An improvement on impedance control performance of an exoskeleton suit in the presence of uncertainty

In recent years, different exoskeleton devices have been developed to provide the users with the mechanical power required in augmentation and rehabilitation applications. The exoskeleton control is one of the most challenging issues causes widely attention of researches during recent decade. Although different methods of control have been presented, there are several issues which have not been answered yet. This paper tries to focus on one of them entitling the robustness of impedance control in the presence of uncertainty for an exoskeleton and finds a solution for it. For this purpose, a RBF neural network with adaptive learning algorithm is employed to compensate the model uncertainty. Unlike other research works, the introduced controller is merely based on the robot kinematics. The convergence of the closed loop system to the desired impedance in the presence of uncertainty is verified by Lyapanov theorem. In the following, its implementation feasibility is evaluated by a simulation of an exoskeleton leg in the swing phase.