A Transfer Games Actor–Critic Learning Framework for Anti-Jamming in Multi-Channel Cognitive Radio Networks

A cognitive radio network (CRN) is a novel solution that promises to solve the spectrum scarcity problem and enhance utilization. However, unsecured CRN can easily be manipulated in order attack legacy users on communication channel. As result, network's performance significantly degrades. Therefore, channel security an important issue needs addressed CRN. In this work, we focus improving of multi-channel CRN, while various jammers try access channels interest prevent SUs from using them. By game-theoretic concepts by defining states, actions, players' rewards, propose game-based schemes find best for secondary (SUs) avoid jammer's attacks channels. Accordingly, finding optimal maximize long-term reward SU where are not used primary (PUs) jammed attackers. addition, idea transfer learning might applied under consideration, thus, Game-Actor-Critic (TGACT) scheme proposed, which uses transferred knowledge double-game period accelerate process provide improvement selection. Finally, proposed simulated with different configurations. The simulation results show quite resistant jammer attacks, achieve better compared other selection schemes.