Optimal Adaptation of 3D Beamformers in UAV Networks

Abstract

The optimal beamformer weight update timing is derived for 3D beamforming between Unmanned Aerial Vehicles (UAVs). The use of optimal update period $(\triangle t^{*})$ ensures that beamforming gain $(G_{tx})$ drops below a required threshold least frequently. When the exact flight path of the UAV receiver (UAV-RX) is available, the proposed optimization problem calculates optimal $\triangle t^{*}$ exactly. It is shown that when the antenna aperture is fixed, the optimal $\triangle t^{*}$ monotonically decreases as the operating frequency increases. As a result, the beamformer needs to be updated more often when using higher frequencies. However, it is shown that the fractional overhead relative to the large bandwidths available in mmWave spectrum can still be lower, justifying the use of mmWave frequencies in UAV-to-UAV communications. When exact flight path knowledge is not available, the proposed algorithm incorporates UAV-RX location predictions to update the beamforming weights in a timely manner. The proposed method was simulated in a UAV communication system and the performance of the system is analyzed in terms of the UAV-RX location prediction error. The findings indicate that when there are location prediction errors, incorporating a gain margin to the minimum gain threshold further enhances the algorithm’s performance by approximately 10-20%.