CN101834651B - Data information linear preprocessing method of multiuser multiple data stream MIMO (Multiple Input Multiple Output) system - Google Patents

Abstract

A data information linear preprocessing method for a multi-user multi-data stream MIMO system in the field of wireless communication technology, comprising the following steps: a base station transmitter determines the data information to be sent to each receiving user; the user channel information that each user will obtain Feedback to the base station transmitter; the base station transmitter obtains interference information and noise power; the base station transmitter performs linear preprocessing and superposition on the data information transmitted to each user; the base station transmitter broadcasts the final transmission information obtained; each user Receive the information sent by the base station transmitter, and perform matching filtering on the received information, and the information after matching filtering is the information required by the user. The present invention can balance the preprocessing gains between multiple data streams of the user, thereby obtaining better system bit error rate performance; at the same time, it is not necessary to separately calculate the covariance matrix of the user channel matrix and the interference channel matrix, and the calculation is simple and the numerical performance stable and thus suitable for practical applications.

Description

Linear preprocessing method of data information in multi-user multi-stream MIMO system

technical field

The invention relates to a processing method in the technical field of wireless communication, in particular to a linear preprocessing method for data information of a multi-user multi-data stream MIMO (Multiple Input Multiple Output) system.

Background technique

Due to its great potential in capacity and performance, MIMO wireless communication system has attracted extensive research in recent years. With the deepening of research, MIMO technology has been extended from the initial point-to-point single-user MIMO (SU-MIMO) system to point-to-multipoint multi-user MIMO (MU-MIMO) system. In the industry, both SU-MIMO and MU-MIMO technologies have been adopted by the LTE and LTE-Advanced standards of the International Organization for Standardization 3GPP.

The MU-MIMO technology can simultaneously transmit data information to each user in the same time, frequency and code domain in the manner of space division multiple access (SDMA), thus further improving the throughput of the system. However, when multiple users share the same time-frequency resource, co-channel interference (CCI) will inevitably be introduced. Due to the existence of CCI, appropriate methods must be adopted to eliminate or inhibit CCI. However, due to the existence of multiple users, the scheme of interference suppression or cancellation at the receiving end in the traditional SU-MIMO system is no longer feasible in MU-MIMO, so in MU-MIMO, it is necessary to pre-check the transmission at the sending end. Data information is processed to suppress or eliminate CCI, which is the preprocessing technique generally described in the literature.

The MU-MIMO schemes for suppressing CCI by preprocessing the transmitted data information are mainly divided into two categories: nonlinear and linear. Nonlinear preprocessing is complex and time-delayed, making it difficult to apply in practice. The implementation complexity of linear preprocessing is low, and as the number of users increases, it can obtain the same capacity as the optimal preprocessing solution - Dirty Paper Theory (DPC), so it is more practical.

After searching the existing literature, it is found that the linear preprocessing methods of data information are mainly divided into two categories:

One is the optimization method based on maximizing the signal-to-interference-to-noise ratio (SINR): this method cannot directly obtain a closed solution, so some suboptimal methods represented by block diagonalization (BD) are derived (see Q.H.Spencer, "Zero-forcing methods for downlink spatial multiplexing in multipleuser MIMO channels (zero-forcing methods in downlink multi-user MIMO spatial multiplexing)," IEEE Transactions on Signal Processing, vol.52, pp.461-471, Feb.2004), BD The method has strict restrictions on the relationship between the number of transmitting antennas and the number of receiving antennas in the system. When the limiting conditions are not met, it is difficult to obtain ideal performance through the BD method.

Another type of design method is the optimization method based on signal-leakage-to-noise ratio (SLNR) proposed in recent years: (see M. Sadek, "Aleakage-based precoding scheme for downlink multi-user MIMO channels (a kind of downlink multi-user MIMO based on Leakage preprocessing method), "IEEE Transactions on Wireless Communications, vol.6, pp.1711-1721, May 2007), this method expects the received signal power of each user to be requested to be as large as possible, and its noise power and leakage The sum of the interference power to other users should be as small as possible. The main advantage of this method is that the objective function avoids the nesting problem of the preprocessing matrix among users under the SINR optimization method, and can directly obtain the optimal closed solution without satisfying the limitation of the number of transmitting and receiving antennas in the BD method, so it has a better Broad application space. However, the SLNR linear preprocessing method still has the following three important defects:

1) When each user transmits multiple data streams, the preprocessing gain distribution among the various data streams is seriously unbalanced, which leads to the degradation of the overall system performance.

2) In the preprocessing design process, the covariance matrix of the user channel matrix and the interference channel matrix needs to be calculated separately. Information loss will occur when the covariance matrix is obtained, and since it is difficult for the transmitting end to obtain accurate channel information in an actual system, the above information loss situation will be more serious.

3) In the preprocessing design process, it is necessary to obtain the inverse matrix of one of the noise-corrected covariance matrices. When the signal-to-noise ratio is large, the matrix is approximately a non-singular matrix (irreversible matrix), and it is difficult to obtain its inverse matrix and solve it. Complex (see: "Matrix Analysis and Application", Zhang Xianda, Tsinghua University Press, published in 2004).

Contents of the invention

The object of the present invention is to overcome the above-mentioned defects in the prior art, and provide a method for linear preprocessing of data information in a multi-user multi-data stream MIMO system. The method of the present invention can further balance the preprocessing gains among the data information streams while effectively suppressing CCI, thereby significantly improving the bit error rate performance of the system; and the method of the present invention avoids the coordination between the user channel matrix and the interference channel matrix. The calculation process of the variance matrix does not need to approximate the inverse matrix of the non-singular matrix, so the method not only has a low degree of information loss, but also has a simple design and lower implementation complexity.

The present invention is achieved through the following technical solutions, comprising the following steps:

In the first step, each user sends a request signal for image and voice to the base station transmitter, and the base station transmitter determines the data information to be sent to each receiving user according to the request signal of each user.

The data information to be sent includes image compression coding information and coefficient parameters of speech autoregressive (AR) filter.

The data information to be sent is expressed in symbol form after bit modulation.

In the second step, each user obtains user channel information H according to the pilot data received by itself, and feeds back the user channel information to the base station transmitter.

The user channel information is in quantized form or codebook form.

In the third step, the base station transmitter obtains the interference information H according to the user channel information, and further measures to obtain the noise power σ ² .

In the fourth step, the base station transmitter performs linear preprocessing on the data information transmitted to each user, and superimposes the preprocessed information to obtain the final transmission information S. At this time, the CCI information in the transmission information S has been effectively inhibition.

Described linear pretreatment, concrete steps are:

1) Longitudinal concatenation of user channel information H _k , interference information H _k and noise power σ ² of the kth user to obtain a co-concatenation matrix T _k , namely T _k =[H _k :H _k : _ασIN ];

其中：U_k是T_k的左奇异矩阵，Λ_k是T_k的正奇异值构成的对角矩阵，V_k是T_k的右奇异矩阵；2) Singular value decomposition is performed on the co-concatenated matrix T _k :

Wherein: U _k is the left singular matrix of T _k , Λ _k is the diagonal matrix that the positive singular value of T _k forms, V _k is the right singular matrix of T _k ;

其中：N是基站发射天线的数目，P_k是U′_k的左奇异矩阵，Ω_k是U′_k的正奇异值构成的对角矩阵，Q_k是U′_k的右奇异矩阵；3) Select the first N columns of the left singular matrix U _k to obtain the matrix U′ _k (U′ _k = U _k (:.1:N)), and perform singular value decomposition on the matrix U′ _k :

Where: N is the number of base station transmitting antennas, P _k is the left singular matrix of U′ _k , Ω _k is a diagonal matrix formed by the positive singular values of U′ _k , and Q _k is the right singular matrix of U′ _k ;

取W′_k的前m列得到线性预处理矩阵W_k，其中：m是每个用户分配的数据流数；4) order

Take the first m columns of W′ _k to get the linear preprocessing matrix W _k , where: m is the number of data streams allocated by each user;

5) Use the linear preprocessing matrix W _k to perform linear preprocessing on the initial information transmitted by the base station to the kth user, and the specific formula is:

S _k =W _k s _k ,

Among them: s _k is the initial information transmitted by the base station to the kth user, and S _k is the actual transmitted information after the base station performs linear preprocessing on the initial information transmitted by the kth user.

The specific formula for the superposition is:

$\mathrm{<span\; class="notranslate">\; S</span>}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; K</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; K</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; W</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{\mathrm{<span\; class="notranslate">\; k</span>}}$

Among them: S is the final transmission information of the base station transmitter, and K is the total number of receiving users.

In the fifth step, the base station transmitter broadcasts the obtained final transmission information S, that is, transmits the information to all receiving users simultaneously on the MIMO channel.

In the sixth step, each user receives the information sent by the base station transmitter, and performs matching filtering on the received information, and the information after the matching filtering is the information required by the user.

Described matched filter, specific formula is:

r′ _k ＝[H _k W _k ) ^H /||H _k W _k || _F ]×r _k

Among them: r′ _k is the filtered information of the k-th user, H _k is the user channel information of the k-th user, W _k is the linear preprocessing matrix of the k-th user, ||·|| _F represents the matrix Frobenius norm , r _k is the received information of the kth user.

Compared with the prior art, the present invention has the following beneficial effects:

1) While effectively suppressing CCI, it can further balance the preprocessing gain between multiple data streams of the user, so as to obtain better system bit error rate performance;

2) The process of obtaining the covariance matrix of the user channel matrix and the interference channel matrix is avoided, and there is no need to find the inverse matrix of one of the approximate singular (irreversible) matrices, so this method has a lower degree of information loss and stronger stability sex;

3) The design of data information preprocessing is simple, the implementation complexity is low, and the method is robust, which is very suitable for application in practical systems.

Description of drawings

FIG. 1 is a comparison chart of bit error performance of the embodiment.

Detailed ways

Below in conjunction with accompanying drawing, the method of the present invention is described in detail: the present embodiment implements on the premise of the technical solution of the present invention, has provided detailed implementation and specific operation process, but protection scope of the present invention is not limited to following implementation example.

Example

In this embodiment, the MIMO system has two users, and each user is assigned 3 data streams. The base station has 12 transmitting antennas, and each user at the receiving end has 3 receiving antennas. The channels of each user are independent, identically distributed, zero A cyclic complex Gaussian random variable with a mean variance of 1, the channels between users are independent of each other, and the noise is complex Gaussian white noise with a variance of ^σ2 . The linear preprocessing of the system specifically includes the following steps:

In the first step, each user sends a request signal for image and voice to the base station transmitter, and the base station transmitter determines the data information to be sent to each receiving user according to the request signal of each user.

The data information to be sent includes image compression coding information and coefficient parameters of the speech autoregressive filter.

The data information to be sent is expressed in the form of bit-modulated symbols, wherein: the data information to be sent by the first user is represented by _s1 , and the data information to be sent by the second user is represented by _s2 .

In the second step, the two users respectively obtain their channel information H ₁ and H ₂ according to the pilot data received by themselves, and respectively feed back H ₁ and H ₂ to the base station transmitter in quantized form.

In the third step, the base station transmitter obtains the interference information H ₁ = H ₂ of the first user and the interference information H ₂ = H ₁ of the second user according to the channel information H 1 and _{H 2 of} the two users respectively, and further The noise power σ ² is measured.

In the fourth step, the base station transmitter performs linear preprocessing on the data information transmitted to the two users, and superimposes the preprocessed information to obtain the final transmission information S. At this time, the CCI information in the transmission information S has been Effective suppression.

In this embodiment, the process of performing linear preprocessing on the initial information of the first user is as follows:

1) According to the channel information H ₁ of the first user, the interference information H ₁ and the noise power σ ² , the matrix T ₁ is obtained by longitudinal concatenation, that is, T ₁ =[H ₁ :H ₁ : _ασIN ];

其中：U₁是T₁的左奇异矩阵，Λ₁是T₁的正奇异值构成的对角矩阵，V₁是T₁的右奇异矩阵；2) Perform singular value decomposition on matrix T ₁ :

Wherein: U ₁ is the left singular matrix of T ₁ , Λ ₁ is the diagonal matrix that the positive singular value of T ₁ forms, V ₁ is the right singular matrix of T ₁ ;

其中：N是基站发射天线的数目，本实施例中N＝12，P₁是U′₁的左奇异矩阵，Ω₁是U′₁的正奇异值构成的对角矩阵，Q₁是U′₁的右奇异矩阵；3) Select the first N columns of the left singular matrix U ₁ to obtain the matrix U′ _k (U′ ₁ =U ₁ (:.1:N)), and perform singular value decomposition on the matrix U′ ₁ :

Wherein: N is the number of base station transmitting antennas, N=12 in the present embodiment, P ₁ is the left singular matrix of U′ ₁ , Ω ₁ is the diagonal matrix that the positive singular value of U′ ₁ forms, Q ₁ is U′ Right singular matrix of ₁ ;

取W′₁的前m列得到线性预处理矩阵W₁，其中：m是每个用户分配的数据流数，本实施例中m＝3；4) order

Get the first m columns of W′ ₁ to obtain a linear preprocessing matrix W ₁ , wherein: m is the number of data streams allocated by each user, and m=3 in this embodiment;

5) Using the linear preprocessing matrix to perform linear preprocessing on the initial information _s1 transmitted by the base station to the first user, the specific formula of the obtained actual transmitted information _S1 from the base station to the first user is:

S ₁ =W ₁ s ₁ .

Using the same method, the actual transmission information S ₂ sent by the base station to the second user is obtained.

The final transmission information S obtained by superposition in this embodiment is:

S=S ₁ +S ₂ .

In the fifth step, the base station transmitter broadcasts the obtained final transmission information S, that is, transmits the information to two receiving users simultaneously on the MIMO channel.

In the sixth step, each user receives the information sent by the base station transmitter, and performs matching filtering on the received information, and the information after the matching filtering is the information required by the user.

The information r ₁ received by the first user in this embodiment is:

r' ₁ ＝H ₁ S+n ₁

Among them: n ₁ is Gaussian white noise with power σ ² .

In this embodiment, the information r′ ₁ after the matched filtering obtained by the first user is:

r′ ₁ ＝[(H ₁ W ₁ ) ^H /||H ₁ W ₁ || _F ]×r ₁

Where: ||·|| _F represents the Frobenius norm of the matrix, and the information r′ ₁ is the information required by the first user.

Using the same method, the matched-filtered information r′ ₂ of the second user is obtained, which is the information required by the second user.

The performance comparison of the bit error rate with the change of the signal-to-noise ratio obtained by the method of this embodiment and the SLNR linear preprocessing method is shown in Figure 1. It can be seen from the figure that when the bit error rate is 10 ^-4 , the implementation Compared with the SLNR linear preprocessing method, the example method obtains a performance gain of more than 3dB. The reason is that the method of the present invention can effectively suppress the CCI while further balancing the preprocessing gains between the data streams, thereby significantly improving the performance of the system. Bit error rate performance; in addition, as the number of transceiver antennas and the number of user data streams increase, the performance gap between the two will become more obvious.

In addition, in the practical application of the method in this embodiment, the process of obtaining the covariance matrix of the user channel matrix and the interference channel matrix is avoided, and there is no need to obtain the inverse matrix of one of the approximate singular (irreversible) matrices, so the method Not only has a lower degree of information loss, but also has a simple design, lower implementation complexity, and stronger method robustness.

Claims (6)

1. the data message linear preprocessing method of multi-user's multiple data stream mimo system is characterized in that, may further comprise the steps:

The first step, each user sends the request signal of its required image and voice to base station transmitter, and base station transmitter is determined to receive user's data message to be sent to each according to each user's request signal;

Second step, the pilot data that each user receives according to self obtain subscriber channel information H, and this subscriber channel information is fed back to base station transmitter;

In the 3rd step, base station transmitter obtains interfere information according to subscriber channel information

The one step surveying of going forward side by side obtains noise power σ ²

In the 4th step, base station transmitter carries out linear preliminary treatment to the data message that is transmitted to each user, and pretreated information superposeed obtains final emission information S, and the CCI information among the emission information S of this moment is by establishment;

In the 5th step, the final emission information S that base station transmitter will obtain broadcasts, and namely this information is sent to simultaneously all reception users on mimo channel;

In the 6th step, each user receives the information that base station transmitter sends, and the docking breath of collecting mail carries out matched filtering, and the information after the matched filtering is the needed information of user;

Linear preliminary treatment described in the 4th step, concrete steps are:

1) with k user's subscriber channel information H _k, interfere information

With noise power σ ²Carry out vertical cascade and obtain common cascade matrix T _k, namely $T_k=[H_k;{\stackrel{\&OverBar;}{H}}_k;{\mathrm{\&alpha;\&sigma;I}}_N];$

2) to common cascade matrix T _kCarry out singular value decomposition:

Wherein: U _kT _kLeft singular matrix, A _kT _kThe diagonal matrix that consists of of positive singular value, V _kT _kRight singular matrix;

3) choose left singular matrix U _kFront N row obtain matrix U ' _k, and to matrix U ' _kCarry out singular value decomposition: Wherein: U ' _k=U _k(:, 1:N), N is the number of base station transmit antennas, P _kU ' _kLeft singular matrix, Ω _kU ' _kThe diagonal matrix that consists of of positive singular value, Q _kU ' _kRight singular matrix;

4) order

Get W ' _kFront m row obtain linear preconditioning matrix W _k, wherein: m is the data fluxion of each user assignment;

5) utilize linear preconditioning matrix W _kCarry out linear preliminary treatment for the initial information of k user's emission to the base station, concrete formula is:

S _k＝W _kq _k，

Wherein: q _kThat the base station is to the initial information of k user's emission, S _kThen be that the base station is to the pretreated actual transmission information of initial information inlet wire of k user's emission.

2. the data message linear preprocessing method of multi-user's multiple data stream mimo system according to claim 1 is characterized in that, the data message to be sent described in the first step comprises the coefficient parameter of image compression encoding information and voice autoregressive filter.

3. the data message linear preprocessing method of multi-user's multiple data stream mimo system according to claim 1 is characterized in that, the sign format of the data message to be sent described in the first step after with bit modulation represents.

4. the data message linear preprocessing method of multi-user's multiple data stream mimo system according to claim 1 is characterized in that, the subscriber channel information described in the second step is quantized versions or code book form.

5. the data message linear preprocessing method of multi-user's multiple data stream mimo system according to claim 1 is characterized in that, the stack described in the 4th step, and concrete formula is:

$S=\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{S}_k=\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{W}_k{q}_k$

Wherein: S is the final emission information of base station transmitter, and K is total reception number of users.

6. the data message linear preprocessing method of multi-user's multiple data stream mimo system according to claim 1 is characterized in that, the matched filtering described in the 6th step, and concrete formula is:

r _k′＝[(H _kW _k) ^H/||H _kW _k|| _F]×r _k

Wherein: r _k' be k the filtered information of user, H _kK user's subscriber channel information, W _kK user's linear preconditioning matrix, || || _FRepresenting matrix Frobenius norm, r _kK user's reception information.

Images