Trajectory Optimization and Power Allocation Scheme Based on DRL in Energy Efficient UAV-Aided Communication Networks

With flexibility, convenience and mobility, unmanned aerial vehicles (UAVS) can provide wireless communication networks with lower costs, easier deployment, higher network scalability and larger coverage. This paper proposes the deep deterministic policy gradient algorithm to jointly optimize the power allocation and flight trajectory of UAV with constrained effective energy to maximize the downlink throughput to ground users. To validate the proposed algorithm, we compare with the random algorithm, Q-learning algorithm and deep Q network algorithm. The simulation results show that the proposed algorithm can effectively improve the communication quality and significantly extend the service time of UAV. In addition, the downlink throughput increases with the number of ground users.