Multi-Layer RIS-Assisted Anti-Jamming Communications: A Hierarchical Game Learning Approach

This letter investigates a multi-layer reconfigurable intelligent surface (RIS)-transmitter assisted anti-jamming communications from a Stackelberg game perspective, where the competitive interaction between the base station (BS) and the jammer is established. Specifically, with the assistance of a multi-layer RIS-transmitter, the BS who acts as the leader wishes to transmit the information to the user with maximum utility, while the smart jammer who acts as the follower attempts to interrupt the legitimate transmission at low cost. Moreover, the practical discrete RIS's phase and the assumption that neither the BS nor the jammer can obtain perfect channel state information (CSI) and the competitor's strategies are considered. To handle the intractable game design, a hierarchical game learning approach is proposed to achieve the Stackelberg equilibrium (SE). To elaborate, after leveraging the discretization method and Cauchy-Schwarz inequality to handle the CSI imperfection and unknown strategies, a Noisy Dueling Double Deep Q-network (Noisy-D3QN) is developed to solve each sub-game. Finally, numerical simulations demonstrate the superiority of our proposed scheme in comparison to the existing schemes.