Distributed Bearing Vector Estimation in Multi-agent Networks and Its Applications

Abstract

In the field of control engineering, formation control and network localization have been in the limelight of interest owing to their usefulness in controlling the individual agents and estimating the states of agents in multi-agent networks. The formation control and the network localization are categorized into several schemes according to types of sensed and controlled/estimated variables. Among them, there are angle-based formation control and network localization schemes using angle-based variables, such as bearing and subtended angle between the agents. Most of existing angle-based formation control and network localization schemes depend on the bearing measurements. If subtended angle measurements can be transformed into bearing vectors, the angle-based formation control or network localization without the bearing measurements can be achieved.

In this dissertation, we mainly focus on the problem of estimating bearing vectors between the agents in a two dimensional multi-agent network based on subtended angle measurements. We propose and analyze a distributed bearing vector estimation method via orientation estimation of virtual agents to solve the problem. We also provide a necessary and sufficient condition for exponential convergence of the estimated bearing vectors to the true bearing vectors.

This dissertation studies to apply the proposed bearing vector estimation to angle-based formation control and network localization. First, we apply the bearing vector estimation to a bearing-based formation control using subtended angle-only measurements. A distributed formation control method is proposed and analyzed. Next, we deal with a subtended angle-based network localization. A network construction algorithm is proposed and an analysis of the localizability of the network constructed by it is provided. We propose and analyze a distributed network localization method based on the bearing vector estimation.