基于深度强化学习的多无人机辅助数据采集方法OA

A complex multi-UAVs-assisted WSN data collection scenario was considered,and a deep reinforcement learning-based mult-UAVs-assisted data collection method was proposed to minimize the total energy consumption of the UAVs.Firstly,an improved K-means++-based 3D clustering algorithm was intoduced which divides the SNs into cluster members(CMs),for data sensing,and cluster heads(CHs),for data collecting and uploading.Subsequently,since UAVs need to collect data from the associated CHs at DCPs,a deep deterministic policy gradient-based data collection points(DCPs)setting algorithm was proposed to minimize the total hovering energy consumption of UAVs.Finally,an improved simulated annealing-based UAVs joint path planning algorithm was proposed to determine the best paths with the lowest flight energy consumption for UAVs data collection.Simulation results show that the proposed method could achieve the minimization of total energy consumption for the UAV-assisted data collection.