Reinforcement Learning-Based Detection for State Estimation Under False Data Injection

We consider the problem of network security under false data injection attacks over wireless sensor networks.To resist the attacks which can inject false data into communication channels according to a certain probability, we formulate the online attack detection problem as a partially observable Markov decision process problem and design a detector for each sensor based on the framework of model-free reinforcement learning. By numerical simulations, we illustrate the effectiveness of the proposed reinforcement learning algorithm and show the performance of the proposed detector compared with the typical detector in the existing works.