Evaluation Modeling And Optimization Of 5G Ultra-dense Networks

Ultra-dense network(UDN)is a research interest and an important way of 5G network distribution planning,which has attracted widespread attention in the 5G field.The technology miniaturizes 5G base stations(BSs)and evolves the communication network into a multi-layer,ultra-dense,heterogeneous form,which reduces the cell size and transmission distance,thereby increasing the density of micro BSs per unit area.UDN can also cooperate with interference coordination technology and bring significant transmission gain.At the same time,with the improvement of edge nodes' computing power and the gradual enhancement of the performance of ports between the base stations,the 5 G network has gradually improved the control and scheduling capabilities of each edge node.Then 5G servers can check the information such as network topology and real-time service quality and continuously optimize the "micro BS-UE","micro BS-micro BS" in coordination.Therefore the deployment of UDN has its computing power basis.However,network densification has led to a continuous increase in the number of BSs which cause co-channel interference to target users.Interference caused by high density BSs has become an important technical problem.Therefore,this thesis includes the following research:(1)Aiming at the performance evaluation of 5G ultra-dense millimeter wave networks,a model based on stochastic geometry is given,which comprehensively considers the beamforming mechanism of the base station,communication access scenarios,micro base stations' density,environment obstacles' density and other factors,and the serving micro base station is no longer placed in a fixed position and modeled as the same distribution with other micro base stations.The model is of high efficient and concise.Through this model,the optimal density of micro base stations in ultra-dense networks for different environments can be calculated,which has practical application capabilities.(2)In order to make up for some disadvantages of the stochastic geometric model and improve the universality of the ultra-dense network performance evaluation model,so that the simulator is more suitable for 5G ultra-dense network simulation scenarios,the "L1 layer 's array antenna feature abstraction module"and the "L2 layer's throughput cache and interface module"are designed and implemented on the basis of an open source simulator.The "L1 layer's array antenna feature abstraction module"implements an array antenna class,which adds the beamforming workflow to the original simulator.This module can obtain the different angles' transmission gain of the array antenna."L2 layer's throughput cache and interface module" implements the throughput functions related to the L2 layer,so that the simulator can access the throughput demand in real time.The modules' development provides platform support for the follow-up research of the thesis.(3)Aiming at the "inner-cell" time-frequency resource scheduling scenarios of 5G ultra-dense networks,with the basis of the 5G ultra-dense network simulation platform,an UE scheduling algorithm for 5G heterogeneous network quality-of-service(QoS)improvement is designed and implemented.The algorithm comprehensively considers multiple QoS characteristic factors such as window duration,data packet delay,logical channel priority,RLC mode,downlink guaranteed rate,and downlink scheduling priority weighting factor.The simulation verifies the advantages of the algorithm in evaluation indicators such as average throughput and maximum spectral efficiency.(4)Aiming at the "inter-cell" optimization of 5G ultra-dense networks,an interference coordination method between ultra-dense micro base stations based on a divide-and-conquer recursion strategy is proposed,which is a multi-domain hybrid interference coordination algorithm.The single-layer clustering structure of base stations is enhanced to a divide-and-conquer recursive structure of network tree topology,which realizes low-time complexity and high-quality dynamic interference coordination between micro base stations.At the same time,the algorithm design framework has general characteristics and can be applied to similar coordination issues.