CAREER: Reverse Engineering Human Leg Mechanics to Transform Control of Robotic Prostheses

近年来，使截肢患者能够行走的辅助设备，如机器人假肢、矫形器和外骨骼，取得了相当大的进展。然而，现有的系统在动态的、现实世界的环境中难以运行，包括意外的干扰。因此，这个CAREER项目的目标是解决对辅助机器人的需求，这些机器人可以通过利用人类运动的潜在动力学来自然地响应干扰，也就是说，通过实时为设备提供相同的能力来改变关节刚度和阻尼特性，人类通过肌肉激活在不知不觉中进行调节。关键的想法是了解神经系统如何在行走过程中调节腿部关节的刚度和阻力，并将这些信息用于设计和控制辅助机器人的新生物灵感蓝图。该项目还将深入了解人体对干扰的反应。这个项目的成果将显著提高生活质量和生产力，仅在美国就有100多万下肢截肢者。在整个项目中，研究和教育的整合将产生广泛的影响：1)通过有针对性的活动，向K-12学生推广STEM和残疾意识，展示工程的正式和“制造者”的一面；2)实施跨学科机制，在工程学生和临床医生之间进行知识转移；3)赞助和指导学生高年级设计项目，通过与临床医生共同监督直接支持本研究。首席研究员的长期研究目标是推进一种变革性的方法来设计和控制辅助机器人，这些机器人包含并利用了人类运动的潜在动力学。为了实现这一目标，该项目将测量行走过程中膝关节的机械阻抗，并利用所获得的信息开发新一代仿生控制系统，使辅助机器人的用户能够在包括意外干扰的动态环境中更自由地移动。研究将利用首席研究员实验室已经开发或正在开发的辅助设备：用于膝盖的扭矩可控外骨骼和开源机器人腿假体（OSL，在国家科学基金会的奖励下开发），为膝盖踝关节提供动力。研究计划有三个目标。第一个目标是研究使用扭矩可控外骨骼测量膝关节机械阻抗的技术，该技术将在健全人的行走过程中进行量化。方法将在弹簧和质量的被动机械系统中进行验证，该系统的阻抗在涉及人体受试者之前是已知的。可用于确定运动过程中膝关节机械阻抗的数据将被记录下来，这些数据来自于穿着膝关节外骨骼的人类受试者，当他们在分离式带跑步机上行走时，可以施加小的扰动（例如增加斜坡角度），并通过外骨骼测量响应扭矩。将应用扰动，以便在步态周期的每个点上获得阻抗的快照。点阻抗将通过将惯性、刚度和阻尼参数拟合到二阶微分方程来量化。第二个目标是推导出仿生阻抗控制的控制方程，该方程将在一个开源库中实现，用于控制OSL。Aim 1下得到的结果将与踝关节阻抗的初步估计工作综合起来，得出一个统一的控制规律，可以同时调节膝关节动力学、运动学和机械阻抗。控制方程将被转换成一个软件库，该软件库能够指挥OSL并实现膝关节阻抗的实时控制，即刚度、阻尼和平衡位置参数将作为步态阶段的函数而改变，因此，不会倾向于单一的速度。OSL实现所需刚度和阻尼系数的能力将在台式测试装置中进行验证。第三个目标是了解基于阻抗的控制系统对使用OSL的膝以上截肢者的影响，在有/无干扰的行走中进行测试。受试者将在使用阻抗控制系统和每个人指定的假肢以自行选择的速度在地面行走时进行研究。随后的一项研究将调查干扰对受试者在仪器跑步机上行走的影响。动力学、运动学、代谢和稳定性测量将在不同条件下进行比较，以更好地了解仿生控制方法如何影响截肢者的活动，以及这些测量与健全身体力学的比较。虽然将应用几种干扰，但该项目将重点关注行走对倾斜和下降的影响。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Advances in assistive devices that enable individuals with leg amputations to walk, e.g., robotic prosthesis, orthoses and exoskeletons, have been considerable in recent years. However, existing systems have trouble operating in dynamic, real-world environments that include unexpected disturbances. Thus, the goal of this CAREER project is to address the need for assistive robots that can naturally respond to disturbances by exploiting the underlying dynamics of human locomotion, i.e., by providing the device, in real time, with the same abilities to change joint stiffness and damping properties that humans unknowingly regulate through muscle activations. The key idea is to understand how the nervous system regulates leg joint stiffness and resistance throughout walking, and use this information for a new, bio-inspired blueprint for design and control of assistive robotics. This project will also provide insight into how the human body responds to disturbances. The results from this project will significantly improve quality of life and productivity for over a million lower-limb amputees in the U.S. alone. Throughout the project, the integration of research and education will have broad impact through the 1) promotion of STEM and disability awareness to K-12 students through targeted events that showcase the formal and "maker" sides of engineering, 2) implementation of interdisciplinary mechanisms for knowledge transfer between engineering students and clinicians, and 3) sponsorship and mentoring of student senior design projects that directly support this research through co-oversight with clinicians.The principal investigator's long-term research goal is to advance a transformative approach to the design and control of assistive robots that embrace and exploit the underlying dynamics of human locomotion. Towards this goal, this project will measure the mechanical impedance of the knee joint during walking and use the information obtained to develop a new generation of bio-inspired control systems that will enable users of assistive robots to move more freely in dynamic environments that include unexpected disturbances. Studies will make use of assistive devices already developed or under development in the principal investigator's lab: a torque-controllable exoskeleton for the knee and an Open-Source Robotic Leg prosthesis (OSL, developed under an NSF award) that has powered knee-ankle joints. The Research Plan is organized under three aims. The FIRST AIM is to investigate techniques for measuring knee mechanical impedance using the torque-controllable exoskeleton, which will be quantified in able-bodied subjects during walking. Methodologies will be validated in a passive mechanical system of springs and masses for which impedances are known before involving human subjects. Data that can be used to determine mechanical impedance of the knee during locomotion will be recorded from human subjects wearing the knee exoskeleton while walking on a split-belt treadmill in which small perturbations (e.g. increased ramp angle) can be applied and response torques measured by the exoskeleton. Perturbations will be applied such that a snapshot of impedance can be obtained at each point during the gait cycle. Point impedances will be quantified by fitting inertia, stiffness and damping parameters to a second-order differential equation. The SECOND AIM is to derive governing equations for bio-inspired impedance control of the leg that will be implemented in an open-source library for control of the OSL. The results obtained under Aim 1 will be synthesized with preliminary work estimating the impedance of the ankle to derive a unified control law that enables simultaneous regulation of knee-ankle kinetics, kinematics and mechanical impedance. The governing equations will be translated into a software library that is able to command the OSL and enables real-time control of knee-ankle impedance, i.e., stiffness, damping and equilibrium position parameters will be altered as a function of gait phase and, thus, not predisposed to a single speed. The ability of the OSL to achieve the desired stiffness and dampening coefficients will be validated in a benchtop testing setup. The THIRD AIM is to understand the effect of the impedance-based control system in above-knee amputees using the OSL, tested during ambulation with/without disturbances. Subjects will be studied while ground walking at a self-selected pace with both the impedance-controlled system and each person's prescribed prosthesis. A subsequent study will investigate the effect of disturbances while subjects are walking on an instrumented treadmill. Kinetics, kinematics, metabolics and stability measures across conditions will be compared to better understand how the bio-inspired control approach affects amputee mobility and how these measures compare to able-body mechanics. Though several disturbances will be applied, this project will focus on understanding the effect of walking on inclines and declines.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.