Design And Analysis Of Heterogeneous Cellular Networks Based On Poisson Cluster Process

For the problem of traffic demand and ultra-high speed,ultra-low latency large-scale user connections in the hot scene,based on the traditional technology,the fifth generation mobile communication network adopts some key technologies,such as millimeter wave,device to device(D2D)and ulter dense heterogeneous networks.However,the dense deployment of multi-layer heterogeneous networks means that users will experience more complex intra-layer or inter-layer interference.In addition,in the hot spot area,users or base stations will be clustered,and there is a certain correlation between different layers of base stations.This yieds the traditional homogeneous poisson point process modeling scheme,which assumes that the sites are uniform distributed and independent,cannot model the spatial point distribution more appropriately.Therefore,this thesis analyzes and designs heterogeneous cellular networks based on the modeling scheme of Poisson clustering process.The specific works are as follows:In this thesis,we first consider a two-layer millimeter wave heterogeneous cellular networks consisting of Pico base station(PBS)and femto base station(FBS).In order to more closely simulate the distribution of users and base stations,we adopt the Poisson cluster modeling scheme.In this scheme,the hotspot center modeled as independent Poisson point process is defined as the parent process,and the nodes distributed around it are modeled as Poisson cluster processes.In order to achieve better quality of service for cluster users,a cluster classification scheme based on the closest distance ratio of PBS is proposed.When the ratio of the distances between user equipment(UE)and its service and interference PBS is less than the classification factor,the user defined as cluster-center UE,otherwise it is defined as cluster-edge UE.In addition,in order to improve the spectrum efficiency,a new spectrum sharing and allocation scheme is proposed,which divides the whole bandwidth into two orthogonal subbands.They are used by cluster-center and cluster-edge UEs associated with PBS,respectively.For users in FBSs,one subband will be randomly allocated according to the bandwidth factor.Then,the association probability and spectrum efficiency of cluster users in different scenarios are derived.The results show that,compared with the unclassified cluster scheme,the combination of cluster classification and spectrum sharing can significantly improve the association probability and spectrum efficiency of cluster users.Secondly,considering the excellent characteristics of D2 D transmission for heterogeneous cellular networks,this thesis analyzes a three-layer heterogeneous cellular network with the assistance of D2 D mixed communication.In this model,according to the user's traffic demand and the spatial correlation of different layers of base stations,PBS is modeled as poisson point process,and UEs and FBSs are modeled as different Poisson cluster processes with PBS coverage radius as the boundary.In this scheme,the user can not only get the request content from the base station,but also from the adjacent UE.Firstly,a new scheme of cluster user classification is proposed,which classifies different cluster users based on the radius of PBS coverage.Secondly,FBSs are divided into ordered and unordered scenarios.In the former case UE will choose the nearest FBS association,and in the latter case UE will randomly choose FBS association,which effectively improves the probability of UE association.Finally,considering new interference management scheme,the total available bandwidth is divided into two orthogonal subbands,which are allocated to PBS and MBS respectively.At the same time,inspired by the essential of low-power FBS,the cluster-center FBS and the cluster-edge FBS share two subbands respectively,and the cluster-center subband is also shared by D2 D.Then,the coverage probability of different cluster users is derived.The results show that different distribution variances of UE and FBS have different effects on the association probability.When the power is fixed,the smaller the radius of PBS and FBS,the better the coverage performance.