A novel cognitive anti-jamming stochastic game

This paper proposes a new cognitive anti-jamming stochastic game model for multi-agent environments in which each wideband autonomous cognitive radio (WACR) attempts to predict and evade the transmissions of other radios as well as a dynamic jammer signal. The cognitive framework is divided into two operations: sensing and transmission. Each is helped by its own learning algorithm based on Q-learning. It is shown, through both analysis and simulations, that the proposed cognitive anti-jamming technique has low computational complexity and significantly outperforms non-cognitive sub-band selection policy.