Energy-Efficient Secure UAV-Enabled Wireless Networks against Flying and Ground Eavesdroppers

Unmanned aerial vehicles (UAVs) are easy to be wiretapped by eavesdroppers as aerial base stations, and the UAV jammers can be used as an aid to cause interference to the eavesdroppers. In this paper, we aim to maximize the minimum secrecy energy efficiency (MSEE) of the UAV-assisted secure communication system. Ground eavesdroppers and UAV eavesdroppers are considered simultaneously in this system. To tackle the problem, we propose an algorithm based on multi-agent deep deterministic policy gradient (MADDPG), which jointly optimizes the trajectories and transmit power of UAVs under the constraints of maximum velocity and maximum transmit power. The architecture of centralized training and decentralized execution enables distributed cooperation among multiple UAVs, allowing for better strategies to be trained. Simulation results show that the proposed method achieves better convergence and higher MSEE performance than the benchmark schemes.