3D Beam Tracking for Cellular-Connected UAV

Three-dimensional (3D) beamforming is an effective technique for performance enhancement in cellular-connected unmanned aerial vehicle (UAV) communications. Specifically, ground base stations (GBSs) equipped with full-dimensional (FD) antenna arrays are able to perform fine-tuned 3D transmit/receive beamforming with enhanced interference mitigation for communicating with UAVs. To fully reap the 3D beamforming gain with a flying UAV, a practically important problem is to efficiently track the beamforming direction over time at its serving GBS. By exploiting the line-of-sight (LoS) dominant path of the UAV-GBS channel, we propose in this letter a new beam tracking algorithm with dynamic pilot insertion based on a closed-loop feedback control. The key of the proposed algorithm lies in estimating the UAV's angular speeds based on the received pilots and thereby predicting its azimuth and elevation angles in subsequent time blocks without the need of sending new pilots, thus greatly saving the pilot overhead. Simulation results show that the proposed algorithm outperforms the conventional periodical pilot transmission based beam tracking scheme in terms of both beamforming performance and pilot efficiency.