CN101453259A

CN101453259A - Pre-encoded transmission method for MIMO system

Info

Publication number: CN101453259A
Application number: CNA200710196093XA
Authority: CN
Inventors: 王玮; 张战; 加山英俊
Original assignee: NTT Docomo Inc
Current assignee: NTT Docomo Inc
Priority date: 2007-12-03
Filing date: 2007-12-03
Publication date: 2009-06-10
Also published as: JP2009141957A

Abstract

The present invention provides a precoding transmission method for a multiple-input multiple-output system, including: the mobile station assumes that the transmission matrix is a unitary matrix, uses an MMSE receiver to combine signals, calculates an equivalent channel, obtains CQI and CDI, and feeds back to the base station The base station obtains the zero-forcing precoding transmission matrix or the adjusted zero-forcing precoding transmission matrix corresponding to some or all possible user combinations according to the channel direction index and channel quality information fed back by the mobile station, and performs multi-user scheduling operation to determine the The scheduled user combination: the base station uses the zero-forcing precoding transmission matrix corresponding to the scheduled user combination or the adjusted zero-forcing precoding transmission matrix to perform precoding and transmit, and the scheduled users perform MMSE reception. The invention reduces the channel feedback error, thus effectively improving the frequency spectrum utilization rate of the system.

Description

A precoding transmission method for a multiple-input multiple-output system

技术领域 technical field

本发明涉及多用户多输入多输出系统中的预编码技术，特别是一种多用户多输入多输出系统的预编码传输方法。The invention relates to a precoding technology in a multi-user MIMO system, in particular to a precoding transmission method of a multi-user MIMO system.

背景技术 Background technique

未来无线通信系统要求能提供越来越高的信息传输速率和通信质量。为了在有限的频谱资源上实现这一目标，MIMO(Multi Input Multi Output，多输入多输出)技术已成为未来无线通信中所采用的必不可少的手段之一。Future wireless communication systems are required to provide higher and higher information transmission rates and communication quality. In order to achieve this goal on limited spectrum resources, MIMO (Multi Input Multi Output) technology has become one of the indispensable means used in future wireless communications.

在MIMO系统中，发送端利用多根天线进行信号的发送，接收端利用多根天线进行信号的接收。研究表明，相比于传统的单天线传输方法，MIMO技术可以显著的提高信道容量，从而提高信息传输速率。In the MIMO system, the transmitting end uses multiple antennas to transmit signals, and the receiving end uses multiple antennas to receive signals. Studies have shown that compared with traditional single-antenna transmission methods, MIMO technology can significantly increase channel capacity, thereby increasing information transmission rates.

在MIMO下行系统中采用预编码(pre-coding)的发送方法可以有效提高MIMO系统的性能。预编码的基本思想是根据当前的信道信息，对待发送数据在发送之前进行预处理，包括线性处理和非线性处理等。Using a pre-coding (pre-coding) transmission method in a MIMO downlink system can effectively improve the performance of the MIMO system. The basic idea of precoding is to preprocess the data to be sent before sending according to the current channel information, including linear processing and nonlinear processing.

下面对采用预编码的数据处理过程描述如下。The data processing process using precoding is described below.

假设小区中基站数目为1，移动台数目为K，基站有M个发射天线，每个移动台有N_k个接收天线。在某一时间和频率上，基站向被调度上的S≤K个用户发射数据流，即向第k(k＝1，2，...，S)个用户发射L_k个数据流，其中：Assume that the number of base stations in the cell is 1, the number of mobile stations is K, the base station has M transmitting antennas, and each mobile station has N _k receiving antennas. At a certain time and frequency, the base station transmits data streams to the scheduled S≤K users, that is, transmits L _k data streams to the kth (k=1, 2, ..., S) users, where :

L_k≤min{N_k，M}，且 $Σ_{k = 1}^{S} L_{k} = M$ L _k ≤ min{N _k , M}, and $Σ_{k = 1}^{S} L_{k} = m$

第k个用户的L_k×1维发射矢量s_k(数据符号)通过一个M×L_k维预编码发射矩阵T_k从M个天线发射出去，用户k的信道特性矩阵为N_k×M维矩阵H_k。The L _k ×1-dimensional transmit vector s _k (data symbol) of the k-th user is transmitted from M antennas through an M×L _k- dimensional precoding transmit matrix T _k , and the channel characteristic matrix of user k is N _k ×M dimensional Matrix H _k .

第k个用户接收到的N_k×1维信号矢量y_k，通过一L_k×N_k接收解码矩阵R_k，产生L_k×1维软输出矢量

如下：The N _k ×1-dimensional signal vector y _k received by the kth user receives and decodes the matrix R _k through a L _k ×N _k to generate an L _k ×1-dimensional soft output vector

as follows:

${\overset{^^}{s the s}}_{k k} = = {R R}_{k k} (({H h}_{k k} {Σ Σ}_{i i = = 11}^{S S} {T T}_{i i} {s the s}_{i i} + + {n no}_{k k})) = = {R R}_{k k} {H h}_{k k} {T T}_{k k} {s the s}_{k k} + + {R R}_{k k} {H h}_{k k} {\underset{i i = = 11}{Σ Σ}}_{i i &NotEqual; &NotEqual; k k}^{S S} {T T}_{i i} {s the s}_{i i} + + {R R}_{k k} {n no}_{k k}$

上式中的右边第二部分为多用户干扰，而n_k为噪音信号。The second part on the right side in the above formula is multi-user interference, and _nk is a noise signal.

假设每个用户都可以精确估计其信道特性矩阵H_k，但只能反馈有限信息给基站。基站与所有移动台拥有共同的码书(Codebook)，如DFT(DiscreteFlorier Transform，离散傅立叶变换)基矩阵，用户将反馈信道量化成为码书中的某个矢量，B比特反馈可以支持的码书最多包含2^B个矢量。Assume that each user can accurately estimate its channel characteristic matrix H _k , but can only feed back limited information to the base station. The base station and all mobile stations have a common codebook (Codebook), such as DFT (DiscreteFlorier Transform, Discrete Fourier Transform) base matrix, the user quantizes the feedback channel into a certain vector in the codebook, and B-bit feedback can support the most codebooks Contains 2 ^B vectors.

当L_k＝1时，每用户只反馈一个等价信道h_k＝r_kH_k的量化信息给基站。第k个用户的接收信号如下所示：When L _k =1, _each user only feeds back quantized information of one equivalent channel h _k =rk H _k to the base station. The received signal of the kth user is as follows:

${\overset{^^}{s the s}}_{k k} = = {r r}_{k k} (({H h}_{k k} Ts Ts + + {n no}_{k k})) = = {r r}_{k k} (({H h}_{k k} {Σ Σ}_{i i = = 11}^{S S} {t t}_{i i} {s the s}_{i i} + + {n no}_{k k})) = = {r r}_{k k} {H h}_{k k} {t t}_{k k} {s the s}_{k k} + + {r r}_{k k} {H h}_{k k} {\underset{i i = = 11}{Σ Σ}}_{i i &NotEqual; &NotEqual; k k}^{S S} {t t}_{i i} {s the s}_{i i} + + {r r}_{k k} {n no}_{k k}$

其中：in:

r_k为接收机合并矢量；r _k is the receiver combining vector;

T＝[t₁，...，t_S]为总的预编码发射矩阵，s＝[s₁，s₂，...，s_S]′为总的发射信号；T=[t ₁ ,...,t _S ] is the total precoding transmission matrix, s=[s ₁ , s ₂ ,...,s _S ]' is the total transmission signal;

n_k为噪音信号，方差为σ²。n _k is a noise signal, and its variance is σ ² .

在现有技术中，采用迫零预编码技术(ZFP(Zero-Forcing Precoding)，下面进行详细描述。In the prior art, a zero-forcing precoding technology (ZFP (Zero-Forcing Precoding)) is adopted, which will be described in detail below.

ZFP方法直接对用户信道H_k进行SVD(Singular-Value Decomposition，奇异值分解)操作，取其最大特征值对应的特征矢量U_k|₁作为接收机合并矢量，记为

即：The ZFP method directly performs SVD (Singular-Value Decomposition, Singular Value Decomposition) operation on the user channel H _k , and takes the eigenvector U _k | ₁ corresponding to its maximum eigenvalue as the receiver combination vector, denoted as

Right now:

SVD(H_k)＝U_kΛ_kV_k SVD(H _k )＝U _k Λ _k V _k

${r r}_{k k}^{SVD SVD} = = {U u}_{k k} {| |}_{11}$

h_k＝r_k ^SVDH_k h _k ＝r _k ^SVD H _k

然后尝试每一个假设的发射码书(Codebook)矢量，求得对应的SINR(Signal to Interference plus Noise Ratio，信号与干扰和噪声的比)。Then try each hypothetical transmit codebook (Codebook) vector to obtain the corresponding SINR (Signal to Interference plus Noise Ratio, the ratio of signal to interference and noise).

假设被调度到的每个用户的发射矢量基本正交，且真正发射波束与假设发射码书矢量误差很小的情况下，SINR下边界的近似值为：Assuming that the transmit vectors of each user to be scheduled are basically orthogonal, and the error between the real transmit beam and the assumed transmit codebook vector is small, the approximate value of the lower boundary of the SINR is:

${SINR SINR}_{k k}^{((i i))} \approx \approx \frac{p p {| | | | {h h}_{k k} | | | |}^{22} {cos cos}^{22} {θ θ}_{i i}}{{Mσ Mσ}^{22} + + p p {| | | | {h h}_{k k} | | | |}^{22} {sin sin}^{22} {θ θ}_{i i}}$

其中： $\cos θ_{i} = | \frac{h_{k}}{| | h_{k} | |} c_{i} |,$ c_i为码书中的码矢量，‖.‖为向量的模；in: $\cos θ_{i} = | \frac{h_{k}}{| | h_{k} | |} c_{i} |,$ c _i is the code vector in the codebook, ‖.‖ is the modulus of the vector;

最后将该用户等价信道量化为使SINR最大的那个码矢量，即：Finally, the user equivalent channel is quantized to the code vector that maximizes the SINR, namely:

${\overset{^^}{h h}}_{k k} = = arg arg \underset{{{{c c}_{i i}}},, i i = = 00,, . . . . . .,, 22^{B B} - - 11}{max max} | | \frac{{h h}_{k k}}{| | | | {h h}_{k k} | | | |} {c c}_{i i} | | = = arg arg \underset{{{{c c}_{i i}}},, i i = = 11,, . . . . . .,, 22^{B B}}{max max} {SINR SINR}_{i i}^{k k}$

最后，得到相应的信道方向索引CDI和信道质量信息CQI如下：Finally, the corresponding channel direction index CDI and channel quality information CQI are obtained as follows:

${CDI CDI}_{k k} = = arg arg \underset{i i = = 00,, . . . . . .,, 22^{B B} - - 11}{max max} | | \frac{{h h}_{k k}}{| | | | {h h}_{k k} | | | |} {c c}_{i i} | | = = arg arg \underset{i i = = 11,, . . . . . .,, 22^{B B} - - 11}{max max} {SINR SINR}_{k k}^{((i i))}$

${CQI CQI}_{k k} = = max max {{{SINR SINR}_{k k}^{((i i))}}}$

并将信道方向索引和信道质量信息反馈给基站。And feed back the channel direction index and channel quality information to the base station.

基站根据所有用户反馈的CQI及CDI，求出所有可能用户组合(K个用户中选S个用户)所对应的发射矩阵T_ZFP，如下所示：Based on the CQI and CDI fed back by all users, the base station calculates the transmit matrix T _ZFP corresponding to all possible user combinations (S users are selected from K users), as follows:

${T T}_{ZFP ZFP} = = [[{t t}_{11},, . . . . . .,, {t t}_{S S}]] = = \overset{^^}{H h} {((S S))}^{H h} {((\overset{^^}{H h} ((S S)) \overset{^^}{H h} {((S S))}^{H h}))}^{- - 11} diag diag {((p p))}^{11 / / 22}$

其中：in:

$\hat{H} (S) = [{\hat{h}}_{1}, \cdot \cdot \cdot, {\hat{h}}_{S}]$ 是被选择的S个用户量化反馈信道的组合； $\hat{h} (S) = [{\hat{h}}_{1}, &Center Dot; &Center Dot; &Center Dot;, {\hat{h}}_{S}]$ is the combination of the selected S user quantization feedback channels;

上标H表示共轭转置；The superscript H indicates the conjugate transpose;

上标-1表示逆；The superscript -1 means inverse;

所述向量p＝[p₁，...，p_S]为各用户发射功率分配。当用户间的能量分布相同时， $p_{k} = \frac{P}{S | | t_{k} | |},$ P为总发射功率。也可以用其他方法，如Water falling算法，进行各用户间的功率分配。The vector p=[p ₁ , . . . , p _S ] is the transmit power allocation for each user. When the energy distribution among users is the same, $p_{k} = \frac{P}{S | | t_{k} | |},$ P is the total transmit power. Other methods, such as the Water falling algorithm, may also be used to allocate power between users.

然后用求出的t_k对每个用户SINR做一个修正：Then use the calculated t _k to make a correction to each user's SINR:

${SINR SINR}_{k k,, est est} \approx \approx {| | {\overset{^^}{h h}}_{k k} {t t}_{k k} | |}^{22} {CQI CQI}_{k k}$

再用修正的SINR估计各个用户组合的容量：Then use the modified SINR to estimate the capacity of each user combination:

$sum sum {rate rate}_{k k} = = {Σ Σ}_{k k = = 11}^{S S} {log log}_{22} ((11 + + {SINR SINR}_{k k,, est est}))$

最后选择最大容量的用户组合进行发射，其发射矩阵为该用户组合所对应的T_ZFP。Finally, the user combination with the largest capacity is selected for transmission, and its transmission matrix is T _ZFP corresponding to the user combination.

由于在有限反馈系统中，量化误差的存在使各用户的干扰不可能完全消除，被选到的用户做MMSE接收以抑制多用户干扰，即：Due to the existence of quantization error in the limited feedback system, the interference of each user cannot be completely eliminated, and the selected user performs MMSE reception to suppress multi-user interference, namely:

${r r}_{k k}^{MMSE MMSE} = = {(({H h}_{k k} {t t}_{k k}))}^{H h} {(({H h}_{k k} {T T}_{ZFP ZFP} {(({H h}_{k k} {T T}_{ZFP ZFP}))}^{H h} + + {σ σ}^{22} {I I}_{N N}))}^{- - 11}$

其真正SINR为：Its true SINR is:

${SINR SINR}_{k k} = = \frac{{p p}_{k k} {| | {r r}_{k k} {h h}_{k k} {t t}_{k k} | |}^{22}}{{σ σ}^{22} {r r}_{k k}^{H h} {r r}_{k k} + + {\underset{i i = = 11}{Σ Σ}}_{i i &NotEqual; &NotEqual; k k}^{K K} {p p}_{i i} {| | {r r}_{k k} {h h}_{k k} {t t}_{i i} | |}^{22}}$

真正容量为：The real capacity is:

$sum rate sum rate = = {Σ Σ}_{k k = = 11}^{M m} {log log}_{22} ((11 + + {SINR SINR}_{k k}))$

然而，采用上述的方案，由于真正接收机为MMSE合并器，而反馈时所用的合并器为每个用户自己信道SVD分解所得的特征矢量，其仅仅考虑了用户自身的信道，没有考虑其他用户的干扰，所以上述方法得到的合并信道与真正的接收信道偏差较大，导致反馈不准确，使得基站调度及预编码矩阵的计算均不准确，所以系统的频谱效率并没有得到最大化的利用。However, with the above scheme, since the real receiver is an MMSE combiner, the combiner used for feedback is the eigenvector obtained by the SVD decomposition of each user's own channel, which only considers the user's own channel and does not consider other users' Therefore, the combined channel obtained by the above method has a large deviation from the real receiving channel, resulting in inaccurate feedback, making base station scheduling and precoding matrix calculation inaccurate, so the spectral efficiency of the system has not been maximized.

发明内容 Contents of the invention

为了解决现有ZFP方法中，由于其使用的反馈合并信道与真正的接收信道偏差大导致频谱效率较低的问题，本发明的目的是提供一种更有效的多输入多输出系统的预编码传输方法，提升系统频谱效率。In order to solve the problem of low spectral efficiency due to the large deviation between the feedback combining channel and the real receiving channel in the existing ZFP method, the purpose of the present invention is to provide a more effective precoding transmission of the MIMO system method to improve the spectral efficiency of the system.

为了实现上述目的，本发明提供了一种多输入多输出系统的预编码传输方法，包括：In order to achieve the above object, the present invention provides a precoding transmission method for a MIMO system, comprising:

步骤A，移动台假设实际的发射矩阵可近似为码书中的某个酉矩阵，利用这个酉矩阵，将其中的一个矢量作为该用户的发射波束矢量，其他M-1个矢量作为所有干扰用户的发射波束矢量，使用最小均方误差(MMSE)接收器进行信号合并，计算出等价信道，并获取量化等价信道，同时获取信道方向索引和信道质量信息，反馈到基站；In step A, the mobile station assumes that the actual transmit matrix can be approximated as a unitary matrix in the codebook. Using this unitary matrix, one of the vectors is used as the transmit beam vector of the user, and the other M-1 vectors are used as all interfering users The transmit beam vector, using the minimum mean square error (MMSE) receiver for signal combination, calculates the equivalent channel, and obtains the quantized equivalent channel, and obtains the channel direction index and channel quality information at the same time, and feeds back to the base station;

步骤B，基站根据移动台反馈的信道方向索引和信道质量信息获取部分或所有可能的用户组合所对应的迫零预编码发射矩阵或调整的迫零预编码发射矩阵，并进行多用户调度操作，确定被调度到的用户组合S；Step B, the base station obtains the zero-forcing precoding transmission matrix or the adjusted zero-forcing precoding transmission matrix corresponding to some or all possible user combinations according to the channel direction index and channel quality information fed back by the mobile station, and performs multi-user scheduling operation, Determine the scheduled user combination S;

步骤C，基站利用所述被调度到的用户组合对应的迫零预编码发射矩阵或调整的迫零预编码发射矩阵进行预编码后发射。Step C, the base station uses the zero-forcing precoding transmission matrix corresponding to the scheduled user combination or the adjusted zero-forcing precoding transmission matrix to perform precoding and then transmit.

上述的方法，其中，还包括：The above method, which also includes:

步骤D，移动台使用最小均方误差接收器进行信号接收。Step D, the mobile station uses the minimum mean square error receiver to receive signals.

上述的方法，其中，所述步骤A具体包括：The above-mentioned method, wherein, the step A specifically includes:

步骤111，移动台计算其对应于码书中酉矩阵中的每一个矢量的等价信道；Step 111, the mobile station calculates its equivalent channel corresponding to each vector in the unitary matrix in the codebook;

步骤112，移动台根据该等价信道获取移动台对应于酉矩阵中的每一个矢量的信号与干扰和噪声之比SINR；Step 112, the mobile station obtains the signal-to-interference and noise ratio SINR of the mobile station corresponding to each vector in the unitary matrix according to the equivalent channel;

步骤113，找到SINR中最大值，根据该最大值得到移动台的量化等价信道，利用该量化的等价信道获取信道方向索引；Step 113, find the maximum value in the SINR, obtain the quantized equivalent channel of the mobile station according to the maximum value, and use the quantized equivalent channel to obtain the channel direction index;

步骤114，根据SINR的最大值获取信道质量信息，Step 114, obtain channel quality information according to the maximum value of SINR,

步骤115，将信道方向索引和信道质量信息反馈给基站。Step 115, feeding back the channel direction index and channel quality information to the base station.

上述的方法，其中，所述用户组合所对应的迫零预编码发射矩阵为： $T_{ZFP} = [t_{1}, . . ., t_{S}] = \hat{H} {(S)}^{H} {(\hat{H} (S) \hat{H} {(S)}^{H})}^{- 1} diag {(p)}^{1 / 2},$ 所述

为被选择的S个用户量化信道的组合，所述向量p＝[p₁，...，p_S]为各用户发射功率分配。In the above method, wherein the zero-forcing precoding transmission matrix corresponding to the user combination is:

T_{ZFP} = [t_{1}, . . ., t_{S}] = \hat{h} {(S)}^{h} {(\hat{h} (S) \hat{h} {(S)}^{h})}^{- 1} diag {(p)}^{1 / 2},

said

The combination of quantized channels for the selected S users, the vector p=[p ₁ , . . . , p _S ] is the transmit power allocation for each user.

上述的方法，其中，所述用户组合S所对应的调整的迫零预编码(又叫MMSE预编码)发射矩阵为： $T_{RZFP} = \hat{H} {(S)}^{H} {(\hat{H} (S) \hat{H} {(S)}^{H} + σ_{0}^{2} I)}^{- 1} diag {(p)}^{1 / 2},$ 所述为被选择的S个用户量化信道的组合，所述向量p＝[p₁，...，p_S]为各用户发射功率分配。The above method, wherein the adjusted zero-forcing precoding (also called MMSE precoding) transmission matrix corresponding to the user combination S is: $T_{RZFP} = \hat{h} {(S)}^{h} {(\hat{h} (S) \hat{h} {(S)}^{h} + σ_{0}^{2} I)}^{- 1} diag {(p)}^{1 / 2},$ said The combination of quantized channels for the selected S users, the vector p=[p ₁ , . . . , p _S ] is the transmit power allocation for each user.

上述的方法，其中，所述步骤C中根据QoS选择相应的调度算法进行多用户调度操作。In the above method, in the step C, a corresponding scheduling algorithm is selected according to QoS to perform multi-user scheduling operations.

由于本发明的方法在反馈阶段利用MMSE合并器进行CQI和CDI的计算，减小了信道反馈的误差，因此有效的提升了系统的频谱利用率。Since the method of the present invention utilizes the MMSE combiner to calculate the CQI and CDI in the feedback stage, the channel feedback error is reduced, thereby effectively improving the spectrum utilization rate of the system.

附图说明 Description of drawings

图1为本发明的方法的流程示意图；Fig. 1 is a schematic flow sheet of the method of the present invention;

图2为本发明的方法步骤11的详细流程示意图；Fig. 2 is the detailed flow diagram of method step 11 of the present invention;

图3为本发明的方法中步骤14采用Max C/I调度算法的处理流程图；Fig. 3 adopts the processing flowchart of Max C/I scheduling algorithm in step 14 in the method of the present invention;

图4和图5为利用本发明的方法的仿真结果示意图。4 and 5 are schematic diagrams of simulation results using the method of the present invention.

具体实施方式 Detailed ways

本发明的一种多输入多输出系统的预编码传输方法，在反馈阶段，移动台利用MMSE合并器计算量化等价信道的信道方向索引(CDI)，以及信道质量信息(CQI)，进而在调度阶段和预编码阶段利用该反馈信息进行相应的处理，提高频谱利用效率。In the precoding transmission method of a MIMO system of the present invention, in the feedback phase, the mobile station uses the MMSE combiner to calculate the channel direction index (CDI) and channel quality information (CQI) of the quantized equivalent channel, and then in the scheduling stage and precoding stage use the feedback information to perform corresponding processing to improve spectrum utilization efficiency.

如图1所示，本发明的多输入多输出系统的预编码传输方法包括：As shown in Figure 1, the precoding transmission method of the MIMO system of the present invention includes:

步骤11，由于在反馈时，每个用户并不知道预编码发射矩阵T，因此就无法得到真正的接收机合并矢量

和等价信道h_k，但是在实际的多用户系统中，基站调度器总是倾向于选择反馈信道相互正交的用户，因为这样可使用户间干扰较小，容量相对较大。若被选择的用户的信道相互正交，则计算出的发射矩阵T也为一正交矩阵。在后面的仿真图中也可以看出，用户数越多，选择相互正交用户组的概率就越大，发射矩阵也就越近似为一正交矩阵。因此，我们将发射矩阵T假设为一酉矩阵，利用下述公式计算MMSE合并矢量

如下：Step 11, since each user does not know the precoding transmission matrix T during feedback, so the real receiver combining vector cannot be obtained

and equivalent channel h _k , but in the actual multi-user system, the base station scheduler always tends to select users whose feedback channels are orthogonal to each other, because it can make the interference between users smaller and the capacity relatively larger. If the channels of the selected users are orthogonal to each other, the calculated transmit matrix T is also an orthogonal matrix. It can also be seen from the following simulation diagram that the more the number of users, the greater the probability of selecting mutually orthogonal user groups, and the more approximate the transmit matrix is to an orthogonal matrix. Therefore, we assume that the transmit matrix T is a unitary matrix, and use the following formula to calculate the MMSE combination vector

as follows:

${r r}_{k k}^{MMSE MMSE} = = {(({H h}_{k k} {t t}_{k k}))}^{H h} {(({H h}_{k k} T T {(({H h}_{k k} T T))}^{H h} + + {σ σ}^{22} {I I}_{N N}))}^{- - 11}$

从而得到等价信道h_k＝r_k ^MMSEH_k，并利用所述等价信道获取信道方向索引和信道质量信息；Thus, the equivalent channel h _k = _rk ^MMSE H _k is obtained, and the channel direction index and channel quality information are obtained by using the equivalent channel;

步骤12，移动台反馈CDI和CQI到基站；Step 12, the mobile station feeds back the CDI and CQI to the base station;

步骤13，基站根据所有移动台反馈的CDI和CQI，求出部分或所有可能的用户组合(K个用户中选S<＝M个用户)对应的迫零预编码发射矩阵T_ZFP或调整的迫零预编码发射矩阵T_RZFP，并对移动台进行调度；Step 13, the base station obtains the zero-forcing precoding transmission matrix T _ZFP or the adjusted zero-forcing matrix corresponding to some or all possible user combinations (select S<=M users among K users) according to the CDI and CQI fed back by all mobile stations Precoding the transmit matrix T _RZFP , and scheduling the mobile station;

在此，是求出部分用户组合对应的迫零预编码发射矩阵T_ZFP或调整的迫零预编码发射矩阵T_RZFP，还是求出所有用户组合对应的迫零预编码发射矩阵T_ZFP或调整的迫零预编码发射矩阵T_RZFP与调度算法相关，将在后面举例说明。Here, whether to obtain the zero-forcing precoding transmission matrix T _ZFP or the adjusted zero-forcing precoding transmission matrix T _RZFP corresponding to some user combinations, or to obtain the zero-forcing precoding transmission matrix T _ZFP or the adjusted zero-forcing precoding transmission matrix corresponding to all user combinations The zero-forcing precoding transmission matrix T _RZFP is related to the scheduling algorithm, and will be described with an example later.

其中迫零预编码发射矩阵T_ZFP或调整的迫零预编码发射矩阵T_RZFP如下所示：Wherein the zero-forcing precoding transmission matrix T _ZFP or the adjusted zero-forcing precoding transmission matrix T _RZFP is as follows:

${T T}_{RZFP RZFP} = = \overset{^^}{H h} {((S S))}^{H h} {((\overset{^^}{H h} ((S S)) \overset{^^}{H h} {((S S))}^{H h} + + {σ σ}_{00}^{22} I I))}^{- - 11} diag diag {((p p))}^{11 / / 22}$

其中： $\hat{H} (S) = [{\hat{h}}_{1}, \cdot \cdot \cdot, {\hat{h}}_{S}]$ 是被选择的S个用户量化等价信道的组合；in: $\hat{h} (S) = [{\hat{h}}_{1}, \cdot \cdot &Center Dot;, {\hat{h}}_{S}]$ is the combination of the selected S user quantized equivalent channels;

向量p＝[p₁，...，p_S]为各用户发射功率分配。The vector p=[p ₁ , ..., p _S ] is the transmit power allocation for each user.

而基站对移动台调度可根据QoS选择相应的调度算法进行多用户调度操作。The base station can select a corresponding scheduling algorithm according to QoS to perform multi-user scheduling operation for mobile station scheduling.

步骤14，对调度到的移动台数据使用对应的迫零预编码发射矩阵TZFP或调整的迫零预编码发射矩阵T_RZFP进行预编码后发射。Step 14: Perform precoding on the scheduled mobile station data using the corresponding zero-forcing precoding transmission matrix TZFP or the adjusted zero-forcing precoding transmission matrix T _RZFP and then transmit.

下面对方法进行进一步详细的描述。The method is described in further detail below.

在步骤11中，假定发射矩阵T是码书中的一个酉矩阵C_g，并假设其中的一个矢量c_i∈C_g为该用户的发射矢量，其他M-1个矢量为干扰用户的发射矢量，为方便对本发明的理解，在对步骤11进行详细说明之前，先对该酉矩阵C_g和矢量c_i进行一定说明。In step 11, assume that the transmit matrix T is a unitary matrix C _g in the codebook, and assume that one of the vectors _ci ∈ C _g is the transmit vector of the user, and the other M-1 vectors are the transmit vectors of the interfering user , in order to facilitate the understanding of the present invention, before step 11 is described in detail, the unitary matrix C _g and the vector _ci are described first.

在本发明的具体实施例可以利用多种类型的码书，但在此仅以基于DFT的预编码码书为例进行说明。Various types of codebooks may be used in specific embodiments of the present invention, but here only a DFT-based precoding codebook is used as an example for illustration.

该码书为一个由2^B个向量组成的DFT矩阵，其中每个向量均包括M个元素，第i个向量的第m个元素表示如下：The codebook is a DFT matrix composed of 2 ^B vectors, each of which includes M elements, and the mth element of the i-th vector is expressed as follows:

c_i(m)＝exp(-j2πim/2^B)c _i (m)＝exp(-j2πim/2 ^B )

m＝0，...，M-1；m=0,...,M-1;

i＝0，...，2^B-1。i=0, . . . , ^2B -1.

这2^B个向量组成了G＝2^B/M个酉矩阵，其中，第g个酉矩阵表示如下：These 2 ^B vectors form G=2 ^B /M unitary matrices, where the g-th unitary matrix is expressed as follows:

C_g＝{c_g，c_g+G，...，c_g+(M-1)G}C _g = {c _g , c _g+G , . . . , c _g+(M-1)G }

g＝0，...，G-1。g=0, . . . , G-1.

步骤11如图2所示，具体包括：Step 11, as shown in Figure 2, specifically includes:

步骤111，移动台计算对应于每一个c_i的等价信道

，如下所示：Step 111, the mobile station calculates the equivalent channel corresponding to each c _i

,As follows:

${h h}_{k k}^{((i i))} = = {r r}_{k k}^{MMSE MMSE ((i i))} {H h}_{k k}$

其中：in:

r_k ^MMSE(i)＝(H_kc_i)^H(H_kC_g(H_kC_g)^H+σ²I_N)^-1 r _k ^MMSE(i) ＝(H _k c _i ) ^H (H _k C _g (H _k C _g ) ^H +σ ² I _N ) ^-1

步骤112，移动台根据该等价信道

获取对应于每一个c_i的

如下所示：Step 112, the mobile station according to the equivalent channel

Get corresponding to each c _i

As follows:

${SINR SINR}_{k k}^{((i i))} = = \frac{P P {| | | | {h h}_{k k}^{((i i))} {c c}_{i i} | | | |}^{22}}{M m {σ σ}^{22} + + {Σ Σ}_{j j &NotEqual; &NotEqual; i i (({c c}_{j j} &Element; &Element; {C C}_{g g}))}^{M m - - 11} P P {| | | | {h h}_{k k}^{((i i))} {c c}_{j j} | | | |}^{22}}$

步骤113，找到

中最大值，根据该最大值得到量化等价信道

如下：Step 113, find

The maximum value, according to which the quantized equivalent channel is obtained

as follows:

${\overset{^^}{h h}}_{k k} = = arg arg \underset{{{{c c}_{i i}}},, i i = = 00,, . . . . . .,, 22^{B B} - - 11}{max max} | | \frac{{h h}_{k k}}{| | | | {h h}_{k k} | | | |} {c c}_{i i} | | = = arg arg \underset{{{{c c}_{i i}}},, i i = = 11,, . . . . . .,, 22^{B B} - - 11}{max max} {SINR SINR}_{k k}^{((i i))}$

步骤114，根据

的最大值和对应的等价信道

获取CDI和CQI，如下：Step 114, according to

The maximum value of and the corresponding equivalent channel

Obtain CDI and CQI, as follows:

${CDI CDI}_{k k} = = arg arg \underset{i i = = 00,, . . . . . .,, 22^{B B} - - 11}{max max} | | \frac{{h h}_{k k}}{| | | | {h h}_{k k} | | | |} {c c}_{i i} | | = = arg arg \underset{i i = = 00,, . . . . . .,, 22^{B B} - - 11}{max max} {SINR SINR}_{k k}^{((i i))}$

${CQI CQI}_{k k} = = max max {{{SINR SINR}_{k k}^{((i i))}}}$

在步骤13中，基站需要对移动台执行调度，在此采用的调度算法可以是最大化信噪比Max C/I调度算法、比例公平PF调度算法和轮询调度算法等，在本发明的具体实施例中以最大化信噪比Max C/I调度算法为例进行详细说明。In step 13, the base station needs to carry out scheduling to the mobile station, and the scheduling algorithm adopted here can be to maximize the signal-to-noise ratio Max C/I scheduling algorithm, proportional fair PF scheduling algorithm and round-robin scheduling algorithm, etc., in the specific embodiment of the present invention In the embodiment, the Max C/I scheduling algorithm for maximizing the signal-to-noise ratio is taken as an example to describe in detail.

如图3所示，该步骤13中的基站对移动台进行调度具体包括：As shown in Figure 3, the scheduling of the mobile station by the base station in step 13 specifically includes:

步骤131，对于每一个可能的用户可能组合S，量化信道组合为 $\hat{H} (S) = [{\hat{h}}_{1}, \cdot \cdot \cdot, {\hat{h}}_{S}],$ 利用公式Step 131, for each possible user combination S, the quantized channel combination is $\hat{h} (S) = [{\hat{h}}_{1}, &Center Dot; &Center Dot; \cdot, {\hat{h}}_{S}],$ use the formula

$T_{ZFP} = [t_{1}, . . ., t_{S}] = \hat{H} {(S)}^{H} {(\hat{H} (S) \hat{H} {(S)}^{H})}^{- 1} diag {(p)}^{1 / 2}$ 或 $T_{ZFP} = [t_{1}, . . ., t_{S}] = \hat{h} {(S)}^{h} {(\hat{h} (S) \hat{h} {(S)}^{h})}^{- 1} diag {(p)}^{1 / 2}$ or

${T T}_{RZFP RZFP} = = [[{t t}_{11},, . . . . . .,, {t t}_{S S}]] = = \overset{^^}{H h} {((S S))}^{H h} {((\overset{^^}{H h} ((S S)) \overset{^^}{H h} {((S S))}^{H h} + + {σ σ}_{00}^{22} I I))}^{- - 11} diag diag {((p p))}^{11 / / 22}$

计算其迫零预编码或调整的迫零预编码发射矩阵T_ZFP或T_RZFP，利用其中的t_k(k＝1，...S)对该组合中的每个用户的SINR_k做一个修正，得到修正后的SINR_k，est表示如下：Calculate its zero-forcing precoding or adjusted zero-forcing precoding transmission matrix T _ZFP or _TRZFP , and use t _k (k=1,...S) to make a correction to the SINR _k of each user in the combination , the corrected SINR _{k,est is} expressed as follows:

步骤132，利用修正后的SINR估计各个用户组合的容量，如下：Step 132, using the revised SINR to estimate the capacity of each user combination, as follows:

$sum sum {rate rate}_{S S} = = {Σ Σ}_{k k = = 11}^{S S} {log log}_{22} ((11 + + {SINR SINR}_{k k,, est est}))$

步骤133，选择最大容量的用户组合作为调度上的用户。Step 133, selecting the user combination with the largest capacity as the scheduled user.

当然，上述的调度算法是一种较为理想的算法，对所有可能的用户组合求出对应的迫零预编码发射矩阵T_ZFP或调整的迫零预编码发射矩阵T_RZFP进行容量估计，进而根据容量估计结果进行选择，但在用户数量较多时，为简化调度算法运算量，其他的一些次最佳算法，如greedy算法等也可应用到本发明进行调度，此时仅需要计算部分可能的用户组合的迫零预编码发射矩阵T_ZFP或调整的迫零预编码发射矩阵T_RZFP进行容量估计，进而根据容量估计结果进行选择。Of course, the above scheduling algorithm is a relatively ideal algorithm. For all possible user combinations, the corresponding zero-forcing precoding transmission matrix T _ZFP or the adjusted zero-forcing precoding transmission matrix T _RZFP is calculated for capacity estimation, and then according to the capacity However, when the number of users is large, in order to simplify the calculation amount of the scheduling algorithm, some other sub-optimal algorithms, such as the greedy algorithm, can also be applied to the present invention for scheduling. At this time, only part of the possible user combinations need to be calculated. The zero-forcing precoding transmission matrix T _ZFP or the adjusted zero-forcing precoding transmission matrix T _RZFP is used for capacity estimation, and then the selection is made according to the capacity estimation result.

在步骤14中，基站利用被调度到的用户组合所对应的发射矩阵(迫零预编码发射矩阵T_ZFP或调整的迫零预编码发射矩阵T_RZFP)进行预编码后发射即可。In step 14, the base station uses the transmit matrix (zero-forcing precoding transmit matrix T _ZFP or adjusted zero-forcing precoding transmit matrix T _RZFP ) corresponding to the scheduled user combination to perform precoding and then transmit.

而在移动台端，由于用户间干扰不可能完全抑制掉，用户做MMSE接收，即：On the mobile station side, since the inter-user interference cannot be completely suppressed, the user performs MMSE reception, namely:

图4、图5和图6为本发明方法应用后的仿真示意图。Fig. 4, Fig. 5 and Fig. 6 are simulation schematic diagrams after the application of the method of the present invention.

第一仿真条件如下所示：发射天线为M＝4，接收天线为N＝4，天线间不相关，用户数为20，反馈3比特CDI及精确的CQI；The first simulation condition is as follows: the transmitting antenna is M=4, the receiving antenna is N=4, the antennas are not correlated, the number of users is 20, and 3-bit CDI and accurate CQI are fed back;

图4为上述第一仿真条件下，频谱效率随SNR变化的曲线，处于上方的曲线为利用本发明的方法的频谱效率随SNR变化的曲线，而处于下方的曲线为传统ZFP方法的频谱效率随SNR变化的曲线，从图4中可以看出，在如图4所示，在SNR＝12dB时，利用本发明方法相对于现有的ZFP方法，频谱效率大概高2.5bps/Hz，而且随着SNR的增大，频谱效率的提升越来越明显。Fig. 4 is under above-mentioned first simulation condition, the curve that spectrum efficiency changes with SNR, the curve that is in the top is the curve that utilizes the spectrum efficiency of the method of the present invention to change with SNR, and the curve that is in the bottom is that the spectrum efficiency of traditional ZFP method changes with The curve of SNR variation, as can be seen from Fig. 4, when as shown in Fig. 4, when SNR=12dB, utilize the method of the present invention relative to existing ZFP method, spectral efficiency is about 2.5bps/Hz higher, and along with With the increase of SNR, the improvement of spectral efficiency becomes more and more obvious.

第二仿真条件如下所示：发射天线为M＝4，接收天线为N＝4，天线间不相关，反馈3比特CDI及精确的CQI，SNR为12dB；The second simulation condition is as follows: the transmitting antenna is M=4, the receiving antenna is N=4, the antennas are not correlated, 3-bit CDI and accurate CQI are fed back, and the SNR is 12dB;

图5为上述第二仿真条件下，频谱效率随用户数的变化曲线，处于上方的曲线为利用本发明的方法的频谱效率随用户数变化的曲线，而处于下方的曲线为传统ZTP方法的频谱效率随用户数变化的曲线，从图可以看出，在用户数为12时，利用本发明方法相对于现有的ZFP方法，频谱效率大概高1.8bps/Hz，而且随着用户数的增加，频谱效率的提升越来越明显。Fig. 5 is under the above-mentioned second simulation condition, the change curve of spectrum efficiency with the number of users, the curve at the top is the curve that utilizes the spectrum efficiency of the method of the present invention to change with the number of users, and the curve at the bottom is the spectrum of the traditional ZTP method The curve of efficiency changing with the number of users can be seen from the figure, when the number of users is 12, using the method of the present invention compared with the existing ZFP method, the spectral efficiency is about 1.8bps/Hz higher, and with the increase of the number of users, The improvement of spectral efficiency is more and more obvious.

以上所述仅是本发明的优选实施方式，应当指出，对于本技术领域的普通技术人员来说，在不脱离本发明原理的前提下，还可以做出若干改进和润饰，这些改进和润饰也应视为本发明的保护范围。The above is only a preferred embodiment of the present invention, it should be pointed out that, for those of ordinary skill in the art, without departing from the principle of the present invention, some improvements and modifications can also be made, and these improvements and modifications can also be made. It should be regarded as the protection scope of the present invention.

Claims

1. A precoding transmission method for a MIMO system, characterized in that, comprising:

In step A, the mobile station assumes that the actual transmit matrix can be approximated as a unitary matrix in the codebook, and using the unitary matrix, one of the vectors is used as the transmit beam vector of the user, and the other vectors are used as transmit beam vectors of the interfering user, Use the minimum mean square error receiver to combine signals, calculate the equivalent channel, and obtain the channel direction index and channel quality information, and feed them back to the base station;

Step B, the base station obtains the zero-forcing precoding transmission matrix or the adjusted zero-forcing precoding transmission matrix corresponding to some or all possible user combinations according to the channel direction index and channel quality information fed back by the mobile station, and performs multi-user scheduling operation, Determine the user combination to be dispatched;

Step C, the base station uses the zero-forcing precoding transmission matrix corresponding to the scheduled user combination or the adjusted zero-forcing precoding transmission matrix to perform precoding and then transmit.

2. The method according to claim 1, further comprising:

Step D, the mobile station uses the minimum mean square error receiver to receive signals.

3. The method according to claim 2, wherein said step A specifically comprises:

Step 111, the mobile station calculates its equivalent channel corresponding to each vector in the unitary matrix;

Step 112, the mobile station obtains the signal-to-interference and noise ratio SINR of the mobile station corresponding to each vector in the unitary matrix according to the equivalent channel;

Step 113, finding the maximum value in the SINR, and obtaining the quantized equivalent channel of the mobile station according to the maximum value;

Step 114, obtain the channel direction index and channel quality information according to the maximum value of the SINR and the corresponding equivalent channel, and then feed them back to the base station.

4. The method according to claim 3, wherein the zero-forcing precoding transmission matrix corresponding to the user combination is:

T_{ZFP} = [t_{1}, . . ., t_{S}] = \hat{h} {(S)}^{h} {(\hat{h} (S) \hat{h} {(S)}^{h})}^{- 1} diag {(p)}^{1 / 2},

said

To be the combination of quantized channels of the selected S users, the vector p=[p ₁ , . . . , p _S ] is the transmission power allocation of each user.

5. The method according to claim 3, wherein the adjusted zero-forcing precoding transmission matrix corresponding to the user combination is:

T_{RZFP} = \hat{h} {(S)}^{h} {(\hat{h} (S) \hat{h} {(S)}^{h} + σ_{0}^{2} I)}^{- 1} diag {(p)}^{1 / 2},

said

6. The method according to any one of claims 1 to 5, wherein in the step B, a corresponding scheduling algorithm is selected according to QoS to perform multi-user scheduling operations.