CISE Postdoctoral Research Associates in Experimental Computer Science: Challenges in Micromanipulation: Massively Parallel MEMS Algorithms and Systems

9901407唐纳德，布鲁斯·兰德尔·达特茅斯学院教授实验计算机科学博士后研究助理微操作：大规模并行MEMS算法和系统随着MEMS(微电子机械系统)制造技术的改进增加了微型机械的可用性和多样性，工程师们正在定义他们可能面临的越来越多的任务。利用MEMS单元的大型协调系统执行任务的想法推动了设计和控制这些设备组的自动化、算法方法的发展。研究的重点是使MEMS能够执行分布式操作任务的算法和系统，以挑战大规模并行微执行器阵列的设计、控制和编程。用于大规模并行分布式操作的新算法和MEMS正在开发中。这需要设计、制造和控制新的MEMS器件，以及用于任务级操作的新控制系统、算法和软件的设计、构造和编程。正在开发用于零件定位、分离、分类和摆正的算法和系统。本研究的主要重点是扩展可编程矢量场理论以利用传感器反馈，并研究纤毛阵列的闭环控制算法、在线算法以及分布式局部控制。目前的准静态矢量场模型正在扩展为全动态模型。还在研究控制不确定性和误差对开发在不确定情况下有效的运动规划范例的影响。这应该导致规划和操作的更现实和更健壮的几何理论，该理论正在通过在物理MEMS阵列设备上实现算法进行测试。最后，正在探索大规模并行微型机器人系统在生物医学系统和人类增强中的应用。

9901407 Donald, Bruce Randall Dartmouth College CISE Postdoctoral Research Associates in Experimental Computer Science Challenges in Micromanipulation: Massively Parallel MEMS Algorithms and SystemsAs improvements in fabrication technology for MEMS (MicroElectoMechanical Systems) increase the availability and diversity of micromachines, engineers are defining a growing number of tasks to which they may be put. The idea of carrying out tasks using large coordinated systems of MEMS units motivates the development of automated, algorithmic methods for designing and controlling these groups of devices. The research focuses on algorithms and systems that enable MEMS to perform distributed manipulation tasks, taking on the challenge of design, control, and programming of massively-parallel arrays of microactuators.New algorithms and MEMS for massively parallel distributed manipulation are being developed. This entails the design, fabrication, and control of new MEMS devices, together with the design, construction, and programming of new control systems, algorithms, and software for task-level manipulation. Algorithms and systems are being developed for part orientation, singulation, sorting, and posing. The primary focus of this research is to extend the theory of programmable vector fields to exploit sensor feedback, and investigate closed loop control algorithms, on-line algorithms, and distributed, local control of ciliary arrays. A current quasi-static vector field model is being extended to a full dynamic model. The effect of control uncertainty and error to develop motion planning paradigms that are effective under uncertainty are also being investigated. This should result in a more realistic and robust geometric theory of planning and manipulation, which is being tested by implementing algorithms on physical MEMS array devices. Finally, the application of massively-parallel micro-robotic systems to biomedical systems and human augmentation is being explored.