Control of SEA Modular Joint for Rehabilitation Exoskeleton Based on Modified Model Predictive Control

The lower limb rigid rehabilitation exoskeleton is research emphasis in lower limb disability assistance and rehabilitation. Modular joints integration in exoskeleton provides high-efficient human-robot interaction. However, it is difficult to achieve miniaturization of structure design. Meanwhile,...

Full description

Saved in:
Bibliographic Details
Main Authors: Chen, Wanxin, Zhang, Bi, Yao, Jie, Zhao, Xingang
Format: Conference Proceeding
Language:English
Subjects:
Control systems
Cost function
Exoskeletons
Predictive models
Switches
Trajectory tracking
Transforms
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The lower limb rigid rehabilitation exoskeleton is research emphasis in lower limb disability assistance and rehabilitation. Modular joints integration in exoskeleton provides high-efficient human-robot interaction. However, it is difficult to achieve miniaturization of structure design. Meanwhile, the lack of constraint design in control makes it difficult to meet the requirements of interaction. A novel SEA modular joint is proposed and the related control problem under constraints are established in this paper. Benefited from interval switching strategy, the proposed SEA modular joint has variable stiffness characteristics while the control complexity introduced by nonlinearity is reduced. On the control of SEA modular joint, the framework of model predictive control (MPC) is applied. The iterative linearization is presented to resolve the nonlinearity of system and the modified cost function in MPC is presented to achieve the prescribed degree of stability. The trajectory tracking and disturbance experiments are conducted. The results show that the proposed method can realize the tracking of reference trajectory while satisfying control constraints. The controller can reject disturbance effectively and modified cost function enhances MPC performance significantly.
ISSN:2642-6633
DOI:10.1109/CYBER53097.2021.9588326