doi abstract bibtex

Instrumented treadmills provide a convenient means for applying horizontal perturbations during gait or standing. However, varying the treadmill belt speed introduces inertial artifacts in the sagittal plane moment component of the ground reaction force. Here we present a compensation method based on a second-order dynamic model that predicts inertial pitch moment from belt acceleration. The method was tested experimentally on an unloaded treadmill at a slow belt speed with small random variations (1.20±0.10m/s) and at a faster belt speed with large random variations (2.00±0.50m/s). Inertial artifacts of up to 12Nm (root-mean-square, RMS) and 30Nm (peak) were observed. Coefficients of the model were calibrated on one trial and then used to predict and compensate the pitch moment of another trial with different random variations. Coefficients of determination (R(2)) were 72.08% and 96.75% for the slow and fast conditions, respectively. After compensation, the root-mean-square (RMS) of the inertial artifact was reduced by 47.37% for the slow speed and 81.98% for fast speed, leaving only 1.5Nm and 2.1Nm of artifact uncorrected, respectively. It was concluded that the compensation technique reduced inertial errors substantially, thereby improving the accuracy in joint moment calculations on an instrumented treadmill with varying belt speed.

@article{hnat_inertial_2014, title = {Inertial compensation for belt acceleration in an instrumented treadmill}, volume = {47}, copyright = {All rights reserved}, issn = {1873-2380}, doi = {10.1016/j.jbiomech.2014.10.014}, abstract = {Instrumented treadmills provide a convenient means for applying horizontal perturbations during gait or standing. However, varying the treadmill belt speed introduces inertial artifacts in the sagittal plane moment component of the ground reaction force. Here we present a compensation method based on a second-order dynamic model that predicts inertial pitch moment from belt acceleration. The method was tested experimentally on an unloaded treadmill at a slow belt speed with small random variations (1.20±0.10m/s) and at a faster belt speed with large random variations (2.00±0.50m/s). Inertial artifacts of up to 12Nm (root-mean-square, RMS) and 30Nm (peak) were observed. Coefficients of the model were calibrated on one trial and then used to predict and compensate the pitch moment of another trial with different random variations. Coefficients of determination (R(2)) were 72.08\% and 96.75\% for the slow and fast conditions, respectively. After compensation, the root-mean-square (RMS) of the inertial artifact was reduced by 47.37\% for the slow speed and 81.98\% for fast speed, leaving only 1.5Nm and 2.1Nm of artifact uncorrected, respectively. It was concluded that the compensation technique reduced inertial errors substantially, thereby improving the accuracy in joint moment calculations on an instrumented treadmill with varying belt speed.}, language = {eng}, number = {15}, journal = {Journal of Biomechanics}, author = {Hnat, Sandra K. and van den Bogert, Antonie J.}, month = nov, year = {2014}, pmid = {25458202}, keywords = {Acceleration, Artifacts, Biomechanics, Calibration, Exercise Test, Gait, Inertial artifacts, Instrumentation, Models, Theoretical, Surface perturbation}, pages = {3758--3761}, }

Downloads: 0