An Augmented Full-Body Model that Improves Upper Body Tracking and Reduces Dynamic Inconsistency in Complex Motion.

Abstract

In recent years, the applications of musculoskeletal simulations have been expanded from simple walking to complex movements in various kinds of sports. The goal of this study was to augment the capability of the currently widely used full-body model (Rajagopal (2016) IEEE Trans. Biomed. Eng. 63:2068-2079) to improve the tracking of the kinematics of the head, shoulder, arms, and torso during complex full-body motion. Based on the testing of different modeling choices of neck, shoulder, and torso segments, the original Rajagopal full-body model was augmented by adding three joints in the spine and two sternoclavicular joints. The inverse kinematics and inverse dynamics of sports-related movements from 16 collegiate athletes were compared between the original Rajagopal and augmented full-body model. Our results showed that the augmented full-body model had significant improvements in tracking errors of the markers on the head, arm, torso, and pelvis during inverse kinematics, which led to reduced dynamic inconsistency in inverse dynamics, compared to the Rajagopal model. With a significant improvement in tracking the kinematics of the upper body, the augmented full-body model is a more suitable model to perform simulations involving complex full-body movements and is available for research use upon request from simtk.org.