not related for the ankle JA variability. For knee joint LyE, the two systems may be comparable but initial results suggest not for ankle joint LyE.
ACKNOWLEDGEMENTS Funding: NIH P20 GM191090, R15 HD094194, U54 GM115458
REFERENCES
[1] Al-Amri M et al. (2018) Sensors. 18,719: 1-29.
[2] Blair S et al. (2018). J BIOMECH. 73: 24-32
[3] Barak Y et al. (2006). Physical Therapy. 86:150