Impedance Control of Rehabilitation Robots for Lower Limbs, Review

Abstract

Impedance control is commonly used in rehabilitation robots to maintain the human-robot interaction force below safe levels for the patient, and simultaneously, to control the limb position according to the desired trajectories determined by the therapy. Impedance control aims to regulate the dynamic relationship between the robot force and the actuator movement. On the other hand, design of rehabilitation robots for lower limbs demands specific features that come from their biomechanical functions as weight bearing and locomotion. This paper aims to review the main features for implementing an impedance control in a lower limb rehabilitation robot. Initially, we present a general formulation of the impedance control starting from the dynamic equations toward the most basic forms of this controller. We emphasize that impedance control enable real-time variations in the amount of assistance via the variation of the impedance parameters (inertia, damping and stiffness) of the controller without significant problems of stability. Control performance does not only depend on the control schemes but also a good selection of all the elements of the control loop, including sensors and actuators. We review on the most commonly used actuators for impedance controlled robots. Finally, we addressed the main challenges and limitations for research and development of most efficient implementations and indicates the most relevant trends that literature evidence.

Towards human-knee orthosis interaction based on adaptive impedance control through stiffness adjustment

“…Based on the literature, proper functioning of an adaptive impedance controlled system implies the hierarchical organization of the control architecture [5], [16]. Thus, this study proposes the adaptive impedance control architecture illustrated in Fig.…”

“…In this sense, several adaptive control strategies, in particular, the assist-as-needed (AAN) control, have been investigated [2]. AAN training strategies can couple controllers capable of modulating the impedance of robots to promote a compliant human-robot interaction [4], [5].…”

Interactive control methods for rehabilitation robot

“…在机器人与人机交互系统中, 阻抗控制策略应用广泛 [36,103,104] . 该控制策略的概念最先由 Hogan [105] 提出, 是阻尼控制和刚性控制的推广, 并且首先在下肢康复机器人 Lokomat 上实现 [106] .…”

“…Ficuciello 等 [110,111] Ibarra 等 [36] 综述了阻抗控制在下肢康复机器人中运用的主要特性. 首先介绍了阻抗控制器中 最基本的动力学方程及其简化形式, 并展示了阻抗控制的不同应用形式, 例如 PD 计算力矩阻抗控制 器 [112,113] 、基于位置阻抗控制器 [114] 、混合阻抗 -力控制器 [115] ; 在康复治疗过程中, 机器人系统通 常需要依据训练形式与患者需要实时调整阻抗参数 (惯性、阻尼和刚度), 调节机器人提供的辅助力或 阻抗力, 且不能对系统稳定性造成本质影响, 基于此, 文中介绍了变参数阻抗控制器, 例如利用有限状 态机修改阻抗参数 [49, 116∼118] , 利用健侧下肢的表面肌电信号调整控制参数 [49] , 基于规则数据库调整 阻抗参数等 [119] ; 由于控制性能的好坏依赖于控制回路中所有环节, 包括传感器与驱动器, 而并不仅仅 依赖于控制方法, 因而作者分析了选择合适的传感器与驱动器的重要性; 最后讨论了研究中目前存在 的局限性以及未来趋势.…”

“…基于虚拟隧道的控制策略允许患肢在隧道内自 由运动, 当患肢处于隧道外部时, 康复机器人将对其施加一个柔顺力来调整患肢的位置与姿态, 并将 患肢拉回正常轨道范围内; 舒适、自然、柔顺的人机交互控制, 可充分调动患者参与康复训练的积极 性, 改善康复训练效果. 阻抗控制参数可以根据患者需要的辅助力或抗阻力的大小而变化 [36] , 保证康 复机器人与患者之间的柔顺性, 有利于激发患者残存肌力 表面肌电信号是骨骼肌产生的电活动信号 [37,38] , 可以通过贴合在皮肤表面的电极采集获取 [39,40] , 它具有非侵入性、易获取、可操作性强、安全性高等特点 [41] , 因而基于 sEMG 的人机交互有以下优 点 [42] : (1) sEMG 可从驱动关节运动的肌肉提取, 与执行运动的肢体无关, 为患肢的康复提供一种新 的渠道;…”