Adaptive Coordinated Control of an Assistive Lower-Limb Exoskeleton for Hemiparetic Patients

Abstract

Lower-limb exoskeletons play an important role in improving gait symmetry and walking ability for hemiparetic patients. However, individual variability in gait symmetry and motor capability limits their assistive performance. To address this challenge, this paper proposes an Adaptive Coordinated Control (ACC) framework for unilateral lower-limb exoskeleton assistance, featuring a hierarchical control architecture. At the high level, a Key-Event-Based Coordinated Gait Mapping (KEB) method is developed to generate personalized reference trajectories online without relying on pre-collected gait databases, by extracting key gait-event features from the unaffected limb and reconstructing continuous trajectories. At the low level, a virtual-tunnel-based adaptive control strategy is designed to guide the affected joints along a time-independent reference path, while the assistance level is updated online according to a spatiotemporal symmetry factor to adapt to the patient's motor capability. Experimental results show that the KEB method maintains high accuracy and stability across different walking speeds. Further experiments with hemiparetic patients demonstrate that the ACC framework effectively adapts to different motor abilities and significantly improves gait symmetry and walking performance.