A Shared Autonomy Approach to Robotic Arm Assistance with Daily Activities

高度麻痹患者依靠他人的照顾和环境的改造来完成日常生活活动。由于需要一定程度的支持，对瘫痪患者的护理费用很高。通过技术手段恢复操作能力将大大改善使用者的生活质量，并将减少瘫痪的长期经济影响，并通过解放护理人员（通常是家庭成员或朋友）的时间和劳动为社会做出贡献。为此，该项目建议设计并验证一个安装在轮椅上的机械臂，该机械臂带有增强现实界面，可实现非触觉人机交互。重要的是，该设计将涉及多模态接口方法，其中包括最近开发的舌驱动接口，头部方向传感器和语音识别系统。以用户为中心和参与式设计方法将确保工程系统将被四肢瘫痪的人看好。该奖项还支持人机协作辅助设备设计的基础研究，包括界面设计和潜在的机器人视觉和规划算法。研究贡献包括提高对如何有效地协调人类的高级推理和思维过程与当前机械臂的自主操作能力和局限性的理解。机器人应用的社会意义将被资本化，以创造引人入胜的教育和推广活动，以促进工程数学。长期目标是改变轮椅使用者的生活，使他们能够独立进行日常生活活动（ADL），而不受环境的影响。满足ADL各种需求的技术往往具有较高的控制复杂性，这阻碍了采用。这些技术必须要求高级（引导）命令而不是低级控制信号，并以与用户对世界的概念相容的方式要求反馈。研究目标是设计和验证具有增强现实界面和多模态用户界面的轮椅机器人手臂，用于将人类命令和意图控制回路与机器人手臂决策和控制回路耦合。该系统的新颖之处在于它是一种共享控制和共享感知辅助技术。该耦合系统需要人工输入来克服机器人视觉的感知限制，并采用机器人的规划和操作来克服用户的物理限制。通过让人类参与场景解读，并向机器人手臂提供以自我为中心的信息，耦合系统在复杂的操作任务中最大限度地减少和简化了用户输入。相关的研究目标是：(I)设计一个人机增强现实界面系统，该系统对两个系统（人类和机器人）进行耦合反馈，以沟通意图并要求复杂操作任务的高级用户反馈；（II）通过参与式设计过程确定操作辅助需求和用户界面设计；（III）根据期望的命令和控制目标评估用户界面。（四）评估操作系统对任务绩效和认知负担的影响。设计独特的共享控制和共享传感系统并实现所提出的研究目标将改变辅助机器人技术将用户与技术结合的方式。将这项研究纳入现有的教育和推广活动，将向学生推广工程和机器人技术，并为他们参与日益自动化的世界做好准备。

Persons with high-level paralysis rely on the care of others and the modification of their environment for accomplishing the activities of daily living (ADL). Due to the degree of support required, paralysis is costly to provide care for. Recovering the ability to manipulate through technological means would lead to significant improvements to the quality of life for the user, and would reduce the long-term economic impact of paralysis and contribute to society by freeing up caregiver time and labor (usually a family-member or friend). To that end, the project proposes to design and validate a wheelchair-mounted robotic-arm with an augmented reality interface for enabling non-tactile human-robot interaction. Importantly, the design will involve a multi-modal interface approach, which includes a recently developed tongue drive interface, a head orientation sensor, and a speech recognition system. User-centric and participatory design methods will ensure that the engineered system will be seen favorably by persons with tetraplegia. The award also supports fundamental research into the design of collaborative human-robot assistive devices, including the interface design and the underlying robot vision and planning algorithms. The research contribution includes improved understanding on how to effectively coordinate the higher level reasoning and thought processes of humans with the autonomous operation capabilities and limitations of current robotic arms. The social significance of the robotics application will be capitalized to create engaging educational and outreach activities in promotion of engineering mathematics.The long-term goal is to transform the lives of wheelchair bound persons with limited to no manipulation abilities by enabling them to independently perform the activities of daily living (ADL) irrespective of their environment. Technology meeting the varied needs of the ADL tends to have high control complexity, which impedes adoption. It is essential for these technologies to request high-level (guiding) commands rather than low-level control signals, and to request feedback in a manner compatible with the user's conception of the world. The research goal is to engineer and validate a wheelchair-fitted robotic-arm with an augmented reality interface and a multi-modal user interface for coupling the human command and intent control loop with the robotic-arm decision and control loop. The novelty of the proposed system is that it is a shared control and a shared sensing assistive technology. The coupled system requires human input to overcome the perceptual limitations of robot vision, and employs robotic planning and manipulation to overcome the physical limitations of the user. By involving the human for scene interpretation and by providing ego-centric information to the robot arm, the coupled system minimizes and simplifies user input during complex manipulation tasks. The associated research objectives are to (I) engineer a human-robot augmented reality interface system with coupled feedback to the two systems (human and robot) for communicating intent and requesting high-level user feedback for complex manipulation tasks, (II) identify the manipulation assistance needs and user interface design through a participatory design process, (III) evaluate the user interface with respect to the desired command and control objectives, and (IV) assess the impact of the operational system with respect to task performance and cognitive burden. Engineering the unique shared control and shared sensing system and achieving the proposed research objectives would transform the way that assistive robotic technologies couple the user to technology. Incorporating the research into existing education and outreach activities would promote engineering and robotics to students and prepare them for participation in an increasingly automated world.