A Discrete Time Consensus Approach for Fault Detection and Recovery in Unreliable Networks

In this paper we address the problem of the fault detection and recovery in networks of agents with discrete time dynamics. In particular, we apply a fault detection and recovery approach, proposed for the standard linear consensus protocol, to a consensus algorithm that has been previously presented by the authors and is based on a triangular splitting of the iteration matrix of the standard consensus algorithm. Moreover, we show that for a particular topology of the communication graph the fault recovery strategy can be implemented by the agents in a fully decentralized and autonomous approach. An example describes how in the considered framework the fault recovery is implemented by each agent.