Timely and covert communications under deep learning-based eavesdropping and jamming effects

This paper explores the concept of timeliness in covert communications when faced with eavesdropping and jamming. Time-sensitive information is to be transmitted through a wireless channel between a transmitter and a receiver, while an adversary seeks to detect the communication attempts with a deep learning-based classifier (using feedforward or convolutional neural networks). The adversary jams any detected transmission, subject to an average power budget. When the transmit power is set at a high level, the outage probability decreases, resulting in more reliable communication. However, this also increases the accuracy of the adversary's detection, making it more likely for the jammer to successfully identify and jam the communication. On the other hand, using a low transmit power leads to a higher outage probability for communication but decreases the accuracy of the adversary in detecting and disrupting a transmission. The trade-off between reliability, timeliness, and stealthiness in wireless communications is analyzed in this paper by characterizing the Age of Information and its behavior under the influence of eavesdropping and jamming effects. Results indicate novel operation modes for timely and covert communications under eavesdropping and jamming effects.