Robust adaptive admittance control of an exoskeleton in the presence of structured and unstructured uncertainties. (4th March 2018)

• Record Type: Journal Article
• Authors: Shahi, Hossein; Yousefi-Koma, Aghil; Mohammadi Moghaddam, Majid
• Is Part Of: Advanced robotics. Volume 32:Number 5(2018)
• Page Start: 242
• Page End: 265
• Publication Date: 2018-03-04
• Subjects: Admittance control -- composite adaptive control -- Lyapunov redesign -- mixed structured and unstructured uncertainties -- locally weighted projection regression (LWPR) -- lower limb exoskeleton; Robotics -- Periodicals; Robotics -- Japan -- Periodicals; Robotics; Japan; Periodicals; 629.89205
• DOI: 10.1080/01691864.2017.1408491
• Languages: English
• ISSNs: 0169-1864

Abstract

Abstract: In this paper, an admittance control scheme for a user-in-charge exoskeleton is presented. The controller basically consists of a composite adaptive controller implementing a feedback law to estimate the structured uncertainties and to modify the apparent dynamics of the robot, and an LWPR estimator which tries to give an appropriate approximation of unmodeled uncertainty along with a robust term aiming to overcome the approximation residue. The control scheme offers a unified general control structure that explains the effect of each control component on the others. It is proved that based on the developed controller, the tracking and estimation errors converge to small boundaries with ultimate boundedness property due to the presence of the unstructured uncertainty. Based on simulations of a 2-DOF leg, the effectiveness of the controller is investigated. The results show the effectiveness of employing a universal approximator alongside a robust adaptive control and the success of the recommended approach in estimating model parameters and unmodeled dynamics simultaneously. Abstract :