Mitigation of Byzantine attacks for target location estimation in wireless sensor networks

In this paper, we propose a Byzantine identification scheme for a target localization scenario in wireless sensor networks (WSN) with binary quantized observations. We first start with identical quantizers at the local sensors. We show with simulations that the proposed scheme identifies `almost' all the Byzantines present in the network. Then, in order to improve the performance of the network, we propose a dynamic non-identical quantizer design scheme at the local sensors. We show that our proposed design not only enhances the system performance by reducing the estimation error of the target location but also makes it robust to the presence of Byzantines by making them ineffective when combined with our Byzantine identification scheme.