Collaborative Research: Legible Co-Adaptation of Wearable Devices for As-Needed Assistance of Arm Motion

合作研究：可穿戴设备的清晰协同适应，以按需协助手臂运动

• 批准号: 2110214
• 负责人: Holly Yanco
• 金额: $ 50.6万
• 依托单位: University of Massachusetts Lowell
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-10-01 至 2024-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2110214&HistoricalAwards=false
• 关键词: Collaborative; Research; Legible; Co; Adaptation

For people who have difficulties moving their arms or grasping objects, an exoskeleton can be used to augment a person’s capabilities and, in some cases, to provide rehabilitation. An arm exoskeleton can take the form of a brace worn from the hand to above the elbow, with motors to assist movement and to provide additional strength. Existing commercial exoskeleton systems have software with control parameters that can be adjusted to an individual’s capabilities, needs, and comfort, but these settings are not modified automatically while the device is being used. In some cases, a person could fatigue over the course of a long task, thus requiring the exoskeleton to help more at the end of a task than at the beginning. In other cases, a person might gain more ability as they are assisted in movements over days or weeks, allowing the exoskeleton to start providing less assistance. The motivating objective of this research is to enable co-adaptation of an arm-worn exoskeleton, where the control parameters of the exoskeleton adapt over time to the person’s changing capabilities, generating synergistic coordinated motion between the human-exoskeleton team. This project promotes the progress of science and advances the national health, prosperity and welfare through the development of an adaptive exoskeleton controller designed to provide as-needed motion assistance based on the user's changing level of muscular fatigue and capabilities. The research considers augmenting the abilities of healthy participants, which aligns with reducing musculoskeletal risk for an industrial or military application, as well as motion assistance for older populations, which can aid with sensorimotor deficits due to aging. The project addresses fatigue detection, legible notifications, and device assistance to enable synergistic physical and cognitive fit between the human and intelligent exoskeleton machine. This convergent research brings together theories from computer science, human factors, robotics, physical therapy, and kinesiology to advance a fundamental understanding of human-robot interactions. The project also includes public engagement through STEM activities for middle school girls, as well as through activities for life-long learning for older adults.The function of the co-adaptive algorithm can be expressed using principles motivated by dynamical systems theory. The algorithm will adapt the exoskeleton controller at the parameter dynamics level (a time scale describing parameters associated with completing the task) for a specified graph dynamics (a longer time scale describing the connectivity architecture representing the system). The algorithm encourages stability of the graph dynamics using a legibility scheme. The project team will use the commercially available Myomo MyoPro powered orthosis as a platform to evaluate co-adaptation of a human-exoskeleton team. The project team plans a control legibility scheme that will convey a series of notifications to the user to inform her/him about what the MyoPro is doing. The legibility scheme is designed to align with enabling a calibrated trust of the user with respect to the device and to prevent bifurcations in the graph dynamics. The specific objectives of the research are: (1) to incorporate measures of muscle fatigue in a legible co-adaptive exoskeleton controller; (2) to evaluate the hypothesis that a co-adaptive controller can improve human-exoskeleton task performance in the presence of fatigue in a healthy young adult population; and (3) to examine the generalizability of the co-adaptive controller to an older adult population. The primary task across all studies is a pick-and-place task where participants grasp an object (e.g., a book, pen, basket, cup) from a table and place it on a bookshelf, then return the object to the table. The use of different object sizes and weights increases the number of grasp types in the dataset, as well as the motor control strategies required, providing increased generalizability of the algorithms across grasping tasks. Participants will perform concurrent cognitive tasks that will be used to assess cognitive load during the primary pick-and-place task. To increase the rate of fatigue, the primary task is alternated with a fatiguing task. Results from these studies will advance the usability of exoskeleton systems by accounting for natural forms of human adaptation within the exoskeleton control policy, enabling exoskeletons to be applied for longer-duration applications that support aging in place and the mitigation of injury.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.

对于难以移动手臂或抓握物体的人，外骨骼可用于增强人的能力，并且在某些情况下，提供康复。手臂外骨骼可以采取从手穿戴到手肘上方的支架的形式，其具有马达以辅助移动并提供额外的力量。现有的商业外骨骼系统具有带有控制参数的软件，这些控制参数可以根据个人的能力、需求和舒适度进行调整，但是这些设置在设备被使用时不会自动修改。在某些情况下，一个人可能会在长时间的任务过程中感到疲劳，因此需要外骨骼在任务结束时比在开始时提供更多的帮助。在其他情况下，一个人可能会获得更多的能力，因为他们在几天或几周的运动中得到帮助，允许外骨骼开始提供更少的帮助。本研究的激励目标是实现臂戴式外骨骼的共同适应，其中外骨骼的控制参数随着时间的推移而适应人的变化能力，从而在人-外骨骼团队之间产生协同的协调运动。该项目通过开发自适应外骨骼控制器来促进科学的进步，并促进国民的健康，繁荣和福利，该控制器旨在根据用户的肌肉疲劳和能力的变化水平提供所需的运动辅助。该研究考虑增强健康参与者的能力，这与降低工业或军事应用的肌肉骨骼风险以及老年人群的运动辅助相一致，这可以帮助由于衰老而导致的感觉运动缺陷。该项目致力于疲劳检测、清晰的通知和设备辅助，以实现人类和智能外骨骼机器之间的协同物理和认知适应。这种融合的研究汇集了计算机科学，人为因素，机器人技术，物理治疗和运动学的理论，以推进对人机交互的基本理解。该项目还包括通过面向中学女生的STEM活动以及面向老年人的终身学习活动来实现公众参与。协同自适应算法的功能可以使用动力系统理论所激发的原理来表达。 该算法将在参数动态水平（描述与完成任务相关联的参数的时间尺度）处使外骨骼控制器适应指定的图形动态（描述表示系统的连接性架构的较长时间尺度）。该算法鼓励稳定的图形动态使用易读性计划。 该项目团队将使用市售的Myomo MyoPro动力矫形器作为平台，以评估人类外骨骼团队的共同适应性。项目团队计划了一个控制易读性方案，该方案将向用户传达一系列通知，以告知她/他MyoPro正在做什么。易读性方案被设计成与实现用户相对于设备的校准信任一致，并防止图形动态中的分叉。该研究的具体目标是：（1）将肌肉疲劳的测量纳入清晰的共适应外骨骼控制器中;（2）评估共适应控制器可以在健康的年轻成年人群体中存在疲劳的情况下改善人类外骨骼任务性能的假设;以及（3）检查共适应控制器对老年人群体的可推广性。所有研究的主要任务是一个拾取和放置任务，参与者抓住一个物体（例如，书、笔、篮子、杯子）放在书架上，然后把物体放回桌子上。使用不同的物体大小和重量增加了数据集中抓取类型的数量，以及所需的运动控制策略，从而提高了算法在抓取任务中的通用性。参与者将同时执行认知任务，这些任务将用于评估主要拾取和放置任务期间的认知负荷。为了增加疲劳率，主要任务与疲劳任务交替进行。这些研究的结果将通过考虑外骨骼控制策略中人类适应的自然形式来提高外骨骼系统的可用性，使外骨骼能够应用更长时间-该奖项反映了NSF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。