A Method to Estimate Low SNR Regions for Optimal Deployment of UAV Relays in Next-Generation Mobile Communications

Abstract

To optimally deploy relays in millimeter wave mobile communications, this paper proposes to predict received signal power by exploiting information on propagation paths. Within a certain distance from observation points, the proposed scheme is superior in terms of the accuracy to conventional Gaussian process (GP) based prediction, while reducing computational complexity more drastically than ray tracing. The integration of GP into the proposed scheme is also investigated, in which GP is applied to prediction points where the proposed one is ineffective. Computer simulations demonstrate that the integration scheme achieves approximate 2% improvement in the F1 score over GP.