Optimal communication strategies for robotic networks

Abstract

This dissertation presents a method to control information flow through a communication network consisting of robots. An emphasis is put on constructing routing and limiting controllers that integrate the requirements of robotic networks with those of traditional communication networks. The routing controller determines optimal paths for information flow between source and destination nodes while the limiting controller controls the rate at which new traffic enters the network. The networks are modelled with flow models while the controllers are constructed using finite time optimal control theory as well as receding horizon controllers. These techniques are applied to both broadcast and point to point networks.