Privacy Preserving Discrete-Time Average Consensus by Injecting Edge-based Perturbations

This paper aims to solve the privacy preserving average consensus problem for discrete-time multi-agent systems over balanced graphs. The main task is to design an algorithm that can guarantee average consensus and prevent agents' initial state values from being disclosed in the meantime. To address the problem, a novel algorithm, which adds edge-based perturbations to the system, is proposed. At the first iteration, the proposed algorithm requires each agent to generate some designed perturbation signals and add them to the edges in the network. From the second iteration, all agents simply adopt a conventional average consensus protocol. A theoretical analysis of privacy preservation properties is then presented, which shows that our algorithm can prevent both internal honest-but-curious agents and external eavesdroppers from obtaining privacy. Finally, the effectiveness of the algorithm is demonstrated by a numeric simulation.