DeepFake: Deep Dueling-Based Deception Strategy to Defeat Reactive Jammers

In this paper, we introduce DeepFake, a novel deep reinforcement learning-based deception strategy to deal with reactive jamming attacks. particular, for smart and attack, the jammer is able sense channel attack if it detects communications from legitimate transmitter. To such attacks, propose an intelligent which allows transmitter transmit “fake” signals attract jammer. Then, attacks channel, can leverage strong data by using ambient backscatter communication technology or harvest energy future use. By doing so, not only undermine ability of jammer, but also utilize improve system performance. effectively learn adapt dynamic uncertainty develop learning algorithm dueling neural network architecture obtain optimal policy thousand times faster than those conventional algorithms. Extensive simulation results reveal that our proposed DeepFake framework superior other anti-jamming strategies in terms throughput, packet loss, rate.