Using Bayesian Optimization to Identify Optimal Exoskeleton Parameters Targeting Propulsion Mechanics: A Simulation Study

In this study, we determined the feasibility of modeling the relationship between robot control parameters and propulsion mechanics as a Gaussian process. Specifically, we used data obtained in a previous experiment that used pulses of torque applied at the hip and knee joint, at early and late stance, to establish the relationship a 3D control parameter space and the resulting changes in hip extension and propulsive impulse. We estimated Gaussian models both at the group level and for each subject. Moreover, we used the estimated subject-specific models to simulate virtual human-in-the-loop optimization (HIL) experiments based on Bayesian optimization to establish their convergence under multiple combinations of acquisition functions and seed point selection methods. Results of the group-level model are in agreement with those obtained with linear mixed effect model, thus establishing the feasibility of Gaussian process modeling. The estimated subject-specific optimal conditions have large between-subject variability in the metric of propulsive impulse, with only 31% of subjects featuring a subject-specific optimal point in the surrounding of the group-level optimal point. Virtual HIL experiments indicate that expected improvement is the most effective acquisition method, while no significant effect of seed point selection method was observed. Our study may have practical effects on the adoption of HIL robot-assisted training methods focused on propulsion.