CAREER: Dynamic Modeling and Fabrication of Compliant Material Systems for On-Demand Specialist Robots

这笔学院早期职业发展(Career)补助金将使开发具有成本效益的“专业”机器人成为可能，这些机器人可以在几个小时内由非专家制作原型。该项目的目标是使机器人对于机器人学的新手以及在工业、教育和学术研究中的应用变得更加无处不在、更容易接触和可调。要实现这一目标，需要转向更实惠的材料、可获得的制造策略和辅助设计软件，该软件考虑到经济高效的电机和材料的特殊动态，以及特定应用的需求。该项目的结果将影响需要专业化的领域，如老年人辅助机器人、定制农业应用和智能城市的垃圾拾取。通过这个项目开发的机器人也将成为一个负担得起的起点，让孩子们参加由当地以年轻人为重点的非营利组织组织的课外机器人比赛。此外，为该项目开发的模型和模板将被整合到大学级别的机器人课程中，允许学生在学习早期更深入地探索高级机器人主题。该项目的基本研究贡献在于考虑并将符合要求的材料系统集成到支持动态机器人系统的专业化和优化的统一设计框架中。研究计划包括为材料系统的非线性力学开发简化模型，以允许更负担得起的机器人设计，并考虑到在模拟中的遵从性，以定制、建模和优化其性能。这种方法将研究机器学习技术，以自动调整参数、模板驱动的设计，以平衡速度、有效负载和效率等相互竞争的性能权衡。该项目的研究目标包括法规遵从性的表示、法规遵从性的使用和理解、法规遵从性系统的优化，以及通过原型、实验验证和范例用例来验证框架。该方法的一般性将通过对两足机器人的几个子系统进行建模和优化来展示。该奖项反映了NSF的法定使命，并已通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

This Faculty Early Career Development (CAREER) grant will make it possible to develop cost-effective, “specialist” robots that can be prototyped by a non-expert in a matter of hours. The goal of this project is to make robots more ubiquitous, accessible, and tunable for newcomers to robotics and for applications in industry, education, and academic research. Achieving this goal requires a shift towards more affordable materials, accessible fabrication strategies, and assistive design software that considers the particular dynamics of cost-efficient motors and materials, as well as the needs of specific applications. The results of this project will impact fields in which specialization is desirable, such as assistive robotics for the elderly, custom agricultural applications, and trash pickup in smart cities. The robots developed through this project will also serve as an affordable starting point for children to compete in after-school robotics competitions organized by local, youth-focused nonprofits. In addition, the models and templates developed for this project will be integrated into college-level robotics curricula, permitting students to venture deeper into advanced robotics topics earlier in their studies.The fundamental research contribution of this project lies within the consideration and integration of compliant material systems into a unified design framework that supports the specialization and optimization of dynamical robotic systems. The research plan involves developing reduced models for the nonlinear mechanics of material systems that permit more affordable robot designs, and considering that compliance in simulations to customize, model, and optimize their performance. This approach will investigate machine learning techniques to automatically tune parametric, template-driven designs in a way that balances competing performance trade-offs like speed, payload, and efficiency. The project’s research objectives include the representation of compliance, the utilization and understanding of compliance, the optimization of compliant systems, and validation of the framework through prototyping, experimental validation, and exemplar use-cases. The generality of the approach will be demonstrated by modeling and optimizing several subsystems of a bipedal robot.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.