REU: CAREER: Physics-Based Robotic Manipulation Planning and Control

职业：基于物理的机器人操纵规划与控制这是四年连续奖的第一年。本研究的目标是开发利用任务机制的机器人操作规划器和控制器。摩擦力、重力和动力的有效利用可以代替额外的致动器，从而开发出廉价、极简的机器人。这些想法将应用于以下问题：(1)工业零件馈送，其中操作规划器将允许一个简单的双关节机器人，它可以推动和滚动零件，以定位和定向输送机上的三维零件，以及(2)人机协作操作的运动规划。协作机器人是一种被动的可编程约束机器，它通过建立虚拟引导面来辅助人类完成操作任务。问题在于找到“符合人体工程学的”导向——在所有将物体引向目标配置的导向中，哪一个导向对人类来说是最容易的？它将在汽车组装中重复任务的背景下进行研究。这项面向可访问的极简机器人的工作将由两门新课程的开发来补充。第一个是一年级/二年级的机电一体化设计课程，将利用目前正在开发的机电一体化设计实验室。本课程和实验室将允许本科生在职业生涯的早期参与研究和独立项目。第二门课程是关于机器人操作和工业自动化的力学和算法的高级/研究生课程。

IIS-9875469Lynch, KevinUniversity of Rochester$50,000.00CAREER: Physics-Based Robotic Manipulation Planning and ControlThis is the first year flitiding of a four-year continuing award. The goal of this research is to develop robot manipulation planners and controllers which exploit task mechanics. The effective use of friction, gravity, and dynamic forces can substitute for extra actuators, allowing the development of inexpensive, minimalist robots. The ideas will be applied to problems in (1) industrial parts feeding, where the manipulation planner will allow a simple two-joint robot which can push and roll parts to position and orient three-dimensional parts on a conveyor, and (2) motion planning for human-robot collaborative manipulation. A collaborative robot is a passive programmable constraint machine which sets up virtual guiding surfaces to assist a human in manipulation tasks. The problem is to find "ergonomically correct" guides --- of all the guides that lead the object to the goal configuration, which makes the task easiest for the human? It will be investigated in the context of repetitive tasks in auto assembly. This work toward accessible, minimalist robotics will be complemented by the development of two new courses. The first is a fresbman/sophomore-level mechatronics design course which will make use of the mechatronics design lab currently under development. This course and the lab will allow undergraduates to get involved in research and independent projects early in their careers. The second course is a senior/graduate-level course on mechanics and algorithms for robotic manipulation and industrial automation.