Research On Grouping Technology In Multi-user FDD Massive MIMO Systems

Massive MIMO technology can significantly improve spectrum efficiency and system capacity,therefore it has become one of the key technologies for future communication.The FDD massive MIMO systems do not have channel reciprocity,thus when the systems need huge overhead of acquiring channel state information,the system performance will be severely limited.This thesis studies the grouping technologies which can be used in FDD massive MIMO systems.The study mainly includes an antenna grouping based feedback compression method,an GSM transmission scheme based on antenna grouping,user grouping algorithms in joint spatial division and multiplexing(JSDM)method.These methods are all suitabe for FDD massive MIMO systems.Firstly,we introduce some related technologies of massive MIMO systems,including the linear precoding technology based on complete CSI,precoding technique based on limited feedback for FDD systems,and a new open loop spatial multiplexing technology: Spatial Modulation(SM).Secondly,we study an antenna grouping based feedback compression method for FDD massive MIMO systems.We summarize the basic idea of antenna grouping of this compression method and the CSI acquirement process at the base station under this feedback scheme.The feedback packet structure of this feedback method is also described.We propose an effective grouping pattern selection scheme under correlated channels,and the performance and complexity analyses of the feedback scheme are also provided.The simulation results show that,when the same feedback performance is obtained,compared to the traditional vector quantization feedback scheme,the feedback compression scheme based on the antenna grouping can save more than 20% feedback overhead.Thirdly,we introduce Generalized Spatial Modulation(GSM),which is suitable for FDD massive MIMO systems.We study a new GSM transmission scheme: gGSM,which separates the antennas into groups,and then performs SM processing for each antenna group.The antenna grouping methods adopted in this scheme are interleaved grouping and block grouping.According to the simulation results,it is shown that compared with the traditional GSM transmission scheme,gGSM has better bit error rate performance when transmit antennas are highly correlated,and is more suitable for the FDD massive MIMO systems.After that,we study the user grouping algorithm which can be used in JSDM method.As a key step in JSDM,user grouping algorithm greatly affects the system performance.In this section,we overview the JSDM technology,including its system model and main ideas.Then we elaborate two kinds of two-level precoding strategies of JSDM,and analyse two typical user grouping algorithms.We also propose a modified K-means user grouping algorithm for the JSDM method which can be adopted in FDD massive MIMO systems.The simulations are implemented to compare the sum rate performance among the different user grouping algorithms in JSDM method.