Distributed Sequential Hypothesis Testing With Byzantine Sensors

This paper considers the problem of sequential binary hypothesis testing based on observations from a network $m$ sensors where subset is compromised by malicious adversary. The asymptotic average sample number required to reach certain level error probability selected as performance metric system. We propose an asymptotically optimal voting algorithm for sensor with fusion center and generalize it fully-distributed networks, stays under weak assumption that connected. Moreover, we prove both proposed algorithms are in presence Byzantine sensors, sense each them forms Nash equilibrium worst-case attack (flip-attack). Compared existing distributed detection strategies, scheme has low message complexity, which independent number, taking advantage sparsity votes. results corroborated numerical simulations.