Detection of the Covert Channel in The Presence of The Channel Estimation Error

In this article we analyze a covert channel models based on superposition of signals. Based on several general models of hidden transmission we propose optimal modulation schemes for covert modulation. Proposed solution are analyzed from the point of view of detection performance. We focus on the problem covert channel demodulation in the presence of Automatic Gain Control errors caused by hardware imperfections or Rayleigh fading channel. This leads us to a improved method based on deep learning (DL) scheme. Finally simulation results are provided to evaluate proposed model with regard to different training sets of Neutral Network.