CHS: Medium: Collaborative Research: Augmenting Human Cognition with Collaborative Robots

协作机器人是机器人技术在制造、采矿、建筑和能源工业环境中的一个日益增长的应用空间。国际机器人联合会最近的一份报告显示，到2025年，全球机器人支出将达到130亿美元。工业机器人的最大消费者一直在亚洲市场，即中国、日本、韩国，美国落后于欧洲（德国、法国、西班牙）和亚洲。确保美国工业和工人重新获得协作制造机器人领域的领导地位，符合美国的经济和战略利益。为了实现这一目标，该项目将发展对技术和社会技术要求的理解，以加速协作机器人的使用。该项目将通过增强人类对更安全、更有效的协作机器人交互的认知，为人机交互和人机交互提供新的知识和理论。协作机器人技术的新设计原则将促进工人和雇主的成长和进步。在这里获得的基础知识将直接适用于其他使用协作机器人的高风险领域，如海上石油钻井平台、军事和建筑。该项目旨在增强新工人群体（如残疾工人）的权能，允许老年工人继续工作，并可能帮助新工人，从而缩小技能差距，提高工作效率。该团队将专注于扩大女性在计算机领域的参与。除了传统的学术场所（如会议、期刊等）外，研究成果还将通过现有的州、区域和国家雇主和行业合作伙伴网络，通过工作场所讲习班和研讨会进一步传播。为了实现这些目标，研究团队计划：(1)在易受系统故障影响的协作机器人环境中开发一种新的HRI任务/场景分类方案；(2)建立基本的神经生理、认知和社会行为模型（工作量、认知负荷、疲劳/压力、影响和信任）来监测和模拟思维-运动-机器联系；(3)利用这些模型来确定人类的认知、社会、行为和环境状态何时以及如何需要通过技术进行调整，以提高人力资源指数以实现高效和安全的工作绩效；最后(4)创建一个创新和变革性的工作4.0架构（AMELIA：增强学习创新），其中包括一层增强现实，供人类和机器人相互学习和交流当前状态。该团队将从他们的生理数据和眼动追踪中推断出工人的认知状态。然后，他们将使用嵌入式传感器读数、错误代码和监控摄像头来描述机器人的状态。通过增强现实，AMELIA将向工人和机器人提供这些数据，以便有效地实时调整行为，以减少故障来源和错误，同时确保最小的额外认知负荷。该团队计划使用人工情感智能设计一种新颖的交流方案，在这种方案中，机器人和人类可以在潜在的危险情况下合作。机器人将使用不同的机器学习技术检测工人的认知状态，然后采取适当的行动。最终，AMELIA旨在使工人能够专注于复杂的认知问题解决任务，安全有效地执行，同时确保它适应工人的态度和认知状态。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。

Collaborative robotics is a growing application space in robot technology used in manufacturing, mining, construction, and energy industrial settings. A recent report by the International Federation of Robots indicates that global robotics spending will reach $13 billion in 2025. The largest consumers of industrial robotics have been in the Asian market, i.e. China, Japan, Republic of Korea, with the U.S. lagging behind both Europe (Germany, France, Spain) and Asia. It is in the national economic and stratgic interest to ensure that US industry and workers regain leadership in collaborative manufacturing robots. Towards that goal, this project will develop understanding of technical and socio-technical requirements to accelerate the use of collaborative robotics. The project will contribute new knowledge and theory of Human-Computer Interaction and Human-Robot Interaction, by augmenting human cognition for safer and more efficient collaborative robot interaction. The new design principles for collaborative robotic technologies will improve both the worker and the employer's growth and progress. Fundamental knowledge gained here will be directly applicable in other high-risk domains that use collaborative robots, such as offshore oil rigs, military, and construction. The project seeks to empower new populations of workers (e.g., workers with disabilities), allow older workers to remain in the workforce, and potentially assist novice workers, thereby reducing skills gaps and improving work efficiency. The team will focus on broadening the participation of females in computing. In addition to traditional academic venues (e.g. conference, journals, etc.), research results will be further disseminated through workplace workshops and seminars through existing state, regional, and national networks of employers and industry partners. To meet these goals, the team of researchers plan to: (1) develop a novel HRI task/scenario classification scheme in collaborative robotics environments that are vulnerable to system failures; (2) establish fundamental neurophysiological, cognitive, and socio-behavioral models (workload, cognitive load, fatigue/stress, affect, and trust) to monitor and model the mind motor machine nexus; (3) use these models to determine when and how a human's cognitive, social, behavioral and environmental states require adjustment via technology to enhance HRI for efficient and safe work performance; and finally (4) create an innovative and transformative Work 4.0 architecture (AMELIA: AugMEnted Learning InnovAtion) that includes a layer of augmented reality for human and robots to mutually learn and communicate current states. The team will characterize worker cognitive states inferred from their physiological data and eye tracking. They will then use embedded sensor readings, error codes, and surveillance cameras to characterize robot states. Through augmented reality, AMELIA will provide this data to both the worker and the robot for effective real-time adjustment in behaviors to mitigate failure sources and errors while ensuring minimal additional cognitive load. The team plans a novel communication scheme using artificial emotional intelligence in which robots and humans collaborate in potentially dangerous situations. The robot will detect the worker's cognitive state using different machine learning techniques, and then take the appropriate action. Ultimately AMELIA seeks to empower the worker to focus on complex, cognitive problem-solving tasks, performed safely and efficiently, while ensuring that it adapts to both the worker's attitudes and cognitive states.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.