PFI:BIC - Adaptive Robotic Nursing Assistants for Physical Tasks in Hospital Environments

创新伙伴关系：建设创新能力（PFI:BIC）项目旨在提供下一代辅助机器人，以支持医院注册护士（RNs）的活动。美国有近300万注册护士，这使她们成为美国最大的医疗保健提供者。影响这一庞大劳动力资源的技术必然会产生影响。由于机器人技术和计算机技术的进步，智能通信、传感和计算硬件的使用正变得越来越普遍——不仅对医疗保健专业人员来说是如此，对患者本身也是如此。该项目由肯塔基大学路易斯维尔分校与德克萨斯大学阿灵顿分校合作，将专注于创造新的设计工具，可以配置自适应机器人护理助理（ARNA）的硬件和软件。ARNA将专门用于帮助医疗机构的护士完成简单的任务，如升降辅助，运送日常轻量级物品（药品，医疗可穿戴设备），以及移动较重物品（如家具，轮床和患者本身）的一些物理辅助。从该项目中获得的见解所产生的设计和工程创新，除了在医院之外，还可以作为产品部署到更广泛的消费者市场中，具有巨大的价值。例子包括家庭服务和辅助机器人，公共场所的辅助机器人，以及人类与机器人工人近距离接触的协同机器人制造。对人-机器人和护士-机器人互动的更好理解可以代表使能技术，这将促进研究突破，提高机器人的生产力和社会接受度。该研究还将促进对机器人设计美学和界面的感知效应的理解。拟议中的自适应机器人护士助理将在杂乱的医院中导航，同时配备多模态皮肤传感器，可以预测护士的意图，自动执行日常的低级任务，但让护士保持决策循环。模块化和强大的硬件将部署在专门为护士物理援助设计的可重构平台上。自适应人机界面将在该项目中发挥关键作用，因为这些界面直接影响机器人在动态、非结构化环境中帮助护士的能力。与预先编程机器人的行为不同，学习算法将被用来使机器人适应人类的偏好。两个领先的应用程序设想的可行性评估通过定量和定性指标，包括病人的保姆和步行者。坐式机器人将测量生命体征，评估患者运动和姿势的风险，并对患者进行持续观察，并向护士提供反馈。行走机器人将帮助护士和病人提供部分平衡支持，导航混乱的环境，并协助医疗设备运输。牵头机构是肯塔基大学路易斯维尔分校与德克萨斯大学阿灵顿分校及其多学科部门合作，包括工程学院、护理学院和德克萨斯大学阿灵顿研究所（UTARI）。主要的工业合作伙伴包括QinetiQ-North America（马萨诸塞州沃尔瑟姆），一家专门从事无人系统的大公司，以及RE2（宾夕法尼亚州匹兹堡），一家专门从事模块化机器人操纵器的小企业，将提供独特的经过实战考验的硬件和系统工程。提议的机器人的住院测试和评估将由德克萨斯大学阿灵顿护理学院和德克萨斯健康资源（达拉斯-沃斯堡，德克萨斯州）的护士研究人员进行，这是一家大型医疗保健提供商。

This Partnerships for Innovation: Building Innovation Capacity (PFI:BIC) project aims to provide next-generation assistive robots to support the activities of hospital-based registered nurses (RNs). There are nearly three million registered nurses employed in the United States, making them the largest pool of healthcare providers in the country. Technology that affects the performance of this large labor pool cannot fail to have impact. Due to advancements in robotics and computer technology, access to intelligent communication, sensing, and computing hardware is on the cusp of becoming common--not only for healthcare professionals, but also for patients themselves. The project led by The University of Kentucky at Louisville in collaboration with the University of Texas at Arlington will focus on the creation of new design tools that can configure the hardware and software of adaptive robotic nursing assistants (ARNA). ARNA will be specifically designed to assist nurses in healthcare facilities with simple tasks such as, lift assistance, delivery of everyday lightweight objects (medicine, medical wearable equipment), and some physical assistance with movement of heavier objects, such as furniture, gurneys, and the patients themselves. The design and engineering innovations resulting from insights gained in this project may have great value deployed as products in broader consumer markets in addition to hospitals. Examples include in-home service and assistive robots, robots for assistance in public venues, and co-Robot manufacturing where humans are in close proximity to robot workers. The improved understanding of human-robot and nurse-robot interaction could represent enabling technology that will facilitate research breakthroughs and increase productivity and social acceptance of robotics. The research will also advance the understanding of the perceptual effects of robot design aesthetics and interfaces.The proposed Adaptive Robotic Nurse Assistants will navigate cluttered hospitals, while equipped with multi-modal skin sensors that can anticipate nurse intent, automate mundane low-level tasks, but keep nurses in the decision loop. Modular and strong hardware will be deployed in reconfigurable platforms specially designed for nurse physical assistance. Adaptive human-machine interfaces will play a key role in this project, as these interfaces directly impact the ability of robots to help nurses in a dynamic, unstructured environment. Rather than pre-programming robot behaviors, learning algorithms will be used so that robots adapt to human preferences. Two leading applications are envisioned for feasibility evaluation by quantitative and qualitative metrics, including patient sitters and walkers. The sitter robot will take vital sign measurements, evaluate risk from patient movement and pose, and provide continuous observation of patients and feedback to and from nurses. The walker robot will assist nurses and patients by providing partial balance support, navigating cluttered environments, and assisting with medical equipment transportation.The lead institution is the University of Kentucky at Louisville in collaboration with the University of Texas at Arlington with its multidisciplinary departments including the College of Engineering, College of Nursing, and the University of Texas at Arlington Research Institute (UTARI). Primary industrial partners include QinetiQ-North America (Waltham, MA), a large corporation specializing in unmanned systems, and RE2 (Pittsburgh, PA), a small business specializing in modular robotic manipulators that will contribute unique battle-tested hardware and systems engineering. In-hospital testing and evaluation of the proposed robots will be carried out by nurse researchers at the University of Texas at Arlington College of Nursing and Texas Health Resources (Dallas-Fort Worth, TX), a large healthcare provider.