NRI: Small: Dexterous Manipulation with Underactuated Hands: Strategies, Control Primitives, and Design for Open-Source Hardware

NRI：小：用欠驱动的手进行灵巧操纵：策略、控制原语和开源硬件设计

• 批准号: 1317976
• 负责人: Aaron Dollar
• 金额: $ 119.76万
• 依托单位: Yale University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2013
• 资助国家: 美国
• 起止时间: 2013-09-15 至 2018-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1317976&HistoricalAwards=false
• 关键词: NRI; Small; Dexterous; Manipulation; Underactuated

The proposed work will contribute to the fundamental understanding of dexterous manipulation, especially in how low-dimensional hands that have traditionally been used for pure grasping tasks can be leveraged for excellent manipulation capabilities. The PI team will investigate how the passive mechanics of compliant and underactuated hands can be utilized for manipulation strategies include vibrational stick-slip, equilibrium point manipulation, and finger gaiting, as well as develop low-level planning and control schemes to implement those for a wide range of manipulation movements. This work will lead to the development of general-purpose robot hands having low-dimensional actuation and sensing. Furthermore, the focus on design for open-source dissemination through rapid prototyping techniques will result in novel fabrication strategies in furtherance of fast innovation in a range of fields.A key thread of the proposed work is that simple hand designs and control schemes will be made freely available through an open-source repository. The hands will be able to be easily and inexpensively fabricated with rapid prototyping techniques, lower the entry barrier from expense hand hardware and encouraging continual design improvements. Additionally, the team will kick-start a user community around those hands through an on-site tutorial at Yale for approximately 20 students and postdocs from top manipulation research groups. For the education and outreach component of this proposal, a major thrust involves enhancing RoboticsCourseWare.org, the PI's open-access repository for robotics teaching materials, in order to highlight open-source and other inexpensive hardware platforms for educational purposes. A strong focus on involving a diverse group of undergraduates and high school students in the research plan is also highlighted.

拟议的工作将有助于灵巧操作的基本理解，特别是在如何低维的手，传统上用于纯抓取任务可以利用优秀的操作能力。PI团队将研究如何将顺应性和欠驱动手部的被动力学用于操作策略，包括振动粘滑，平衡点操作和手指步态，以及开发低层次的规划和控制方案，以实现广泛的操作运动。这项工作将导致具有低维驱动和传感的通用机器人手的发展。此外，通过快速原型技术进行开源传播的设计将导致新的制造策略，以促进一系列领域的快速创新。拟议工作的一个关键线程是，简单的手设计和控制方案将通过开源存储库免费提供。手将能够容易和廉价地制造快速原型技术，降低进入门槛，从昂贵的手硬件和鼓励不断的设计改进。此外，该团队还将通过在耶鲁大学为来自顶级操纵研究小组的大约20名学生和博士后提供现场教程，围绕这些手启动一个用户社区。对于该提案的教育和推广部分，一个主要的推动力涉及加强RoboticsCourseWare.org，PI的机器人教学材料开放访问存储库，以突出开源和其他廉价的硬件平台用于教育目的。还强调了在研究计划中涉及不同群体的本科生和高中生的强烈关注。