Resilient Distributed Parameter Estimation for Sensor Networks Against Sparse-Varying Attacks

This article investigates resilient distributed parameter estimation (RDPE) against sensor attacks with variable sparsity. First, the fixed sparsity of attacks is relaxed to variable sparsity over a specific time scale, with a sparse-varying sensor attack model proposed. Then, to counteract such attacks, an improved resilient distributed parameter observer is constructed by following the concept of sliding windows. Without altering the redundancy condition of sensor measurements, a sufficient condition to resist the sparsity-varying attacks is presented. Furthermore, under the assumption of accessible global historical attack detection information, the performance of RDPE is improved. Finally, some numerical simulation examples are presented to demonstrate the effectiveness of the proposed design.