Research On Sum Rate And Precoder In Massive Distributed MIMO Systems

As a key technology for the next generation communication systems,Massive Multiple Input Multiple Output(Massive MIMO)is a hot topic due to its advantages in data rate and energy efficiency.However,these benefits are affected much by the user distribution,channel fading and the number of antennas,such that it is crucial to develop a direct and clear way analyzing the system sum rate over composite fading channels,and further design precoders of low overhead and complexity which is based on a concise expression of the sum rate.This dissertation focuses on the sum rate analysis and precoder design with the consideration of user density,precoding complexity,sum rate under composite fading channel and its asymptotic feature for the infinite antennas case.Divided into three points of ’Massive’,’Distributed’ and ’Massive Distributed’,this dissertation subsequently studies the precoder design based on interference classification and cancellation of isolated users in Massive MIMO systems,the sum rate analysis for precoder design in distributed MIMO(D-MIMO)systems over composite fading channels,and the uplink sum rate in massive distributed MIMO systems over Rayleigh-Lognormal(RLN)fading channels.Accordingly,the research work lies in three aspects:1.Propose a precoding algorithm based on interference classification and cancellation in Massive MIMO systems.By noticing the aggregation of user distribution,this work defines user crowd and isolated user as two kinds of user units which are based on the dominant eigen-space of the corresponding channel matrices.Afterwards,Chordal Distance and Fubini-Study Distance are introduced from the subspace packing problem to group user units.As such,the inter-group interference is mitigated by the approximated block diagonalization(BD)pre-beamforming,while the intra-group interference is handled via interference cancellation(IC)assist zero forcing(ZF)or regularized zero forcing(RZF)precoding.Finally,the simulation results validate the algorithm and give some insights for the network design.In sum,this part focuses on different user unit,studies the interference classification and mitigation of isolated users,and design a two-stage precoder based on it.2.Propose a direct and effective way to analyze the sum rate in Distributed MIMO systems over composite fading channel.Considering the importance of composite fading in distributed systems,this research achieves the sum rate of both ZF and decision feedback ZF(DF-ZF)receivers in a series form by tackling with a complex integral based on the random matrix theory.Next,a three-item expansion of the sum rate is used as a direct evaluation of the system performance to design the precoder owe to its rapid convergency.Moreover,a full-matrix precoder with optimal power allocation is designed.At last,simulation validates the correctness and effectiveness of the sum rate expression and the advantage of the designed precoder.Note that the outer precoder requires only the second order statistics of channel,and the inner precoder realizes the optimal power loading maximizing the sum rate.This research attains an analysis framework of the sum rate with regard to RLN fading,designs an optimal power loading algorithm to maximize the sum rate,and further achieves a precoder of full matrix.3.Propose a general method to analyze the sum rate and its asymptotic behavior in Massive Distributed MIMO systems.In such a system,the effect of small-scale fading and noise diminishes as the number of antennas goes large,while the effect of the large-scale fading becomes important in a distributed structure.This research first establishes the signal transmission model of both ZF and DF receivers over composite RLN fading channels,and then transform the sum rate to a incomplete-gamma-function related problem through the Toeplitz Matrix Theory(TMT),and then discusses the asymptotic behavior by exploiting the good structure in mathematics.In brief,on the one hand,this work acquires the sum rate from the original complex form to a series with regard to the negative power of antennas which converges quickly;on the other,the asymptotic feature is meaningful as a reference to future system design.