Low Cost Anti-Blockage Uav 3D Trajectory Design in Millimeter-Wave Communication Networks

When acting as the dynamic aerial millimeter-wave (mmWave) base station, unmanned aerial vehicles (UAVs) can provide ubiquitous service to ground moving users (UEs). In this paper, we propose a low cost anti-blockage UAV 3D trajectory design algorithm that enables real-time blockage prediction and trajectory optimization without detailed building geometric information or additional equipment, with the aim of maximizing the throughput of UEs. Specifically, the UAV is conducted to collect UE locations and corresponding blockage status data and update the blockage prediction model periodically. A joint optimization algorithm is designed, integrating the soft actorcritic (SAC) method with the blockage prediction model. The SAC based algorithm alternates between blockage prediction and position optimization, effectively planning the UAV 3D path to improve UE throughput. This ensures that the UAV's position optimization aligns more closely with the actual blockage environment, thereby significantly enhancing the system's anti-blockage capabilities. The results of numerical simulation demonstrate that the proposed algorithm can considerably improve the accuracy of blockage prediction and throughput of UEs.