A Multi-Domain Anti-Jamming Scheme Based on Bayesian Stackelberg Game With Imperfect Information

To deal with the smart jammer which can sense the legitimate communication and adjust its jamming policy, an anti-jamming game scheme is proposed in this paper to jointly optimize the transmission power and frequency hopping period. Considering that both parties in the game can only get the other's channel gains in probabilistic form, the interaction between the legitimate transmitter and the jammer is modeled as Bayesian Stackelberg game, where the legitimate transmitter is set as the leader while the jammer acts as the follower. The transmitter and the jammer determine their optimal strategies in power domain and time domain to maximize their own utilities which are formulated based on the spectrum efficiency. Besides, the imperfect information including observation errors and the bounded rationality of the jammer is considered when formulating the utility functions. By using backward induction, the optimal solutions in time-power domain are obtained for the transmitter and the jammer. Simulation results show that the proposed multi-domain game scheme outperforms the single-domain game schemes and the multi-domain random scheme. Moreover, the impact of imperfect information is discussed through simulations.