Research On User Clustering And Pilot Allocation Technology Based On Massive MIMO

With the development of 5th Generation Mobile Communication(5G),Massive Multiple-Input and Multiple-Output(M-MIMO),as one of the key technologies of 5G,has greatly increased the capacity of wireless communication system by increasing the degree of freedom of space,thus becoming a research hotspot at present.However,in massive MIMO systems,with the increasing number of users and antennas,the correlation between users and the feedback overhead of system acquisition channel state information(CSI)increase,and the pilot pollution will not disappear because of the increase of antenna number.All of these are the key factors that limit the performance improvement of massive MIMO systems.Therefore,this paper studies how to use clustering algorithm to reduce the interference of massive MIMO system and improve system capacity.Firstly,this paper introduces the basic concepts and related technical principles of massive MIMO system,focusing on the problems faced by macro-cell and macro-small cell heterogeneous networks(HeNet)with massive MIMO.Then,in the massive MIMO macro-cell scenario,a user clustering scheme based on user multi-dimensional feature similarity is proposed to reduce user interference and improve the capacity of massive MIMO system.This scheme designs an Affinity Propagation(AP)clustering algorithm based on multi-dimensional feature similarity by analyzing the multi-dimensional feature attributes of users.Also,in order to adapt to the dynamic user status and system environment,Q-learning algorithm is used to improve the system's adaptability.The simulation results show that the scheme reduces the system processing dimension and improve the system capacity and network flexibility.Secondly,in order to meet the practical application requirements of 5G,this paper extends the research scenario to macro-small cell HetNet with massive MIMO.To solve the pilot pollution problem that limits the capacity improvement of massive MIMO systems,a macro-small cell HetNet pilot allocation scheme based on micro-cluster is proposed.By measuring the potential pilot contamination,the scheme uses graph theory-based algorithm to cluster microcells,and designs macro-small cell HetNet pilot adjustment algorithm based on potential pilot pollution to achieve optimal macro-small cell HetNet pilot allocation.The simulation results show that the scheme improves the accuracy of channel estimation and the utilization of pilot resources,and increases the number of users serviced by the system.Finally,this paper puts forward the prospect of the research:for the multi-dimensional characteristics of users,we can study the performance difference caused by different weight ratios;for the pilot pollution problem under macro-small cell HeNet,it can be extended to multiple macro-cells.And pilot pollution can also be further studied from channel estimation,precoding and other aspects..