Privacy-Preserving Push-Sum Average Consensus Via State Decomposition

Average consensus is extensively used in distributed networks for computation and control, where all the agents constantly communicate with each other update their states order to reach an agreement. Under a general average algorithm, information exchanged through wireless or wired communication could lead disclosure of sensitive private information. In this paper, we propose privacy-preserving push-sum approach directed that can protect privacy while achieving simultaneously. Each node decomposes its initial state arbitrarily into two substates, equals state, guaranteeing agent's will converge accurate consensus. Only one substate by neighbours over time, reserved. That say, only would be visible adversary, preventing from leakage. Different existing state-decomposition which applies undirected graphs, our proposed applicable strongly connected digraphs. addition, direct contrast offset-adding based vulnerable external eavesdropper, ensure against both honest-but-curious eavesdropper. A numerical simulation provided illustrate effectiveness approach.