A Flexible Robotics-Inspired Computational Model of Compressive Loading on the Human Spine

Abstract

The use of powerful robotics tools in modeling and analysis of complex biomechanical systems has led to the development of computationally effective and scalable human models. Our early work in this regard focused on a humanoids-based rigid-body modeling of human kinodynamics to monitor instantaneous and accumulating effects of external loads on body joints. Nevertheless, despite their high-computational efficiency, the flexible nature of the spine and its characteristic were not taken into account, resulting in less accurate estimations of the spinal compressive forces. Accordingly, in this work, we propose a flexible model of the human spine mechanics for assessing compressive loading, and integrate it in our robotics-based whole-body model. Such a model can quantify the compressive force distribution along the spine and the muscles' activity for a measured back configuration and known external forces, which both contribute to increasing the ergonomic risk level. The muscles' activity predictions are validated through an experimental analysis on six human subjects. Three different tests are conducted considering different loading conditions. Results demonstrate the potential of the proposed approach in monitoring the spine compressive loading and predicting a muscles activity with an average relative error of 10% against experimental data. Minimizing the required number of sensors and the amount of computational resources, the presented approach is particularly suitable for online risk evaluation in a real working scenario.