Research On Statistical Performance Of Multi-antenna Cellular Networks In Hotspot Deployment

Driven by the rapid development of mobile internet,mobile data traffics generated by terminals and machines have grown rapidly,and there emerge diverse communication scenarios.To cope with the rapid growth of data and support for different scenarios,Researches on the fifth-generation(5G)wireless communication system have been fully carried out,and the technical requirement of achieving 1000 times growth in the system capacity has been proposed by IMT-2020.Multi-antenna dense network is one of the key technologies for 5G to increase the system capacity.For dense network deployments,hotspot is a typical deployment scenario,where base stations(BSs)are usually clustered due to the geographical factors or population factors.Compared with the traditional cellular network,the dense network under hotspots scenario faces more challenges.Firstly,BSs present clustering distributions around hotspots,which bring change to the network topology.Secondly,the high-density deployment of BSs complicates the distribution of network interference.Therefore,it is of great significance to establish an appropriate network model,and analyze the key factors affecting the network performance under hotspots scenario.In addition,effective interference cancellation strategies should also be explored to improve network performance.This thesis focuses on the study of multi-antenna dense networks under hotspots scenario.Based on the theory of stochastic geometry,the expressions of network performance indicators are derived,and the impact of network clustering characteristics on the system performance is analyzed.Moreover,combined with the network interference distribution characteristic,we further explore the intra-cluster interference coordination strategy to improve users'rate performance.The purpose of this thesis is to provide useful reference for 5G dense network deployment and optimization.Firstly,for the single-tier multi-antenna small cells,we model the locations of BSs as a Gaussian Poisson Point Process(GPP).We derive the approximate and asymptotic expression for the area spectral efficiency(ASE),and we further study the impact of different system parameters on ASE.It is observed that ASE increases with the clustering probability and decreases with the cluster radius.Configuring multiple antennas at BSs can bring significant gain to the improvement of ASE.When the number of antennas is fixed,there exists an optimal number of scheduled users for BSs to maximize ASE.In addition,compared with the uniform deployment,the network interference distribution under hotspots scenario is clustered.Because of the small distance between neighboring BSs in each cluster,serious intra-cluster interference is a major bottleneck that restricts the network capacity.In multi-antenna dense networks,multi-point cooperation is an effective way to deal with the inter-cell interference.Considering that intra-cluster interference is the main factor limiting system performance under hotspots scenario,it is beneficial to consider some interference coordination within single cluster.On the one hand,it can improve the system performance effectively by eliminate the severe intra-cluster interference.On the other hand,the intra-cluster BSs can form a small-scale cooperative cluster,which avoids the signal interaction of the whole network and has low complexity and less overhead.Based on the intra-cluster interference coordination scheme,this thesis further further analyzes and optimizes users'rate performance under hotspot scenarios.We firstly derives the relationship between the average data rate and the number of degrees of freedom(DoF)used for coordinated beamforming at the BS.Based on the expression,we prove that there exists an optimal DoF for interference cancelling at the BS to maximize the average data rate.The optimal solution can be quickly obtained by one-dimensional data search.Theoretical analysis and simulation results show that in the case of a small cluster radius,the intra-cluster interference is serious,using as many DoF as possible to eliminate intra-cluster interference can bring great gain for the users' rate performance improvement.However,as the cluster radius increases,the performance gain is gradually reduced.It means that there is a trade-off between interference suppression and transmission of useful signals when the number of antennas is fixed.The optimal DoF provides an optimal compromise between transmitting useful signals and interference suppression,which maximizes the average data rate.