Analysis And Optimization Of Statistical Performance In Multi-Antenna Ultra Dense Networks

The research and standardization on 5G have been launched inspired by the concepts of mobile Internet and interconnectedness of all things.The most crucial goal of 5G is to improve the network throughput to meet the mobile users’ ultimate demand for data rates.Compared to the concurrent cellular networks,5G should enhance the network throughput more than a thousand times.To this end,ultra dense networks(UND)and multi-antenna have been recognized as two basic techniques to cope with the extremely demand for data rates.Specially,UDN improves the gains of spectrum reuse through the dense deployment of small cell base stations(SBSs).On the other hand,multi-antenna can further boost the network capacity by the techniques of diversity and multiplexing.In order to understand the behaviors of the network combining these two techniques,this dissertation focuses on the aspects of area spectral efficiency(ASE),network deployment and interference coordination from the perspective of statistical performance,thus providing guidelines for the deployment and optimization of future 5G systems.The main contributions of this dissertation are summarized as follows:1)Considering the characteristics of higher base station density and more antennas in future networks,this dissertation first focuses on single tier multi-antenna UDNs and studies the functional relationships between ASE and other system parameters such as base station density,the number of antennas and so on.The optimal combination of network deployment parameters is also provided to minimize the average network energy consumption(NEC).By modeling the single tier network as a Poisson point process(PPP),this dissertation derives the ASE and its lower-bound.It is found that ASE increases linearly with respect to the base station(BS)density.For the purpose of ASE maximization,the ratio between the optimal number of active users and the number of antennas is demonstrated to remain consistent.Under the optimal configurations,ASE can increase linearly with the number of antennas.Besides,this dissertation proves that the problem of NEC minimization is equivalent to the problem of network energy efficiency(NEE)maximization.To solve the formulated problem,the optimal algorithm and low-complexity suboptimal algorithm are both provided.Numerical results show that the suboptimal algorithms can achieve near optimal performance.This dissertation finds that the optimal ASE increase linearly with respect to the ASH target.Furthermore,it is preferred to equip the macro base stations(MBSs)with more antennas to reduce NEC.2)Taking into consideration the characteristics of multi-tier structure and new architecture in future networks,this dissertation pursues the theoretical relationships between ASE and other system parameters in two-tier multi-antenna UDNs.This dissertation also optimizes the deployment strategy to minimize the average NEC under control/data separation architecture.By modeling the MBSs and SBSs as two PPPs,this dissertation derives the expressions for ASE and its approximation and demonstrates the theoretical gains of deploying more SBSs.When range expansion is adopted,a higher transmit power of the SBSs is demonstrated to bring in a larger ASE.For the purpose of ASE maximization,the optimal number of active users in each tier has a linear relationship with the number of antennas and is insensitive to different parameter configurations.Under the optimal configurations,ASE can increase linearly with the numbers of antennas in both tiers.Different from single antenna networks,if the number of antennas of MBSs is sufficiently larger than that of SBSs,range expansion can improve ASE.Moreover,in order to minimize the NEC while guaranteeing the performance of coverage and throughput,the dissertation studies the optimal deployment strategy for two-tier multi-antenna UDNs under control/data separation architecture.A NEC minimization problem is formulated to jointly optimize BS density,the number of antennas and spectrum allocation ratio in different tiers.An optimal algorithm and a low-complexity suboptimal algorithm are proposed to solve the formulated problem.Numerical results show that the proposed suboptimal algorithm can achieve near optimal performance.Besides,the optimal NEC is insensitive to spectrum allocation ratio.3)As inter-cell interference is still the bottleneck of future cellular networks,this dissertation further studies the interference coordination schemes in two-tier multi-antenna UDNs.The functional relationships between users’ average achievable data rate and interference coordination parameters are characterized.The optimal interference coordination scheme is also provided from the perspective of statistical performance.This dissertation focuses on the scenario of dense deployment of closed access SBSs,in which inter-cell interference is quite strong.The SBSs are assumed to adopt power control,access control and coordinated beamforming to suppress the inter-tier interference to the macro users(MUEs).Leveraging two approximations,this dissertation derives the expressions for users’ average data rate to reveal the functional relationships between users’ performance and interference coordination parameters.Based on the analysis results,this dissertation further discusses the optimal interference coordination scheme to maximize the small cell users’(SUEs)data rate while guaranteeing the MUEs’ performance.In the special case without coordinated beamforming,it is demonstrated that the optimal strategy is to only adopt power control.In general cases,the optimal strategy is the combination of power control and coordinated beamforming.If the target of MUEs’ data rate is small,numerical results show that coordinated beamforming can achieve near optimal performance.When different schemes are utilized individually,from the perspective of the performance of SUEs,coordinated beamforming is superior to the others while access control is the worst.