CN110635836A - A MMSE-PCA Channel Estimation Method for Millimeter-Wave Massive MIMO System Based on Beam Selection - Google Patents

Abstract

Translated fromChinese

本发明提出一种基于波束选择的毫米波大规模MIMO系统MMSE‑PCA(最小均方误差‑主成分分析)信道估计方法。该方法首先在基站侧以MMSE预编码为基准的幅度最大化标准选择波束信号。其次在接收端使用PCA信道压缩方法，将信道进行压缩降维，最后使用经典的最小二乘法(LS)进行信道估计。MMSE预编码可以减小系统中信道噪声和各个用户之间的干扰，减少基站使用射频链路数目，因此可降低系统的实现成本和能量损耗。同时，PCA利用原始高维数据的相关性，将高维的数据压缩到低维。该方法在降低系统复杂度基础上，达到有效优化系统传输性能和提高能量效率的目的。

The present invention proposes a MMSE-PCA (Minimum Mean Square Error-Principal Component Analysis) channel estimation method for a millimeter-wave massive MIMO system based on beam selection. The method first selects beam signals at the base station side based on the amplitude maximization criterion based on MMSE precoding. Secondly, the PCA channel compression method is used at the receiving end to compress the channel for dimension reduction, and finally the classical least squares (LS) method is used for channel estimation. MMSE precoding can reduce channel noise and interference between users in the system, reduce the number of radio frequency links used by the base station, and thus reduce the implementation cost and energy consumption of the system. At the same time, PCA utilizes the correlation of the original high-dimensional data to compress high-dimensional data into low-dimensional data. The method achieves the purpose of effectively optimizing the system transmission performance and improving the energy efficiency on the basis of reducing the system complexity.

Description

Translated fromChinese

一种基于波束选择的毫米波大规模MIMO系统MMSE–PCA信道估计方法A beam selection-based MMSE-PCA channel estimation for millimeter-wave massive MIMO systemcounting method

技术领域：Technical field:

本发明涉及无线通信领域，尤其涉及一种基于波束选择的毫米波大规模MIMO系统MMSE-PCA信道估计的方法。The invention relates to the field of wireless communication, in particular to a method for MMSE-PCA channel estimation of a millimeter-wave massive MIMO system based on beam selection.

背景技术：Background technique:

随着现代电子信息技术的不断发展，我国的移动通信技术取得了卓有成效的成果，4G移动网络在我国的大力推行，极大地改善了人们的生活体验与生产方式。目前，面向2020年的第五代移动通信技术(5G)还处于起步阶段。毫米波大规模多输入多输出(MIMO)是未来5G无线通信的一项关键技术，因为它的带宽更宽和频谱效率更高，可以显著提高数据速率。然而随着移动通信用户数量及无线数据传输速率的飞速增长，现有的频谱资源变得拥挤不堪，已无法达到5G的通信指标要求，因此未被完全开发的毫米波频谱资源在未来5G通信中的应用研究得到了国内外研究学者的关注。一方面，毫米波的带宽可达10GHz，可以为通信系统提供丰富的带宽资源，另一方面，由于无线通信系统中天线尺寸与信号波长成正比，毫米波的波长使得其对应的天线尺寸大大降低，适合在发送端和接收端部署大量的天线，从而获得较高的天线阵列增益。因此，毫米波与大规模MIMO技术的完美结合，将会成为当前通信领域的研究热点。With the continuous development of modern electronic information technology, my country's mobile communication technology has achieved fruitful results. The vigorous implementation of 4G mobile network in my country has greatly improved people's life experience and production methods. At present, the fifth-generation mobile communication technology (5G) for 2020 is still in its infancy. Millimeter-wave massive multiple-input multiple-output (MIMO) is a key technology for future 5G wireless communications because of its wider bandwidth and higher spectral efficiency, which can significantly increase data rates. However, with the rapid growth of the number of mobile communication users and the wireless data transmission rate, the existing spectrum resources have become congested and cannot meet the communication index requirements of 5G. Therefore, the millimeter-wave spectrum resources that have not been fully developed will be used in future 5G communication. The applied research has attracted the attention of domestic and foreign researchers. On the one hand, the bandwidth of millimeter waves can reach 10 GHz, which can provide rich bandwidth resources for communication systems. On the other hand, since the antenna size in wireless communication systems is proportional to the signal wavelength, the wavelength of millimeter waves greatly reduces the corresponding antenna size. , it is suitable to deploy a large number of antennas at the transmitting end and the receiving end, so as to obtain a higher antenna array gain. Therefore, the perfect combination of millimeter wave and massive MIMO technology will become a research hotspot in the current communication field.

然而，在实际应用中实现毫米波大规模MIMO并不是一项简单的工作。一个关键的挑战是，在传统的MIMO系统中的每个天线通常需要一个专用的射频(RF)链(包括数模转换器、上变频器等)。在基带部分一般采用数字预编码技术对发射信号进行预处理，经过预处理的信号可以大大降低系统中的干扰从而使系统性能得到大幅度提升。然而，在全数字预编码方案中，每根天线都要对应一条RF链路，随着基站天线数及用户数的不断增加，系统所需RF链路数也随之增多，导致系统实现成本增加，并且造成巨大的能量损耗。这导致毫米波大规模MIMO系统的硬件成本和能耗难以负担，因为天线数量变得巨大(例如，256根天线)，并且RF链的能耗很高(例如，毫米波频率下每根RF链约250mW)。However, implementing mmWave massive MIMO in practical applications is not a simple task. A key challenge is that each antenna in conventional MIMO systems typically requires a dedicated radio frequency (RF) chain (including digital-to-analog converters, upconverters, etc.). In the baseband part, digital precoding technology is generally used to preprocess the transmitted signal. The preprocessed signal can greatly reduce the interference in the system and thus greatly improve the system performance. However, in the all-digital precoding scheme, each antenna must correspond to an RF link. With the continuous increase of the number of base station antennas and the number of users, the number of RF links required by the system also increases, resulting in an increase in system implementation costs. , and cause huge energy loss. This makes the hardware cost and energy consumption of mmWave massive MIMO systems unaffordable, as the number of antennas becomes huge (eg, 256 antennas) and the energy consumption of the RF chains is high (eg, each RF chain at mmWave frequencies about 250mW).

为了减少所需的RF链数量，最近提出采用透镜线形(ULA)天线阵列(一种具有能量聚焦能力的电磁透镜和一种与位于透镜焦面上的元件相匹配的天线阵)的毫米波大规模MIMO系统。通过采用ULA天线阵，通过将来自不同方向的信号集中在不同天线上，可以将空间信道转换为波束空间信道。由于毫米波频率下的散射并不丰富，因此毫米波通信中的有效路径数目十分有限，仅占用少量波束。因此，毫米波波束空间信道是稀疏的，我们可以根据稀疏的选择少量的主波束。在毫米波大规模MIMO系统中，基站端配置大量的天线阵列元，信号以利用波束成形技术将信号集中在一个块区域空间，可以使得毫米波大规模MIMO路径存在有一定的稀疏特性。利用这一特性，采用近些年研究较为广泛的压缩感知来对信道进行处理。首先依据相关的压缩感知理论的研究用混合预编码器获得毫米波系统的测量矩阵，然后毫米波系统的信道估计问题可以作为一个典型的稀疏信号恢复问题来研究。To reduce the number of RF chains required, recent proposals for millimeter-wave large antennas using a lensed linear (ULA) antenna array (an electromagnetic lens with energy-focusing capabilities and an antenna array matched to elements located on the focal plane of the lens) have been proposed. Massive MIMO system. By employing ULA antenna arrays, spatial channels can be converted to beam spatial channels by concentrating signals from different directions on different antennas. Since the scattering at mmWave frequencies is not abundant, the number of effective paths in mmWave communication is very limited, occupying only a small number of beams. Therefore, the mmWave beam spatial channel is sparse, and we can select a small number of main beams according to the sparseness. In a millimeter-wave massive MIMO system, the base station is equipped with a large number of antenna array elements, and the signals are concentrated in a block space by using beamforming technology, which can make the millimeter-wave massive MIMO path have a certain sparse characteristic. Taking advantage of this characteristic, compressed sensing, which has been widely studied in recent years, is used to process the channel. Firstly, the measurement matrix of the millimeter-wave system is obtained by using the hybrid precoder according to the related compressed sensing theory, and then the channel estimation problem of the millimeter-wave system can be studied as a typical sparse signal recovery problem.

总而言之，解决高能耗，并如何达到降低系统复杂度基础上有效提高系统性能和能量效率的目的，是目前毫米波大规模MIMO系统的信道估计研究所面临的挑战。All in all, how to solve the high energy consumption and how to effectively improve the system performance and energy efficiency on the basis of reducing the system complexity are the challenges faced by the current channel estimation research of millimeter-wave massive MIMO systems.

发明内容：Invention content:

本发明旨在至少解决现有技术中存在的技术问题，特别创新地提出了一种基于波束选择的毫米波大规模MIMO系统MMSE-PCA信道估计的方法。The present invention aims to at least solve the technical problems existing in the prior art, and particularly innovatively proposes a method for MMSE-PCA channel estimation of a millimeter-wave massive MIMO system based on beam selection.

为了实现本发明的上述目的，本发明提供了一种基于波束选择的毫米波大规模MIMO系统MMSE-PCA信道估计方法，其特征在于，包括：In order to achieve the above objects of the present invention, the present invention provides a beam selection-based MMSE-PCA channel estimation method for a millimeter-wave massive MIMO system, which is characterized by comprising:

S1，在基站侧采用以最小均方误差预编码为基准的幅度最大化(MM)标准选择波束信号，并引入最小均方误差线性预编码技术，以减弱噪声的影响和用户间干扰；S1, at the base station side, the amplitude maximization (MM) standard based on the minimum mean square error precoding is used to select the beam signal, and the minimum mean square error linear precoding technology is introduced to reduce the influence of noise and inter-user interference;

S2，采用时分双工(TDD)大规模MIMO系统，根据TDD系统中的信道互易，在上行链路通过最小二乘法(LS)信道估计来获得信道状态信息(CSI)；S2, adopting a Time Division Duplex (TDD) massive MIMO system, according to channel reciprocity in the TDD system, obtains channel state information (CSI) through least squares (LS) channel estimation in the uplink;

S3，采用Saleh-Valenzuela信道模型体现信道稀疏特性，在接收端使用主成分分析(PCA)信道压缩方法，把CSI从高维映射到低维，用于降低特征维度；S3, use the Saleh-Valenzuela channel model to reflect the channel sparse characteristics, and use the principal component analysis (PCA) channel compression method at the receiving end to map the CSI from high-dimensional to low-dimensional to reduce the feature dimension;

S4，接收端将信道进行压缩降维之后，再采用LS进行信道估计。S4, after the receiving end compresses and reduces the dimension of the channel, LS is used to estimate the channel.

所述的一种基于波束选择的毫米波大规模MIMO系统MMSE-PCA信道估计的方法，其特征在于，所述S1包括：The method for MMSE-PCA channel estimation of a millimeter-wave massive MIMO system based on beam selection is characterized in that, the S1 includes:

基站侧采用以MMSE预编码为基准的MM标准选择波束信号，在迫零(ZF)算法的基础上引入了MMSE线性预编码技术，MMSE预编码矩阵表达式为：The base station side uses the MM standard based on MMSE precoding to select beam signals, and introduces the MMSE linear precoding technology based on the zero-forcing (ZF) algorithm. The MMSE precoding matrix expression is:

其中，β表示功率控制因子，||·||₂表示求2范数，E(·)表示求期望；为计算简单，MMSE预编码算法的优化问题可以看作是在一定的功率约束条件下求解接收信号与发送信号的最小均方误差的问题；在此基础上，建立目标函数：Among them, β represents the power control factor, ||·||₂ represents the 2-norm, and E( ) represents the expectation; for the sake of simplicity, the optimization problem of the MMSE precoding algorithm can be regarded as a certain power constraint condition. Solve the problem of the minimum mean square error of the received signal and the transmitted signal; on this basis, establish the objective function:

其中，P表示信号的最大发射功率；根据MMSE准则，得到预编码矩阵为：Among them, P represents the maximum transmit power of the signal; according to the MMSE criterion, the obtained precoding matrix is:

其中，σ²为噪声功率，功率控制因子β为：Among them, σ² is the noise power, and the power control factor β is:

其中，Tr(H)表示矩阵的迹，(H)^-1表示矩阵的逆，H^H表示矩阵的共轭转置。Among them, Tr(H) represents the trace of the matrix, (H)^-1 represents the inverse of the matrix, and H^H represents the conjugate transpose of the matrix.

所述的一种于波束选择的毫米波大规模MIMO系统MMSE-PCA信道估计方法，其特征在于，所述S2包括：The method for MMSE-PCA channel estimation in a millimeter-wave massive MIMO system based on beam selection is characterized in that, the S2 includes:

天线矩阵U表达式为：The antenna matrix U is expressed as:

式中：

N表示空间方位角，基于3D波束空间的大规模MIMO的系统模型接收信号可表示为：where:

N represents the spatial azimuth, and the received signal of the massive MIMO system model based on 3D beam space can be expressed as:

式中

为波束空间的接收信号矢量，将信道矢量与转化成波束空间的信道矢量后，转化方式为：

包含了h_k的所有信道信息，可用于估计整个CSI；那么波束空间的信道矩阵可定义为：

表示下行波束空间信道矩阵；in the formula

is the received signal vector in the beam space, after converting the channel vector and into the channel vector in the beam space, the conversion method is:

Contains all the channel information of h_k and can be used to estimate the entire CSI; then the channel matrix of the beam space can be defined as:

represents the downlink beam space channel matrix;

系统模型如图1所示，上行链路通过LS信道估计来获得CSI，这个过程中每个用户需要在Q时刻向基站发送正交导频序列ψ_m，假设将Q时刻分成为M个块，每个块由K个时刻组成，根据TDD系统中的信道互易，在第m块的基站处接收的上行链路信号矢量

可以表示为：The system model is shown in Figure 1. The uplink obtains CSI through LS channel estimation. In this process, each user needs to send an orthogonal pilot sequence ψ_m to the base station at time Q. It is assumed that the time Q is divided into M blocks, Each block consists of K time instants, the uplink signal vector received at the base station of the mth block according to the channel reciprocity in the TDD system

It can be expressed as:

通过自适应选择网络，基站端用维数为K×N的模拟组合器W_m来组合出并通过射频链在基带采样中获得维数为K×K的采样信号R_m，其中R_m表达式为：Through the adaptive selection of the network, the base station uses an analog combiner W_m with a dimension of K × N to combine And the sampling signal R_m with dimension K×K is obtained in the baseband sampling through the radio frequency chain, where the expression of R_m is:

最后，将降低维度的信号与正交导频矩阵相乘，获得波束空间信道

的检测矩阵Z_m，Finally, combine the reduced-dimensional signal with the orthogonal pilot matrix Multiply to get the beam space channel

The detection matrix Z_m of ,

其中表示有效噪声矩阵，在TDD系统中，根据上行链路估计得到的CSI，由于信道互易性，可以作为下行信道的CSI。in Represents the effective noise matrix. In the TDD system, the CSI estimated according to the uplink can be used as the CSI of the downlink channel due to channel reciprocity.

所述的一种基于波束选择的毫米波大规模MIMO系统MMSE-PCA信道估计的方法，其特征在于，所述S3包括：The method for MMSE-PCA channel estimation of a millimeter-wave massive MIMO system based on beam selection is characterized in that, the S3 includes:

由于毫米波通信中有效路径数量有限，因此H具有稀疏结构特性，毫米波Saleh-Valenzuela信道模型如图2所示；Due to the limited number of effective paths in millimeter-wave communication, H has a sparse structure, and the millimeter-wave Saleh-Valenzuela channel model is shown in Figure 2;

那么基于波束空间的大规模MIMO系统降低维度的信号可以表示为：Then the dimensionally reduced signal of a massive MIMO system based on beam space can be expressed as:

式中，

B表示所选波束的集合，P_r为已降低维数的预编码矩阵；为了实现近乎最佳的性能，基站需要获得具有有限数量的射频链的3D波東空间信道，为了保证K个用户的空间复用增益，所需射频链的最小数目应为N_RF＝K，所以考虑射频链数目为N_RF＝K，且不损失一般性；In the formula,

B represents the set of selected beams, and P_r is the precoding matrix with reduced dimensionality; in order to achieve near-optimal performance, the base station needs to obtain 3D Bodong spatial channels with a limited number of radio frequency chains, in order to ensure the For spatial multiplexing gain, the minimum number of RF chains required should be N_RF =K, so consider the number of RF chains as N_RF =K without loss of generality;

信号通过MM-MMSE进行预编码形成波束发送出去后，接收端接收到预编码信号，然后通过PCA对CSI进行降维；After the signal is precoded by MM-MMSE to form a beam and sent out, the receiver receives the precoded signal, and then uses PCA to reduce the dimensionality of the CSI;

基于低复杂度PCA的CSI算法中，首先对协方差矩阵进行特征值分解，求出C_H的特征值和特征向量，表示为：In the CSI algorithm based on low-complexity PCA, the covariance matrix is first decomposed into eigenvalues, and the eigenvalues and eigenvectors of_CH are obtained, which are expressed as:

其中，

是对角矩阵，其对角元素是协方差矩阵的特征值，

的列向量是协方差矩阵C_H的特征向量；in,

is a diagonal matrix whose diagonal elements are the eigenvalues of the covariance matrix,

The column vector of is the eigenvector of the covariance matrix_CH ;

然后将特征值按照由大到小进行排列，选取特征值贡献率超过阈值γ的前l个特征值所对应的特征向量，组成压缩矩阵

Then the eigenvalues are arranged in descending order, and the eigenvectors corresponding to the first l eigenvalues whose eigenvalue contribution rate exceeds the threshold γ are selected to form a compression matrix

其次利用压缩矩阵

将高维下行信道信息矩阵H_r压缩到低维空间，表示为：Second, use the compression matrix

The high-dimensional downlink channel information matrix H_r is compressed into a low-dimensional space, expressed as:

其中表示降维后的信道矩阵；in represents the channel matrix after dimension reduction;

基于PCA算法的反馈量压缩比为：The feedback compression ratio based on the PCA algorithm is:

r_PCA＝l(N_r+N_t)/(N_r×N_t)r_PCA =l(N_r +N_t )/(N_r ×N_t )

采用非码本的反馈，反馈量是通过压缩比来定义，压缩比越小，所需反馈开销越低；Using non-codebook feedback, the amount of feedback is defined by the compression ratio. The smaller the compression ratio, the lower the required feedback overhead;

最后用户将

以及压缩矩阵

反馈给基站端，基站端接收到反馈信息后利用同样的压缩矩阵恢复原始信道，表示为；The end user will

and the compression matrix

Feedback to the base station, after receiving the feedback information, the base station uses the same compression matrix to restore the original channel, which is expressed as;

其中

表示信道的恢复值。in

Represents the recovery value of the channel.

所述的一种基于波束选择的毫米波大规模MIMO系统MMSE-PCA信道估计方法，其特征在于，所述S4包括：The beam selection-based MMSE-PCA channel estimation method for a millimeter-wave massive MIMO system is characterized in that, the S4 includes:

接收端用LS算法估计出CSI；The receiver uses the LS algorithm to estimate the CSI;

令信道矩阵为H，接收信号矩阵为发送信号矩阵为X，其估计可表示为：Let the channel matrix be H and the received signal matrix be The transmitted signal matrix is X, and its estimate can be expressed as:

为得到具体表达式，对上式求偏导，并令偏导为0，可得：In order to obtain the specific expression, take the partial derivative of the above formula, and set the partial derivative to 0, we can get:

解之，得到信道估计为：Solving it, the channel estimate is obtained as:

若X是非奇异矩阵，可以使用X的穆尔-彭罗斯逆。If X is a nonsingular matrix, the Moore-Penrose inverse of X can be used.

综上所述，由于采用了上述技术方案，本发明的有益效果是：To sum up, due to the adoption of the above-mentioned technical solutions, the beneficial effects of the present invention are:

通过采用时分双工(TDD)的3D波束空间系统模型，联合波束成形技术并提出一种基于波束选择的毫米波大规模MIMO系统MMSE-PCA(最小均方误差-主成分分析)信道估计算法。最终可达到改善大规模MIMO系统中存在多用户干扰较为严重等不利于信道估计的问题的目的，在减少系统复杂度的基础上提高信道的信息传输效率，并优化系统性能。A MMSE-PCA (Minimum Mean Square Error-Principal Component Analysis) channel estimation algorithm for millimeter-wave massive MIMO system based on beam selection is proposed by adopting the 3D beam space system model of Time Division Duplex (TDD), combined with beamforming technology. In the end, the purpose of improving the problem of serious multi-user interference in the massive MIMO system that is not conducive to channel estimation can be achieved, and the information transmission efficiency of the channel can be improved on the basis of reducing the system complexity, and the system performance can be optimized.

本发明的附加方面和优点将在下面的描述中部分给出，部分将从下面的描述中变得明显，或通过本发明的实践了解到。Additional aspects and advantages of the present invention will be set forth, in part, from the following description, and in part will be apparent from the following description, or may be learned by practice of the invention.

附图说明Description of drawings

本发明的上述和/或附加的方面和优点从结合下面附图对实施例的描述中将变得明显和容易理解，其中：The above and/or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of embodiments taken in conjunction with the accompanying drawings, wherein:

图1是毫米波Saleh-Valenzuela信道模型；Figure 1 is the millimeter-wave Saleh-Valenzuela channel model;

图2是3D波束空间大规模MIMO系统模型；Figure 2 is a 3D beam space massive MIMO system model;

图3是本发明总体流程图。Figure 3 is a general flow chart of the present invention.

具体实施方式Detailed ways

下面详细描述本发明的实施例，所述实施例的示例在附图中示出，其中自始至终相同或类似的标号表示相同或类似的元件或具有相同或类似功能的元件。下面通过参考附图描述的实施例是示例性的，仅用于解释本发明，而不能理解为对本发明的限制。The following describes in detail the embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals refer to the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary, only used to explain the present invention, and should not be construed as a limitation of the present invention.

在本发明的描述中，需要理解的是，术语“纵向”、“横向”、“上”、“下”、“前”、“后”、“左”、“右”、“竖直”、“水平”、“顶”、“底”“内”、“外”等指示的方位或位置关系为基于附图所示的方位或位置关系，仅是为了便于描述本发明和简化描述，而不是指示或暗示所指的装置或元件必须具有特定的方位、以特定的方位构造和操作，因此不能理解为对本发明的限制。In the description of the present invention, it should be understood that the terms "portrait", "horizontal", "upper", "lower", "front", "rear", "left", "right", "vertical", The orientations or positional relationships indicated by "horizontal", "top", "bottom", "inside", "outside", etc. are based on the orientations or positional relationships shown in the accompanying drawings, which are only for the convenience of describing the present invention and simplifying the description, rather than An indication or implication that the referred device or element must have a particular orientation, be constructed and operate in a particular orientation, is not to be construed as a limitation of the invention.

在本发明的描述中，除非另有规定和限定，需要说明的是，术语“安装”、“相连”、“连接”应做广义理解，例如，可以是机械连接或电连接，也可以是两个元件内部的连通，可以是直接相连，也可以通过中间媒介间接相连，对于本领域的普通技术人员而言，可以根据具体情况理解上述术语的具体含义。In the description of the present invention, unless otherwise specified and limited, it should be noted that the terms "installed", "connected" and "connected" should be understood in a broad sense, for example, it may be a mechanical connection or an electrical connection, or two The internal communication between the elements may be directly connected or indirectly connected through an intermediate medium, and those of ordinary skill in the art can understand the specific meanings of the above terms according to specific situations.

本发明通过基于波束选择的毫米波大规模MIMO系统MMSE-PCA信道估计的方法，能够有效地提升毫米波波束选择方案的速率和性能，先用在基站侧以MMSE预编码为基准的幅度最大化标准选择波束信号，以此减轻多用户干扰。之后在接收端使用PCA信道压缩方法，将信道进行压缩降维，降低信道估计的复杂度，最后使用经典的最小二乘法进行信道估计。The present invention can effectively improve the rate and performance of the millimeter-wave beam selection scheme through the MMSE-PCA channel estimation method of the millimeter-wave massive MIMO system based on beam selection, and is first used on the base station side to maximize the amplitude based on MMSE precoding Standard selection of beam signals to mitigate multi-user interference. Afterwards, the PCA channel compression method is used at the receiving end to compress the channel to reduce the dimension and reduce the complexity of channel estimation. Finally, the classical least squares method is used for channel estimation.

结合附图3对本发明进行详细说明，主要包括以下步骤：The present invention is described in detail in conjunction with accompanying drawing 3, mainly comprises the following steps:

步骤1：开始；Step 1: start;

步骤2：以MMSE预编码为基准的幅度最大化标准选择波束信号；Step 2: Select the beam signal based on the amplitude maximization criterion based on MMSE precoding;

先采用以MMSE预编码为基准的幅度最大化标准选择波束信号，在ZF算法的基础上引入了MMSE线性预编码技术,因而更加有效的平面噪音和用户之间的干扰。MMSE预编码矩阵表达式为:Firstly, the amplitude maximization standard based on MMSE precoding is used to select beam signals, and MMSE linear precoding technology is introduced on the basis of ZF algorithm, so the plane noise and interference between users are more effective. The MMSE precoding matrix expression is:

其中，β表示功率控制因子，||·||₂表示求2范数，E(·)表示求期望。为计算简单，MMSE预编码算法的优化问题可以看做是在一定的功率约束条件下求解接收信号与发送信号的最小均方误差的问题。在此基础上，建立目标函数：Among them, β represents the power control factor, ||·||₂ represents the 2-norm, and E(·) represents the expectation. In order to simplify the calculation, the optimization problem of the MMSE precoding algorithm can be regarded as the problem of solving the minimum mean square error between the received signal and the transmitted signal under certain power constraints. On this basis, establish the objective function:

其中，P表示信号的最大发射功率。根据MMSE准则，得到预编码矩阵为：Among them, P represents the maximum transmit power of the signal. According to the MMSE criterion, the obtained precoding matrix is:

其中，σ²为噪声功率。功率控制因子β为：where σ² is the noise power. The power control factor β is:

步骤3：接收端使用PCA信道压缩方法，将信道进行压缩降维；Step 3: The receiver uses the PCA channel compression method to compress the channel to reduce dimension;

采用PCA的压缩信道估计算法，降低信道矩阵的维度，减小了信道估计的计算复杂度。基于低复杂度PCA的CSI反馈算法中，首先对协方差矩阵进行特征值分解，求出C_H的特征值和特征向量，表示为：The compressed channel estimation algorithm of PCA is adopted to reduce the dimension of the channel matrix and reduce the computational complexity of the channel estimation. In the low-complexity PCA-based CSI feedback algorithm, the covariance matrix is first decomposed into eigenvalues, and the eigenvalues and eigenvectors of_CH are obtained, which are expressed as:

其中，

是对角矩阵，其对角元素是协方差矩阵的特征值，

的列向量是协方差矩阵C_H的特征向量。in,

is a diagonal matrix whose diagonal elements are the eigenvalues of the covariance matrix,

The column vectors of are the eigenvectors of the covariance matrix_CH .

然后将特征值按照由大到小进行排列，选取特征值贡献率超过阈值γ的前l个特征值所对应的特征向量，组成压缩矩阵

Then the eigenvalues are arranged in descending order, and the eigenvectors corresponding to the first l eigenvalues whose eigenvalue contribution rate exceeds the threshold γ are selected to form a compression matrix

其次利用压缩矩阵

将高维下行信道信息矩阵H_r压缩到低维空间，表示为：Second, use the compression matrix

The high-dimensional downlink channel information matrix H_r is compressed into a low-dimensional space, expressed as:

其中

表示降维后的信道矩阵。in

Represents the channel matrix after dimensionality reduction.

基于PCA算法的反馈量压缩比为：The feedback compression ratio based on the PCA algorithm is:

r_PCA＝l(N_r+N_t)/(N_r×N_t)r_PCA =l(N_r +N_t )/(N_r ×N_t )

本项目拟采用非码本的反馈，反馈量是通过压缩比来定义，压缩比越小，所需反馈开销越低。This project intends to use non-codebook feedback, and the amount of feedback is defined by the compression ratio. The smaller the compression ratio, the lower the required feedback overhead.

最后用户将

以及压缩矩阵

反馈给基站端，基站端接收到反馈信息后利用同样的压缩矩阵恢复原始信道，表示为：The end user will

and the compression matrix

Feedback to the base station, after receiving the feedback information, the base station uses the same compression matrix to restore the original channel, which is expressed as:

其中

表示信道的恢复值。in

Represents the recovery value of the channel.

步骤4：用LS算法估算出信道状态信息；Step 4: Use the LS algorithm to estimate the channel state information;

令信道矩阵为H，接收信号矩阵为

发送信号矩阵为X，其估计可表示为：Let the channel matrix be H and the received signal matrix be

The transmitted signal matrix is X, and its estimate can be expressed as:

为了了解LS信道估计结果，对上式求偏导，并令偏导为0，可得In order to understand the LS channel estimation result, take the partial derivative of the above formula, and set the partial derivative to be 0, we can get

解之，得到信道估计为：Solving it, the channel estimate is obtained as:

若X是非奇异矩阵，可以使用X的穆尔-彭罗斯逆。If X is a nonsingular matrix, the Moore-Penrose inverse of X can be used.

步骤5：结束。Step 5: End.

在本说明书的描述中，参考术语“一个实施例”、“一些实施例”、“示例”、“具体示例”、或“一些示例”等的描述意指结合该实施例或示例描述的具体特征、结构、材料或者特点包含于本发明的至少一个实施例或示例中。在本说明书中，对上述术语的示意性表述不一定指的是相同的实施例或示例。而且，描述的具体特征、结构、材料或者特点可以在任何的一个或多个实施例或示例中以合适的方式结合。In the description of this specification, description with reference to the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples", etc., mean specific features described in connection with the embodiment or example , structure, material or feature is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

尽管已经示出和描述了本发明的实施例，本领域的普通技术人员可以理解：在不脱离本发明的原理和宗旨的情况下可以对这些实施例进行多种变化、修改、替换和变型，本发明的范围由权利要求及其等同物限定。Although embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, The scope of the invention is defined by the claims and their equivalents.

Claims (5)

1. A millimeter wave massive MIMO system MMSE-PCA channel estimation method based on beam selection is characterized by comprising the following steps:

s1, selecting beam signals by adopting an amplitude maximization (MM) standard with minimum mean square error precoding as a reference at the base station side, and introducing a minimum mean square error linear precoding technology to weaken the influence of noise and interference among users;

s2, adopting a Time Division Duplex (TDD) large-scale MIMO system, and obtaining Channel State Information (CSI) by Least Square (LS) channel estimation in an uplink according to channel reciprocity in the TDD system;

s3, a Saleh-Vallenzuela channel model is adopted to reflect the channel sparse characteristic, and a Principal Component Analysis (PCA) channel compression method is used at a receiving end to map CSI from a high dimension to a low dimension for reducing the characteristic dimension;

s4, the receiving end compresses the channel to reduce dimension, and then adopts LS to estimate the channel.

2. The method for MMSE-PCA channel estimation in a MMSE massive MIMO system based on beam selection as claimed in claim 1, wherein the S1 comprises:

the base station side adopts MM standard with MMSE precoding as reference to select beam signals, MMSE linear precoding technology is introduced on the basis of Zero Forcing (ZF) algorithm, and MMSE precoding matrix expression is as follows:

wherein beta represents a power control factor, | · |. non-woven phosphor₂2 norm is obtained, and E (-) is expected; in order to achieve simple calculation, the optimization problem of the MMSE precoding algorithm can be regarded as the problem of solving the minimum mean square error of a received signal and a transmitted signal under a certain power constraint condition; on the basis, an objective function is established:

wherein P represents the maximum transmit power of the signal; according to the MMSE criterion, the obtained precoding matrix is as follows:

wherein σ²For noise power, the power control factor β is:

wherein Tr (H) represents the trace of the matrix (H)^-1Representing the inverse of the matrix, H^HRepresenting the conjugate transpose of the matrix.

3. The MMSE-PCA channel estimation method for the beam-selection MMSE massive MIMO system of claim 1, wherein the S2 comprises:

the antenna matrix U expression is:

in the formula:

representing the attitude, the system model received signal of massive MIMO based on 3D beam space can be represented as:

in the formula

For the received signal vector of the beam space, after converting the channel vector and the channel vector into the channel vector of the beam space, the conversion mode is as follows:

comprises h_kCan be used to estimate the entire CSI; then the beamSpatial channel matrix

Can be defined as:

representing a downlink beam space channel matrix;

the uplink obtains the CSI by LS channel estimation, in which each user needs to send orthogonal pilot sequence psi to the base station at time Q_mAssuming that the Q time is divided into M blocks each consisting of K times, the uplink signal vector received at the base station of the M-th block according to channel reciprocity in the TDD system

Can be expressed as:

by adaptively selecting a network, the base station uses an analog combiner W with dimension K N_mTo be combined out

And obtaining a sampling signal R with dimension K multiplied by K in baseband sampling through a radio frequency chain_mWherein R is_mThe expression is as follows:

finally, the signals with reduced dimensionality and the orthogonal pilot frequency matrix are combined

Multiplying to obtain beam space channel

Is detected by the detection matrix Z_m，

Wherein

And the effective noise matrix is represented, and in a TDD system, the CSI obtained according to uplink estimation can be used as the CSI of a downlink channel due to channel reciprocity.

4. The method for MMSE-PCA channel estimation in a MMSE massive MIMO system based on beam selection as claimed in claim 1, wherein the S3 comprises:

because the number of effective paths in millimeter wave communication is limited, H has the characteristic of a sparse structure;

then the reduced dimensionality signal of the massive MIMO system based on beam space can be expressed as:

in the formula,b denotes the set of selected beams, P_rA reduced dimension precoding matrix; in order to achieve near-optimal performance, the base station needs to obtain a 3D wave-imperial spatial channel with a limited number of radio frequency chains, and in order to guarantee spatial multiplexing gain for K users, the minimum number of required radio frequency chains should be N_RFK, so consider the number of radio frequency chains to be N_RFK, and without loss of generality;

after signals are precoded through MM-MMSE to form beams and are sent out, a receiving end receives the precoded signals, and then dimension reduction is carried out on CSI through PCA;

in the CSI algorithm based on low-complexity PCA, firstly, eigenvalue decomposition is carried out on a covariance matrix to obtain C_HIs represented as:

wherein,

is a diagonal matrix, whose diagonal elements are eigenvalues of a covariance matrix,

is a covariance matrix C_HThe feature vector of (2);

then, the eigenvalues are arranged from big to small, and eigenvectors corresponding to the first eigenvalues with eigenvalue contribution rates exceeding a threshold gamma are selected to form a compression matrix

Second using the compression matrix

A high-dimensional downlink channel information matrix H_rCompressed into a low dimensional space, represented as:

whereinRepresenting the channel matrix after dimensionality reduction;

the feedback quantity compression ratio based on the PCA algorithm is as follows:

r_PCA＝l(N_r+N_t)/(N_r×N_t)

non-codebook feedback is adopted, the feedback quantity is defined by a compression ratio, and the smaller the compression ratio is, the lower the required feedback overhead is;

finally, the user will

And compressing the matrix

Feeding back the feedback information to a base station end, and recovering an original channel by using the same compression matrix after the base station end receives the feedback information, wherein the original channel is represented as the reference channel;

wherein

Representing the recovery value of the channel.

5. The MMSE-PCA channel estimation method for the MMSE-MMSE system based on the beam selection as claimed in claim 1, wherein the S4 comprises:

the receiving end estimates the CSI by an LS algorithm;

let the channel matrix be H and the received signal matrix be

The transmit signal matrix is X, and its estimate can be expressed as:

to obtain a specific expression, the partial derivatives of the above formula are solved, and the partial derivatives are made to be 0, so that:

to solve this, the channel estimate is obtained as:

if X is a non-singular matrix, the Mueller-Penrose inverse of X may be used.

Images