WO2021196726A1 - 一种快速搜索方法 - Google Patents

一种快速搜索方法 Download PDF

Info

Publication number
WO2021196726A1
WO2021196726A1 PCT/CN2020/135753 CN2020135753W WO2021196726A1 WO 2021196726 A1 WO2021196726 A1 WO 2021196726A1 CN 2020135753 W CN2020135753 W CN 2020135753W WO 2021196726 A1 WO2021196726 A1 WO 2021196726A1
Authority
WO
WIPO (PCT)
Prior art keywords
channel
user
vector
energy
polynomial
Prior art date
Application number
PCT/CN2020/135753
Other languages
English (en)
French (fr)
Inventor
杜清河
欧奕杰
申宁
Original Assignee
西安交通大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 西安交通大学 filed Critical 西安交通大学
Publication of WO2021196726A1 publication Critical patent/WO2021196726A1/zh

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04BTRANSMISSION
    • H04B7/00Radio transmission systems, i.e. using radiation field
    • H04B7/02Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/0413MIMO systems
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00Baseband systems
    • H04L25/02Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/0202Channel estimation

Definitions

  • This application belongs to the field of wireless communication technology, and particularly relates to a fast search method.
  • Massive MIMO Multiple-Input Multiple-Output
  • 5G Fifth-Generation, fifth-generation mobile communication technology
  • the channels of each user tend to be orthogonal, and multi-user interference tends to disappear.
  • the use of low-complexity signal detection technology at the receiving end can achieve good performance.
  • Massive MIMO technology has shown great potential in increasing system capacity, spectrum efficiency and energy efficiency. The performance of massive MIMO systems largely depends on perfect CSI (Channel State Information).
  • a unified transmission strategy can be designed in TDD and FDD (Frequency Division Duplexing) massive MIMO systems, which can effectively solve the pilot pollution and the training overhead that increases with the increase in the number of antennas. problem.
  • spatial rotation is an innovative method to improve channel estimation performance.
  • spatial rotation can be regarded as searching for a fixed number of orthogonal DFT (Discrete Fourier Transform, Discrete Fourier Transform) basis vectors to more accurately represent the channel.
  • orthogonal DFT Discrete Fourier Transform, Discrete Fourier Transform
  • the orthogonal space basis of each user channel needs to be updated. Without any prior knowledge, the high complexity brought by calculating spatial features and searching for rotation parameters has become an obstacle to realizing spatial rotation operations.
  • spatial rotation is an innovative method to improve the performance of channel estimation.
  • spatial rotation can be regarded as searching for a fixed number of orthogonal DFT (Discrete Fourier Transform, Discrete Fourier Transform) basis vectors to more accurately represent the channel.
  • orthogonal DFT Discrete Fourier Transform, Discrete Fourier Transform
  • the orthogonal space basis of each user channel needs to be updated.
  • This application provides a fast search method.
  • this application provides a quick search method, which includes the following steps:
  • Step 1) The base station calculates the discrete Fourier transform according to the uplink channel vector h k of the kth user, and determines the spatial characteristics of the kth user
  • Step 2) Initialize N angle values, respectively calculate the discrete Fourier transform of the channel vector at N rotation angles, and calculate the sparse channel energy at N angles according to the discrete Fourier transform;
  • Step 3) Take the rotation angle as the independent variable and the sparse channel energy as the dependent variable, and use the m-order polynomial to perform polynomial fitting on the data of N points;
  • Step 4) Calculate the maximum value of the polynomial in the angle range as the rotation angle of the k-th user, and combine its spatial characteristics to obtain the optimal orthogonal spatial basis of the channel;
  • Step 5) Repeat steps 1) to 3) for other K-1 users, and update the channel optimal orthogonal space base of each user in the current coherent time slot.
  • the uplink channel vector h k of the k-th user can be represented by a vector of M ⁇ 1.
  • the incident angle range of the diameter of the upstream channel for the k-th user [ ⁇ k - ⁇ k, ⁇ k + ⁇ k], the incident angle of each diameter in the range satisfying a uniform distribution, channel
  • the incident angle domain has sparse characteristics.
  • the discrete Fourier transform of the channel in step 1) can be expressed as Where F is an M ⁇ M matrix, and the elements in the p-th row and q-th column are When the number of base station antennas tends to infinity, the energy of each path is concentrated at one point in the DFT domain; when the number of base station antennas is limited, energy leakage occurs, and the DFT points around the center point contain a small amount of channel energy, and the channel energy remains in the DFT domain. Presents a highly concentrated characteristic; according to the characteristic of energy leakage, calculate the subscript set for:
  • q max and q min are subscript sets respectively
  • the maximum and minimum in n is equal to Round up
  • ⁇ M is the number of orthogonal space bases set in advance.
  • step 3 Another implementation manner provided by this application is: the specific method of polynomial fitting in step 3) is as follows:
  • the polynomial order m satisfies m ⁇ N.
  • the order vector (a total of m) as:
  • the least square method is used to obtain the minimum variance solution as in Is the pseudo-inverse of matrix ⁇ m.
  • step 4) the determination of the k-th user's rotation angle and the optimal orthogonal space base is as follows:
  • the optimal orthogonal space basis set of the k-th user uplink channel vector is determined as:
  • ⁇ ( ⁇ k, opt ) H represents the conjugate transposed matrix of the matrix ⁇ ( ⁇ k, opt ), and the vector v q represents the qth column of the discrete Fourier transform matrix F after the conjugate transpose.
  • the fast search method provided in this application is a fast search method for the optimal orthogonal space base in a massive MIMO system.
  • the fast search method provided in the present application proposes a fast method for searching orthogonal space bases, which is used to realize an efficient sparse representation of uplink and downlink channels in a massive MIMO system.
  • the method of determining the user space characteristic parameters is improved, and a low-complexity method for searching the user's optimal spatial rotation angle is proposed.
  • This method has almost no space when the AS is narrow. Performance loss. Therefore, this method is of great significance for reducing the computational complexity of TDD and FDD massive MIMO systems and improving the channel estimation performance. .
  • the fast search method provided in this application greatly reduces the computational complexity of the massive MIMO system while ensuring the improvement of channel estimation accuracy.
  • the user's spatial characteristic parameters are more accurately extracted under the premise of low complexity, and a more efficient channel sparse representation in the spatial domain is realized.
  • the fast search method provided by this application selects orthogonal space bases by calculating the user's spatial characteristics and rotation parameters; under the same number of space bases, it improves the uplink and downlink channel estimation accuracy of TDD and FDD massive MIMO systems, and reduces The computational complexity of the system.
  • the fast search method provided by this application is based on the user's DOA and AS estimation.
  • the AS is small, there is almost no performance loss, and when the AS is large It has better channel estimation performance than other methods.
  • the base station can extract the current spatial information of each user, which helps the base station to adjust the number of user training sequences in real time and ensure the quality of channel estimation.
  • the fast search method provided by this application is based on the study of the energy change of the sparse channel, and it can approach the best through a small number of high-dimensional calculations.
  • the complexity of the spatial rotation operation in a single coherent time slot is greatly reduced, and the requirements for hardware implementation are reduced.
  • the optimal orthogonal space base search method proposed in this application is not only suitable for massive MIMO systems under SBEM, but also can achieve significant results for two-dimensional space base extended models (2D-SBEM).
  • 2D-SBEM two-dimensional space base extended models
  • FIG. 1 is a schematic diagram of the space-based expansion model (SBEM) of the massive MIMO system of the present application;
  • Figure 2 is a performance comparison diagram of each spatial feature method of the present application.
  • Figure 3 is a performance comparison of the orthogonal space-based search method under SBEM of the present application.
  • Fig. 4 is a performance comparison of the orthogonal space-based search method under 2D-SBEM of the present application.
  • K single-antenna users of the massive MIMO system are randomly distributed in the coverage area of the base station.
  • the base station obtains the channel vector of each user through uplink channel estimation in the current coherent time slot.
  • it is necessary to update the orthogonal space base corresponding to each user channel so that the virtual beam can be Target users more accurately.
  • the number of optimal orthogonal space bases finally obtained is much smaller than the number of base station antennas, which contains most of the energy of the channel, and is a highly sparse representation of the channel.
  • the optimal orthogonal space basis of the downlink channel can be directly obtained according to the search result of the optimal orthogonal space basis of the uplink channel, which greatly improves the performance of uplink and downlink channel estimation under the premise of low complexity.
  • the base station calculates the discrete Fourier transform according to the uplink channel vector h k of the kth user to determine the spatial characteristics of the kth user
  • the uplink channel vector h k of the k-th user can be represented by a vector of M ⁇ 1.
  • the incident angle range of the diameter of the upstream channel for the k-th user [ ⁇ k - ⁇ k, ⁇ k + ⁇ k], the incident angle of each diameter in the range satisfying a uniform distribution, channel
  • the incident angle domain has sparse characteristics.
  • the discrete Fourier transform of the channel can be expressed as Where F is an M ⁇ M matrix, and the elements in the p-th row and q-th column are When the number of base station antennas tends to infinity, the energy of each path is concentrated at one point in the DFT domain; when the number of base station antennas is limited, energy leakage occurs, and the DFT points around the center point contain a small amount of channel energy, and the channel energy remains in the DFT domain. Presents a highly concentrated characteristic. According to the characteristics of energy leakage, calculate the subscript set for:
  • q max and q min are subscript sets respectively
  • the maximum and minimum in n is equal to Round up
  • ⁇ M is the number of orthogonal space bases set in advance.
  • step 3 the specific method of polynomial fitting in step 3) is as follows:
  • the polynomial order m satisfies m ⁇ N.
  • the order vector (a total of m) as:
  • the least square method is used to obtain the minimum variance solution as in Is the pseudo-inverse of matrix ⁇ m.
  • step 4 determination of the k-th user's rotation angle and the optimal orthogonal space basis in step 4) is as follows:
  • the optimal orthogonal space basis set of the k-th user uplink channel vector is determined as:
  • ⁇ ( ⁇ k, opt ) H represents the conjugate transposed matrix of the matrix ⁇ ( ⁇ k, opt ), and the vector v q represents the qth column of the discrete Fourier transform matrix F after the conjugate transpose.
  • Figures 2, 3, and 4 compare the system performance achieved by the solution proposed by this application with the existing technical solutions, show the performance of the spatial feature calculation and rotation angle search methods proposed by this application, and reflect the large-scale performance of this application.
  • Fig. 2 compares the curve change of the mean square error of the uplink channel achieved by the spatial feature calculation scheme proposed by this application and the existing scheme with the angle expansion.
  • Configuration in the simulation the number of base station antennas is 128, the orthogonal base is set to 16, and the signal-to-noise ratio is 20 decibels. From the simulation results, it can be seen that the existing low-complexity implementation scheme (with the maximum energy point as the center point) deteriorates with the increase in angle expansion, and the spatial feature calculation scheme proposed in this application is under the premise of low complexity. , It is close to the optimal performance under any angle of expansion.
  • the spatial feature calculation scheme proposed in this application is based on DOA and AS estimation, and can deal with the randomness of the gain and incidence angle distribution in the multipath.
  • the base station can more accurately extract the current spatial information of each user in each coherent time slot, which helps the base station to adjust the number of user training sequences in real time and ensure the quality of channel estimation.
  • Fig. 3 compares the curve variation of the mean square error of the uplink channel achieved with the signal-to-noise ratio under the SBEM and the existing schemes of the orthogonal space base selection scheme proposed in this application.
  • the optimal orthogonal space basis of the downlink channel can be directly derived from the optimal orthogonal space basis of the uplink channel.
  • the cross-space basis selection scheme is of great significance for improving the uplink and downlink channel estimation performance of massive MIMO systems under SBEM.
  • Fig. 4 compares the curve variation of the mean square error of the uplink channel achieved with the signal-to-noise ratio under the 2D-SBEM of the orthogonal space base selection scheme proposed by this application and the existing scheme.

Abstract

计算空间特征和搜索旋转参数带来的高复杂度成为实现空间旋转操作的障碍。本申请的快速搜索方法,1)基站根据第k个用户的上行信道向量hk计算其离散傅里叶变换,确定第k个用户的空间特征(aa);2)初始化N个角度值,分别计算N个旋转角度下信道向量的离散傅里叶变换,根据离散傅里叶变换计算N个角度下稀疏信道能量;3)以旋转角度为自变量,稀疏信道能量为因变量,用m阶多项式对N个点的数据进行多项式拟合;4)计算在角度范围内多项式的最大值作为第k个用户的旋转角度,结合其空间特征获得信道最优正交空间基;5)对其他K-1个用户重复步骤1)至步骤3),更新当前相干时间内每个用户的信道最优正交空间基。降低了整体系统设计难度。

Description

一种快速搜索方法 技术领域
本申请属于无线通信技术领域,特别是涉及一种快速搜索方法。
背景技术
大规模MIMO(Multiple-Input Multiple-Output,多输入多输出)技术被视作5G(5th-Generation,第五代移动通信技术)物理层中最有前景的技术。随着天线数量的增加,各用户的信道趋于正交,多用户干扰趋于消失,在接收端采用低复杂度的信号检测技术就能实现良好的性能。大规模MIMO技术在增加系统容量,频谱效率和能效方面显示出巨大潜力。大规模MIMO系统的性能在很大程度上依赖于完美的CSI(Channel State Information,信道状态信息)。对于TDD(Time Division Duplexing,时分双工)大规模MIMO系统中的传统正交训练策略,随着发射天线数量的增加,计算复杂度的升高将增加获取信道状态信息的开销;同时随着小区内用户数量的增加,导频污染将削弱系统性能和效率的提高。
大多数针对大规模MIMO系统的研究要求基站建立在高层建筑或专用楼塔的顶部,这样在基站端附近基本没有散射体,入射信号的AS(Angle Spread,角度扩展)被认为是足够窄的,因此可以利用用户不重叠的空间信息来实现正交传输,利用信道的稀疏性可以创建空域上的低秩模型。SBEM(Spatial Basis Expansion Model,空间基扩展模型)是一种典型的基于大规模均匀线性阵列的低秩模型,它依赖于入射信号的平均DOA(Direction of Arrivals,波达方向)和AS。利用信道的角度互易,可以在TDD和FDD(Frequency Division Duplexing,频分双工)大规模MIMO系统中设计统一的传输策略,有效地解决导频污染和随天线数目上升而升高的训练开销问题。
在SBEM中,空间旋转是一种改进信道估计性能的创新方法。对于在空域上稀疏的信道,空间旋转可以看作是搜索固定数量的正交的DFT(Discrete Fourier Transform,离散傅里叶变换)基向量以更准确地表示信道。但是,由于在不同的相干时隙内信道多径增益发生变化,为了获得更好的性能,每个用户信道的正交空间基需要更新。在没有任何先验知识的情况下,计算空间特征和搜索旋转参数带来的高复杂度成为实现空间旋转操作的障碍。
发明内容
1.要解决的技术问题
基于在SBEM中,空间旋转是一种改进信道估计性能的创新方法。对于在空域上稀疏的信道,空间旋转可以看作是搜索固定数量的正交的DFT(Discrete Fourier Transform,离散傅 里叶变换)基向量以更准确地表示信道。但是,由于在不同的相干时隙内信道多径增益发生变化,为了获得更好的性能,每个用户信道的正交空间基需要更新。在没有任何先验知识的情况下,计算空间特征和搜索旋转参数带来的高复杂度成为实现空间旋转操作的障碍的问题,本申请提供了一种快速搜索方法。
2.技术方案
为了达到上述的目的,本申请提供了一种快速搜索方法,所述方法包括如下步骤:
步骤1):基站根据第k个用户的上行信道向量h k计算其离散傅里叶变换,确定第k个用户的空间特征
Figure PCTCN2020135753-appb-000001
步骤2):初始化N个角度值,分别计算N个旋转角度下信道向量的离散傅里叶变换,根据离散傅里叶变换计算N个角度下稀疏信道能量;
步骤3):以旋转角度为自变量,稀疏信道能量为因变量,用m阶多项式对N个点的数据进行多项式拟合;
步骤4):计算在角度范围内多项式的最大值作为第k个用户的旋转角度,结合其空间特征获得信道最优正交空间基;
步骤5):对其他K-1个用户重复步骤1)至步骤3),更新当前相干时隙内每个用户的信道最优正交空间基。
本申请提供的另一种实施方式为:所述步骤1)中第k个用户的空间特征
Figure PCTCN2020135753-appb-000002
确定如下:
设基站端有M>>1根天线且为均匀线阵,K个用户均为单天线用户,则第k个用户的上行信道向量h k可用M×1的向量表示。在大规模MIMO系统中,第k个用户的上行信道的径的入射角范围为[θ k-Δθ kk+Δθ k],每条径的入射角在这个范围满足均匀分布,信道在入射角域上具有稀疏的特性。
本申请提供的另一种实施方式为:所述步骤1)中信道的离散傅里叶变换可以表示为
Figure PCTCN2020135753-appb-000003
其中F是M×M的矩阵,其第p行第q列的元素为
Figure PCTCN2020135753-appb-000004
当基站天线数量趋于无穷时,每条径的能量集中在DFT域的一个点;当基站天线数量有限时,出现能量泄漏,中心点周围的DFT点包含少量信道能量,信道能量在DFT域仍呈现高度集中的特性;根据能量泄漏的特性,计算下标集合
Figure PCTCN2020135753-appb-000005
为:
Figure PCTCN2020135753-appb-000006
其中,系数β定义为函数h(x)=|sin(M/2x)/sin(1/2x)|在一个周期内次最大极值和最大极值的比值;
Figure PCTCN2020135753-appb-000007
表示向量
Figure PCTCN2020135753-appb-000008
第q 0个元素;计算空间特征
Figure PCTCN2020135753-appb-000009
为:
Figure PCTCN2020135753-appb-000010
其中,
Figure PCTCN2020135753-appb-000011
q max和q min分别为下标集合
Figure PCTCN2020135753-appb-000012
中的最大值和最小值,n等于
Figure PCTCN2020135753-appb-000013
向上取整,τ<<M为预先设定的正交空间基的数量。
本申请提供的另一种实施方式为:所述步骤2)中第k个用户旋转角度的初始化计算如下:
Figure PCTCN2020135753-appb-000014
为间隙大小在角度范围
Figure PCTCN2020135753-appb-000015
均匀选取N个角度,即第n个角度选取为
Figure PCTCN2020135753-appb-000016
信道旋转该角度后的离散傅里叶变换为
Figure PCTCN2020135753-appb-000017
其中矩阵Φ(φ n)是由元素
Figure PCTCN2020135753-appb-000018
组成的对角矩阵。
本申请提供的另一种实施方式为:所述步骤2)中第k个用户旋转角度的稀疏信道能量的计算如下:
结合用户的空间特征,第k个用户的上行信道在φ n的旋转角度下经稀疏表示后的能量为:
Figure PCTCN2020135753-appb-000019
其中,
Figure PCTCN2020135753-appb-000020
表示FΦ(φ n)h k的子向量,该向量包含了以
Figure PCTCN2020135753-appb-000021
中元素作为索引的FΦ(φ n)h k中的元素;
Figure PCTCN2020135753-appb-000022
表示向量
Figure PCTCN2020135753-appb-000023
的2-范数的平方,即稀疏信道的能量值;
分别对n个角度重复以上步骤,即可得到第k个用户的旋转角度和稀疏信道的能量的对应关系,记作ε k=f(φ k)。
本申请提供的另一种实施方式为:所述步骤3)中多项式拟合的具体方法如下:
设目标多项式为
Figure PCTCN2020135753-appb-000024
其中,为了保证解的唯一性,多项式阶数m满足m<N。定义阶次向量(共m个)为:
Figure PCTCN2020135753-appb-000025
定义目标向量为
Figure PCTCN2020135753-appb-000026
将多项式系数对应的向量记作a m=[a 0,a 1,...,a m] T,根据阶次向量生成矩阵Φ m=[φ 01,...,φ m],最优多项式拟合问题可转化为求解系数向量的最优解:
Figure PCTCN2020135753-appb-000027
采用最小二乘法获得最小方差解为
Figure PCTCN2020135753-appb-000028
其中
Figure PCTCN2020135753-appb-000029
是矩阵Φ m的伪逆矩阵。
本申请提供的另一种实施方式为:所述步骤4)第k个用户旋转角度和最优正交空间基的确定如下:
已知多项式
Figure PCTCN2020135753-appb-000030
对其求导可得:
Figure PCTCN2020135753-appb-000031
令f'(φ k)=0,在自变量区间
Figure PCTCN2020135753-appb-000032
方程有一或多个解。将解集分别代入f(φ k),取最大的f(φ k)对应的φ k作为第k个用户的旋转角度φ k,opt
结合第k个用户的旋转角度φ k,opt和空间特征
Figure PCTCN2020135753-appb-000033
第k个用户上行信道向量的最优正交空间基集合确定为:
Figure PCTCN2020135753-appb-000034
其中,Φ(φ k,opt) H表示矩阵Φ(φ k,opt)的共轭转置矩阵,向量v q表示离散傅里叶变换矩阵F共轭转置后的第q列。
3.有益效果
与现有技术相比,本申请提供的一种快速搜索方法的有益效果在于:
本申请提供的快速搜索方法,为一种大规模MIMO系统中最优正交空间基的快速搜索方法。
本申请提供的快速搜索方法,提出了一种用于搜索正交空间基的快速方法,用于实现大规模MIMO系统中上下行信道的高效稀疏表示。基于对信道在DFT域中的研究,改进了确定用户空间特征参数的方法,并提出了一种用于搜索用户最优空间旋转角度的低复杂度方法,该方法在AS足较窄时几乎没有性能损失。因此,该方法对于降低TDD和FDD大规模MIMO系统的计算复杂度,提升信道估计性能具有重要意义。。
本申请提供的快速搜索方法,在保证信道估计精度提升的同时大大降低大规模MIMO系统的计算复杂度。基于空间旋转操作,在低复杂度的前提下更精确地提取了用户的空间特征参数,实现了空域上一种更高效的信道稀疏表示。
本申请提供的快速搜索方法,通过计算用户的空间特征和旋转参数来选取正交空间基;在空间基数量相同的条件下,提升了TDD和FDD大规模MIMO系统上下行信道估计精度,降低了系统的运算复杂度。
本申请提供的快速搜索方法,与现有的用户空间特征计算方法相比,本申请提出的空间特征计算方法基于用户的DOA和AS估计,当AS较小时几乎没有性能损失,当AS较大时具有相比去其他方法更优异的信道估计性能。同时通过该方法。在每个相干时隙内基站能提取每个用户当前的空间信息,有助于基站即时调整用户的训练序列数量,保证信道估计质量。
本申请提供的快速搜索方法,与现有的用户空间特征计算方法相比,本申请提出的旋转角度计算方法基于对稀疏信道能量变化的研究,通过较少次数的高维运算便能趋近最佳结果,极大降低了单个相干时隙内空间旋转操作的复杂度,降低了硬件实现的要求。
本申请提供的快速搜索方法,本申请提出的最优正交空间基搜索方法不仅适用于SBEM下的大规模MIMO系统,对于二维空间基扩展模型(2D-SBEM)应用此方法也能取得显著的复杂度降低和信道估计性能提升的效果。
附图说明
图1是本申请的大规模MIMO系统空间基扩展模型(SBEM)示意图;
图2是本申请的各空间特征方法性能对比图;
图3是本申请的SBEM下正交空间基搜索方法性能对比;
图4是本申请的2D-SBEM下正交空间基搜索方法性能对比。
具体实施方式
在下文中,将参考附图对本申请的具体实施例进行详细地描述,依照这些详细的描述, 所属领域技术人员能够清楚地理解本申请,并能够实施本申请。在不违背本申请原理的情况下,各个不同的实施例中的特征可以进行组合以获得新的实施方式,或者替代某些实施例中的某些特征,获得其它优选的实施方式。
参考图1,本申请中,大规模MIMO系统的K个单天线用户随机分布在基站的覆盖范围。基站周围散射物较少,基站端入射角极窄,径的相关性很强。基站在当前相干时隙通过上行信道估计获得了每个用户的信道向量,为了保证下一相干时隙的信道估计质量,需要对每个用户信道对应的正交空间基进行更新,使虚拟波束能够更准确地对准用户。经过对用户空间特征和旋转参数的计算,最后获取的最优正交空间基的数量远小于基站端天线数量,包含了信道的大部分能量,是对信道的高度稀疏化表示。根据角度互易性,根据上行信道最优正交空间基搜索结果能直接获取下行信道最优正交空间基,在低复杂度的前提下大大提升了上下行信道估计性能。
针对以上系统模型,本申请的具体步骤包括:
1)基站根据第k个用户的上行信道向量h k计算其离散傅里叶变换,确定第k个用户的空间特征
Figure PCTCN2020135753-appb-000035
2)初始化N个角度值,分别计算N个旋转角度下信道向量的离散傅里叶变换,根据离散傅里叶变换计算N个角度下稀疏信道能量;
3)以旋转角度为自变量,稀疏信道能量为因变量,用m阶多项式对N个点的数据进行多项式拟合;
4)计算在角度范围内多项式的最大值作为第k个用户的旋转角度,结合其空间特征获得信道最优正交空间基。
5)对其他K-1个用户重复步骤1)至步骤3),更新当前相干时隙内每个用户的信道最优正交空间基。
进一步的,步骤1)中第k个用户的空间特征
Figure PCTCN2020135753-appb-000036
确定如下:
设基站端有M>>1根天线且为均匀线阵,K个用户均为单天线用户,则第k个用户的上行信道向量h k可用M×1的向量表示。在大规模MIMO系统中,第k个用户的上行信道的径的入射角范围为[θ k-Δθ kk+Δθ k],每条径的入射角在这个范围满足均匀分布,信道在入射角域上具有稀疏的特性。
信道的离散傅里叶变换可以表示为
Figure PCTCN2020135753-appb-000037
其中F是M×M的矩阵,其第p行第q 列的元素为
Figure PCTCN2020135753-appb-000038
当基站天线数量趋于无穷时,每条径的能量集中在DFT域的一个点;当基站天线数量有限时,出现能量泄漏,中心点周围的DFT点包含少量信道能量,信道能量在DFT域仍呈现高度集中的特性。根据能量泄漏的特性,计算下标集合
Figure PCTCN2020135753-appb-000039
为:
Figure PCTCN2020135753-appb-000040
其中,系数β定义为函数h(x)=|sin(M/2x)/sin(1/2x)|在一个周期内次最大极值和最大极值的比值;
Figure PCTCN2020135753-appb-000041
表示向量
Figure PCTCN2020135753-appb-000042
第q 0个元素;计算空间特征
Figure PCTCN2020135753-appb-000043
为:
Figure PCTCN2020135753-appb-000044
其中,
Figure PCTCN2020135753-appb-000045
q max和q min分别为下标集合
Figure PCTCN2020135753-appb-000046
中的最大值和最小值,n等于
Figure PCTCN2020135753-appb-000047
向上取整,τ<<M为预先设定的正交空间基的数量。
进一步的,步骤2)中第k个用户旋转角度的初始化和稀疏信道能量的计算如下:
Figure PCTCN2020135753-appb-000048
为间隙大小在角度范围
Figure PCTCN2020135753-appb-000049
均匀选取N个点,即第n个角度选取为
Figure PCTCN2020135753-appb-000050
信道旋转该角度后的离散傅里叶变换为
Figure PCTCN2020135753-appb-000051
其中矩阵Φ(φ n)是由元素
Figure PCTCN2020135753-appb-000052
组成的对角矩阵。结合用户的空间特征,第k个用户的上行信道在φ n的旋转角度下经稀疏表示后的能量为:
Figure PCTCN2020135753-appb-000053
其中,
Figure PCTCN2020135753-appb-000054
表示FΦ(φ n)h k的子向量,该向量包含了以
Figure PCTCN2020135753-appb-000055
中元素作为索引的FΦ(φ n)h k中的元素;
Figure PCTCN2020135753-appb-000056
表示向量
Figure PCTCN2020135753-appb-000057
的2-范数的平方,即稀疏信道的能量值。
分别对n个角度重复以上步骤,即可得到第k个用户的旋转角度和稀疏信道的能量的对应关系,记作ε k=f(φ k)。
进一步的,步骤3)中多项式拟合的具体方法如下:
设目标多项式为
Figure PCTCN2020135753-appb-000058
其中,为了保证解的唯一性,多项式阶数m满足m<N。定义阶次向量(共m个)为:
Figure PCTCN2020135753-appb-000059
定义目标向量为
Figure PCTCN2020135753-appb-000060
将多项式系数对应的向量记作a m=[a 0,a 1,...,a m] T,根据阶次向量生成矩阵Φ m=[φ 01,...,φ m],最优多项式拟合问题可转化为求解系数向量的最优解:
Figure PCTCN2020135753-appb-000061
采用最小二乘法获得最小方差解为
Figure PCTCN2020135753-appb-000062
其中
Figure PCTCN2020135753-appb-000063
是矩阵Φ m的伪逆矩阵。
进一步的,步骤4)中第k个用户旋转角度和最优正交空间基的确定如下:
已知多项式
Figure PCTCN2020135753-appb-000064
对其求导可得:
Figure PCTCN2020135753-appb-000065
令f'(φ k)=0,在自变量区间
Figure PCTCN2020135753-appb-000066
方程有一或多个解。将解集分别代入f(φ k),取最大的f(φ k)对应的φ k作为第k个用户的旋转角度φ k,opt
结合第k个用户的旋转角度φ k,opt和空间特征
Figure PCTCN2020135753-appb-000067
第k个用户上行信道向量的最优正交空间基集合确定为:
Figure PCTCN2020135753-appb-000068
其中,Φ(φ k,opt) H表示矩阵Φ(φ k,opt)的共轭转置矩阵,向量v q表示离散傅里叶变换矩阵F共轭转置后的第q列。
图2、3、4对比了本申请所提出方案与现有的技术方案所实现的系统性能,展示了本申请所提出的空间特征计算和旋转角度搜索方法的性能,反映了本申请在大规模MIMO系统中 对信道估计性能的提升和系统复杂度的降低效果。
图2对比了本申请所提出的空间特征计算方案与现有方案所实现的上行信道均方误差随角度扩展的曲线变化。仿真中配置:基站天线数为128,设定正交基数为16,信噪比为20分贝。从仿真结果可以看出:已有的低复杂度实现方案(以能量最大点为中心点)随着角度扩展的增大性能恶化,本申请所提出的空间特征计算方案在低复杂度的前提下,在任意角度扩展下都趋近于最优性能。这是由于本申请所提出的空间特征计算方案基于DOA和AS估计,能够应对多径中增益和入射角分布的随机性。除此之外,应用本方案,在每个相干时隙内基站能更精确地提取每个用户当前的空间信息,有助于基站即时调整用户的训练序列数量,保证信道估计质量。
图3对比了本申请所提出正交空间基选取方案在SBEM下与现有方案所实现的上行信道均方误差随信噪比的曲线变化。仿真中配置:单边角度扩展Δθ k=2°。从仿真结果可以看出:在信噪比较低时,噪声影响限制了信道估计性能;在信噪比较高时,空间旋转是提升上行信道估计性能的一种有效手段。本申请所提出正交空间基选取方案在离散傅里叶变换次数为N=4时就能趋于最优性能,约等于现有方案N=16时的性能,复杂度降低了约75%,这为空间旋转的实现提供了基础。除此之外,由于信道的角度互易性,在同一相干时隙内,下行信道的最优正交空间基可以由上行信道的最优正交空间基直接得出,因此本申请所提出正交空间基选取方案对于提升SBEM下大规模MIMO系统上下行信道估计性能有重要意义。
图4对比了本申请所提出正交空间基选取方案在2D-SBEM下与现有方案所实现的上行信道均方误差随信噪比的曲线变化。仿真中配置:基站端为128×128的均匀面阵,径的单边角度扩展Δθ k=Δγ k=2°。从仿真结果可以看出:在信噪比较低时,噪声影响限制了信道估计性能;在信噪比较高时,空间旋转是提升上行信道估计性能的一种有效手段。对两个角度域分别应用本申请所提出的正交空间基选取方案,在二维离散傅里叶变换次数为N=4+4时就能趋于最优性能,约等于现有方案N=9*9时的性能,复杂度降低了约90%,几乎没有性能损失。由于信道的角度互易性,在同一相干时隙内,下行信道的最优正交空间基可以由上行信道的最优正交空间基直接得出,因此本申请所提出正交空间基选取方案对于提升2D-SBEM下大规模MIMO系统上下行信道估计性能有重要意义。
尽管在上文中参考特定的实施例对本申请进行了描述,但是所属领域技术人员应当理解,在本申请公开的原理和范围内,可以针对本申请公开的配置和细节做出许多修改。本申请的保护范围由所附的权利要求来确定,并且权利要求意在涵盖权利要求中技术特征的等同物文字意义或范围所包含的全部修改。

Claims (7)

  1. 一种快速搜索方法,其特征在于:所述方法包括如下步骤:
    步骤1):基站根据第k个用户的上行信道向量h k计算其离散傅里叶变换,确定第k个用户的空间特征
    Figure PCTCN2020135753-appb-100001
    步骤2):初始化N个角度值,分别计算N个旋转角度下信道向量的离散傅里叶变换,根据离散傅里叶变换计算N个角度下稀疏信道能量;
    步骤3):以旋转角度为自变量,稀疏信道能量为因变量,用m阶多项式对N个点的数据进行多项式拟合;
    步骤4):计算在角度范围内多项式的最大值作为第k个用户的旋转角度,结合其空间特征获得信道最优正交空间基;
    步骤5):对其他K-1个用户重复步骤1)至步骤3),更新当前相干时隙内每个用户的信道最优正交空间基。
  2. 如权利要求1所述的快速搜索方法,其特征在于:所述步骤1)中第k个用户的空间特征
    Figure PCTCN2020135753-appb-100002
    确定如下:
    设基站端有M>>1根天线且为均匀线阵,K个用户均为单天线用户,则第k个用户的上行信道向量h k可用M×1的向量表示。在大规模MIMO系统中,第k个用户的上行信道的径的入射角范围为[θ k-Δθ kk+Δθ k],每条径的入射角在这个范围满足均匀分布,信道在入射角域上具有稀疏的特性。
  3. 如权利要求2所述的快速搜索方法,其特征在于:所述步骤1)中信道的离散傅里叶变换可以表示为
    Figure PCTCN2020135753-appb-100003
    其中F是M×M的矩阵,其第p行第q列的元素为
    Figure PCTCN2020135753-appb-100004
    当基站天线数量趋于无穷时,每条径的能量集中在DFT域的一个点;当基站天线数量有限时,出现能量泄漏,中心点周围的DFT点包含少量信道能量,信道能量在DFT域仍呈现高度集中的特性;根据能量泄漏的特性,计算下标集合
    Figure PCTCN2020135753-appb-100005
    为:
    Figure PCTCN2020135753-appb-100006
    其中,系数β定义为函数h(x)=|sin(M/2x)/sin(1/2x)|在一个周期内次最大极值和最大极值的比值;
    Figure PCTCN2020135753-appb-100007
    表示向量
    Figure PCTCN2020135753-appb-100008
    第q 0个元素;计算空间特征
    Figure PCTCN2020135753-appb-100009
    为:
    Figure PCTCN2020135753-appb-100010
    其中,
    Figure PCTCN2020135753-appb-100011
    q max和q min分别为下标集合
    Figure PCTCN2020135753-appb-100012
    中的最大值和最小值,n等于
    Figure PCTCN2020135753-appb-100013
    向上取整,τ<<M为预先设定的正交空间基的数量。
  4. 如权利要求1所述的快速搜索方法,其特征在于:所述步骤2)中第k个用户旋转角度的初始化计算如下:
    Figure PCTCN2020135753-appb-100014
    为间隙大小在角度范围
    Figure PCTCN2020135753-appb-100015
    均匀选取N个点,即第n个角度选取为
    Figure PCTCN2020135753-appb-100016
    信道旋转该角度后的离散傅里叶变换为
    Figure PCTCN2020135753-appb-100017
    其中矩阵Φ(φ n)是由元素
    Figure PCTCN2020135753-appb-100018
    组成的对角矩阵。
  5. 如权利要求4所述的快速搜索方法,其特征在于:所述步骤2)中第k个用户旋转角度的稀疏信道能量的计算如下:
    结合用户的空间特征,第k个用户的上行信道在φ n的旋转角度下经稀疏表示后的能量为:
    Figure PCTCN2020135753-appb-100019
    其中,
    Figure PCTCN2020135753-appb-100020
    表示FΦ(φ n)h k的子向量,该向量包含了以
    Figure PCTCN2020135753-appb-100021
    中元素作为索引的FΦ(φ n)h k中的元素;
    Figure PCTCN2020135753-appb-100022
    表示向量
    Figure PCTCN2020135753-appb-100023
    的2-范数的平方,即稀疏信道的能量值;
    分别对n个角度重复以上步骤,即可得到第k个用户的旋转角度和稀疏信道的能量的对应关系,记作ε k=f(φ k)。
  6. 如权利要求5所述的快速搜索方法,其特征在于:所述步骤3)中多项式拟合的具体方法如下:
    设目标多项式为
    Figure PCTCN2020135753-appb-100024
    其中,为了保证解的唯一性,多项式阶数m满足m<N。定义阶次向量(共m个)为:
    Figure PCTCN2020135753-appb-100025
    定义目标向量为
    Figure PCTCN2020135753-appb-100026
    将多项式系数对应的向量记作a m=[a 0,a 1,...,a m] T,根据阶次向量生成矩阵Φ m=[φ 01,...,φ m],最优多项式拟合问题可转化为求解系数向量的最优解:
    Figure PCTCN2020135753-appb-100027
    采用最小二乘法获得最小方差解为
    Figure PCTCN2020135753-appb-100028
    其中
    Figure PCTCN2020135753-appb-100029
    是矩阵Φ m的伪逆矩阵。
  7. 如权利要求5所述的快速搜索方法,其特征在于:所述步骤4)第k个用户旋转角度和最优正交空间基的确定如下:
    已知多项式
    Figure PCTCN2020135753-appb-100030
    对其求导可得:
    Figure PCTCN2020135753-appb-100031
    令f'(φ k)=0,在自变量区间
    Figure PCTCN2020135753-appb-100032
    方程有一或多个解。将解集分别代入f(φ k),取最大的f(φ k)对应的φ k作为第k个用户的旋转角度φ k,opt
    结合第k个用户的旋转角度φ k,opt和空间特征
    Figure PCTCN2020135753-appb-100033
    第k个用户上行信道向量的最优正交空间基集合确定为:
    Figure PCTCN2020135753-appb-100034
    其中,Φ(φ k,opt) H表示矩阵Φ(φ k,opt)的共轭转置矩阵,向量v q表示离散傅里叶变换矩阵F共轭转置后的第q列。
PCT/CN2020/135753 2020-04-03 2020-12-11 一种快速搜索方法 WO2021196726A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010260247.2A CN111614386B (zh) 2020-04-03 2020-04-03 一种快速搜索方法
CN202010260247.2 2020-04-03

Publications (1)

Publication Number Publication Date
WO2021196726A1 true WO2021196726A1 (zh) 2021-10-07

Family

ID=72195674

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/135753 WO2021196726A1 (zh) 2020-04-03 2020-12-11 一种快速搜索方法

Country Status (2)

Country Link
CN (1) CN111614386B (zh)
WO (1) WO2021196726A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111614386B (zh) * 2020-04-03 2021-04-27 西安交通大学 一种快速搜索方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8416896B2 (en) * 2004-10-06 2013-04-09 Broadcom Corporation Method and system for channel estimation in a single channel MIMO system with multiple RF chains for WCDMA/HSDPA
CN110380994A (zh) * 2019-05-13 2019-10-25 上海海事大学 快速贝叶斯匹配追踪海上稀疏信道估计方法
CN110460549A (zh) * 2019-08-02 2019-11-15 南京邮电大学 一种新颖的多用户3d mimo系统的信道估计方法
CN110636018A (zh) * 2019-09-29 2019-12-31 哈尔滨工程大学 一种网格补偿大规模mimo信道估计方法
CN111614386A (zh) * 2020-04-03 2020-09-01 西安交通大学 一种快速搜索方法

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103684700B (zh) * 2013-12-31 2017-05-24 重庆邮电大学 一种基于正交联合码本集的3d mu‑mimo预编码方法
US11374635B2 (en) * 2018-06-22 2022-06-28 Samsung Electronics Co., Ltd. Method and apparatus for sensor assisted beam selection, beam tracking, and antenna module selection
CN110166383B (zh) * 2019-05-15 2021-07-02 南京邮电大学 一种基于树状随机搜索导频设计方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8416896B2 (en) * 2004-10-06 2013-04-09 Broadcom Corporation Method and system for channel estimation in a single channel MIMO system with multiple RF chains for WCDMA/HSDPA
CN110380994A (zh) * 2019-05-13 2019-10-25 上海海事大学 快速贝叶斯匹配追踪海上稀疏信道估计方法
CN110460549A (zh) * 2019-08-02 2019-11-15 南京邮电大学 一种新颖的多用户3d mimo系统的信道估计方法
CN110636018A (zh) * 2019-09-29 2019-12-31 哈尔滨工程大学 一种网格补偿大规模mimo信道估计方法
CN111614386A (zh) * 2020-04-03 2020-09-01 西安交通大学 一种快速搜索方法

Also Published As

Publication number Publication date
CN111614386B (zh) 2021-04-27
CN111614386A (zh) 2020-09-01

Similar Documents

Publication Publication Date Title
CN112910807B (zh) 一种基于空间随机采样的智能超表面信道估计方法及系统
Xie et al. Channel estimation for TDD/FDD massive MIMO systems with channel covariance computing
Ma et al. High-resolution channel estimation for frequency-selective mmWave massive MIMO systems
US8032184B2 (en) Method for generating downlink beamforming weighting vectors
CN102175989B (zh) 一种非相干分布式信号二维波达角的测定方法
WO2021227482A1 (zh) 一种大规模天线系统中的安全传输方法
CN109490850A (zh) 主瓣干扰下宽带阵列自适应波束形成方法
WO2022104993A1 (zh) 天波大规模mimo通信方法、模型及系统
CN110971547B (zh) 一种基于毫米波/太赫兹的宽带大规模终端检测和信道估计方法
Zhang et al. Channel estimation for massive MIMO-OFDM systems by tracking the joint angle-delay subspace
CN108667493A (zh) 一种面向大规模mimo中nlos场景下的波束赋形方法
CN106788655A (zh) 互耦条件下未知互耦信息的干扰相干稳健波束形成方法
WO2023185373A1 (zh) 天波大规模mimo三重波束基信道建模及信道信息获取
CN110166401A (zh) 大规模mimo正交频分复用系统的相位噪声抑制方法
WO2021196726A1 (zh) 一种快速搜索方法
CN110212951B (zh) 一种基于Butler矩阵的大规模MIMO信道估计方法
Li et al. Reconfigurable intelligent surface aided position and orientation estimation based on joint beamforming with limited feedback
CN108572347A (zh) 基于通信信号信道状态响应信息的面阵二维测角方法及系统
CN105847196B (zh) 一种基于DOA的多小区Massive MIMO系统的信道估计方法
Lin et al. Nonorthogonal waveform assisted DOA estimation for joint MIMO sensing and communication
Ji et al. Multipath extraction based UL/DL channel estimation for FDD massive MIMO-OFDM systems
Han et al. Superdirectivity-enhanced wireless communications: A multi-user perspective
CN108900290B (zh) 一种基于定位信息的导频分配方法
Wu et al. Time-Frequency-Space Transmit Design and Signal Processing with Dynamic Subarray for Terahertz Integrated Sensing and Communication
Xu et al. Channel Feature Projection Clustering Based Joint Channel and DoA Estimation for ISAC Massive MIMO OFDM System

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20928932

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20928932

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 20928932

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 17/05/2023)

122 Ep: pct application non-entry in european phase

Ref document number: 20928932

Country of ref document: EP

Kind code of ref document: A1