Distributed control of automonous agents

主要的兴趣是与分布式机器人网络，非常简单的机器人有一些集体的能力相关的理论问题。建议继续以前的研究，针对学生的兴趣和能力。提出了两个特别的领域：1）研究机器人交会问题，考虑更现实的方面：图像捕获的时间延迟，相机分辨率的误差，采样数据控制器，机器人之间的处理器的非同步性。具有有限可见性的独轮车的会合问题是一个完全开放的问题，将成为一个很好的博士课题。2)尝试将曲线缩短理论应用于机器人队形演化。有限数量机器人的编队可以用多边形来描述。我们的目标是创建一个分布式控制策略的多边形，表现出属性类似于盖奇-汉密尔顿-格雷森理论的曲线在平面上。

The primary interest is with theoretical problems associated with distributed robot networks where very simple robots have some collective capability. It is proposed to continue previous research, tailored to student interest and ability. Two particular areas are proposed: 1) To study the robot rendezvous problem with more realistic aspects taken into account:  time delay for image capture, errors in camera resolution, sampled data controllers, non-synchroneity in processors from robot to robot. The rendezvous problem for unicycles with limited visibility is a completely open problem and would make an excellent PhD topic. 2) To try to adapt curve-shortening theory to robot formation evolution. The formation of a finite number of robots can be described by a polygon. The goal is  to create a distributed control strategy for polygons that exhibits attributes similar to Gage-Hamilton-Grayson theory for curves in the plane.