Contribution to development of a fully predictive Simulation Framework for Crutch-assisted Gait

Abstract

After a traumatic event such as a spinal cord injury or a stroke, patients need to undergo a recovery process to maximize their post-injury mobility, which usually involves using an assistive gait device. This thesis is part of a broader effort to develop a tool to aid clinicians in creating highly personalized rehabilitation plans, using individualized multibody simulations of the human body and trajectory optimization techniques. Among the different assistive devices that are available, this thesis focuses on the forearm crutches, more specifically on the three-point and swing-through crutch gait patterns. It builds on a previously developed optimal control formulation for a similar gait pattern, increasing its complexity by adding muscle torque generators (MTGs) to model muscle actuation, aiming towards a 3D fully predictive simulation framework. For that purpose, MATLAB is used as the programming interface, the optimal control problem is solved with GPOPS-II and OpenSim is used to host the multibody model. The developed problem formulation, explained in detail in the thesis, produces simulations that show realistic behaviours and the joint variables predicted are within reasonable values. The results demonstrate that incorporating MTGs in this kind of simulations is feasible; however, they also show that finding an optimal solution for such a complex problem is difficult, as it was not fully achieved during this thesis. Nevertheless, various hypotheses are presented to explain this outcome, along with proposed approaches to achieve convergence in future projects.