Performance Analysis For User-Centric Uumanned Aerial Vehicles(Base Stations)Networks Via Stochastic Geometry

Unmanned aerial vehicles(UAVs)have controllable mobility and high probability of line-of-sight(LoS),and hence improving the coverage and capacity of the networks.User-centric UAV networks break through traditional cell-centric architecture.In user-centric UAV networks,UAVs are used as base stations(BSs),several potential serving BSs form a dynamic BS group(BSG)for each user,so that the user can fully exploits the diversity gain of potential serving BSs within BSG.However,in user-centric UAV networks,the spatial correlation of BS scheduling is induced.Using tools of stochastic geometry,this paper theoretically analyzes the diversity gain of potential BSs' backhaul and access capacities for backhaul-aware BSG selection in user-centric UAV networks.Furthermore,based on the modelling of spatial correlation,analytical expressions of coverage probability are derived for user-centric UAV networks.The analytical results provide guidance to practical deployments of UAV networks.The contributions of this paper are categorized as follows:The theoretical analysis of backhaul-aware cell selection is not studies in the literature.This paper models the relations between backhaul capacity and user association bias,and analyzes the performance of backhaul-aware cell selection by obtaining the analytical results of outage probability using stochastic geometry.More specifically,in the BSG selection,high-backhaul-capacity BS has a higher bias value to attract the users from low-backhaul-capacity BS.The serving BS is always the BS with the largest transmission rate(minimum of access and backhaul rate),therefore fully utilizing the diversity gain within BSG.The results show that the outage probability is decreased by 30.6%when the BSG size equals 5,compared to the distance-based BSG selection.The spatial correlation in the user-centric UAV networks is not studied in the literature.This paper models spatial correlation for multi-beam joint transmission of multiple BSs in user-centric UAV networks,and obtains analytical results of coverage probability.More specifically,we model the spatial correlation induced by the degree-of-freedom of BS beams,as well as the competition of user scheduling.Also,we model the LoS and none-line-of-sight(NLoS)probability and their path losses.Based of limited feedback of channel direction information(CDI),the performance bound of coverage probability is obtained using double gamma ratio(DGR).