Trajectory planning for UAV-enabled multi-user systems in presence of eavesdroppers

Unmanned-aerial-vehicle (UAV)-enabled wireless communication is a promising paradigm in extending signal coverage and enhancing system capacity due to its superiority in low cost and flexible deployment. However, the performance of such systems heavily depends on the UAV flight trajectories, especially in the presence of eavesdroppers. In fact, for such scenarios, UAV trajectories can be optimized to enhance the security of communications with grounded legitimate users. In this paper, for the scenarios with coordinated multi-point receptions (CoMP) and multiple colluding eavesdroppers, a novel scheme on UAV trajectory is proposed to maximize the achievable average secrecy rate (ASR) based on reinforcement learning techniques by action oriented optimization on reward weights. Numerical results show that the proposed solution can obtain higher average secrecy rate compared to the existing benchmark schemes.