Optimal False Data Injection Attacks in a Wireless Sensor Network

A false data injection attack in a wireless sensor network for cyber physical systems is designed. System state estimation is analyzed by remote distributed consensus estimators, and the Kullback-Leibler divergence is utilized as an indicator of the attack's stealthiness. The attack studied in this paper is injected into the wireless network channel between sensors and is Gaussian distributed with an arbitrary mean. Based on the relationship between the system performance under attack and stealthiness, a constrained optimization problem is given, and the optimal attack strategy can be calculated by solving the problem through the Lagrange multiplier method. Then based on theoretical research, the algorithm for generating the attack sequence is summarized. Finally, the theoretical analysis is justified through a numerical simulation example.