Intelligent Covert Communications: Hiding Transmission via Learning

With growing security threats to increasingly sophisticated wireless systems, it is challenging to protect the confidential information. Although the methods of encrypting transmitted information and limiting potential wiretap channels have been developed, the confidential information remains vulnerable to interception if the adversaries can detect the transmission. To tackle this issue, covert communication presents a promising solution by hiding the transmission behavior. Nonetheless, heterogeneous network and dynamic environment pose difficulties to realize covert transmission via traditional optimization approaches. Fortunately, artificial intelligence (AI) can be leveraged to improve the covert transmission. In this article, we first provide preliminaries of covert communications. Then, we comprehensively discuss the motivation of AI for covert communications. Moreover, we analyze four AI techniques for covert transmission and then investigate a case study of intelligent covert communications. Finally, several challenges and future directions are discussed.