Toward Task-Independent Optimal Adaptive Control of a Hip Exoskeleton for Locomotion Assistance in Neurorehabilitation

Personalized robotic exoskeleton control is essential in assisting individuals with motor deficits. However, current research still lacks a solution from the end of a practical need of the problem to the end of its successful demonstration in physical environments, namely an end-to-end solution, that enables stable and continuous walking across different tasks. This study addresses this challenge by introducing a hierarchical control framework for the purpose. At the low level, impedance control ensures joint compliance without causing injury to users. At the high level, a reinforcement learning (RL)-based optimal adaptive controller automatically personalizes assistance to both hip extension and flexion (namely, bi-directional) to reach a target range of motion (ROM) under multiple walking conditions. As the first potentially feasible approach to this challenging problem and to meet practical use requirements, we developed a least-square policy iteration-based solution to configure the intrinsic parameters within the well-established finite state machine impedance control (FSM-IC). We successfully tested the control solution on eight young unimpaired participants and one participant post-stroke wearing a hip exoskeleton while walking on an instrumented treadmill. The proposed method can be applied to solving for optimal impedance parameters for individual users and different task scenarios to increase joint ROM. Our next step is to further evaluate this solution framework on additional people with hemiparesis who may benefit from hip joint assistance in therapy or daily activities to restore normative or improve gait patterns.

I. Introduction

Advances in wearable lower-limb exoskeletons have shown great promise in augmenting the movement capability of human wearers [1], [2]. These systems have been developed for different applications, such as military [3], industry [4], and medical rehabilitation [5], [6]. Focusing on rehabilitation implementations, lower-limb exoskeletons have been designed to deliver active assistance to patients with neurological deficits for improving locomotor performance, including people with paraplegia (no motor functions), such as complete spinal cord injury (SCI) [7], [8] and individuals with reduced force-generating capacity caused by incomplete SCI, stroke, cerebral palsy, and multiple sclerosis, etc. [9], [10], [11]. The appeal of modern wearable robotic exoskeletons in physical rehabilitation is their intelligent, active components that can be programmed toward the needs of different patient populations [5]. For example, early designs of powered exoskeletons were for patients with paraplegia, where exoskeletons were programmed to take over the entire lower-limb movement control. Typical control strategies often focused on tracking the kinematics of individual joints during walking or other locomotion tasks [12], [13]. This control technique is quite mature and has been used in the majority of commercial exoskeletons for rehabilitation. Nevertheless, for individuals who still have voluntary motor ability, joint position control is inappropriate and can potentially cause injuries to patients. Instead, compliance is essential to ensure safe human-exoskeleton interactions. Currently, there remains an open question as to how to provide the desired mechanical assistance, tailored to the individual patient, task context, and the environment, despite the emergence of many engineering efforts by the research community [2], [9], [14], [15], [16], [17].