Resilient Consensus-Based Distributed Signal Reconstruction Against Byzantine Attacks

In this paper, we propose a distributed signal reconstruction protocol based on resilient consensus framework under a Byzantine adversary model. Specifically, a compromised node possesses complete network knowledge and can deviate arbitrarily from any preset algorithm protocol or sends different values to different neighbors. Under the above attack scenario, a version of the resilient distributed algorithm is constructed based on the data decomposition mechanism, which achieves the localization and isolation of malicious attack signals, and ensures the consensus of global estimation. Moreover, sufficient conditions for regular nodes to complete the signal reconstruction are given. Finally, a simulation experiment is presented to verify the proposed technique.