Title: Distributed Decentralized Estimation
Speaker: Maksim Subbotin (UCSB)

Abstract
We propose a solution to the problem of synthesizing distributed decentralized estimators for a formation of agents. The collected dynamics of the formation are modeled by a discrete linear time-invariant system. In the considered estimation structure each agent of the formation carries an estimate of the entire formation state. Agents of the formation can communicate between each other through digital unidirectional links modeled with two-state Markov chains which results in a stochastic communication topology. The design procedure is based on convex optimization problem formulated with linear matrix inequalities.