Augmented Consensus Algorithm for Discrete-time Dynamical Systems

Abstract

We propose a novel state estimation algorithm for consensus dynamics subject to measurement error. We first demonstrate that with properly tuned parameters, our algorithm attains the same equilibrium value that would be attained using the traditional algorithm based on local state feedback (nominal consensus). We then show that our approach improves consensus performance in a particular class of problems by reducing the state error (i.e., the difference between the agent states and the consensus value). A numerical example compares the performance of the distributed algorithm we propose to that of the traditional local feedback scheme. The results show that the proposed algorithm significantly reduces the state error.