Cross-Layer Anti-Jamming Scheme: A Hierarchical Learning Approach

This paper investigates the cross-layer optimization for anti-jamming in the network and MAC layers, in which the jammer can adjust the jamming policies to maximize the jamming effectiveness. The joint problem of routing selection, channel allocation, and power control is formulated as a Stackelberg game. The jammer leads the game by choosing the optimal jamming power and channels. The user follows by selecting the optimal nodes and corresponding channels, and adjusts its transmitting power to meet the communication requirement. Then, based on Q-learning, a cross-layer anti-jamming learning algorithm is proposed to obtain the Stackelberg equilibrium. Finally, simulation results are presented to verify the effectiveness of the proposed algorithm.