NRI: FND: COLLAB: A Foundational Approach to Muscle Actuators that Lowers Barriers to Muscle-Powered Robotics Research

这个国家机器人计划（NRI）项目将通过开发一种使用仿生肌肉致动器的肌肉动力机器的基本方法来促进科学的进步和国家的繁荣。仿生肌肉致动器，或机器人肌肉，是一种密切模仿自然界生物肌肉特性的致动器。不同的肌肉致动器在致动器的选择和设计、建模和控制行为和特性方面都面临着共同的挑战。该奖项支持识别和统一建模、控制和设计策略，这将降低肌肉动力机器人研究的障碍。为设计师降低机器人肌肉的风险，一个重要的努力是创建一个机器人肌肉工具包，一个可访问的软件和硬件工具包，用于设计肌肉驱动的机器，并使这些技术的发展和扩散在大范围内民主化。设计挑战和竞赛将激发所有年龄和经历的对机器人肌肉感兴趣的人的兴趣和社区建设，从K-12到大学生和研究人员。计划在圣地亚哥和大波士顿地区社区安排外展活动，以激发未被充分代表的少数民族对机器人的兴趣。为了在所谓的性能和配置空间中锚定建模、控制和设计策略，将开发一种新的、数据驱动的过程和结构，用于参数化定义肌肉驱动的一组常见仿生特征。这些方法可以跨肌肉类型和推广超越目前可用的肌肉致动器。有一种基于多维性能标准的肌肉选择策略，该策略结合了经验数据和物理模型。设计策略的肌肉配置放大沿着这些性能尺寸，跨肌肉推广，将进行调查。新技术将能够很容易地加入到这个框架中，在不复制金标准执行器的情况下进行定量比较研究。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。

This National Robotics Initiative (NRI) project will promote the progress of science and advance the national prosperity by developing a foundational approach towards muscle-powered machines that move using biomimetic muscle actuators. Biomimetic muscle actuators, or robot muscles, are actuators that closely mimic the properties of biological muscles in nature. A common set of challenges in actuator selection and design, and modeling and controlling behaviors and properties emerge among different muscle actuators. This award supports identifying and unifying modeling, control, and design strategies that would lower the barrier to muscle-powered robotics research. A significant effort to de-risk robot muscles for designers is the creation of a robot muscles toolkit, an accessible software and hardware toolkit for designing muscle-powered machines and for democratizing the development and proliferation of these technologies at a broad scale. Design challenges and competitions will spur the interests and community building for those interested in robot muscles across all ages and experiences, from K-12 to college students and researchers. Local arrangements for outreach in the San Diego and Greater Boston Area communities are planned for stirring interest in robotics for underrepresented minorities.A novel, data-driven process and structure for parameterizing a common set of biomimetic features that define muscle actuation will be developed in order to anchor modeling, control, and design strategies across a so-called performance and configuration space. These methods can be employed across muscle types and generalize beyond currently available muscle actuators. There is a strategy for muscle selection over multi-dimensional performance criteria that incorporates both empirical data and physics models. Design strategies for muscle configurations for amplification along these performance dimensions, generalizable across muscles, will be investigated. New technologies will be able to be easily added into this framework to perform quantitative comparison studies without replication of gold standard actuators.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.