Average Consensus with Asynchronous Updates and Unreliable Communication

In this work we introduce an algorithm for distributed average consensus which is able to deal with asynchronous and unreliable communication systems. It is inspired by two algorithms for average consensus already present in the literature, one which deals with asynchronous but reliable communication and the other which deals with unreliable but synchronous communication. We show that the proposed algorithm is exponentially convergent under mild assumptions regarding the nodes update frequency and the link failures. The theoretical results are complemented with numerical simulations.