Facilitating Bidirectional Communication for Fast and Flexible Human-Robot Collaboration

协作机器人通过将机器精度与人的灵活性相结合，可以改善和增强各种行业的运营，如制造业和医疗保健。然而，在开发能够与人类助手一起灵活、安全和智能地在结构较少的环境中操作的智能机器人助手方面仍然存在巨大的挑战；在理解机器人应该如何导航以及弥合机器和人类之间的交互裂缝方面存在差距。对于流畅的协作，必须开发方法和算法来增强机器人和人类之间的理解、学习和交流。我拟议的研究旨在通过四种方法解决在人-机器人交互(HRI)环境中对更紧密沟通渠道的需求：(A)使用机器学习对人类行为和不可预测性进行建模，以使机器人的反应情景化；(B)在社会和物理HRI中探索双向非交流线索以支持流畅的协作；(C)制定策略，使机器人能够随着时间的推移学习改善行为；以及(D)使用虚拟现实(VR)和增强现实(AR)来增强用户对机器人上下文和行为的理解。这项工作旨在为智能协作机器人助理走出实验室，进入家庭和工作场所铺平道路，在那里他们可以安全有效地支持和帮助人们。我希望通过这个研究项目培养三名本科生、三名硕士和两名博士生。这些HQP将在HRI、机器学习和AR/VR等尖端领域获得实践经验，传授工业界和学术界追求的相关技能和知识。

Collaborative robotics can improve and enhance operations in various industries, such as manufacturing and healthcare, by merging machine precision with human flexibility. However, there are still tremendous challenges in developing intelligent robot assistants that can operate in less-structured contexts alongside human counterparts flexibly, safely, and intelligently; gaps exist in understanding how robots are supposed to navigate and bridge the interaction rift between machines and humans. For fluid collaboration, methods and algorithms must be developed to enhance understanding, learning, and communication between robots and humans. My proposed research aims to address the need for tighter communication channels in human-robot interaction (HRI) contexts through four methods: (a) modelling of human behaviour and unpredictability to contextualize robot responses using machine learning, (b) exploring bidirectional non-verbal communicative cues in both social and physical HRI to support fluent collaboration, (c) developing strategies to allow robots to learn to improve behaviours over time, and (d) using virtual reality (VR) and augmented reality (AR) to enhance user understanding of robot contexts and behaviours. This work aims to pave the way for intelligent collaborative robotic assistants to move out of the lab and into homes and workspaces where they can support and assist people safely and effectively. I expect to train three undergraduate, three master's, and two doctoral students through this research program. These HQP will be provided with hands-on experience in the cutting-edge domains of HRI, machine learning, and AR/VR, imparting relevant skills and knowledge sought after by industry and academia.