A Human-Robot Visual-Inertial Monitoring System for Discoveries on Safe Human-Autonomy Interactions in Dynamic Environments

人机视觉惯性监测系统，用于发现动态环境中安全的人机自主交互

• 批准号: RTI-2022-00697
• 负责人: Hashemi, Ehsan
• 金额: $ 10.37万
• 依托单位: University of Alberta
• 依托单位国家: 加拿大
• 项目类别: Research Tools and Instruments
• 财政年份: 2021
• 资助国家: 加拿大
• 起止时间: 2021-01-01 至 2022-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=741750
• 关键词: Human; Robot; Visual; Inertial; Monitoring

To develop safe and computationally efficient cooperative controls in highly dynamic environments by enhancing our understanding of human interactive cognition and attention, we are requesting funding to purchase a state-of-the-art human-robot monitoring system. The system includes remote and on-board sensing units and a wireless electroencephalogram (EEG). It will enable studies on interactive cognition for human-autonomy shared control (e.g., in assistive robotics and automated driving) using EEG at 1kHz, and will allow detection (within 200m) and augmented perception through fusion of the LiDAR and radar data at 20 Hz at each unit (i.e., Edge Computing). The integrated monitoring system along with our existing infrastructure is unique to our cross-university team and establishes a world-class test facility for research on cooperative perception and human-autonomy interaction in dynamic environments in Canada. The equipment will leverage +$0.75 million CFI- and NSERC-funded facilities (on mobile robotics and neuroscience), will support +18 HQP, and will enable a diverse cross-university research program addressing scientific and applied challenges in three main Themes: 1) Networked Robotics: Enhance robot navigation for perceptually degraded conditions by designing learning-aided distributed estimators inspired by human attention. 2) Human-centered teleoperation: Develop safe teleoperation of mobile robots/vehicles in dynamic environments by studying operator behavior in such complex scenarios, subject to delay and loss of perception, and by forming cooperative controls around the human intent. The recent partnership between TELUS and the University of Alberta for a 5G-enabled living lab will be leveraged. 3) Assistive robotics: Improve personalization and safety by estimating user intent, exploring interactive cognition (a genuine breakthrough for co-designing learning modules and state observers), and developing situation-aware controls in rollators for users with cognitive impairment. The monitoring system will further enable a host of interdisciplinary research activities to expand infrastructure usage as a core facility, through the current and new faculty members and HQP from control engineering, psychology, and computer science disciplines, to sustain future growth of research activities, and to support Canadian robotic and AI industries. The strong integration of the program with social and cognitive sciences and our established EDI policies will provide strong means for unbiased and equitable hiring in those disciplines. The program will offer strong training to HQP through: advanced theoretical/technical and professional development training and standardized internships. This will promote rigorous methodologies, enhanced research culture, and impactful outcomes in the program. A delay in acquiring such crucial equipment will constrain scientific potentials for the current cross-university projects and will reduce HQP training.

为了通过提高我们对人类交互认知和注意力的理解，在高度动态的环境中开发安全和计算效率高的合作控制，我们正在申请资金购买最先进的人机监控系统。该系统包括远程和机载传感单元以及无线脑电图（EEG）。它将使人类自主共享控制的交互认知研究成为可能（例如，在辅助机器人和自动驾驶）使用EEG在1 kHz，并将允许检测（200米内）和增强感知通过融合激光雷达和雷达数据在20赫兹在每个单元（即，边缘计算）。集成监控系统沿着我们现有的基础设施是我们跨大学团队所独有的，并建立了一个世界级的测试设施，用于研究加拿大动态环境中的合作感知和人类自主互动。该设备将利用超过75万美元的CFI和NSERC资助的设施（移动的机器人和神经科学），将支持+18 HQP，并将实现一个多样化的跨大学研究计划，解决三个主要主题的科学和应用挑战：1）网络机器人：通过设计受人类注意力启发的学习辅助分布式估计器，增强机器人导航感知退化的条件。2)以人为中心的遥操作：通过研究操作员在这种复杂场景中的行为，在动态环境中开发移动的机器人/车辆的安全遥操作，受到延迟和感知损失，并围绕人类意图形成合作控制。TELUS与阿尔伯塔大学最近就支持5G的生活实验室建立的合作伙伴关系将得到利用。3)辅助机器人：通过估计用户意图、探索交互式认知（共同设计学习模块和状态观察器的真正突破）以及为有认知障碍的用户开发助行车中的情境感知控制，提高个性化和安全性。该监测系统将进一步使一系列跨学科研究活动能够通过来自控制工程，心理学和计算机科学学科的现任和新任教师和HQP扩大基础设施的使用，以维持研究活动的未来增长，并支持加拿大机器人和人工智能行业。该计划与社会和认知科学的紧密结合以及我们既定的EDI政策将为这些学科的公正和公平招聘提供强有力的手段。该计划将通过高级理论/技术和专业发展培训以及标准化实习为HQP提供强有力的培训。这将促进严格的方法，增强研究文化，并在该计划的影响力的结果。延迟获得这些关键设备将限制当前跨大学项目的科学潜力，并将减少HQP培训。