CN105429688A - Multi-cell precoding method for suppressing pilot pollution in large-scale distributed antenna system - Google Patents

Abstract

The invention relates to a multi-cell pre-coding method for inhibiting pilot pollution in a large-scale distributed antenna system, and belongs to the technical field of wireless communication. The method comprises the following steps: calculating an estimated value of an uplink channel from one user to one RAU in a central cell according to signals received by antennas on the RAU in the central cell; calculating a covariance matrix of errors in the estimated value of the uplink channel on the RAU in the central cell; calculating estimated values of channel vectors from the user to all RAUs of the central cell; calculating estimated values of channel matrixes from all users to all the RAUs of the central cell; calculating transmitting signal vectors of the central cell by conducting multi-cell pre-coding on data vectors of all the users in the central cell; and performing normalization to obtain actual vectors of signals which are finally transmitted by antennas on all the RAUs in the central cell. Through adoption of the method, the impact of pilot pollution on the downlink pre-coding performance can be effectively inhibited, the system throughput can be improved, and the pre-coding calculation complexity can be greatly reduced.

Description

Translated fromChinese

大规模分布天线系统中抑制导频污染的多小区预编码方法Multi-cell precoding method for suppressing pilot pollution in large-scale distributed antenna system

技术领域technical field

本发明属于无线通信技术领域，涉及一种大规模分布式天线系统中抑制导频污染的多小区预编码方法，尤其涉及一种基于共轭梯度法的最小均方误差多小区预编码方法。The invention belongs to the technical field of wireless communication, and relates to a multi-cell precoding method for suppressing pilot pollution in a large-scale distributed antenna system, in particular to a minimum mean square error multi-cell precoding method based on a conjugate gradient method.

背景技术Background technique

随着移动终端数的飞速增长、物联网的发展，以及无线数据业务种类的不断增多，用户对无线数据业务的需求不断对现有的移动通信技术提出新的挑战。作为一种具有满足未来无线移动通信对频谱效率、能量效率和信道容量需求潜质的技术，大规模多输入多输出(以下简称MassiveMIMO)技术已经在学术界和产业界引起了广泛的关注。With the rapid increase in the number of mobile terminals, the development of the Internet of Things, and the increasing types of wireless data services, users' demands for wireless data services continue to pose new challenges to existing mobile communication technologies. As a technology with the potential to meet the spectrum efficiency, energy efficiency and channel capacity requirements of future wireless mobile communications, massive multiple-input multiple-output (hereinafter referred to as MassiveMIMO) technology has attracted widespread attention in academia and industry.

MassiveMIMO技术通过在基站(以下简称BS)侧部署数量众多的天线，来为与通信用户之间的通信提供足够多的自由度以提高分集增益和复用增益。小区边缘用户由于较高的路径损耗导致其接收到的有用信号功率很低，在实际的多小区场景下，小区边缘用户接收到的干扰信号的强度可与有用信号相比，甚至超过了有用信号的强度，这使得小区边缘用户的吞吐率性能严重恶化。为了解决小区边缘覆盖差的问题，学术界和产业界考虑将原来集中在BS的天线以分布式的方式部署在小区中，通过减小信号的传输距离来减小路径损耗，提高有用信号的功率。通过大量的以分布式方式部署在小区中的天线联合为小区中用户提供服务的系统称为大规模分布式天线系统。在大规模分布式天线系统中，分布式的无线接入点叫做远端接入单元(以下简称RAU)，每个RAU上有若干根天线，且每个RAU通过有线或无线的回程链路连接到基带处理单元上。由于不同RAU的地理位置不同，因此，不同RAU上的天线到用户的大尺度衰落也不同，这使得大规模分布式天线系统的信道建模与分析很困难。目前，大多数关于分布式天线系统的下行预编码研究都假设各RAU上的各天线与各用户之间的信道状态信息(以下简称CSI)是已知的。然而在实际的多小区时分双工系统中，考虑最大化频谱利用率的导频全复用方式，不同小区中的用户使用同一套导频序列。每一个小区中的天线都会收到来自不同小区用户发送的导频序列，从而使得信道估计产生较大误差，这就是导频污染造成的。在分布式天线系统中，部分靠近小区边缘的RAU可能遭受更加严重的导频污染。这使得理想CSI的假设不再成立。现有的多小区下行预编码方案或假设已知理想CSI，或针对集中式多天线系统，不能直接应用到大规模分布式天线系统中，且其计算复杂度随着天线数的增多而急剧增加。因此，使用估计的信道对下行传输数据进行预编码，同时能最大限度地抑制导频污染影响的低复杂度算法仍有待研究。The Massive MIMO technology provides enough degrees of freedom for communication with communication users to improve diversity gain and multiplexing gain by deploying a large number of antennas on the base station (hereinafter referred to as BS) side. Due to the high path loss of the cell edge users, the useful signal power received by them is very low. In the actual multi-cell scenario, the strength of the interference signal received by the cell edge users can be compared with the useful signal, or even exceeds the useful signal. strength, which seriously deteriorates the throughput performance of cell edge users. In order to solve the problem of poor coverage at the edge of the cell, the academic and industrial circles consider deploying the antennas that were originally concentrated in the BS in a distributed manner in the cell, reducing the path loss by reducing the signal transmission distance, and increasing the power of useful signals. . A system that jointly provides services to users in a cell through a large number of antennas deployed in a distributed manner is called a large-scale distributed antenna system. In a large-scale distributed antenna system, the distributed wireless access point is called a remote access unit (hereinafter referred to as RAU), each RAU has several antennas, and each RAU is connected by a wired or wireless backhaul link to the baseband processing unit. Due to the different geographic locations of different RAUs, the large-scale fading from antennas to users on different RAUs is also different, which makes channel modeling and analysis of large-scale distributed antenna systems difficult. At present, most studies on downlink precoding of distributed antenna systems assume that channel state information (hereinafter referred to as CSI) between each antenna on each RAU and each user is known. However, in an actual multi-cell time-division duplex system, considering the full pilot multiplexing method to maximize spectrum utilization, users in different cells use the same set of pilot sequences. The antennas in each cell will receive pilot sequences sent by users in different cells, resulting in large errors in channel estimation, which is caused by pilot pollution. In a distributed antenna system, some RAUs close to the cell edge may suffer from more severe pilot pollution. This makes the assumption of ideal CSI no longer valid. The existing multi-cell downlink precoding schemes either assume that the ideal CSI is known, or are aimed at a centralized multi-antenna system, and cannot be directly applied to a large-scale distributed antenna system, and its computational complexity increases sharply with the increase in the number of antennas . Therefore, a low-complexity algorithm that uses the estimated channel to precode downlink transmission data while minimizing the influence of pilot pollution remains to be studied.

发明内容Contents of the invention

本发明的目的是为已有技术存在的问题，提出一种大规模分布天线系统中抑制导频污染的多小区预编码方法，该方法首先分别对用户到每个RAU的信道进行最小均方误差估计，然后在用估计的信道对待发送的下行数据进行预编码，本发明能充分利用估计出的本小区用户和其他小区用户的信道设计预编码矩阵，以最大限度抑制导频污染对系统性能的影响。The purpose of the present invention is to solve the problems existing in the prior art, and propose a multi-cell precoding method that suppresses pilot pollution in a large-scale distributed antenna system. The method first performs the minimum mean square error on the channel from the user to each RAU respectively. estimate, and then use the estimated channel to precode the downlink data to be sent. The present invention can make full use of the estimated channels of the users in this cell and users in other cells to design a precoding matrix, so as to suppress the influence of pilot pollution on system performance to the greatest extent. influences.

本发明提出的大规模分布天线系统中抑制导频污染的多小区预编码方法，包括以下步骤：The multi-cell precoding method for suppressing pilot pollution in the large-scale distributed antenna system proposed by the present invention comprises the following steps:

(1)考虑以要进行预编码的小区(以下简称中心小区)为中心的包括L个小区的大规模分布式天线系统，设中心小区中的RAU数为M，每个RAU配备的天线数为N，中心小区中的用户数为K，中心小区周围L-1个小区中的总用户数为R，每个用户配置一根天线；对大规模分布式天线系统中的用户进行编号，中心小区的用户编号为1,2,…,K，中心小区周围L-1个小区中的用户编号为K+1,K+2,…,K+R，则大规模分布式天线系统中u第个用户到中心小区的第m个RAU的上行信道为h_mu＝λ_mug_mu，其中λ_mu表示大规模分布式天线系统中第u个用户到中心小区的第m个RAU的大尺度衰落，g_mu表示大规模分布式天线系统中第u个用户到中心小区的第m个RAU的小尺度信道，代表复数域，1≤u≤K+R，1≤m≤M；设大规模分布式天线系统中第u个用户在信道估计阶段发射的导频序列为其中τ为导频序列的长度，该长度大于大规模分布式天线系统中所有小区用户数的最大值，第u个用户在信道估计阶段发射导频序列的功率为ρ_u；各导频序列通过用户的天线发射，经过信道，在中心小区的第m个RAU的N根天线得到RAU接收信号，记为其中，表u个用户在信道估计阶段发射的导频序列的转置，N_m表示中心小区的第m个RAU的接收信号中的加性高斯白噪声，矩阵N_m中的元素独立同分布于均值为0、方差为σ²的循环复高斯分布，1≤m≤M；(1) Considering a large-scale distributed antenna system including L cells centered on the cell to be precoded (hereinafter referred to as the central cell), the number of RAUs in the central cell is M, and the number of antennas each RAU is equipped with is N, the number of users in the central cell is K, the total number of users in the L-1 cells around the central cell is R, and each user is equipped with an antenna; the users in the large-scale distributed antenna system are numbered, and the central cell The user numbers of the user are 1, 2,..., K, and the user numbers in the L-1 cells around the central cell are K+1, K+2,...,K+R, then the uth in the large-scale distributed antenna system The uplink channel from the user to the mth RAU of the central cell is h_mu = λ_mu g_mu , where λ_mu represents the large-scale fading from the u-th user to the m-th RAU of the central cell in the large-scale distributed antenna system, and g_mu represents the m-th RAU from the u-th user to the central cell in the large-scale distributed antenna system The small-scale channel of Represents the complex domain, 1≤u≤K+R, 1≤m≤M; suppose the pilot sequence transmitted by the uth user in the channel estimation stage in the large-scale distributed antenna system is Where τ is the length of the pilot sequence, which is greater than the maximum number of users in all cells in the large-scale distributed antenna system. The power of the uth user transmitting the pilot sequence in the channel estimation stage is ρ_u ; each pilot sequence passes The user's antenna transmits, passes through the channel, and receives the signal received by the RAU from the N antennas of the mth RAU in the central cell, denoted as in, Table the transposition of the pilot sequence transmitted by u users in the channel estimation phase, N_m represents the additive white Gaussian noise in the received signal of the mth RAU of the central cell, The elements in the matrix N_m are independently and identically distributed in a cyclic complex Gaussian distribution with a mean of 0 and a variance of σ² , 1≤m≤M;

(2)计算上述中心小区的第m个RAU的N根天线接收到的信号Y_m各列的协方差矩阵Φ_m，其中1≤m≤M，表示第k用户在信道估计阶段发射的导频序列的共轭转置，I_τ为τ阶单位矩阵；(2) Calculate the covariance matrix Φ_m of each column of the signal Y_m received by the N antennas of the m-th RAU of the central cell, where 1≤m≤M, Represents the conjugate transpose of the pilot sequence transmitted by the kth user in the channel estimation stage, I_τ is the identity matrix of order τ;

(3)用共轭梯度法计算大规模分布式天线系统中第u个用户到中心小区的第m个RAU的上行信道h_mu的估计值具体过程如下：(3) Calculate the estimated value of the uplink channel_hmu from the uth user to the mth RAU of the central cell in the large-scale distributed antenna system by using the conjugate gradient method The specific process is as follows:

(3-1)计算大规模分布式天线系统中第u个用户使用的导频序列π_u的共轭经过上述Φ_m的逆矩阵滤波得到的向量f_mu，其中表示π_u的共轭，是Φ_m的逆矩阵；(3-1) Calculate the vector f_mu obtained by the conjugate of the pilot sequence π_u used by the uth user in the large-scale distributed antenna system after the inverse matrix filtering of the above Φ_m , in represents the conjugate of π_u , is the inverse matrix of Φ_m ;

(3-2)建立一个上述滤波过程的等效线性方程组模型采用共轭梯度法，求解等效线性方程组模型，得到大规模分布式天线系统中第u个用户使用的导频序列的共轭经过滤波的向量f_mu，具体步骤如下：(3-2) Establish an equivalent linear equation model of the above filtering process Use the conjugate gradient method to solve the equivalent linear equation model to obtain the conjugate of the pilot sequence used by the uth user in the large-scale distributed antenna system go through Filtered vector f_mu , the specific steps are as follows:

(3-2-1)初始化：设置迭代次数阈值P，设向量f_mu的初始值为则上述模型的初始余向量为r₀，初始化共轭基向量为p₀，p₀＝r₀，设置迭代变量t＝0；(3-2-1) Initialization: set the iteration threshold P, and set the initial value of the vector f_mu to Then the above model The initial covector of is r₀ , Initialize the conjugate basis vector as p₀ , p₀ =r₀ , and set the iteration variable t=0;

(3-2-2)更新迭代变量使t＝t+1，计算第t次迭代过程中共轭基向量系数α_t，其中表示共轭基向量p_t-1的共轭转置，并根据该系数α_t，计算第t次迭代过程中的向量f_mu的值(3-2-2) Update the iterative variable so that t=t+1, and calculate the conjugate basis vector coefficient α_t in the tth iteration process, in Represents the conjugate transpose of the conjugate base vector p_t-1 , and according to the coefficient α_t , calculate the value of the vector f_mu during the tth iteration

(3-2-3)根据计算第t次迭代后上述模型的余向量r_t，r_t＝r_t-1-α_tΦ_mp_t-1；(3-2-3) According to Calculate the covector r_t of the above model after the tth iteration, r_t = r_t-1 -α_t Φ_m p_t-1 ;

(3-2-4)计算第t次迭代过程中共轭基向量的调节系数β_t，并根据该系数计算第t+1次迭代过程中的共轭基向量p_t，p_t＝r_t+β_tp_t-1；(3-2-4) Calculate the adjustment coefficient β_t of the conjugate basis vector in the t-th iteration process, And calculate the conjugate basis vector p_t in the t+1th iteration process according to the coefficient, p_t = r_t +β_t p_t-1 ;

(3-2-5)根据上述迭代阈值P对迭代变量t进行判断，若t≥P，则停止迭代，并使进行步骤(3-3)，若t＜P，则进行步骤(3-2-2)-(3-2-5)；(3-2-5) Judge the iteration variable t according to the above iteration threshold P, if t≥P, stop the iteration, and use Carry out step (3-3), if t<P, then carry out step (3-2-2)-(3-2-5);

(3-3)根据上述步骤(3-2)得到的向量计算大规模分布式天线系统中第u个用户到中心小区的第m个RAU的上行信道h_mu的估计值其中，1≤u≤K+R，，1≤m≤M；(3-3) According to the vector obtained in the above step (3-2), calculate the estimated value of the uplink channel_hmu from the uth user to the mth RAU of the central cell in the large-scale distributed antenna system Among them, 1≤u≤K+R, 1≤m≤M;

(4)计算上述中心小区的第m个RAU的上行信道h_jmlk的估计误差的协方差矩阵C_mu，$C_{mu}={\mathrm{\&lambda;}}_{mu}(1-\frac{{\mathrm{\&rho;}}_u{\mathrm{\&tau;\&lambda;}}_{mu}}{{\mathrm{\&pi;}}_u^H\left(\underset{k=1}{\overset{K+R}{\&Sigma;}}{\mathrm{\&rho;}}_k{\mathrm{\&lambda;}}_{mk}{\mathrm{\&pi;}}_k\right)+{\mathrm{\&sigma;}}^2})I_N,$其中，1≤u≤K+R，1≤m≤M；(4) Calculate the covariance matrix C_mu of the estimation error of the uplink channel h_jmlk of the mth RAU of the above-mentioned central cell,$C_{mu}={\mathrm{\&lambda;}}_{mu}(1-\frac{{\mathrm{\&rho;}}_u{\mathrm{\&tau;\&lambda;}}_{mu}}{{\mathrm{\&pi;}}_u^h\left(\underset{k=1}{\overset{K+R}{\&Sigma;}}{\mathrm{\&rho;}}_k{\mathrm{\&lambda;}}_{mk}{\mathrm{\&pi;}}_k\right)+{\mathrm{\&sigma;}}^2})I_N,$ Among them, 1≤u≤K+R, 1≤m≤M;

(5)构造大规模分布式天线系统中第u个用户到中心小区的所有M个RAU的信道向量h_u的估计值其中，1≤u≤K+R，表示步骤(3)中获得的大规模分布式天线系统中第u个用户到中心小区的第m个RAU的上行信道h_mu的估计值的转置，1≤m≤M；(5) Construct the estimated value of the channel vector h_u of all M RAUs from the uth user to the central cell in the large-scale distributed antenna system Among them, 1≤u≤K+R, Indicates the estimated value of the uplink channel_hmu from the uth user to the mth RAU of the central cell in the large-scale distributed antenna system obtained in step (3) Transpose of , 1≤m≤M;

(6)构造大规模分布式天线系统中所有K+R个用户到中心小区的所有M个RAU的信道矩阵H₀的估计值计算H₀的估计误差的协方差矩阵C₀，中心小区的所有K个用户到中心小区的所有M个RAU的信道矩阵H的估计值为其中为的前K列，(6) Construct the estimated value of the channel matrix_H0 of all K+R users to all M RAUs in the central cell in the large-scale distributed antenna system Compute the covariance matrix C₀ of the estimation error of H₀ , The estimated channel matrix H from all K users in the central cell to all M RAUs in the central cell is in for The first K columns of ,

(7)用共轭梯度法计算中心小区的所有K个用户的数据向量s经过多小区预编码后得到的中心小区的发射信号向量x，具体过程如下：(7) Use the conjugate gradient method to calculate the data vector s of all K users in the central cell and obtain the transmitted signal vector x of the central cell after multi-cell precoding. The specific process is as follows:

(7-1)利用步骤(6)中得到的大规模分布式天线系统中心小区所有K个用户到中心小区的所有M个RAU的信道矩阵H的估计值计算s经过匹配滤波后的向量z，(7-1) Using the estimated value of the channel matrix H of all K users in the central cell of the large-scale distributed antenna system obtained in step (6) to all M RAUs in the central cell Calculate s after Matched filtered vector z,

(7-2)计算中心小区的多小区最小均方误差均衡矩阵W，其中表示的共轭转置，I_MN表示MN阶单位矩阵；(7-2) Calculate the multi-cell minimum mean square error equalization matrix W of the central cell, in express The conjugate transpose of , I_MN represents the MN order identity matrix;

(7-3)建立一个大规模分布式天线系统中心小区进行多小区最小均方误差预编码的等效线性方程组模型z＝Wx，采用共轭梯度法，求解该等效线性方程组模型，得到中心小区的发射信号向量x，具体步骤如下：(7-3) Establish a large-scale distributed antenna system central cell to carry out the equivalent linear equation model z=Wx of multi-cell minimum mean square error precoding, adopt the conjugate gradient method to solve the equivalent linear equation model, To obtain the transmit signal vector x of the central cell, the specific steps are as follows:

(7-3-1)初始化：设置迭代次数阈值V，设中心小区的发射信号向量x的初始值为x⁽⁰⁾，则上述模型z＝Wx的初始余向量为v₀，v₀＝z，初始化共轭基向量为q₀，q₀＝r₀，设置迭代变量ω＝0；(7-3-1) Initialization: set the iteration threshold V, set the initial value of the transmitted signal vector x of the central cell to x⁽⁰⁾ , then the initial covector of the above model z=Wx is v₀ , v₀ =z , initialize the conjugate basis vector as q₀ , q₀ =r₀ , set the iteration variable ω=0;

(7-3-2)更新迭代变量使ω＝ω+1，计算第ω次迭代过程中共轭基向量系数μ_t，其中表示共轭基向量q_t-1的共轭转置，并根据该系数μ_t，计算第ω次迭代过程中的中心小区的发射信号向量x的值x^(t)，x^(t)＝x^(t-1)+μ_tq_t-1；(7-3-2) Update the iterative variables to make ω=ω+1, calculate the conjugate basis vector coefficient μ_t in the ωth iteration process, in Represents the conjugate transpose of the conjugate base vector q_t-1 , and according to the coefficient μ_t , calculate the value x^(t) of the transmitted signal vector x of the central cell in the ωth iteration process, x^(t) = x^(t-1) + μ_t q_t-1 ;

(7-3-3)根据x^(t)计算第ω次迭代后上述模型的余向量v_t，v_t＝v_t-1-μ_tWq_t-1；(7-3-3) Calculate the covector v_t of the above-mentioned model after the ωth iteration according to x^(t) , v_t =v_t-1- μ_t Wq_t-1 ;

(7-3-4)计算第ω次迭代过程中共轭基向量的调节系数θ_t，并根据该系数计算第ω+1次迭代过程中的共轭基向量q_t，q_t＝v_t+β_tq_t-1；(7-3-4) Calculate the adjustment coefficient θ_t of the conjugate basis vector in the ωth iteration process, And calculate the conjugate basis vector q_t in the iterative process of ω+1 according to this coefficient, q_t =v_t +β_t q_t-1 ;

(7-3-5)根据上述迭代阈值V对迭代变量ω进行判断，若ω≥V，则停止迭代，并使x＝x^(t-1)，进行步骤(7-4)，若ω＜V，则进行步骤(7-3-2)-(7-3-5)；(7-3-5) Judge the iterative variable ω according to the above-mentioned iteration threshold V, if ω≥V, then stop the iteration, and set x=x^(t-1) , proceed to step (7-4), if ω< V, then carry out steps (7-3-2)-(7-3-5);

(8)对上述步骤(7)得到的得到中心小区的发射信号向量x进行归一化，使得到中心小区的各RAU上的天线最终发射的实际信号向量，实现大规模分布式天线系统中抑制导频污染的多小区预编码。(8) normalize the transmitted signal vector x of the central cell obtained in the above step (7), so that The actual signal vector finally transmitted by the antennas on each RAU of the central cell is obtained, and multi-cell precoding that suppresses pilot pollution in a large-scale distributed antenna system is realized.

本发明提出的大规模分布式天线系统中抑制导频污染的多小区预编码方法，其优点是，在导频全复用的时分双工系统中，能充分利用基站估计出的信道信息，有效抑制导频污染对下行预编码性能的影响，提升系统吞吐量；此外，本方法还通过共轭梯度法求解线性方程组，避免了在预编码过程中直接对大型矩阵求逆，能大大减少预编码的计算复杂度。The multi-cell precoding method for suppressing pilot pollution in the large-scale distributed antenna system proposed by the present invention has the advantage that in the time division duplex system with full pilot multiplexing, the channel information estimated by the base station can be fully utilized, effectively Inhibit the impact of pilot pollution on downlink precoding performance and improve system throughput; in addition, this method also solves linear equations through the conjugate gradient method, which avoids directly inverting large matrices during the precoding process and can greatly reduce precoding. Computational complexity of encoding.

附图说明Description of drawings

图1是本发明方法涉及的大规模分布式天线系统场景示意图。FIG. 1 is a schematic diagram of a scene of a large-scale distributed antenna system involved in the method of the present invention.

图2是本发明方法涉及的大规模分布式天线系统中抑制导频污染的多小区预编码方法流程框图。Fig. 2 is a flowchart of a multi-cell precoding method for suppressing pilot pollution in a large-scale distributed antenna system involved in the method of the present invention.

具体实施方式detailed description

本发明提出的大规模分布式天线系统中抑制导频污染的多小区预编码方法结合附图及实施例说明如下：The multi-cell precoding method for suppressing pilot pollution in the large-scale distributed antenna system proposed by the present invention is described as follows in conjunction with the accompanying drawings and embodiments:

本发明提出的大规模分布式天线系统中抑制导频污染的多小区预编码方法，The multi-cell precoding method for suppressing pilot pollution in the large-scale distributed antenna system proposed by the present invention,

本方法涉及L个小区组成的大规模分布式天线系统，设地理位置处于中心的中心小区中的远端接入单元RAU数为M，每个RAU配备的天线数为N，中心小区中的用户数为K，中心小区周围L-1个小区中的总用户数为R，每个用户配置一根天线，其中L、M、N、K、R均为正整数；对大规模分布式天线系统中的用户进行编号，中心小区的用户编号为1,2,…,K，中心小区周围L-1个小区中的用户编号为K+1,K+2,…,K+R，则大规模分布式天线系统中u第个用户到中心小区的第m个RAU的上行信道为h_mu＝λ_mug_mu，其中λ_mu表示大规模分布式天线系统中第u个用户到中心小区的第m个RAU的大尺度衰落，g_mu表示大规模分布式天线系统中第u个用户到中心小区的第m个RAU的小尺度信道，代表复数域，1≤u≤K+R，1≤m≤M；设大规模分布式天线系统中第u个用户在信道估计阶段发射的导频序列为其中τ为导频序列的长度，该长度大于大规模分布式天线系统中所有小区用户数的最大值，第u个用户在信道估计阶段发射导频序列的功率为ρ_u；各导频序列通过用户的天线发射，经过信道，在中心小区的第m个RAU的N根天线得到RAU接收信号，记为其中，表u个用户在信道估计阶段发射的导频序列的转置，N_m表示中心小区的第m个RAU的接收信号中的加性高斯白噪声，矩阵N_m中的元素独立同分布于均值为0、方差为σ²的循环复高斯分布，1≤m≤M；This method involves a large-scale distributed antenna system composed of L cells, assuming that the number of remote access units RAUs in the center cell in the center is M, the number of antennas each RAU is equipped with is N, and the users in the center cell The number is K, the total number of users in the L-1 cells around the central cell is R, and each user is configured with an antenna, where L, M, N, K, and R are all positive integers; for large-scale distributed antenna systems The users in the central cell are numbered, the user numbers in the central cell are 1, 2,...,K, and the user numbers in the L-1 cells around the central cell are K+1, K+2,...,K+R, then the large-scale In the distributed antenna system, the uplink channel from the uth user to the mth RAU of the central cell is h_mu = λ_mu g_mu , where λ_mu represents the large-scale fading from the u-th user to the m-th RAU of the central cell in the large-scale distributed antenna system, and g_mu represents the m-th RAU from the u-th user to the central cell in the large-scale distributed antenna system The small-scale channel of Represents the complex domain, 1≤u≤K+R, 1≤m≤M; suppose the pilot sequence transmitted by the uth user in the channel estimation stage in the large-scale distributed antenna system is Where τ is the length of the pilot sequence, which is greater than the maximum number of users in all cells in the large-scale distributed antenna system. The power of the uth user transmitting the pilot sequence in the channel estimation stage is ρ_u ; each pilot sequence passes The user's antenna transmits, passes through the channel, and receives the signal received by the RAU from the N antennas of the mth RAU in the central cell, denoted as in, Table the transposition of the pilot sequence transmitted by u users in the channel estimation phase, N_m represents the additive white Gaussian noise in the received signal of the mth RAU of the central cell, The elements in the matrix N_m are independently and identically distributed in a cyclic complex Gaussian distribution with a mean of 0 and a variance of σ² , 1≤m≤M;

该方法包括以下步骤：The method includes the following steps:

(1)计算所述中心小区的第m个RAU的N根天线接收到的信号Y_m各列的协方差矩阵Φ_m，其中1≤m≤M，表示第k用户在信道估计阶段发射的导频序列的共轭转置，I_τ为τ阶单位矩阵；(1) Calculate the covariance matrix Φ_m of each column of the signal Y_m received by the N antennas of the m-th RAU of the central cell, where 1≤m≤M, Represents the conjugate transpose of the pilot sequence transmitted by the kth user in the channel estimation stage, I_τ is the identity matrix of order τ;

(2)用共轭梯度法计算大规模分布式天线系统中第u个用户到中心小区的第m个RAU的上行信道h_mu的估计值具体过程如下：(2) Calculate the estimated value of the uplink channel_hmu from the uth user to the mth RAU of the central cell in the large-scale distributed antenna system by using the conjugate gradient method The specific process is as follows:

(2-1)计算大规模分布式天线系统中第u个用户使用的导频序列π_u的共轭经过所述协方差矩阵Φ_m的逆矩阵的滤波过程得到的向量f_mu，其中表示π_u的共轭，是Φ_m的逆矩阵；(2-1) Calculate the vector f_mu obtained by the conjugate of the pilot sequence π_u used by the uth user in the large-scale distributed antenna system through the filtering process of the inverse matrix of the covariance matrix Φ_m , in represents the conjugate of π_u , is the inverse matrix of Φ_m ;

(2-2)建立一个所述滤波过程的等效线性方程组模型采用共轭梯度法，求解等效线性方程组模型，得到大规模分布式天线系统中第u个用户使用的导频序列的共轭经过滤波的向量f_mu，具体步骤如下：(2-2) Set up an equivalent linear equations model of the filtering process Use the conjugate gradient method to solve the equivalent linear equation model to obtain the conjugate of the pilot sequence used by the uth user in the large-scale distributed antenna system go through Filtered vector f_mu , the specific steps are as follows:

(2-2-1)初始化：设置迭代次数阈值P，设向量f_mu的初始值为则所述模型的初始余向量为r₀，初始化共轭基向量为p₀，p₀＝r₀，设置迭代变量t＝0；(2-2-1) Initialization: set the iteration threshold P, Let the initial value of the vector f_mu be then the model The initial covector of is r₀ , Initialize the conjugate basis vector as p₀ , p₀ =r₀ , and set the iteration variable t=0;

(2-2-2)更新迭代变量使t＝t+1，计算第t次迭代过程中共轭基向量系数α_t，其中表示共轭基向量p_t-1的共轭转置，并根据该系数α_t，计算第t次迭代过程中的向量f_mu的值(2-2-2) Update the iterative variables to make t=t+1, and calculate the conjugate basis vector coefficient α_t in the t-th iteration process, in Represents the conjugate transpose of the conjugate base vector p_t-1 , and according to the coefficient α_t , calculate the value of the vector f_mu during the tth iteration

(2-2-3)根据计算第t次迭代后所述模型的余向量r_t，r_t＝r_t-1-α_tΦ_mp_t-1；(2-2-3) According to Calculating the covector r_t of the model after the tth iteration, r_t = r_t-1 -α_t Φ_m p_t-1 ;

(2-2-4)计算第t次迭代过程中共轭基向量的调节系数β_t，并根据该系数计算第t+1次迭代过程中的共轭基向量p_t，p_t＝r_t+β_tp_t-1；(2-2-4) Calculate the adjustment coefficient β_t of the conjugate basis vector in the t-th iteration process, And calculate the conjugate basis vector p_t in the t+1th iteration process according to the coefficient, p_t = r_t +β_t p_t-1 ;

(2-2-5)根据所述迭代阈值P对迭代变量t进行判断，若t≥P，则停止迭代，并使进行步骤(2-3)，若t＜P，则进行步骤(2-2-2)-(2-2-5)；(2-2-5) judge the iteration variable t according to the iteration threshold P, if t≥P, then stop the iteration, and make Carry out step (2-3), if t<P, then carry out step (2-2-2)-(2-2-5);

(2-3)根据所述步骤(2-2)得到的向量计算大规模分布式天线系统中第u个用户到中心小区的第m个RAU的上行信道h_mu的估计值其中，1≤u≤K+R，，1≤m≤M；(2-3) Calculate the estimated value of the uplink channel h_mu of the uth user to the mth RAU of the central cell in the large-scale distributed antenna system according to the vector obtained in the step (2-2) Among them, 1≤u≤K+R, 1≤m≤M;

(3)计算所述中心小区的第m个RAU的上行信道h_jmlk的估计误差的协方差矩阵C_mu，$C_{mu}={\mathrm{\&lambda;}}_{mu}(1-\frac{{\mathrm{\&rho;}}_u{\mathrm{\&tau;\&lambda;}}_{mu}}{{\mathrm{\&pi;}}_u^H\left(\underset{k=1}{\overset{K+R}{\&Sigma;}}{\mathrm{\&rho;}}_k{\mathrm{\&lambda;}}_{mk}{\mathrm{\&pi;}}_k\right)+{\mathrm{\&sigma;}}^2})I_N,$其中，1≤u≤K+R，1≤m≤M；(3) Calculating the covariance matrix C_mu of the estimation error of the uplink channel h_jmlk of the mth RAU of the central cell,$C_{mu}={\mathrm{\&lambda;}}_{mu}(1-\frac{{\mathrm{\&rho;}}_u{\mathrm{\&tau;\&lambda;}}_{mu}}{{\mathrm{\&pi;}}_u^h\left(\underset{k=1}{\overset{K+R}{\&Sigma;}}{\mathrm{\&rho;}}_k{\mathrm{\&lambda;}}_{mk}{\mathrm{\&pi;}}_k\right)+{\mathrm{\&sigma;}}^2})I_N,$ Among them, 1≤u≤K+R, 1≤m≤M;

(4)构造大规模分布式天线系统中第u个用户到中心小区的所有M个RAU的信道向量h_u的估计值其中，1≤u≤K+R，表示步骤(2)中获得的大规模分布式天线系统中第u个用户到中心小区的第m个RAU的上行信道h_mu的估计值的转置，1≤m≤M；(4) Construct the estimated value of the channel vector h_u of all M RAUs from the uth user to the central cell in the large-scale distributed antenna system Among them, 1≤u≤K+R, Indicates the estimated value of the uplink channel_hmu from the uth user to the mth RAU of the central cell in the large-scale distributed antenna system obtained in step (2) Transpose of , 1≤m≤M;

(5)构造大规模分布式天线系统中所有K+R个用户到中心小区的所有M个RAU的信道矩阵H₀的估计值计算H₀的估计误差的协方差矩阵C₀，中心小区的所有K个用户到中心小区的所有M个RAU的信道矩阵H的估计值为其中为的前K列，(5) Construct the estimated value of the channel matrix H₀ of all K+R users to all M RAUs in the central cell in the large-scale distributed antenna system Compute the covariance matrix C₀ of the estimation error of H₀ , The estimated channel matrix H from all K users in the central cell to all M RAUs in the central cell is in for The first K columns of ,

(6)用共轭梯度法计算中心小区的所有K个用户的数据向量s经过多小区预编码后得到的中心小区的发射信号向量x，具体过程如下：(6) Use the conjugate gradient method to calculate the data vector s of all K users in the central cell and obtain the transmitted signal vector x of the central cell after multi-cell precoding. The specific process is as follows:

(6-1)利用步骤(5)中得到的大规模分布式天线系统中心小区所有K个用户到中心小区的所有M个RAU的信道矩阵H的估计值计算s经过匹配滤波后的向量z，(6-1) Using the estimated value of the channel matrix H of all K users in the central cell of the large-scale distributed antenna system obtained in step (5) to all M RAUs in the central cell Calculate s after Matched filtered vector z,

(6-2)计算中心小区的多小区最小均方误差均衡矩阵W，其中表示的共轭转置，I_MN表示MN阶单位矩阵；(6-2) Calculate the multi-cell minimum mean square error equalization matrix W of the central cell, in express The conjugate transpose of , I_MN represents the MN order identity matrix;

(6-3)建立一个大规模分布式天线系统中心小区进行多小区最小均方误差预编码的等效线性方程组模型z＝Wx，采用共轭梯度法，求解该等效线性方程组模型，得到中心小区的发射信号向量x，具体步骤如下：(6-3) Establish a large-scale distributed antenna system center cell to carry out the equivalent linear equation model z=Wx of multi-cell minimum mean square error precoding, adopt the conjugate gradient method to solve the equivalent linear equation model, To obtain the transmit signal vector x of the central cell, the specific steps are as follows:

(6-3-1)初始化：设置迭代次数阈值V，设中心小区的发射信号向量x的初始值为x⁽⁰⁾＝0，则所述模型z＝Wx的初始余向量为v₀，v₀＝z，初始化共轭基向量为q₀，q₀＝r₀，设置迭代变量ω＝0；(6-3-1) Initialization: set the iteration threshold V, Assuming that the initial value of the transmit signal vector x of the central cell is x⁽⁰⁾ =0, then the initial covector of the model z=Wx is v₀ , v₀ =z, and the initial conjugate basis vector is q₀ , q₀ ＝r₀ , set iteration variable ω＝0;

(6-3-2)更新迭代变量使ω＝ω+1，计算第ω次迭代过程中共轭基向量系数μ_t，其中表示共轭基向量q_t-1的共轭转置，并根据该系数μ_t，计算第ω次迭代过程中的中心小区的发射信号向量x的值x^(t)，x^(t)＝x^(t-1)+μ_tq_t-1；(6-3-2) Update the iterative variable so that ω=ω+1, calculate the conjugate basis vector coefficient μ_t in the ωth iteration process, in Represents the conjugate transpose of the conjugate base vector q_t-1 , and according to the coefficient μ_t , calculate the value x^(t) of the transmitted signal vector x of the central cell in the ωth iteration process, x^(t) = x^(t-1) + μ_t q_t-1 ;

(6-3-3)根据x^(t)计算第ω次迭代后所述模型的余向量v_t，v_t＝v_t-1-μ_tWq_t-1；(6-3-3) Calculate the covector v_t of the model after the ωth iteration according to x^(t) , v_t =v_t-1- μ_t Wq_t-1 ;

(6-3-4)计算第ω次迭代过程中共轭基向量的调节系数θ_t，并根据该系数计算第ω+1次迭代过程中的共轭基向量q_t，q_t＝v_t+β_tq_t-1；(6-3-4) Calculate the adjustment coefficient θ_t of the conjugate basis vector in the ωth iteration process, And calculate the conjugate basis vector q_t in the iterative process of ω+1 according to this coefficient, q_t =v_t +β_t q_t-1 ;

(6-3-5)根据所述迭代阈值V对迭代变量ω进行判断，若ω≥V，则停止迭代，并使x＝x^(t-1)，进行步骤(6-4)，若ω＜V，则进行步骤(6-3-2)-(6-3-5)；(6-3-5) Judge the iteration variable ω according to the iteration threshold V, if ω≥V, then stop the iteration, and make x=x^(t-1) , go to step (6-4), if ω <V, then proceed to steps (6-3-2)-(6-3-5);

(7)对所述步骤(6)得到的得到中心小区的发射信号向量x进行归一化，使得到中心小区的各RAU上的天线最终发射的实际信号向量，实现大规模分布式天线系统中抑制导频污染的多小区预编码。(7) normalize the transmitted signal vector x obtained by the central cell obtained in the step (6), so that The actual signal vector finally transmitted by the antennas on each RAU of the central cell is obtained, and multi-cell precoding that suppresses pilot pollution in a large-scale distributed antenna system is realized.

Claims (5)

Translated fromChinese

1.一种大规模分布式天线系统中抑制导频污染的多小区预编码方法，其特征在于，包括以下步骤：1. a multi-cell precoding method for suppressing pilot pollution in a large-scale distributed antenna system, is characterized in that, comprises the following steps:(1)设在以要进行预编码的中心小区为中心的包括L个小区的大规模分布式天线系统中，中心小区中的RAU数为M，每个RAU配备的天线数为N，中心小区中的用户数为K，中心小区周围L-1个小区中的总用户数为R，每个用户配置一根天线；对大规模分布式天线系统中的用户进行编号，中心小区的用户编号为1,2,…,K，中心小区周围L-1个小区中的用户编号为K+1,K+2,…,K+R，则大规模分布式天线系统中u第个用户到中心小区的第m个RAU的上行信道为其中λ_mu表示大规模分布式天线系统中第u个用户到中心小区的第m个RAU的大尺度衰落，g_mu表示大规模分布式天线系统中第u个用户到中心小区的第m个RAU的小尺度信道，代表复数域，1≤u≤K+R，1≤m≤M；设大规模分布式天线系统中第u个用户在信道估计阶段发射的导频序列为其中τ为导频序列的长度，该长度大于大规模分布式天线系统中所有小区用户数的最大值，第u个用户在信道估计阶段发射导频序列的功率为ρ_u；各导频序列通过用户的天线发射，经过信道，在中心小区的第m个RAU的N根天线得到RAU接收信号，记为其中，表u个用户在信道估计阶段发射的导频序列的转置，N_m表示中心小区的第m个RAU的接收信号中的加性高斯白噪声，矩阵N_m中的元素独立同分布于均值为0、方差为σ²的循环复高斯分布，1≤m≤M；(1) Set in a large-scale distributed antenna system including L cells centered on the center cell to be precoded, the number of RAUs in the center cell is M, the number of antennas each RAU is equipped with is N, and the center cell The number of users in the system is K, the total number of users in the L-1 cells around the central cell is R, and each user is configured with an antenna; the users in the large-scale distributed antenna system are numbered, and the user number of the central cell is 1,2,...,K, the number of users in the L-1 cells around the central cell is K+1, K+2,...,K+R, then the uth user in the large-scale distributed antenna system goes to the central cell The uplink channel of the mth RAU is where λ_mu represents the large-scale fading from the u-th user to the m-th RAU of the central cell in the large-scale distributed antenna system, and g_mu represents the m-th RAU from the u-th user to the central cell in the large-scale distributed antenna system The small-scale channel of Represents the complex domain, 1≤u≤K+R, 1≤m≤M; suppose the pilot sequence transmitted by the uth user in the channel estimation stage in the large-scale distributed antenna system is Where τ is the length of the pilot sequence, which is greater than the maximum number of users in all cells in the large-scale distributed antenna system. The power of the uth user transmitting the pilot sequence in the channel estimation stage is ρ_u ; each pilot sequence passes The user's antenna transmits, passes through the channel, and receives the signal received by the RAU from the N antennas of the mth RAU in the central cell, denoted as in, Table the transposition of the pilot sequence transmitted by u users in the channel estimation phase, N_m represents the additive white Gaussian noise in the received signal of the mth RAU of the central cell, The elements in the matrix N_m are independently and identically distributed in a cyclic complex Gaussian distribution with a mean of 0 and a variance of σ² , 1≤m≤M;(2)计算所述中心小区的第m个RAU的N根天线接收到的信号Y_m各列的协方差矩阵Φ_m，其中1≤m≤M，表示第k个用户在信道估计阶段发射的导频序列的共轭转置，I_τ为τ阶单位矩阵；(2) Calculate the covariance matrix Φ_m of each column of the signal Y_m received by the N antennas of the m-th RAU of the central cell, where 1≤m≤M, Represents the conjugate transpose of the pilot sequence transmitted by the kth user in the channel estimation stage, I_τ is the identity matrix of order τ;(3)用共轭梯度法计算大规模分布式天线系统中第u个用户到中心小区的第m个RAU的上行信道h_mu的估计值(3) Calculate the estimated value of the uplink channel_hmu from the uth user to the mth RAU of the central cell in the large-scale distributed antenna system by using the conjugate gradient method(4)计算上述中心小区的第m个RAU的上行信道h_jmlk的估计误差的协方差矩阵C_mu，其中，1≤u≤K+R，1≤m≤M，I_N为N阶单位矩阵，表示第k个用户在信道估计阶段发射的导频序列的共轭转置；(4) Calculate the covariance matrix C_mu of the estimation error of the uplink channel h_jmlk of the mth RAU of the above-mentioned central cell, Among them, 1≤u≤K+R, 1≤m≤M, I_N is the N-order identity matrix, Represents the conjugate transpose of the pilot sequence transmitted by the kth user in the channel estimation phase;(5)构造大规模分布式天线系统中第u个用户到中心小区的所有M个RAU的信道向量h_u的估计值其中，1≤u≤K+R，表示步骤(3)中获得的大规模分布式天线系统中第u个用户到中心小区的第m个RAU的上行信道h_mu的估计值的转置，1≤m≤M；(5) Construct the estimated value of the channel vector h_u of all M RAUs from the uth user to the central cell in the large-scale distributed antenna system Among them, 1≤u≤K+R, Indicates the estimated value of the uplink channel_hmu from the uth user to the mth RAU of the central cell in the large-scale distributed antenna system obtained in step (3) Transpose of , 1≤m≤M;(6)构造大规模分布式天线系统中所有K+R个用户到中心小区的所有M个RAU的信道矩阵H₀的估计值计算H₀的估计误差的协方差矩阵C₀，中心小区的所有K个用户到中心小区的所有M个RAU的信道矩阵H的估计值为其中为的前K列，(6) Construct the estimated value of the channel matrix_H0 of all K+R users to all M RAUs in the central cell in the large-scale distributed antenna system Compute the covariance matrix C₀ of the estimation error of H₀ , The estimated channel matrix H from all K users in the central cell to all M RAUs in the central cell is in for The first K columns of ,(7)用共轭梯度法计算中心小区的所有K个用户的数据向量s经过多小区预编码后得到的中心小区的发射信号向量x；(7) Calculate the data vector s of all K users of the center cell with the conjugate gradient method and obtain the transmit signal vector x of the center cell after multi-cell precoding;(8)对上述步骤(7)得到的得到中心小区的发射信号向量x进行归一化，使得到中心小区的各RAU上的天线最终发射的实际信号向量，实现大规模分布式天线系统中抑制导频污染的多小区预编码。(8) normalize the transmitted signal vector x of the central cell obtained in the above step (7), so that The actual signal vector finally transmitted by the antennas on each RAU of the central cell is obtained, and multi-cell precoding that suppresses pilot pollution in a large-scale distributed antenna system is realized.2.如权利要求1所述方法，其特征在于，所述步骤(3)具体步骤如下：2. method as claimed in claim 1, is characterized in that, described step (3) concrete steps are as follows:(3-1)计算大规模分布式天线系统中第u个用户使用的导频序列π_u的共轭经过所述Φ_m的逆矩阵滤波得到的向量f_mu，其中表示π_u的共轭，是Φ_m的逆矩阵；(3-1) Calculate the vector f_mu obtained by filtering the conjugate of the pilot sequence π_u used by the uth user in the large-scale distributed antenna system through the inverse matrix filtering of Φ_m , in represents the conjugate of π_u , is the inverse matrix of Φ_m ;(3-2)建立一个所述滤波过程的等效线性方程组模型采用共轭梯度法，求解等效线性方程组模型，得到大规模分布式天线系统中第u个用户使用的导频序列的共轭经过滤波的向量f_mu；(3-2) Set up an equivalent linear equations model of the filtering process Use the conjugate gradient method to solve the equivalent linear equation model to obtain the conjugate of the pilot sequence used by the uth user in the large-scale distributed antenna system go through Filtered vector f_mu ;(3-3)根据上述步骤(3-2)得到的向量计算大规模分布式天线系统中第u个用户到中心小区的第m个RAU的上行信道h_mu的估计值其中，1≤u≤K+R，1≤m≤M。(3-3) According to the vector obtained in the above step (3-2), calculate the estimated value of the uplink channel_hmu from the uth user to the mth RAU of the central cell in the large-scale distributed antenna system Among them, 1≤u≤K+R, 1≤m≤M.3.如权利要求2所述方法，其特征在于，所述步骤(3-2)具体步骤如下：3. method as claimed in claim 2, is characterized in that, described step (3-2) concrete steps are as follows:(3-2-1)初始化：设置迭代次数阈值P，设向量f_mu的初始值为则上述模型的初始余向量为r₀，初始化共轭基向量为p₀，p₀＝r₀，设置迭代变量t＝0；(3-2-1) Initialization: set the iteration threshold P, and set the initial value of the vector f_mu to Then the above model The initial covector of is r₀ , Initialize the conjugate basis vector as p₀ , p₀ =r₀ , and set the iteration variable t=0;(3-2-2)更新迭代变量使t＝t+1，计算第t次迭代过程中共轭基向量系数α_t，其中表示共轭基向量p_t-1的共轭转置，并根据该系数α_t，计算第t次迭代过程中的向量f_mu的值(3-2-2) Update the iterative variable so that t=t+1, and calculate the conjugate basis vector coefficient α_t in the tth iteration process, in Represents the conjugate transpose of the conjugate base vector p_t-1 , and according to the coefficient α_t , calculate the value of the vector f_mu during the tth iteration(3-2-3)根据计算第t次迭代后上述模型的余向量r_t，r_t＝r_t-1-α_tΦ_mp_t-1；(3-2-3) According to Calculate the covector r_t of the above model after the tth iteration, r_t = r_t-1 -α_t Φ_m p_t-1 ;(3-2-4)计算第t次迭代过程中共轭基向量的调节系数β_t，并根据该系数计算第t+1次迭代过程中的共轭基向量p_t，p_t＝r_t+β_tp_t-1；(3-2-4) Calculate the adjustment coefficient β_t of the conjugate basis vector in the t-th iteration process, And calculate the conjugate basis vector p_t in the t+1th iteration process according to the coefficient, p_t = r_t +β_t p_t-1 ;(3-2-5)根据上述迭代阈值P对迭代变量t进行判断，若t≥P，则停止迭代，并使进行步骤(3-3)，若t＜P，则进行步骤(3-2-2)-(3-2-5)。(3-2-5) Judge the iteration variable t according to the above iteration threshold P, if t≥P, stop the iteration, and use Go to step (3-3), if t<P, go to steps (3-2-2)-(3-2-5).4.如权利要求1所述方法，其特征在于，所述步骤(7)具体过程如下：4. method as claimed in claim 1, is characterized in that, described step (7) specific process is as follows:(7-1)利用步骤(6)中得到的大规模分布式天线系统中心小区所有K个用户到中心小区的所有M个RAU的信道矩阵H的估计值计算s经过匹配滤波后的向量z，(7-1) Using the estimated value of the channel matrix H of all K users in the central cell of the large-scale distributed antenna system obtained in step (6) to all M RAUs in the central cell Calculate s after Matched filtered vector z,(7-2)计算中心小区的多小区最小均方误差均衡矩阵W，其中表示的共轭转置，I_MN表示MN阶单位矩阵；(7-2) Calculate the multi-cell minimum mean square error equalization matrix W of the central cell, in express The conjugate transpose of , I_MN represents the MN order identity matrix;(7-3)建立一个大规模分布式天线系统中心小区进行多小区最小均方误差预编码的等效线性方程组模型z＝Wx，采用共轭梯度法，求解该等效线性方程组模型，得到中心小区的发射信号向量x。(7-3) Establish a large-scale distributed antenna system central cell to carry out the equivalent linear equation model z=Wx of multi-cell minimum mean square error precoding, adopt the conjugate gradient method to solve the equivalent linear equation model, Get the transmit signal vector x of the central cell.5.如权利要求4所述方法，其特征在于，所述步骤(7-3)具体步骤如下：5. method as claimed in claim 4, is characterized in that, described step (7-3) concrete steps are as follows:(7-3-1)初始化：设置迭代次数阈值V，设中心小区的发射信号向量x的初始值为x⁽⁰⁾，则上述模型z＝Wx的初始余向量为v₀，v₀＝z，初始化共轭基向量为q₀，q₀＝r₀，设置迭代变量ω＝0；(7-3-1) Initialization: set the iteration threshold V, set the initial value of the transmitted signal vector x of the central cell to x⁽⁰⁾ , then the initial covector of the above model z=Wx is v₀ , v₀ =z , initialize the conjugate basis vector as q₀ , q₀ =r₀ , set the iteration variable ω=0;(7-3-2)更新迭代变量使ω＝ω+1，计算第ω次迭代过程中共轭基向量系数μ_t，其中表示共轭基向量q_t-1的共轭转置，并根据该系数μ_t，计算第ω次迭代过程中的中心小区的发射信号向量x的值x^(t)，x^(t)＝x^(t-1)+μ_tq_t-1；(7-3-2) Update the iterative variables to make ω=ω+1, calculate the conjugate basis vector coefficient μ_t in the ωth iteration process, in Represents the conjugate transpose of the conjugate base vector q_t-1 , and according to the coefficient μ_t , calculate the value x^(t) of the transmitted signal vector x of the central cell in the ωth iteration process, x^(t) = x^(t-1) + μ_t q_t-1 ;(7-3-3)根据x^(t)计算第ω次迭代后上述模型的余向量v_t，v_t＝v_t-1-μ_tWq_t-1；(7-3-3) Calculate the covector v_t of the above-mentioned model after the ωth iteration according to x^(t) , v_t =v_t-1- μ_t Wq_t-1 ;(7-3-4)计算第ω次迭代过程中共轭基向量的调节系数θ_t，并根据该系数计算第ω+1次迭代过程中的共轭基向量q_t，q_t＝v_t+β_tq_t-1；(7-3-4) Calculate the adjustment coefficient θ_t of the conjugate basis vector in the ωth iteration process, And calculate the conjugate basis vector q_t in the iterative process of ω+1 according to this coefficient, q_t =v_t +β_t q_t-1 ;(7-3-5)根据上述迭代阈值V对迭代变量ω进行判断，若ω≥V，则停止迭代，并使x＝x^(t-1)，进行步骤(7-4)，若ω＜V，则进行步骤(7-3-2)-(7-3-5)。(7-3-5) Judge the iterative variable ω according to the above-mentioned iteration threshold V, if ω≥V, then stop the iteration, and set x=x^(t-1) , proceed to step (7-4), if ω< V, then proceed to steps (7-3-2)-(7-3-5).