A Unified Switching Control Framework for Continuous Robot-Assisted Training

Variant human conditions require the adaptation of robot-assisted training tasks, which involves the switching of robot's roles and training challenges and thus inevitably leads to the change of robotic controllers. While efforts have been devoted to either switching robot's roles or training difficulty, integrating both into a unified control framework for seamless smooth training remains elusive, as well as the corresponding system stability. In this work, we present a unified control framework to achieve smooth switching of robotic operation modes and training challenges. The unified framework consists of 1) a mode controller with two motion-dependent switching functions to achieve smooth switching of robotic operation modes, and 2) a task controller via a time-dependent switching function to adjust training challenge parameters. The overall system passivity and stability were analyzed using the Lyapunov methods. Experiments with human participants were conducted in a robot-assisted training scenario. Results validated the feasibility of the proposed unified control framework in the switching of robot-assisted training, as well as the training smoothness and safety.