NRI: INT: Collaborative Research: An Open-Source Framework for Continuous Torque Control of Intuitive Robotic Prosthetic Legs

NRI：INT：协作研究：直观机器人假腿连续扭矩控制的开源框架

基本信息

批准号：
2024446
负责人：
Eric Rombokas
金额：
$ 56万
依托单位：
University of Washington
依托单位国家：
美国
项目类别：
Standard Grant
财政年份：
2020
资助国家：
美国
起止时间：
2020-10-01 至 2024-09-30
项目状态：
已结题

来源：
https://www.nsf.gov/awardsearch/showAward?AWD_ID=2024446&HistoricalAwards=false
关键词：
NRI INT Collaborative Research Open

项目摘要

This project will establish an open source set of software control algorithms that will allow an open source robotic prosthetic leg to facilitate rhythmic and non-rhythmic interactions between the human user and the environment. This project builds upon the Open-Source Leg, which is a robust, inexpensive, robotic leg platform that can be easily manufactured, assembled, and programmed. The project's overarching goal is to enable customizable behaviors that are continuously cued by the movement of the user’s body. The project promotes the progress of science by creating open source control hardware and software for compliant actuators that extend the capabilities of the Open-Source Leg. The advantages of compliant torque control, combined with intuitive, expressive control from the user, represents a significant improvement over currently-available prosthetic legs. The project will advance the national health by developing and testing high-level control software that will allow users of the Open Source Leg to seamlessly navigate around obstacles and perform dynamic activities. The improved mobility provided by these technologies will improve the quality of life and functional capabilities of many people living with mobility impairment. Open source hardware and software lower barriers to access for robotic technologies, which makes these robots great candidates not only as assistive co-robots in healthcare and other applications but also as educational tools for undergraduate and graduate students.Emerging powered prostheses such as the NSF-funded Open-Source Leg have motors that can restore normative biomechanics to above-knee amputees, but these devices are limited by their control strategies to a small set of pre-defined, steady-state activities. Each activity is typically divided into a discrete progression of gait periods called phases, resulting in a large set of distinct controllers that struggle to continuously coordinate prosthetic limb motion with the user. Discrete control paradigms have not been able to facilitate transient behaviors like transitions between activities or non-rhythmic motions like stepping backwards or stepping over obstacles. Recently, a new control paradigm has emerged that continuously synchronizes or coordinates prosthetic limb motion to the user based on inertial measurements from the user’s body (e.g., the residual limb). However, prior implementations have been limited to lab-specific prosthetic leg designs with stiff actuators that rigidly enforce the kinematic mappings from user motion to prosthetic joint position rather than complying to varying environmental interactions. The recently developed Open-Source Leg presents a unique opportunity to integrate this state-of-the-art control paradigm in a universally accessible testbed with series elastic actuators that soften interactions between the user, prosthesis, and environment. The overall goals of this project are to 1) understand how to achieve closed-loop torque and impedance control in the series elastic actuator of the open-source leg despite unmodeled dynamics from its low-cost design, and 2) understand how to integrate high-fidelity joint impedance control with two novel continuous controllers that promise to allow users to flexibly and seamlessly navigate obstacles and perform dynamic activities.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.

该项目将建立一套开源软件控制算法，使开源机器人假肢能够促进人类用户与环境之间的有节奏和非节奏交互。该项目建立在开源腿的基础上。该项目的总体目标是实现由用户身体运动不断提示的可定制行为，该项目通过创造开放性来促进科学的进步。用于扩展开源腿功能的兼容执行器的源控制硬件和软件兼容扭矩控制的优点与用户直观、富有表现力的控制相结合，代表了该项目将比现有的假肢取得重大改进。通过开发和测试高级控制软件来促进国民健康，这些软件将使开源腿的用户能够无缝地绕过障碍并执行动态活动，这些技术提供的改进的移动性将提高许多人的生活质量和功能能力。行动不便的人。开源硬件和软件降低了获取机器人技术的障碍，这使得这些机器人不仅成为医疗保健和其他应用中的辅助协作机器人的绝佳候选者，而且还可以作为本科生和研究生的教育工具。新兴的动力假肢，例如 NSF-资助的开源腿部具有可以使膝盖以上截肢者恢复正常生物力学的电机，但这些设备受到其控制策略的限制，只能进行一小组预先定义的稳态活动，每个活动通常分为一个离散的进程。的称为阶段的步态周期，导致大量不同的控制器难以与用户持续协调假肢运动，离散控制范式无法促进瞬态行为，例如活动之间的转换或非节律性运动（例如后退或迈步）。最近，出现了一种新的控制范例，它根据用户身体（例如残肢）的惯性测量连续同步或协调用户的假肢运动。实现仅限于实验室特定的假肢设计，其刚性执行器严格执行从用户运动到假肢关节位置的运动学映射，而不是遵守不同的环境交互，最近开发的开源腿提供了整合这种状态的独特机会。具有系列弹性执行器的通用测试台中最先进的控制范例，可软化用户、假肢和环境之间的交互。该项目的总体目标是 1) 了解如何实现闭环扭矩和。尽管其低成本设计未建模动态，但开源腿的串联弹性执行器中的阻抗控制，以及2）了解如何将高保真联合阻抗控制与两个新颖的连续控制器集成，这两个控制器有望让用户灵活、无缝地该奖项反映了 NSF 的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。

项目成果

期刊论文数量（7）

专著数量（0）

科研奖励数量（0）

会议论文数量（0）

专利数量（0）

Optimizing Representations of Multiple Simultaneous Attributes for Gait Generation Using Deep Learning

使用深度学习优化步态生成的多个同时属性的表示

DOI：
10.1109/tnsre.2023.3273599
发表时间：
2023-01
期刊：
IEEE Transactions on Neural Systems and Rehabilitation Engineering
影响因子：
4.9
作者：
Sharma, Abhishek;Rombokas, Eric
通讯作者：
Rombokas, Eric

Coordinated Movement for Prosthesis Reference Trajectory Generation: Temporal Factors and Attention

假肢参考轨迹生成的协调运动：时间因素和注意力

DOI：
10.1109/biorob49111.2020.9224435
发表时间：
2020-11-01
期刊：
2020 8th IEEE RAS/EMBS International Conference for Biomedical Robotics and Biomechatronics (BioRob)
影响因子：
0
作者：
Vijeth Rai;Abhishek Sharma;Pornthep Preechayasomboon;E. Rombokas
通讯作者：
E. Rombokas

Improving IMU-Based Prediction of Lower Limb Kinematics in Natural Environments Using Egocentric Optical Flow

使用自我中心光流改进自然环境中基于 IMU 的下肢运动学预测

DOI：
10.1109/tnsre.2022.3156884
发表时间：
2022-01
期刊：
IEEE Transactions on Neural Systems and Rehabilitation Engineering
影响因子：
4.9
作者：
Sharma, Abhishek;Rombokas, Eric
通讯作者：
Rombokas, Eric

Complexity of locomotion activities in an outside-of-the-lab wearable motion capture dataset

实验室外可穿戴运动捕捉数据集中运动活动的复杂性

DOI：
10.3389/fbioe.2022.918939
发表时间：
2022
期刊：
Frontiers in Bioengineering and Biotechnology
影响因子：
5.7
作者：
Sharma, Abhishek;Rombokas, Eric
通讯作者：
Rombokas, Eric

Continuous and Unified Modeling of Joint Kinematics for Multiple Activities

多种活动的关节运动学的连续统一建模

DOI：
10.1109/access.2022.3171246
发表时间：
2022-01
期刊：
IEEE Access
影响因子：
3.9
作者：
Rai, Vijeth;Sharma, Abhishek;Boe, David;Preechayasomboon, Pornthep;Rombokas, Eric
通讯作者：
Rombokas, Eric