CN102664835A

CN102664835A - Multicell channel estimation method based on interference elimination route by route

Info

Publication number: CN102664835A
Application number: CN2011102823029A
Authority: CN
Inventors: 黄敏; 张润福; 周春晖; 赵明; 许希斌; 张秀军
Original assignee: Tsinghua University
Current assignee: Tsinghua University
Priority date: 2011-09-21
Filing date: 2011-09-21
Publication date: 2012-09-12
Anticipated expiration: 2031-09-21
Also published as: CN102664835B

Abstract

The invention relates to a multicell channel estimation method based on interference elimination route by route, provides a thought of interference elimination based ona route, and in this thought framework, provides two specific methods which can be used to eliminate interference. After analysis, the multicell channel estimation method based on interference elimination route by route not only can get a channel impulse response with high accuracy, but also has a low computational complexity. Therefore, the method provided by the invention can getthe channel impulse response with high accuracy in the case of low computational complexity, and a good basis for joint detection is provided.

Description

Multi-cell Channel Estimation Method Based on Path-by-Path Interference Cancellation

技术领域 technical field

本发明涉及码分多址移动通信技术领域，具体涉及一种基于逐径干扰消除的多小区信道估计方法。 The invention relates to the technical field of code division multiple access mobile communication, in particular to a multi-cell channel estimation method based on path-by-path interference elimination. the

背景技术 Background technique

TD-SCDMA（Time Division-Synchronous Code Division Multiple-Access）是第三代移动通信技术标准之一，采用时分双工。TD-SCDMA采用了诸多先进的通信技术，如智能天线、动态信道分配、联合检测等。 TD-SCDMA (Time Division-Synchronous Code Division Multiple-Access) is one of the third-generation mobile communication technology standards, using time division duplex. TD-SCDMA adopts many advanced communication technologies, such as smart antenna, dynamic channel allocation, joint detection and so on. the

在3GPP规范的TD-SCDMA系统中，数据以突发形式传送。midamble码（也被称为训练序列）长度为144chips，是插在两个352chips数据域中间，用于接收端进行信道估计的。TD-SCDMA系统中共有128个准正交而非完全正交的基本midamble序列，它们被分为32组。在每个小区中，基站从给定组中的四个基本midamble序列中选取一个。邻近小区使用准正交而非完全正交的midamble序列，同一个小区不同用户的midamble码是由某个基本的midamble序列经过循环移位得到的。在同频组网的多小区下行传输场景中，用户端（UE）收到的信号不仅包含本小区基站发送的信号，还包含了邻近小区基站发送的信号。因此，当UE进行信道估计时，midamble序列的非正交性就引起了小区间干扰。同理，在同频组网的多小区上行传输场景中，基站端收到的信号不仅包含本小区用户发送的信号，并且包含了邻近小区用户发送的信号，当基站进行信道估计时，midamble序列的非正交性同样会引起小区间干扰。 In the TD-SCDMA system specified by 3GPP, data is transmitted in bursts. The midamble code (also known as the training sequence) has a length of 144chips and is inserted between two 352chips data fields for channel estimation at the receiving end. There are 128 quasi-orthogonal rather than completely orthogonal basic midamble sequences in the TD-SCDMA system, and they are divided into 32 groups. In each cell, the base station selects one of the four basic midamble sequences in a given group. Adjacent cells use quasi-orthogonal instead of completely orthogonal midamble sequences, and the midamble codes of different users in the same cell are obtained by cyclic shifting of a basic midamble sequence. In the multi-cell downlink transmission scenario of the same-frequency networking, the signal received by the user end (UE) includes not only the signal sent by the base station of the current cell, but also the signal sent by the base station of the adjacent cell. Therefore, when the UE performs channel estimation, the non-orthogonality of the midamble sequence causes inter-cell interference. Similarly, in the multi-cell uplink transmission scenario of the same frequency network, the signal received by the base station includes not only the signal sent by the user in the local cell, but also the signal sent by the user in the adjacent cell. When the base station performs channel estimation, the midamble sequence The non-orthogonality of will also cause inter-cell interference. the

传统的Steiner信道估计方法，可采用FFT/IFFT进行简化计算。由于在估计某一小区的信道冲击响应时，其他小区用户的信号都作为白噪声处理，噪声和干扰的总功率过大，因此影响了信道估计的准确性。从而使用Steiner算法仅仅可以得到初始的准确性较差的信道冲击响应（CIR），但这一初始的CIR可以作为更加精确的信道估计算法的输入。 The traditional Steiner channel estimation method can use FFT/IFFT to simplify the calculation. When estimating the channel impulse response of a certain cell, the signals of users in other cells are treated as white noise, and the total power of noise and interference is too large, thus affecting the accuracy of channel estimation. Therefore, only an initial channel impulse response (CIR) with poor accuracy can be obtained by using the Steiner algorithm, but this initial CIR can be used as an input of a more accurate channel estimation algorithm. the

基于码片的并行干扰消除算法就是使用传统Steiner算法得到初始的CIR，并对初始的CIR进行降噪处理。然后重构邻近小区的midamble信号，并将其从降噪后的CIR中逐个码片地去除。由于此并行干扰消除方法是逐个码片进行的，因此计算复杂度较高。在另一种传统的联合信道估计算法中，信道估计矩阵由邻近小区中的强干扰用户组成，虽然信道估计精度得到了提高，但是估计过程中需要进行矩阵求逆运算，计算复杂度高。 The chip-based parallel interference elimination algorithm is to use the traditional Steiner algorithm to obtain the initial CIR, and perform noise reduction processing on the initial CIR. The midamble signals of neighboring cells are then reconstructed and removed chip-by-chip from the denoised CIR. Since this parallel interference cancellation method is performed chip by chip, the computational complexity is relatively high. In another traditional joint channel estimation algorithm, the channel estimation matrix is composed of strong interfering users in adjacent cells. Although the accuracy of channel estimation is improved, matrix inversion operation is required in the estimation process, which has high computational complexity. the

发明内容 Contents of the invention

为了克服上述现有技术存在的不足，本发明的目的在于提供一种基于逐径干扰消除的多小区信道估计方法，真正实现了提供一种高精度、低复杂度的多小区信道估计方法，即基于逐径干扰消除的多小区信道估计方法。 In order to overcome the deficiencies in the above-mentioned prior art, the purpose of the present invention is to provide a multi-cell channel estimation method based on path-by-path interference cancellation, which truly provides a high-precision, low-complexity multi-cell channel estimation method, namely Multi-cell channel estimation method based on path-by-path interference cancellation. the

为了达到上述目的，本发明所采用的技术方案是： In order to achieve the above object, the technical scheme adopted in the present invention is:

一种基于逐径干扰消除的多小区信道估计方法，步骤如下： A multi-cell channel estimation method based on path-by-path interference cancellation, the steps are as follows:

步骤1：采用如下公式 Step 1: Use the following formula

$\overset{^^}{h h} = = IFFT IFFT ((\frac{FFT FFT (({e e}_{mid middle}))}{FFT FFT (({m m}_{basic basic}))}))$

得到当前服务小区的初始CIR值

其中，e_mid表示接收端收到的midamble序列，m_basic表示小区使用的基本midamble序列； Get the initial CIR value of the current serving cell

Among them, e _mid represents the midamble sequence received by the receiving end, and m _basic represents the basic midamble sequence used by the cell;

步骤2：对初始CIR进行降噪处理，选取出初始CIR中功率超过设定门限的元素，从而分离提取出各小区的主要路径； Step 2: Perform noise reduction processing on the initial CIR, select the elements in the initial CIR whose power exceeds the set threshold, and then separate and extract the main paths of each cell;

步骤3：对当前服务小区相邻的小区也依次采用步骤（1）和（2）的方法，得到相邻的小区经过降噪处理后的CIR和其主要路径； Step 3: Use steps (1) and (2) in sequence for the cells adjacent to the current serving cell to obtain the CIR and the main path of the adjacent cells after noise reduction processing;

步骤4：逐径地消除相邻的小区干扰，进而得到最终CIR。 Step 4: Eliminate adjacent cell interference path by path, and then obtain the final CIR. the

所述的消除相邻的小区干扰的方法为并行干扰消除方法时，进行干扰消除的步骤如下： When the described method for eliminating adjacent cell interference is a parallel interference elimination method, the steps for interference elimination are as follows:

步骤1：首先设定当前服务小区所使用的midamble码记为m₁，相邻的小区使用的midamble分别记为m₂，m₃，…，m_n，当前服务小区以及相邻的小区的主要路径数依次分别为N₁，N₂，…，N_n，其中n为大于1的整数； Step 1: First set the midamble code used by the current serving cell as m ₁ , and the midamble codes used by adjacent cells are respectively recorded as m ₂ , m ₃ ,..., m _n , the main codes of the current serving cell and adjacent cells The number of paths is respectively N ₁ , N ₂ , ..., N _n , where n is an integer greater than 1;

步骤2：得到包含小区间干扰的信道冲击响应y，即由N＝N₁+N₂+…+N_n个选出的元素组成，表示为： Step 2: Obtain the channel impulse response y including inter-cell interference, which is composed of N=N ₁ +N ₂ +...+N _n selected elements, expressed as:

$y the y = = {[[{\overset{~ ~}{h h}}_{11} ((00)),, {\overset{~ ~}{h h}}_{11} ((11)),, \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot;,, {\overset{~ ~}{h h}}_{11} (({N N}_{11} - - 11)),, {\overset{~ ~}{h h}}_{22} ((00)),, {\overset{~ ~}{h h}}_{22} ((11)),, \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot;,, {\overset{~ ~}{h h}}_{22} (({N N}_{22} - - 11)) \cdot \cdot \cdot &Center Dot; \cdot &Center Dot;,, {\overset{~ ~}{h h}}_{n no} ((00)),, {\overset{~ ~}{h h}}_{n no} ((11)),, \cdot \cdot \cdot \cdot \cdot &Center Dot;,, {\overset{~ ~}{h h}}_{n no} (({N N}_{n no} - - 11))]]}^{T T}$

其中依次为当前服务小区以及相邻的小区的主要路径，i为大于等于1小于等于n的整数； in In turn, they are the main paths of the current serving cell and adjacent cells, and i is an integer greater than or equal to 1 and less than or equal to n;

步骤3：获得N×N的干扰矩阵B，B＝R-E，矩阵R表示当前服务小区以及相邻的小区间各径之间的影响权重矩阵，其中E为单位阵，R_ij表示路径

对路径

的影响权重，j为大于等于1小于等于n的整数,其中第i个小区对第j个小区的影响权重矩阵由如下公式 Step 3: Obtain N×N interference matrix B, B=RE, matrix R represents the influence weight matrix between the current serving cell and the paths between adjacent cells, where E is the identity matrix, and R _ij represents the path

on the path

j is an integer greater than or equal to 1 and less than or equal to n, where the influence weight matrix of the i-th cell on the j-th cell is given by the following formula

${R R}_{ij ij} = = IFFT IFFT ((\frac{FFT FFT (({m m}_{basicj basicj}))}{FFT FFT (({m m}_{basici basici}))}))$

计算得到； calculated;

步骤4：设置迭代的初始值β⁽⁰⁾＝y，设定预定义迭代终止条件，包括最大迭代次数，信道估计精度要求； Step 4: Set the initial value of the iteration β ⁽⁰⁾ = y, set the predefined iteration termination conditions, including the maximum number of iterations, channel estimation accuracy requirements;

步骤5：根据如下迭代公式得到β^(l)，l=0,1… Step 5: Get β ^(l) according to the following iterative formula, l=0,1...

β^(l)=y-Bβ^(l-1),l＝1,2,…,∞. β ^(l) =y-Bβ ^(l-1) ,l=1,2,…,∞.

步骤6：如果β^(l)符合预定义的迭代终止条件，则迭代结束，否则返回步骤（3）； Step 6: If β ^(l) meets the predefined iteration termination condition, then the iteration ends, otherwise return to step (3);

步骤7：将迭代结果β^(l)中各小区的主要路径代入经过降噪处理的CIR中，得到多个小区的经过干扰消除的最终CIR。 Step 7: Substituting the main path of each cell in the iteration result β ^(l) into the CIR after noise reduction processing, to obtain the final CIR of multiple cells after interference elimination.

所述的消除相邻的小区干扰的方法为串行干扰消除方法时，进行干扰消除的步骤如下： When the described method for eliminating adjacent cell interference is a serial interference elimination method, the steps for interference elimination are as follows:

$y the y = = {[[{\overset{~ ~}{h h}}_{11} ((00)),, {\overset{~ ~}{h h}}_{11} ((11)),, \cdot \cdot \cdot \cdot \cdot &Center Dot;,, {\overset{~ ~}{h h}}_{11} (({N N}_{11} - - 11)),, {\overset{~ ~}{h h}}_{22} ((00)),, {\overset{~ ~}{h h}}_{22} ((11)),, \cdot \cdot \cdot \cdot \cdot \cdot,, {\overset{~ ~}{h h}}_{22} (({N N}_{22} - - 11)) \cdot \cdot \cdot \cdot \cdot \cdot,, {\overset{~ ~}{h h}}_{n no} ((00)),, {\overset{~ ~}{h h}}_{n no} ((11)),, \cdot &Center Dot; \cdot &Center Dot; \cdot \cdot,, {\overset{~ ~}{h h}}_{n no} (({N N}_{n no} - - 11))]]}^{T T}$

其中

依次为当前服务小区以及相邻的小区的主要路径，i为大于等于1小于等于n的整数； in

In turn, they are the main paths of the current serving cell and adjacent cells, and i is an integer greater than or equal to 1 and less than or equal to n;

步骤3：针对并行干扰消除中步骤1得到的y矩阵，按功率进行从大到小排序并选出功率最大径h_i，i为大于等于1小于等于n的整数； Step 3: For the y matrix obtained in step 1 in parallel interference cancellation, sort according to power from large to small and select the maximum power path h _i , where i is an integer greater than or equal to 1 and less than or equal to n;

步骤4：获得N×N的干扰矩阵B，B＝R-E，矩阵R表示当前服务小区以及相邻的小区间各径之间的影响权重矩阵，其中E为单位阵，R_ij表示路径

对路径

的影响权重，j为大于等于1小于等于n的整数,其中第i个小区对第j个小区的影响权重矩阵由如下公式 Step 4: Obtain N×N interference matrix B, B=RE, matrix R represents the influence weight matrix between the current serving cell and the paths between adjacent cells, where E is the unit matrix, and R _ij represents the path

on the path

计算得到； calculated;

步骤5：根据如下公式，逐径地消除h_i对其他径h_j的干扰。 Step 5: Eliminate the interference of h _i on other paths h _j path by path according to the following formula.

h_j＝h_j-R_jih_i h _j ＝h _j -R _ji h _i

步骤6：对y矩阵重新排序，除去已选择过的径，选出功率最大径，返回步骤4，直到y矩阵中所有径均已被选择过为止，得到多个小区的经过干扰消除的最终CIR。 Step 6: Reorder the y matrix, remove the selected paths, select the maximum power path, return to step 4, until all the paths in the y matrix have been selected, and get the final interference-eliminated results of multiple cells CIR. the

本发明的优点在于提供了一种基于路径的干扰消除的思路，在这一思路框架下，提供了两种可以使用的干扰消除的具体方法。经过分析，基于逐径干扰消除的多小区信道估计方法不仅可以得到高精度的信道冲击响应，而且具有较低的计算复杂度。因此本发明提供的方法可以在较低计算复杂度的情况下，获得高精度的信道冲击响应，进而为联合检测提供良好的基础。 The advantage of the present invention is that it provides an idea of path-based interference elimination, and under the framework of this idea, two specific methods of interference elimination that can be used are provided. After analysis, the multi-cell channel estimation method based on path-by-path interference cancellation can not only obtain high-precision channel impulse response, but also has low computational complexity. Therefore, the method provided by the present invention can obtain a high-precision channel impulse response with relatively low computational complexity, thereby providing a good basis for joint detection. the

附图说明 Description of drawings

图1是TD-SCDMA突发结构。 Fig. 1 is TD-SCDMA burst structure. the

图2是两小区系统模型。 Figure 2 is a two-cell system model. the

具体实施方式 Detailed ways

下面结合附图和实施例对本发明作更详细的说明。 The present invention will be described in more detail below in conjunction with the accompanying drawings and embodiments. the

基于逐径干扰消除的多小区信道估计方法，步骤如下： The multi-cell channel estimation method based on path-by-path interference cancellation, the steps are as follows:

步骤1：采用如下公式 Step 1: Use the following formula

得到当前服务小区的初始CIR值

所述的消除相邻的小区干扰的方法并行干扰消除方法时，进行干扰消除的步骤如下： During the parallel interference elimination method of the method for eliminating adjacent cell interference, the steps for interference elimination are as follows:

$y the y = = {[[{\overset{~ ~}{h h}}_{11} ((00)),, {\overset{~ ~}{h h}}_{11} ((11)),, \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot;,, {\overset{~ ~}{h h}}_{11} (({N N}_{11} - - 11)),, {\overset{~ ~}{h h}}_{22} ((00)),, {\overset{~ ~}{h h}}_{22} ((11)),, \cdot &Center Dot; \cdot &Center Dot; \cdot \cdot,, {\overset{~ ~}{h h}}_{22} (({N N}_{22} - - 11)) \cdot \cdot \cdot \cdot \cdot &Center Dot;,, {\overset{~ ~}{h h}}_{n no} ((00)),, {\overset{~ ~}{h h}}_{n no} ((11)),, \cdot \cdot \cdot \cdot \cdot &Center Dot;,, {\overset{~ ~}{h h}}_{n no} (({N N}_{n no} - - 11))]]}^{T T}$

其中

对路径的影响权重，j为大于等于1小于等于n的整数,其中第i个小区对第j个小区的影响权重矩阵由如下公式 Step 3: Obtain N×N interference matrix B, B=RE, matrix R represents the influence weight matrix between the current serving cell and the paths between adjacent cells, where E is the identity matrix, and R _ij represents the path

on the path j is an integer greater than or equal to 1 and less than or equal to n, where the influence weight matrix of the i-th cell on the j-th cell is given by the following formula

计算得到； calculated;

β^(l)=y-Bβ^(l-1),l＝1,2,…,∞. β ^(l) =y-Bβ ^(l-1) ,l=1,2,…,∞.

$y the y = = {[[{\overset{~ ~}{h h}}_{11} ((00)),, {\overset{~ ~}{h h}}_{11} ((11)),, \cdot \cdot \cdot &Center Dot; \cdot \cdot,, {\overset{~ ~}{h h}}_{11} (({N N}_{11} - - 11)),, {\overset{~ ~}{h h}}_{22} ((00)),, {\overset{~ ~}{h h}}_{22} ((11)),, \cdot \cdot \cdot &Center Dot; \cdot &Center Dot;,, {\overset{~ ~}{h h}}_{22} (({N N}_{22} - - 11)) \cdot \cdot \cdot &Center Dot; \cdot \cdot,, {\overset{~ ~}{h h}}_{n no} ((00)),, {\overset{~ ~}{h h}}_{n no} ((11)),, \cdot \cdot \cdot &Center Dot; \cdot &Center Dot;,, {\overset{~ ~}{h h}}_{n no} (({N N}_{n no} - - 11))]]}^{T T}$

其中

对路径的影响权重，j为大于等于1小于等于n的整数,其中第i个小区对第j个小区的影响权重矩阵由如下公式 Step 4: Obtain N×N interference matrix B, B=RE, matrix R represents the influence weight matrix between the current serving cell and the paths between adjacent cells, where E is the unit matrix, and R _ij represents the path

计算得到； calculated;

h_j＝h_j-R_jih_i h _j ＝h _j -R _ji h _i

步骤6：对y矩阵重新排序，除去已选择过的径，选出功率最大径，返回步骤4，直到y矩阵中所有径均已被选择过为止，得到多个小区的经过干扰消除的最终CIR。 Step 6: Reorder the y matrix, remove the selected paths, select the maximum power path, return to step 4, until all the paths in the y matrix have been selected, and obtain the final CIR of multiple cells after interference cancellation . the

具体说来以图2所示的两小区模型为例，其中，当前服务小区所适用的midamble码为m₁，邻近小区使用的midamble为m₂。假设在当前服务小区有4个用户，邻近服务小区也有4个用户。在某一时隙内最大用户数为8（K_cell=8），CIR的长度为16（W＝16）。分别在Rec.ITU-R M.1225定义的Pedestrian A（PedA）和Vehi cular A(VehA)两种信道下,采用基于逐径干扰消除的多小区信道估计方法。两种信道的信道参数如表1所示。仿真场景其它参数如表2所示。 Specifically, take the two-cell model shown in FIG. 2 as an example, where the midamble code used by the current serving cell is m ₁ , and the midamble code used by the neighboring cell is m ₂ . Assume that there are 4 users in the current serving cell, and there are 4 users in the adjacent serving cell. The maximum number of users in a time slot is 8 (K _cell =8), and the length of the CIR is 16 (W = 16). Under the two channels of Pedestrian A (PedA) and Vehicular A (VehA) defined in Rec.ITU-R M.1225 respectively, a multi-cell channel estimation method based on path-by-path interference cancellation is adopted. The channel parameters of the two channels are shown in Table 1. Other parameters of the simulation scene are shown in Table 2.

表1 Table 1

表2 Table 2

Parameter Parameter Value value Noise Power Noise Power -104dBm -104dBm Noise Figure Noise Figure 9dB 9dB NodeB Power NodeB Power 34dBm 34dBm PhyCH Power PhyCH Power 22dBm 22dBm

[0081] [0081] Antenna Gain Antenna Gain 11dB 11dB Cell Radius Cell Radius 866m 866m

假设经过提取，当前服务小区的主要路径数为N₁，邻小区的主要路径数为N₂。 Assume that after extraction, the number of main paths in the current serving cell is N ₁ , and the number of main paths in neighboring cells is N ₂ .

采用并行干扰消除方法进行干扰消除，步骤如下： Using the parallel interference elimination method to eliminate interference, the steps are as follows:

（1）得到包含小区间干扰的信道冲击响应y，由N₁+N₂个选出的元素可表示为： (1) To obtain the channel impulse response y including inter-cell interference, the elements selected from N ₁ +N ₂ can be expressed as:

$y the y = = {[[{\overset{~ ~}{h h}}_{11} ((00)),, {\overset{~ ~}{h h}}_{11} ((11)),, \cdot &Center Dot; \cdot &Center Dot; \cdot &Center Dot;,, {\overset{~ ~}{h h}}_{11} (({N N}_{11} - - 11)),, {\overset{~ ~}{h h}}_{22} ((00)),, {\overset{~ ~}{h h}}_{22} ((11)),, \cdot \cdot \cdot \cdot \cdot &Center Dot;,, {\overset{~ ~}{h h}}_{22} (({N N}_{22} - - 11))]]}^{T T}$

其中

为服务小区的主要路径，

为邻小区的主要路径。 in

is the main path serving the cell,

is the main path of the neighboring cell.

（2）获得(N₁+N₂)×(N₁+N₂)的干扰矩阵B，B＝R-E，。矩阵R表示小区间各径之间的影响权重，其中r_ij表示路径j对路径i的影响权重。因为只存在着小区间干扰，而没有小区内的干扰。因此R矩阵中的对角块N₁×N₁矩阵和N₂×N₂矩阵除了对角线元素为1，其他元素均为0。逆对角线块分别表示服务小区对邻小区的影响权重矩阵R₁₂及邻小区对服务小区的影响权重矩阵R₂₁，可分别如下公式计算得到。 (2) Obtain an interference matrix B of (N ₁ +N ₂ )×(N ₁ +N ₂ ), where B=RE. The matrix R represents the influence weight of each path between the cells, where r _ij represents the influence weight of path j on path i. Because there is only inter-cell interference, but no intra-cell interference. Therefore, the diagonal blocks N ₁ ×N ₁ matrix and N ₂ ×N ₂ matrix in the R matrix are 0 except for the diagonal elements which are 1. The anti-diagonal blocks respectively represent the influence weight matrix R ₁₂ of the serving cell on the neighboring cell and the influence weight matrix R ₂₁ of the neighboring cell on the serving cell, which can be calculated by the following formulas respectively.

${R R}_{1212} = = IFFT IFFT ((\frac{FFT FFT (({m m}_{basic basic 22}))}{FFT FFT (({m m}_{basic basic 11}))}))$

${R R}_{21 twenty one} = = IFFT IFFT ((\frac{FFT FFT (({m m}_{basic basic 11}))}{FFT FFT (({m m}_{basic basic 22}))}))$

（3）设置迭代的初始值β⁽⁰⁾=y，设定预定义迭代终止条件，如最大迭代次数，信道估计精度要求等 (3) Set the initial value of the iteration β ⁽⁰⁾ =y, and set the predefined iteration termination conditions, such as the maximum number of iterations, channel estimation accuracy requirements, etc.

（4）根据并行干扰消除步骤（4）中的迭代公式得到β^(l)，l＝0,1…。 (4) Obtain β ^(l) according to the iterative formula in the parallel interference elimination step (4), l=0,1....

（5）如果β^(l)符合预定义的迭代终止条件，则迭代结束，否则返回步骤（4） (5) If β ^(l) meets the predefined iteration termination condition, the iteration ends, otherwise return to step (4)

（6）将迭代结果β^(l)中前N₁个元素对应的代入经过降噪处理的服务小区的CIR中，得到服务小区的最终的CIR。同理，将迭代结果β^(l)中后N₂个元素对应的代入经过降噪处理的邻小区的CIR中，得到邻小区的最终的CIR。 (6) Substitute the first N ₁ elements in the iteration result β ^(l) into the CIR of the serving cell after noise reduction processing, and obtain the final CIR of the serving cell. Similarly, the last N ₂ elements in the iteration result β ⁽¹⁾ are correspondingly substituted into the CIR of the neighboring cell after noise reduction processing, to obtain the final CIR of the neighboring cell.

采用串行干扰消除方法进行干扰消除，步骤如下： Using the serial interference elimination method to eliminate interference, the steps are as follows:

（1）针对并行干扰消除中步骤1得到的y矩阵，按功率进行从大到小排序；、 (1) For the y matrix obtained in step 1 in parallel interference cancellation, sort according to power from large to small;,

（2）选出功率最大径h_i (2) Select the maximum power diameter h _i

（3）由并行干扰消除中步骤2得到的R矩阵，根据串行干扰消除步骤（3）中的公式，逐径地消除h_i对其他径h_j的干扰。 (3) From the R matrix obtained in step 2 of parallel interference elimination, according to the formula in step (3) of serial interference elimination, eliminate the interference of h _i to other paths h _j path by path.

（4）对y矩阵重新排序，除去已选择过的径，选出功率最大径，返回步骤3，直到y矩阵中所有径均已被选择过为止。 (4) Reorder the y matrix, remove the selected path, select the path with the maximum power, and return to step 3 until all the paths in the y matrix have been selected. the

采用MSE（mean square error）和相关系数ρ来衡量该方案中信道估计的精度。 MSE (mean square error) and correlation coefficient ρ are used to measure the accuracy of channel estimation in this scheme. the

其中 in

$mse mse = = \sqrt{E E. (({((\overset{^^}{h h} - - h h)) . .}^{22}))}$

$ρ ρ = = | | \frac{cov cov ((\overset{^^}{h h},, h h))}{\sqrt{cov cov ((\overset{^^}{h h},, h h)) cov cov ((h h,, h h))}} | | . .$

本发明的优点在于提供了一种基于路径的干扰消除的思想，在这一思想框架下，提供了两种逐径并行干扰消除和逐径串行干扰消除方法。经过仿真分析，基于逐径干扰消除的多小区信道估计方法不仅可以提高信道估计的准确性，而且与逐码片干扰消除方法、迫零联合方法和最小均方误差联合算法相比，具有较低的计算复杂度。因此本发明提供的方法可以在低计算复杂度的情况下，获得高精度的信道冲击响应，进而为联合检测提供良好的基础。 The advantage of the present invention is that it provides an idea of path-based interference elimination, under the framework of this idea, two methods of path-by-path parallel interference elimination and path-by-path serial interference elimination are provided. After simulation analysis, the multi-cell channel estimation method based on path-by-path interference cancellation can not only improve the accuracy of channel estimation, but also has lower computational complexity. Therefore, the method provided by the present invention can obtain a high-precision channel impulse response with low computational complexity, thereby providing a good basis for joint detection. the

Claims

1. A multi-cell channel estimation method based on path-by-path interference cancellation, characterized in that the steps are as follows:

Step 1: Use the following formula

Get the initial CIR value of the current serving cell

Step 2: Perform noise reduction processing on the initial CIR, select the elements in the initial CIR whose power exceeds the set threshold, so as to separate and extract the main paths of each cell;

Step 3: The methods of steps (1) and (2) are also adopted in turn for the cells adjacent to the current serving cell to obtain the CIR and the main path of the adjacent cells after noise reduction processing;

Step 4: Eliminate adjacent cell interference path by path, and then obtain the final CIR.

2. The multi-cell channel estimation method based on path-by-path interference elimination according to claim 1, wherein the method for eliminating adjacent cell interference is a parallel interference elimination method for interference elimination, and the steps are as follows:

Step 1: First set the midamble code used by the current serving cell as m ₁ , and the midamble codes used by adjacent cells are respectively recorded as m ₂ , m ₃ ,..., m _n , the main codes of the current serving cell and adjacent cells The number of paths is respectively N ₁ , N ₂ , ..., N _n , where n is an integer greater than 1;

Step 2: Obtain the channel impulse response y including inter-cell interference, which is composed of N=N ₁ +N ₂ +...+N _n selected elements, expressed as:

in

Step 3: Obtain N×N interference matrix B, B=RE, matrix R represents the influence weight matrix between the current serving cell and the paths between adjacent cells, where E is the identity matrix, and R _ij represents the path on the path

calculated;

Step 4: Set the initial value of the iteration β ⁽⁰⁾ = y, set the predefined iteration termination conditions, including the maximum number of iterations, channel estimation accuracy requirements;

Step 5: Get β ^(l) according to the following iterative formula, l=0,1...

β ^(l) =y-Bβ ^(l-1) ,l=1,2,…,∞.

Step 6: If β ^(l) meets the predefined iteration termination condition, then the iteration ends, otherwise return to step (3);

Step 7: Substituting the main path of each cell in the iteration result β ^(l) into the CIR after noise reduction processing, to obtain the final CIR of multiple cells after interference elimination.

3. the multi-cell channel estimation method based on path-by-path interference elimination according to claim 1, wherein the method for eliminating the interference of adjacent cells is to adopt the serial interference elimination method to carry out interference elimination, and the steps are as follows:

in It is the main path of the current serving cell and the adjacent cell in turn, and i is an integer greater than or equal to 1 and less than or equal to n;

Step 3: For the y matrix obtained in step 1 in parallel interference cancellation, sort according to power from large to small and select the maximum power path h _i , where i is an integer greater than or equal to 1 and less than or equal to n;

Step 4: Obtain N×N interference matrix B, B=RE, matrix R represents the influence weight matrix between the current serving cell and the paths between adjacent cells, where E is the unit matrix, and R _ij represents the path

on the path

calculated;

Step 5: Eliminate the interference of h _i on other paths h _j path by path according to the following formula.

h _j ＝h _j -R _ji h _i

Step 6: Reorder the y matrix, remove the selected paths, select the maximum power path, return to step 4, until all the paths in the y matrix have been selected, and obtain the final CIR of multiple cells after interference cancellation . the