CHS: Medium: Collaborative Research: Novel Optimal Control for Co-Adaptation of Human and Powered Lower Limb Prosthesis

CHS：媒介：协作研究：人类和动力下肢假肢共同适应的新型最优控制

• 批准号: 1563454
• 负责人: He Huang
• 金额: $ 74.22万
• 依托单位: North Carolina State University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2016
• 资助国家: 美国
• 起止时间: 2016-07-15 至 2022-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1563454&HistoricalAwards=false
• 关键词: CHS; Medium; Collaborative; Research; Novel

Emerging powered lower limb prostheses hold great promise for restoring normative locomotion in amputees. However, these robotic devices currently lack inter- and intra-wearer adaptability to cope with wearers' physical variations and changes. Frequent manual and heuristic adjustment in clinics is required, which limits the practical use of these advanced prostheses. A new generation of prosthesis control that is intelligent, adaptable, and interactive is needed to better support walking function and improve the quality of life of lower limb amputees. The PIs' long-term research goal is to create bionic legs that can adapt to the individual amputee's physical and cognitive capabilities, coordinate with the wearer's movement and intent, adjust to changing environments, and essentially restore the full function of patients with lower limb impairments. To this end, the PIs' objective in this project is to create a novel optimal control framework for these prostheses. They will systematically address the challenge of supporting automatic adaptation to the wearer's physical capability while achieving desired gait performance for the integrated amputee-prosthesis system. And they will provide a preliminary design and evaluation for an interactive interface that would allow wearers to personalize prosthesis control safely and easily. Project outcomes will open up a new frontier of wearable robotics and lay the foundation for clinical translations of these innovative devices, which will impact not only the prosthetics and orthotics industry but also the robotics community by providing new knowledge relating to human-robot interaction, the biomechanics and neuromotor control community by elucidating the control mechanism of amputee locomotion, and healthcare in general by providing innovative and cost-effective prosthesis solutions.The new amputee-prosthesis performance-based framework for control of powered lower limb prostheses which this work will introduce represents a departure from existing approaches that mainly focus on design for the prosthesis (a local machine), in that it adopts a global approach by accounting for co-adaptation between amputees and prostheses in order to provide optimal, personalized assistance based on wearers' physical conditions. The PIs will use approximate dynamic programming (ADP) to achieve the global control goal. Such innovative use of ADP will provide an opportunity to demonstrate its optimal adaptive control capability in a new test domain of co-adaptive robotic prosthesis, a unique and significant challenge only seen in human wearable robotics but not in lifeless robots. The ADP scheme is based on approximation and learning that alleviate problems associated with the requirement of accurately modeling wearers' neuromuscular control and dynamics that is difficult, if not impossible, to achieve. Additionally, the PIs will conduct an experimental investigation on subjects with transfemoral amputations of the interactions between amputees and prostheses, including evaluations of the compensatory strategies of amputees, and discrepancy assessment between subjective (human) and objective (machine) preferences in prosthesis control.

新兴的动力下肢假肢对恢复截肢者的规范运动有着巨大的希望。 但是，这些机器人设备目前缺乏磨损和内部内部适应性的适应性，无法应对佩戴者的身体变化和变化。 需要在诊所进行频繁的手动和启发式调整，这限制了这些高级假体的实际使用。 需要智能，适应性和互动的新一代假体控制，以更好地支持步行功能并改善下肢截肢者的生活质量。 PIS的长期研究目标是创建仿生的腿，以适应单个截肢者的身体和认知能力，与佩戴者的运动和意图协调，适应不断变化的环境，并基本上恢复下肢损伤患者的全部功能。 为此，PIS的目标是为这些假体创建一个新颖的最佳控制框架。 他们将系统地应对支持对佩戴者身体能力的自动适应的挑战，同时为集成的截肢者 - 预疾病系统实现所需的步态性能。 他们将为交互式接口提供初步的设计和评估，该界面将使佩戴者可以安全，轻松地个性化假体控制。 项目成果将为这些创新设备的临床翻译奠定新的前沿，并为这些创新设备的临床翻译奠定了基础，这不仅会影响假体和矫形器行业，而且还会影响机器人技术社区，通过提供与人类机器人相互作用相关的新知识，通过提供与神经机械的社区相关的新知识，并通过对全面的机械进行控制和保健机械的控制，并提供了整体机械，并提供了机械性机械，并提供了Ampution and Putpution，并提供了对Ampution的机械性，并提供了对Amputise and Putpution and Putpution的控制。具有成本效益的假体解决方案。基于截肢的基于绩效的基于控制的下肢假体框架，这项工作将引入这些框架，这将引入这些框架，这代表着与现有方法的背离，主要关注假体设计（本地机器）（本地机器），因为它通过对临时性的依从性进行了良好的效果，从而在全球范围内采用全球方法，从而提供了良好的效果，从而提供了最佳的效果。 PI将使用近似动态编程（ADP）实现全球控制目标。 ADP的这种创新使用将提供一个机会，以证明其在共同自适应机器人假体的新测试领域中的最佳自适应控制能力，这是一个独特而重大的挑战，仅在人类可穿戴机器人中看到，而在无生理机器人中则不能。 ADP方案基于近似和学习，可以减轻与精确建模佩戴者的神经肌肉控制和动态相关的问题，这是困难（即使不是不可能）实现的。 此外，PI将对截肢者与假体之间相互作用的受试者进行实验研究，包括对截肢者补偿性策略的评估，以及主观（人类）和客观（机器）偏好之间的差异评估。

项目成果

Offline Policy Iteration Based Reinforcement Learning Controller for Online Robotic Knee Prosthesis Parameter Tuning

基于离线策略迭代的强化学习控制器，用于在线机器人膝关节假体参数调整

• DOI: 10.1109/icra.2019.8794212
• 发表时间: 2019
• 期刊: IEEE International Conference on Robotics and Automation (ICRA
• 作者: Li, Minhan;Gao, Xiang;Wen, Yue;Si, Jennie;Huang, He Helen
• 通讯作者: Huang, He Helen

Environmental Context Prediction for Lower Limb Prostheses With Uncertainty Quantification

• DOI: 10.1109/tase.2020.2993399
• 发表时间: 2020-05
• 期刊: IEEE Transactions on Automation Science and Engineering
• 影响因子: 5.6
• 作者: Boxuan Zhong;R. L. D. Silva;Minhan Li;H. Huang;E. Lobaton

Wearer-Prosthesis Interaction for Symmetrical Gait: A Study Enabled by Reinforcement Learning Prosthesis Control

• DOI: 10.1109/tnsre.2020.2979033
• 发表时间: 2020-04-01
• 期刊: IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING
• 影响因子: 4.9
• 作者: Wen, Yue;Li, Minhan;Huang, He
• 通讯作者: Huang, He

Online Reinforcement Learning Control for the Personalization of a Robotic Knee Prosthesis

• DOI: 10.1109/tcyb.2019.2890974
• 发表时间: 2020-06-01
• 期刊: IEEE TRANSACTIONS ON CYBERNETICS
• 影响因子: 11.8
• 作者: Wen, Yue;Si, Jennie;Huang, He (Helen)
• 通讯作者: Huang, He (Helen)