W-HTF-RL: Collaborative Research: Improving the Future of Retail and Warehouse Workers with Upper Limb Disabilities via Perceptive and Adaptive Soft Wearable Robots

W-HTF-RL：协作研究：通过感知和自适应软可穿戴机器人改善上肢残疾的零售和仓库工人的未来

• 批准号: 2026479
• 负责人: Chen Feng
• 金额: $ 89.99万
• 依托单位: New York University
• 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: 2020-08-15 至 2025-07-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2026479&HistoricalAwards=false
• 关键词: HTF; RL; Collaborative; Research; Improving

This project will investigate modeling, perception, and control of soft wearable robots to provide physical assistance and skill training for older workers and workers with physical disabilities in jobs involving picking, placing, and assembly tasks. If successful, the project will enhance their employment, inclusion, and integration in work that is relevant to retail, warehouse, and manufacturing. It is estimated that this technology can directly benefit nearly 20 million people in the U.S. with upper limb impairments due to neurological and musculoskeletal disorders. The long term goal of the project is to improve the quality of work, productivity, and employment of people with disabilities, who are the nation’s largest minority and untapped labor force. To do so, the project will deploy artificial intelligence-powered, soft assistive robots to support workers and understand the resulting impact on economics and policy making. The proposed work has the potential to contribute to national economic growth and health by broadening participation of people with disabilities in the workforce. The assessment of economic impacts will provide the first econometric data-driven understanding of the productivity and labor market effects of artificial intelligence- and robotics-driven augmentation, with a specific focus on the underrepresented population of individuals with disabilities. This project brings together multiple disciplines, including soft robotics, computer vision, learning and control, occupational therapy, worker training and labor economics. This convergent research team represents a collaboration with Rutgers New Jersey Medical School, New York University, and assistive device manufacturers. The team and project activities are structured to achieve multiple convergent goals and deliverables, including: 1) A model of interactions between a lightweight and complainant soft wearable robot with human workers; 2) An interactive visual perception framework that enables multimodal intention detection and action monitoring in a semantic 3D map of the dynamic workspace and provides in-context visual feedback for collaborative robot manipulation; 3) A framework for the transfer of demonstrated skills through cost function learning and model predictive control with perceptual feedback integration for online movement and impedance adaptation of exoskeletons; 4) Training programs for occupational therapists and people with upper-limb disabilities will entail a multi-pronged strategy to enhance awareness and knowledge regarding the scope and effectiveness of assistive robots and improve perception towards their use in collaborative workspaces; 5) Increased understanding the economics of assistive technologies using federal data on occupational ability requirements and then using these estimates in conjunction with productivity results on assistive technologies to construct a range of cost/benefit estimates of these technologies; and 6) An evidence-based policy approach that uses quantitative and qualitative data from field experiments, interviews, and focus groups with employers and employees to determine attitudes, barriers, and best practices for adoption and acceptance of assistive wearable technologies in the workplace. This project has been funded by the Future of Work at the Human-Technology Frontier cross-directorate program to promote deeper basic understanding of the interdependent human-technology partnership in work contexts by advancing design of intelligent work technologies that operate in harmony with human workers.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.

该项目将研究软可穿戴机器人的建模、感知和控制，以在涉及拾取、放置和组装任务的工作中为年长的工人和肢体残疾的工人提供物理帮助和技能培训。如果成功，该项目将促进他们在与零售、仓库和制造业相关的工作中的就业、包容和整合。据估计，这项技术可以直接惠及美国近2000万因神经和肌肉骨骼疾病而导致上肢残疾的人。该项目的长期目标是提高残疾人的工作质量、生产率和就业，他们是美国最大的少数民族和未开发的劳动力。为此，该项目将部署人工智能驱动的软辅助机器人来支持工人，并了解由此产生的对经济和政策制定的影响。拟议的工作有可能通过扩大残疾人在劳动力中的参与，为国家经济增长和健康作出贡献。对经济影响的评估将提供对人工智能和机器人驱动的增强的生产率和劳动力市场影响的首次计量经济学数据驱动的理解，并特别关注未被充分代表的残疾人群体。该项目汇集了多个学科，包括软机器人、计算机视觉、学习与控制、职业治疗、工人培训和劳动经济学。这个融合的研究团队代表着与新泽西罗格斯医学院、纽约大学和辅助设备制造商的合作。团队和项目活动的结构是为了实现多个融合的目标和交付内容，包括：1)轻型和原告软可穿戴机器人与人类工人之间的交互模型；2)交互式视觉感知框架，其允许在动态工作空间的语义3D地图中进行多模式意图检测和动作监控，并为协作机器人操作提供上下文中的视觉反馈；3)通过成本函数学习和模型预测控制来转移所展示的技能的框架，具有用于在线运动和外骨骼的阻抗适应的感知反馈集成；4)职业治疗师和上肢残障人士的培训计划将需要多管齐下的战略，以提高对辅助机器人的范围和有效性的认识和知识，并改善对其在协作工作空间中使用的感知；5)使用关于职业能力要求的联邦数据，增加对辅助技术的经济学的理解，然后将这些估计与辅助技术的生产率结果结合使用，以构建这些技术的一系列成本/收益估计；6)基于证据的政策方法，使用来自现场试验、访谈和与雇主和员工的焦点小组的定量和定性数据，以确定在工作场所采用和接受辅助可穿戴技术的态度、障碍和最佳实践。该项目由人类-技术前沿交叉部门工作未来计划资助，旨在通过推进与人类工人和谐运作的智能工作技术的设计，促进对工作环境中相互依赖的人类-技术伙伴关系的更深层次的基本了解。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势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

Egocentric Prediction of Action Target in 3D

• DOI: 10.1109/cvpr52688.2022.02033
• 发表时间: 2022-03
• 期刊: 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
• 作者: Yiming Li;Ziang Cao;Andrew Liang;Benjamin Liang;Luoyao Chen;Hang Zhao;Chen Feng

High-Frequency Nonlinear Model Predictive Control of a Manipulator

机械手的高频非线性模型预测控制

• DOI: 10.1109/icra48506.2021.9560990
• 发表时间: 2021
• 期刊: 2021 IEEE-RAS International Conference on Robotics and Automation (ICRA
• 作者: Kleff, Sebastien;Meduri, Avadesh;Budhiraja, Rohan;Mansard, Nicolas;Righetti, Ludovic
• 通讯作者: Righetti, Ludovic

Fooling LiDAR Perception via Adversarial Trajectory Perturbation

• DOI: 10.1109/iccv48922.2021.00780
• 发表时间: 2021-03
• 期刊: 2021 IEEE/CVF International Conference on Computer Vision (ICCV)
• 作者: Yiming Li;Congcong Wen;Felix Juefei-Xu;Chen Feng

Leveraging Forward Model Prediction Error for Learning Control

利用前向模型预测误差进行学习控制

• DOI: 10.1109/icra48506.2021.9561396
• 发表时间: 2021
• 期刊: 2021 IEEE-RAS International Conference on Robotics and Automation (ICRA
• 作者: Bechtle, Sarah;Hammoud, Bilal;Rai, Akshara;Meier, Franziska;Righetti, Ludovic
• 通讯作者: Righetti, Ludovic

On the Derivation of the Contact Dynamics in Arbitrary Frames: Application to Polishing with Talos

任意坐标系中接触动力学的推导：Talos 抛光的应用

• DOI: 10.1109/humanoids53995.2022.10000208
• 发表时间: 2022
• 期刊: IEEE-RAS International Conference on Humanoid Robots
• 作者: Kleff, Sebastien;Carpentier, Justin;Mansard, Nicolas;Righetti, Ludovic
• 通讯作者: Righetti, Ludovic