Utveckling av reglermetod och optimering av tröghetsmoment för ett bärbart exoskelett

Wearable and untethered exoskeletons need to be designed to accommodate the user without issues regarding comfort or functionality. Subjects of pathological gait, in particular, can be especially sensitive to additional moments of inertia. Exoskeletons are commonly used by those who require assistance through medical rehabilitation but are also used in industrial and military applications. One of the purposes of the study is finding the placement configuration of hardware modules which minimizes any negative effects of the exoskeleton driving unit on walking dynamics. The second purpose of the study is to evaluate the impact of an incorporated forgetting method on the control performance of proportional control with damping injection and iterative learning. The additional moments of inertia affecting the wearer's center of mass was studied on a driving unit replicated from previous research using iterative placement simulations of individual modules on the lower back. Simulations were performed in SolidWorks using VBA macros and algorithmically searched through a backtracking algorithm. The optimal placement configuration of individual modules within the driving unit was then evaluated based on literature on the effects of moments of inertia on dynamics, stability, and balance when walking. The control performance with an incorporated forgetting method was compared to one without. The methods were implemented on the same replicated driving unit attached to a test rig that applied a quasistatic mechanical resistance. The accuracy of the method was evaluated using normalized RMSE against reference data describing the torque load on an ankle joint over time. The optimal placement configuration was found to be the one that minimizes moment of inertia about the vertical axis. The controller performance evaluation showed statistically significant results that the inclusion of a forgetting method improves torque-following accuracy of proportional control and iterative learning control.