Estimating Center of Mass Kinematics During Perturbed Human Standing Using Accelerometers

Abstract

Estimating center of mass (COM) through sensor measurements is done to maintain walking and standing stability with exoskeletons. The authors present a method for estimating COM kinematics through an artificial neural network, which was trained by minimizing the mean squared error between COM displacements measured by a gold-standard motion capture system and recorded acceleration signals from body-mounted accelerometers. A total of 5 able-bodied participants were destabilized during standing through: (1) unexpected perturbations caused by 4 linear actuators pulling on the waist and (2) volitionally moving weighted jars on a shelf. Each movement type was averaged across all participants. The algorithm’s performance was quantified by the root mean square error and coefficient of determination (R2) calculated from both the entire trial and during each perturbation type. Throughout the trials and movement types, the average coefficient of determination was 0.83, with 89% of the movements with R2 > .70, while the average root mean square error ranged between 7.3% and 22.0%, corresponding to 0.5- and 0.94-cm error in both the coronal and sagittal planes. COM can be estimated in real time for balance control of exoskeletons for individuals with a spinal cord injury, and the procedure can be generalized for other gait studies.

Use of Lower Limb Exoskeletons as an Assessment Tool for Human Motor Performance: A Systematic Review

“…Nevertheless, balance and fall prediction can be evaluated with wearable sensors such as IMUs, and the results may also be transferable to exoskeletons [ 89 , 90 ]. Some exoskeletons already are being prepared or systems for exoskeletons are being developed to measure stability for control purposes [ 91 , 92 ]. Thus, more research is needed that uses exoskeletons to assess a broad range of motor performance variables, including balance which is associated with risk of falling.…”

Center of Mass Estimation for Impaired Gait Assessment Using Inertial Measurement Units

“…The foundation of the neural network estimation algorithm stemmed from the work of Hnat et al [ 40 ] who employed a feed-forward neural network to estimate CoM during bipedal standing perturbations. Upon initial exploration, this neural network structure failed to estimate CoM for the various walking conditions.…”

“…The latter generated more accurate estimates; thus we had three networks estimating either the ML, AP or IS CoM position. We compared network estimates with the motion capture CoM which served as the gold standard [ 40 ].…”

“…Within each training/validation data set, a conventional k-fold cross validation test with a k of 5 was performed [ 44 ]. These internal validation tests followed the same procedure set by Hnat et al [ 40 ] in which the trials for each k-fold were shuffled. Each cross-validation’s RMSE was calculated, and the median RMSE network was saved to account for the cost and benefit of each k-fold.…”