LAUSR

Unmanned aerial vehicle (UAV) networks hold tremendous potential to provide on-demand wireless services for hotspots in city scenario. However, densely located buildings will cause the spatial constraints for UAV distribution and affect the network connectivity. This letter proposes a tractable model of UAV networks with 3D blockage effects and analyzes system coverage probability via stochastic geometry, where buildings are distributed according to Poisson point process (PPP) with heights ( $H_{k}$ ) followed Rayleigh distribution. Specifically, cell selection is considered with the blockage effects instead of distance-based, thus distance distribution of typical UE to the closest visible UAV is modified. In addition, the probability of establishing a line-of-sight (LOS) link is proven to decay exponentially with the horizontal distance between the typical user and the object UAV. Theoretical results show that with the deployment of UAVs, coverage probability can be improved by $5.3\times $ in ultra dense city scenario ( ${p}=$ 0.8) when signal-to-interference ratio (SIR) threshold equals 0dB, compared with terrestrial BSs.