Force Estimation-Based Adaptive Impedance Control for Lower-Limb Exoskeleton Robots With Disturbance

Abstract

In this paper, considering that many robotic systems in rehabilitation applications that are not equipped with force/torque sensors at their physical interaction points, an adaptive impedance control strategy for exoskeleton robots is proposed based on the interaction force estimation. The external disturbance is considered, and a novel controller consisting of a disturbance observer is given. The radial basis function neural network (RBFNN) is used to estimate the interaction forces, an update law for online adjustment of the neural network weights is proposed, which allows the estimation accuracy to be guaranteed even when the interaction forces change. The system's stability is examined through the application of Lyapunov methods, followed by the execution of simulations for validation.