Research On Clustering And Cooperative Transmission Technology In Dense Mobile Communication Networks

The rapid growth of total data traffic in mobile communication systems in recent years has promoted research on 5G mobile communication systems(5G).As one of the key technologies of 5G,ultra-dense network deployment can effectively increase the data transmission rate and increase the system capacity,but the dense deployment also brings stronger interference.How to perform effective interference management in dense networks is a problem that needs to be solved urgently.In this paper,in the context of combating a dense network with complex jamming environment,we first studied basestation-centric clustering under a dense multi-cell network,and then considered the user-centric clustering under a large-scale distributed antenna system.The study was carried out in two directions: dynamic clustering and static clustering..First,this paper introduced a large-scale distributed antenna system and performed a simple performance analysis.The basics of stochastic geometric modeling,MIMO system model,and common precoding methods are also introduced as the basis of the relevant algorithms in this paper.Then,this paper studies the basestation-centric clustering problem in dense multi-cell networks.Basestation-centric clustering refers to dividing all base stations in an area into several non-overlapping cooperative clusters.The base stations in a cluster cooperate to transmit user data in its coverage area.The algorithm is divided into centralized and distributed depending on whether clustering is done through sharing global CSI.This paper presents a centralized clustering method based on hierarchical clustering and a distributed base station clustering method based on game theory.The simulation results show that the proposed algorithm can get close to optimal clustering performance under low computational complexity.Then,this paper studies the user-centric dynamic clustering in large-scale distributed antenna systems.Dynamic clustering uses instantaneous channel state information for clustering and precoding design.For the disadvantages of high dynamic clustering CSI overhead and high computational complexity,this paper proposes a pre-clustering method using fuzzy mean clustering,which uses the statistical channel information to divide the candidate RAUs set for each user,and then performs dynamic clustering.At the cost of a slight decrease in the average user rate,the required CSI overhead and computational complexity can be drastically reduced.As the network scale continues to expand,the complexity of the above dynamic clustering scheme is hard to reduce.The user-centric static clustering in the large-scale distributed antenna system studied lastly can solve this problem effectively.Static clustering first uses the statistical channel information for clustering design,and then applies simple linear precoding in each cluster.The research in this part is based on the stochastic geometry modeling.First,the clustering is determined by the statistical properties of the network topology,and then a simple MRT precoding cooperative transmission is used in the cluster.Two kinds of stochastic geometric modeling are discussed in detail.The first one is user-centered clustering under uniform point distribution,and the optimal number of intra-cluster RAUs is derived using the characteristics of uniform point distribution.The second is the user-centric clustering under the Poisson point distribution,and the optimal cluster radius R is derived using the characteristics of the Poisson point distribution.Simulation results show that the proposed algorithm can achieve good performance.