FRR: A new strategy for task-agnostic control of robotic exoskeletons by estimating underlying biological effort using deep learning

人类的活动能力和身体活动与生活质量、参与、抑郁和死亡率密切相关。对于体力要求较高的工作人员或由于疾病、年老或受伤导致下肢损伤的人员，迫切需要改善他们在行走任务中的运动表现。机器人外骨骼具有帮助人们提高社区流动性、独立性和生活质量的潜力。下肢机器人外骨骼在增强和恢复人体运动方面具有突出的潜力；然而，目前的解决方案缺乏对用户的个性化和跨活动的泛化，极大地限制了外骨骼技术。该项目推进了外骨骼技术，以支持广泛和多样化环境中的人类用户。该研究项目采用了一种独特的方法，将大数据驱动的方法与AI（人工智能）结合起来，使外骨骼控制能够在日常生活的许多任务中实现泛化，包括循环行走任务以及更离散的任务，如切割、跳跃和跨越障碍。该方法包括在许多不同的任务中收集佩戴外骨骼的自然人类数据，并创建识别人类关键内部状态（例如下肢联合活动）的人工智能系统。这创造了一个控制系统，既可以扩展到新的人类用户，也可以扩展到新的下肢任务。这对整个社会产生了巨大的广泛影响，例如改善了临床人群的社区和户外活动，提高了人的能力。机器人基础研究项目的这项研究得到了计算机和信息科学与工程理事会计算机和网络系统部门（CISE/CNS）以及工程理事会土木、机械和制造创新部门（ENG/CMMI）的支持。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。

Human mobility and physical activity are strongly correlated with quality of life, participation, depression, and mortality. There is a critical need for improving locomotor performance in ambulation tasks for individuals with physically demanding jobs or individuals with lower limb impairment due to disease, old age, or injury. Robotic exoskeletons have the potential to assist people to increase their community mobility, independence, and quality of life. Lower-limb robotic exoskeletons have outstanding potential to augment and restore human movement; however, current solutions lack personalization to the user and generalization across activities, constraining exoskeleton technology significantly. This project advances exoskeleton technology to support human users in broad and diverse environments.This research project takes a unique approach of using large data driven approaches with AI (artificial intelligence) to enable generalization of exoskeleton control across numerous tasks of daily living including both cyclic walking tasks as well as more discrete tasks such as cutting, jumping, and stepping over an obstacle. The approach includes collecting natural human data wearing exoskeletons across numerous different tasks and creating AI systems that recognize key internal states of the human, such as their lower limb joint efforts. This creates a control system that is extensible across both new human users as well as novel lower limb tasks. This has tremendously broad impact across society, such as improving community and outdoor ambulation in clinical populations, enhancing human capability.This research in the program for Foundational Research in Robotics is supported by the Computer and Information Science and Engineering Directorate's Division of Computer and Network Systems (CISE/CNS), and the Engineering Directorate's Division of Civil, Mechanical and Manufacturing Innovation (ENG/CMMI).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.