This academic article discusses a newly developed Bayesian optimization framework for auto-tuning shared controllers. The shared controllers are defined as a Model Predictive Control (MPC) problem. The framework includes the design of performance metrics and the representation of user inputs for simulation-based optimization. The framework is used for optimizing a shared controller for an Image Guided Therapy (IGT) robot. The article reveals that VR-based user experiments confirm the increase in performance and generalization ability of the automatically tuned MPC shared controller compared to a hand-tuned baseline version.
Publication date: 3 Nov 2023
Project Page: Not provided
Paper: https://arxiv.org/pdf/2311.01133
