Hamiltonian prediction as a diagnostic metric in the upper limb using assistive robotics

The dynamics of robotic systems, which involve human participation in the control loop, represent high levels of uncertainty that can not only affect the stability of the interaction system but also lead to poor performance in the robot’s task execution. Stability is considered an intervened system; it can be assessed through the measurement of the total intervened energy. To this end, a regression model is proposed for the prediction of total energy, based on the robot’s motion (position and speed) controlled with the human operator in the loop. The purpose is to verify, through energy, the level of training in patients with motor limitations. The case study focuses on a patient with Guillain-Barré syndrome, and the proposed method uses the Hamiltonian function to model the energy behavior of the system during patient interaction.