Date of Publication :25th May 2018

Abstract: Several spectrum management schemes have been proposed in recent years to improve the spectrum use in cognitive radio networks. Few consider the existence of cognitive attackers who can adapt their attack strategy to the environment with a different time spectrum and the secondary consumer strategy In this article, we investigate the security mechanism when secondary users face jamming attack and offer a stochastic game environment to protect against jamming. At each stage of the players, the secondary users observe the availability of the radio spectrum, the quality of the channel and the strategy of attack through the state of the channels detected. Based on this observation, they will decide how many channels are you must reserve for the transmission of control and data messages and how to switch between different channels. By using "mini-learning" training, secondary users can gradually learn the optimal policy that maximizes the expected amount of discounted wages, defined as spectral efficiency. The proposal fixes the anti-jamming policy has shown that it achieves much better results than those achieved of myopic learning, which only maximizes the payment of each phase and a strategy of random defense since successfully assume the dynamics of the environment and the strategic behavior of cognitive aggressors

Reference :