CRI: IAD: Adaptive Manipulation in Human Populated Environments

提案编号：CNS 07-07601 PI(s): Rus, Daniela Kaelbling, Leslie P.；Lozano-Perez,托马斯;史蒂夫·g·马萨夸伊；Roy， Nicholas机构：麻省理工学院剑桥，MA 02139-4307标题：IAD：人类环境中的自适应操纵项目建议：该项目，支持机器人操纵器的现有和发展研究，旨在提供一个基础设施，教师和学生可以在机器人操纵和低成本操纵平台的设计方面协同进步。这项工作开发了新的低成本硬件设计，其特点是简单和可预测性，以及稳定和强大的规划和控制算法，这些算法将使用传感和学习来适应操纵器的不同（非理想）能力。该设备将促进不同研究项目之间的紧密集成，使算法的进步可以快速部署到新的硬件上，新的硬件功能可以推动新的算法开发。该项目解决了算法和设备级的挑战，使机器人能够自适应地与物理世界和非结构化动态环境中的人类进行交互，以应对目前缺乏严格和实用的操纵科学基础的问题。将建模、控制、规划和适应的算法解决方案与机器人操作平台相结合，预计将在几年内产生新的低成本操作平台。这些设计包括兼容的和有能力的机器人操纵器，以及模块化合作、神经模拟和水下操纵器。更广泛的影响：负担得起的，可教的，移动操作器应该使老年人和/或无行为能力的家庭助理成为可能。下一代机器人操作器可能会支持新的工厂自动化水平，其中机器人将与人类协同工作以降低制造成本。这个多元化的跨学科小组将继续培养扎实的学生，并通过波士顿科学博物馆和波士顿水族馆参与广泛的推广活动。教育外展包括机器人教育研讨会。

Proposal #: CNS 07-07601 PI(s): Rus, Daniela Kaelbling, Leslie P.; Lozano-Perez, Tomas; Massaquoi, Steve G.; Roy, Nicholas Institution: Massachusetts Institute of Technology Cambridge, MA 02139-4307 Title: IAD:Adaptive Manipulation in Human Populated Environments Project Proposed:This project, supporting existing and developing research in robotic manipulators, aims to provide an infrastructure where faculty and students can work towards synergistic advances in robot manipulation and in the design of low-cost manipulation platforms. The work develops both new, low-cost hardware designs characterized by simplicity and predictability and stable and robust planning and control algorithms that will use sensing and learning to adapt to the different (non-ideal) capabilities of manipulators. The equipment will foster tight integration between different research programs such that algorithmic advances can be rapidly deployed on new hardware, and new hardware capabilities can drive new algorithmic development. Addressing both the algorithmic and device-level challenges that enable robots to interact adaptively with the physical world and with humans in unstructured dynamic environments, the project responds to the current lack of rigorous and practical scientific foundation for manipulation. The combination of algorithmic solutions to modeling, control, planning, and adaptation, and a robotic manipulation platform is expected to lead to new low-cost manipulation platforms within a few years. The designs include compliant and capable robot manipulators, as well as modular cooperative, neuromimetic, and underwater manipulators.Broader Impacts: Affordable, teachable, mobile manipulators should enable in-home assistants for the elderly and/or incapacitated. The next-generation of robot manipulators is likely to support new levels of factory automation, in which robots will work synergistically with humans to reduce manufacturing cost. This diverse interdisciplinary group will continue the training of solid students and participate in broad outreach activities through the Boston Museum of Science and the Boston Aquarium. Education outreach includes Robotic Education Workshops.