Secure UAV Communication Under Cooperative Adaptive Eavesdroppers with Incomplete Information.

The adaptive eavesdroppers (AEs) bring more serious attacks and exacerbate the uncertainty of wireless communication due to their adaptivity of acting as either passive eavesdroppers or active jammer by learning the communication environment. In this paper, we investigate secure unmanned aerial vehicle communication under multiple cooperative AEs with incomplete information. A hierarchical game framework based on observation accuracies and channel estimation coefficients is proposed, in which a cooperative attack game (CAG) is modeled to get the optimal joint attack action of multiple AEs by proving CAG is an exact potential game and a Stackelberg game is applied to derive the aerial base station's (ABS's) optimal position according to AEs’ strategies. Then, a hierarchical learning algorithm is proposed to search the ABS's optimal position for enhancing the secrecy rate under the cooperative AEs. Finally, the simulation results validate the effectiveness of the proposed hierarchical game framework.