Human-Robot Interaction Control using a Sliding Mode Controller with Disturbance Observer

摘要

Successful performance of an assistive exoskeleton robot relies on the proper adjustment of human-robot interaction torques. Arbitrary human movements, modeling uncertainties, and available disturbances are some of the challenges in achieving a high-precision torque control. In this paper, a sliding mode controller (SMC) is proposed for the adjustment of human-robot interaction torques in an assistive hip exoskeleton. The controller is equipped with a disturbance observer (DOB) that estimates the modeling errors and available disturbances. Some compensation terms are also added to the controller to compensate for the observed disturbances and other factors like gravity, friction and human movements. Performance of the proposed controller is experimentally evaluated by implementing on a custom-made hip exoskeleton robot, HEXA-I, worn by a healthy subject. The tests are conducted in different situations of stance leg, swinging leg, and walking in-phase/out-of-phase with the torque trajectory. The results show that the interaction torque successfully tracks the desired torque trajectory in all conditions. The results also verify the superior performance of the proposed controller in comparison with a PI controller and a SMC controller with no DOB.