Human-in-the-Loop Control Strategy of Unilateral Exoskeleton Robots for Gait Rehabilitation

In this article, a human-in-the-loop control methodology is proposed for the gait rehabilitation of patients with hemiplegia. It utilizes a unilateral exoskeleton system consisting of a unilateral lower limb exoskeleton and a real-time robot follower, such that the affected legs can be coordinated with the healthy legs with the assistance of the exoskeleton robot. In order to achieve immersive training during the physical therapy, the human-in-the-loop controller is developed. Furthermore, a region-based barrier Lyapunov function (BLF) is designed to separate the task workspace of the exoskeleton into a human region and a robot region, enabling the human leg to follow the desired motion trajectory in a compliant region, and the motion control of the exoskeleton is determined by humans; while in the robot region, the exoskeleton dominates the movement of human subjects. In order to make the motion control transit smoothly between the robot region and the human region, an adaptive controller is exploited to counteract the system’s nonlinear uncertainties. Both the theoretical analysis and experimental results support the effectiveness and practicability on hemiplegic patients of our control strategy.

I. Introduction

Stroke is a common disease usually leaving patients with lower limb hemiparesis [1], [2], which brings the impaired balance and increases the incidence of falls and fall damage [3]–[7]. In order to overcome these gait deficiencies, patients require intensive rehabilitation training with coordinated leg motion during physical therapy.