Theoretical and Methodological Foundations for the Design and Implementation of Effective Soft Robots

“软”机器人的出现从根本上改变了机器人系统的设计和建造方式。“软”机器人的身体由可伸缩和柔韧的材料制成。尽管社会对安全、协作的机器人设备的需求和前景看好的设计概念的激增，软机器人还没有出现在人类协作和人类交互的应用中，在这些应用中，它们可能是最有用的。这项提案的研究目标是发现方法和原则，以解决导致软机器人承诺和实施之间的差距的根本问题。这项拟议的研究将使用一种结合了连续介质力学、机器学习和新的形状测量技术的混合方法来回答有关在各种环境中部署的具有各种体系结构的软机器人的设计和控制的公开研究问题。软机器人的设计和实现尤其困难，因为对它们的行为尚不存在基本的了解。例如，软机器人拥有连续可变形的物体，而这些物体不能用基于刚体运动学的传统方法建模。短期内，这项研究计划将研究软机器人身体的柔软、可伸展和顺应材料和结构的行为。它将遵循四个目标：测量软机器人的表面形态，对软机器人的应力松弛、蠕变和滞后进行建模，开发受益于物理原理和机器人数据的混合建模方法，并研究联合优化软机器人的设计、传感和控制的计算框架。从长远来看，该项目中发现的新原理和新算法将为具有韧性的软机器人在现实世界中安全可靠地运行提供方法论基础。这项研究的具体成果将是紧凑、有弹性的设备，能够进行复杂的运动，用于制造自动化、辅助技术、触觉、微创手术和非结构化环境的远程导航。拟议的研究将培养具有“T”字形技能组合的高素质人才(HQP)，将跨学科的广度与特定的专业领域相结合。该研究计划的成员将熟悉机械设计、理论和计算力学、机器学习和人工智能、物理实验和其他属于机器人学背景的技能。公平、多样性和包容性原则(EDI)将是寻找、选择和培训HQP的核心。该项目的发现将形成辅助和协作机器人设备的基础，使加拿大人受益，同时支持加拿大的机器人产业。从这个研究项目中脱颖而出的HQP将拥有在加拿大各地的学术和工业环境中创新的专业知识和领导技能。

The emergence of "soft" robots, whose bodies are made from stretchable and compliant materials, has fundamentally changed the way robotic systems are designed and constructed. Despite societal need for safe, collaborative robotic devices and a proliferation of promising design concepts, soft robots are not yet present in the human-collaborative and human-interactive applications where they might be most useful. The research objective of this proposal is to discover methodologies and principles that solve fundamental problems responsible for the gap between promise and implementation in soft robotics. The proposed research will use a hybrid approach combining continuum mechanics, machine learning, and novel shape measurement techniques to answer open research questions about design and control of soft robots with a variety of architectures deployed in a variety of environments. Soft robots are particularly difficult to design and implement because fundamental understanding of their behaviour does not yet exist. For instance, soft robots have continuously deformable bodies that cannot be modelled with a traditional approach based on rigid-body kinematics. In the short term, this research program will investigate the behaviour of the soft, stretchable and compliant materials and structures that make up the bodies of soft robots. It will follow four objectives: to measure the surface morphology of soft robots, to model soft robot stress relaxation, creep and hysteresis, develop hybridized modelling approaches that benefit from both physical principles and robot data, and investigate computational frameworks for jointly optimizing the design, sensing and control of soft robots. In the long term, the new principles and algorithms discovered in this program will form a methodological basis for resilient soft robots that perform safely and reliably in the real world. The tangible output of this research will be compact, resilient devices capable of sophisticated motions for manufacturing automation, assistive technology, haptics, minimally invasive surgery, and remote navigation of unstructured environments. The proposed research will develop Highly Qualified Personnel (HQP) with "T-shaped" skill sets combining interdisciplinary breadth with particular areas of expertise. Members of this research program will become familiar with mechanical design, theoretical and computational mechanics, machine learning and artificial intelligence, physical experimentation, and other skills that are part of the context of robotics. Principles of Equity, Diversity and Inclusion (EDI) will be central to the search for, selection of, and training of HQP. The findings of this program will form the basis of assistive and collaborative robotic devices that benefit Canadians while bolstering Canada's robotics industry. HQP emerging from this research program will have the expertise and leadership skills to innovate in academic and industrial settings across Canada.