Uncertainty propagation in multibody human model dynamics

Abstract

In biomechanics, calibration of body segment inertial parameters (BSIP) is crucial to take into account subject morphological specificities. To avoid strenuous protocols, identification methods based on rigid body dynamics laws have been proposed. Thanks to a motion capture system and force platforms, these methods optimize BSIP by minimizing errors in the equations of motion. These errors can be defined as the dynamic residuals reflecting inaccuracies arising from estimated BSIP, as well as from kinematics and force plate measurements. The current study aims at evaluating the part of uncertainty on the dynamic residuals directly related to kinematics and force plate measurements. To answer this question, we captured the movements of 10 participants performing a standardized motion. We then applied a Monte Carlo-based approach to introduce variations in the kinematics and force plate measurements, and evaluated the reconstructed difference on the dynamic residuals. Results show that, first, the BSIP evaluation using a regression method seemed to be an acceptable estimate for the studied subjects. Second, the part of uncertainty in the dynamic residuals was significantly higher than the dynamic residuals obtained. In conclusion, a subject-specific calibration of the BSIP based on dynamic residuals, for this model and protocol, seems irrelevant and prone to overfitting of BSIP.

Manual wheelchair biomechanics while overcoming various environmental barriers: A systematic review

“…Yet, when reported, methodological differences in kinetic and kinematic acquisition (opto-electronic motion capture system, system based on inertial measurement units) and in data processing (musculoskeletal model used for computation of joint angles and moments [ 66 , 67 ], point and basis of expression of net joint moments [ 68 , 69 ]) hinder rigorous comparisons of studies on the same barrier, and prevent the formulation of a reliable evidence-based synthesis of the propulsion biomechanics for each barrier. Lastly, poor data acquisition accuracy may lead to improper conclusions, especially for kinematics and kinetics quantities [ 70 , 71 ]. This observation may explain some of the contradictory results reported in the studies such as those involving slopes.…”

Back loading estimation during team handling: Is the use of only motion data sufficient?

“…Fourthly, BSIP were extracted from anthropometric tables. Errors in BSIP estimation induce errors in GRF prediction [ 38 ] and in joint torque estimation [ 39 , 40 ]. Improving the estimation of BSIP using calibration methods could improve the joint torque estimation, especially for subjects distant from the 50th percentile.…”

“…An interesting extension of this current study would be to adapt prediction methods to wearable technologies such as magnetic and inertial measurement units [ 41 – 43 ], depth cameras [ 44 , 45 ] or video cameras [ 46 ]. Since the accuracy of kinematic data plays a major role in prediction methods [ 38 ], studies must be carried out to verify the accuracy levels that can be obtained with this type of instrumentation.…”

Experimental recommendations for estimating lower extremity loading based on joint and activity

“…Pioneering work by Zatsiorsky and Seluyanov (1983) – which was later advanced by De Leva (1996) – developed mathematical equations to characterize the inertial properties of each body segment. Efforts continue to develop new methods to improve model assumptions and reduce measurement errors (El Habachi et al, 2015; Muller et al, 2017; Rao et al, 2006; Reinbolt et al, 2007). Advanced imaging is leveraged to develop subject specific musculoskeletal models (Davidson et al, 2008; Sheets et al, 2010), and techniques like the residual reduction algorithm reconcile dynamic inconsistencies between mass properties and experimental measurements (Delp et al, 2007; Langenderfer et al, 2008).…”