HUMAN-EXOSKELETON COMPUTATIONAL MODEL: AN APPROACH TO THE HUMAN-MACHINE INTERACTION PROBLEM IN ROBOTIC ASSISTED THERAPY

Objetive: The purpose of this work was to develop a human-exoskeleton of lower limbs interaction model and a forward dynamics-based algorithm able to simulate and test interaction controls applied to robotic neurorehabilitation. Methods: The interaction model was developed using the neuromusculoskeletal model gait2392 provided by OpenSim, with coordinate actuators coupled to the joints of the hip, knee and ankle, in order to representing an exoskeleton. A fuzzy-based adaptive impedance control was developed and used to test the proposed system. Results: Two simulations were performed with the interaction model assisted and not assisted by the rehabilitation robot. The results obtained were the movement, the torques related to it (both from the human and the robot) and the virtual stiffness of the robot, determined through the fuzzy-based adaptive impedance control. Conclusion: The results proved that the use of interaction models and simulation is a powerful resource to develop and test interaction controls applied to the robotic assisted therapy.