CN104868945A - Multi-user downlink three-dimensional statistics beam forming adaptive transmission method using channel mean information - Google Patents

Abstract

The invention discloses a multi-user downlink three-dimensional statistical beamforming adaptive transmission method using channel mean value information. The base station adopts a uniform planar antenna array; Its own channel mean value information, that is, Rice factor, channel mean value vertical direction main mode and horizontal direction main mode, is fed back to the base station. For the TDD system, the base station uses the result of the uplink channel estimation to calculate the downlink channel average information of each user. The base station selects serving users according to the obtained channel mean value information, and performs precoded transmission on the serving users. The multi-user downlink three-dimensional statistical beamforming adaptive transmission method using channel mean value information provided by the present invention effectively improves the robustness of the system, and can obtain higher average mutual information with lower computational complexity.

Description

Multi-user Downlink 3D Statistical Beamforming Adaptive Transmission Method Using Channel Mean Information

technical field

The invention relates to a multi-user downlink transmission system for transmitting high-speed data by using a uniform planar antenna array, in particular to an adaptive transmission method for a multi-user downlink transmission system using channel mean value information.

Background technique

In recent years, research on information theory has shown that multi-antenna technology can significantly improve the transmission rate of wireless communication systems. At present, the research on the point-to-point single-user system has basically reached a conclusion, while the research on the capacity and optimal transmission scheme of the multi-user system has attracted extensive attention of international scholars. The multi-user multi-antenna system can obtain good performance and large capacity due to its diversity gain, multiplexing gain and multi-user diversity gain, and will become one of the key technologies of the new generation wireless communication network.

Compared with the traditional single-user multi-antenna system, the multi-user multi-antenna system has the following outstanding advantages: Due to the use of the so-called multi-user multiplexing, the multi-user multi-antenna technology can bring a direct gain in multiple access capacity ( proportional to the number of base station antennas); using multi-user diversity and scheduling, multi-user multi-antenna technology can break through many propagation limitations that plague single-user multi-antenna communication, such as channel matrix rank deficit or antenna correlation; In the case of MISO (multiple-input single-output), the multi-user system can still obtain spatial multiplexing gain, so it is conducive to the development of small and cheap terminals.

Unfortunately, for multi-user systems, reaping the benefits of multi-antenna technology comes at a price. For a single-user multi-antenna communication system, it is not necessary for the transmitter to know the channel information, but it is crucial for many multi-user multi-antenna downlink precoding technologies. Many existing multi-user multi-antenna downlink transmission systems assume that ideal channel information is known by the base station. In actual communication, the channel information of the base station is provided by the user through the uplink limited feedback channel or obtained by uplink channel estimation. Since there is inevitably a feedback delay in the transmission of feedback information, and there may be errors in channel estimation, it is often unrealistic to assume that the ideal channel information is known at the sending end, especially when the number of users and the number of transmitting antennas are large and the channel when the state changes rapidly. In addition, for the FDD system, the feedback of channel information causes a great burden on the uplink capacity, and this problem becomes more serious in the broadband (such as OFDM) system and the system with high mobility. Therefore, it is an appropriate choice to use statistical channel state information, such as transceiver correlation matrix, mean value information, etc., for adaptive transmission.

Contents of the invention

Purpose of the invention: In order to overcome the deficiencies in the prior art, the present invention provides a three-dimensional statistical beamforming adaptive transmission scheme using channel mean value information for a multi-user downlink transmission system using a uniform planar antenna array in the base station, which can be based on the channel mean value Features Adjust sending parameters to obtain higher sum rate and lower complexity.

Technical scheme: in order to achieve the above object, the technical scheme adopted in the present invention is:

A multi-user downlink three-dimensional statistical beamforming adaptive transmission method utilizing channel mean value information, comprising the following steps:

(1) Acquisition of base station channel mean value information: the base station adopts a uniform planar antenna array (UPA), which has M rows in the vertical direction and N antenna elements in each row in the horizontal direction, totaling M×N antenna elements, each The row and each column of antenna elements form a uniform linear antenna array (ULA). The distance between adjacent antenna elements in the horizontal and vertical directions is half of the carrier wavelength. The number of receiving antennas for each user is 1, and the total number of users is L;

The normalized channel row vector between base station and user i is h _i , satisfying The (m-1)N+nth element of h _i (h _i ) _(m-1)N+n is the channel coefficient between the antenna element in the mth row and the nth column of the base station and the user i, superscript Represents conjugate transpose, E{ } represents mean value, (α) _i represents the i-th element of vector α;

For the TDD system, first, the base station obtains the channel h _i from the base station to user i by using the result of uplink channel estimation and channel reciprocity; then, the base station calculates the channel mean value of user i and Rice factor And then get the channel mean the vertical component of $a_i^{\left(v\right)}=[{\left({\stackrel{\‾}{h}}_i\right)}_1,{\left({\stackrel{\‾}{h}}_i\right)}_{N+1},{\left({\stackrel{\‾}{h}}_i\right)}_{2N+1},...,{\left({\stackrel{\‾}{h}}_i\right)}_{(m-1)N+1}]$ and horizontal component $a_i^{\left(h\right)}=[{\left({\stackrel{\‾}{h}}_i\right)}_1,{\left({\stackrel{\‾}{h}}_i\right)}_2,{\left({\stackrel{\‾}{h}}_i\right)}_3,...,{\left({\stackrel{\‾}{h}}_i\right)}_N],$ and calculate separately and Among them, F _M and F _N are the DFT matrices of M×M and N×N respectively, and the elements of the mth row and nth column of F _M and F _N are respectively ${\left[f_m\right]}_{m,\mathrm{no}}=\frac{1}{\sqrt{m}}{e}^{j2\mathrm{\π}(m-1)(\mathrm{no}-1)/m}$ and ${\left[f_N\right]}_{m,\mathrm{no}}=\frac{1}{\sqrt{N}}{e}^{j2\mathrm{\π}(m-1)(\mathrm{no}-1)/N};$ Finally, the base station finds out its main mode in the vertical direction and the main mode in the horizontal direction respectively, that is, finds out the largest element on the diagonal elements of Λ _V,i and Λ _H,i respectively and The corresponding indices l _i and j _i , namely ${\mathrm{\λ}}_{V,i}^{\max }={\mathrm{\λ}}_{V,i}^{\left(l_i\right)},$ ${\mathrm{\λ}}_{h,i}^{\max }={\mathrm{\λ}}_{h,i}^{\left(j_i\right)},$ in and Respectively Λ _{V, i} and Λ _H, the l _i and j _i diagonal elements of i;

For the FDD system, first, user i uses the result of downlink channel estimation to obtain the normalized channel vector h _{i from the base station to user i} ; then, user i calculates its channel mean and Rice factor And then get its channel mean the vertical component of $a_i^{\left(v\right)}=[{\left({\stackrel{\‾}{h}}_i\right)}_1,{\left({\stackrel{\‾}{h}}_i\right)}_{N+1},{\left({\stackrel{\‾}{h}}_i\right)}_{2N+1},...,{\left({\stackrel{\‾}{h}}_i\right)}_{(m-1)N+1}]$ and horizontal component $a_i^{\left(h\right)}=[{\left({\stackrel{\‾}{h}}_i\right)}_1,{\left({\stackrel{\‾}{h}}_i\right)}_2,{\left({\stackrel{\‾}{h}}_i\right)}_3,...,{\left({\stackrel{\‾}{h}}_i\right)}_N],$ and calculate separately and Among them, F _M and F _N are the DFT matrices of M×M and N×N respectively, and the elements of the mth row and nth column of F _M and F _N are respectively ${\left[f_m\right]}_{m,\mathrm{no}}=\frac{1}{\sqrt{m}}{e}^{j2\mathrm{\π}(m-1)(\mathrm{no}-1)/m}$ and ${\left[f_N\right]}_{m,\mathrm{no}}=\frac{1}{\sqrt{N}}{e}^{j2\mathrm{\π}(m-1)(\mathrm{no}-1)/N};$ Finally, user i finds its main mode in the vertical direction and the main mode in the horizontal direction respectively, that is, finds the largest element on the diagonal elements of Λ _V,i and Λ _H,i respectively and corresponding index l _i and j _i , and feed back K _i , l _i and j _i to the base station;

(2) User classification: the base station divides users into M×N categories according to l _i and j _i of each user, i=1,...,L, and the classification criterion is: if l _i =m and j _i =n, then Classify user i into the (m, n)th category;

(3) User scheduling: the base station selects the user with the largest Rice factor from each type of user, that is, the user with the largest _Ki . If the number of selected users is greater than the maximum number of users U in the system, the previously selected Select the U users with the largest Rice factor as the service users among the users, denoted as user 1, user 2, ..., user U;

(4) For the users selected in step (3), the base station calculates the precoding vector of each user: the precoding vector of user i is Among them, the total transmit power of P, for the matrix column l _i of , for the matrix The j _i column of ;

(5) Use the precoding vector calculated in step (4) to perform precoding transmission to the user.

Beneficial effects: the multi-user downlink three-dimensional statistical beamforming adaptive transmission method using channel mean value information provided by the present invention has the following advantages: 1. This method only requires channel mean value information, and is applicable to various typical wireless communication systems; 2. . The adaptive transmission method in the method has low complexity and is easy to implement; 3. The method can obtain a higher sum rate.

Detailed ways

Below in conjunction with embodiment the present invention will be further described.

Consider a multi-user downlink. The base station uses a uniform planar antenna array. The antenna array has M rows in the vertical direction and N antenna elements in each row in the horizontal direction. There are M×N antenna elements in total. Each row and each column are Uniform linear antenna array (ULA), the number of receiving antennas for each user is 1, the total number of users is L, and the base station sends independent signals to each user. Based on the analysis of its sum rate, the following precoding transmission scheme is constructed:

At the user end: if it is an FDD system, user i performs channel estimation on the digital baseband received signal y ⁽ⁱ⁾ (n), uses the channel estimation result to calculate channel mean information, and feeds the channel mean information back to the base station.

On the base station side: if it is an FDD system, the base station receives the channel average information fed back by each user; if it is a TDD system, the base station uses the channel estimation results of the uplink to calculate the channel average information of each user's downlink. Then, use the obtained channel mean value information to classify users, divide all users into M×N categories, schedule the classified users, and select U service users. Then, calculate the transmission precoding vector of each user (the appropriate amount of transmission precoding of user i is w _i ), and perform linear precoding on the input symbol stream of the serving user (the input symbol stream of user i is d _i (n)), Get the sent signal s(n)=[s _1,1 (n),s _1,2 (n),…,s _1,N (n),s _2,1 (n),s _2,2 (n) ,…,s _2,N (n),……,s _M,1 (n),s _M,2 (n),…,s _M,N (n)] ^T , where s _i,j (n ) indicates that the antenna unit in row i and column j of the base station transmits a signal. The relationship between d _i (n) and s(n) satisfies the following:

$\mathrm{the\; s}\left(\mathrm{no}\right)=\underset{i=1}{\overset{u}{\mathrm{\Σ}}}{w}_i{d}_i\left(\mathrm{no}\right)$ 【1】

In order to make the technical solution among the present invention clearer, this solution is described in detail below:

1. Obtaining channel mean information

The channel mean value information in the scheme is the Rice factor of each user, the vertical direction main mode and the horizontal direction main mode of the channel mean value. For the TDD system, first, the base station uses the result of uplink channel estimation and channel reciprocity to obtain user i's normalized downlink channel row vector h _i , and calculates its channel mean value and the Rice factor K _i :

${\stackrel{\‾}{h}}_i=\mathrm{E.}\left\{h_i\right\}$ 【2】

【3】

Among them, E{·} represents expectation. Next, get the vertical component of the channel mean of user i and horizontal component

$a_i^{\left(v\right)}=[{\left({\stackrel{\‾}{h}}_i\right)}_1,{\left({\stackrel{\‾}{h}}_i\right)}_{N+1},{\left({\stackrel{\‾}{h}}_i\right)}_{2N+1},...,{\left({\stackrel{\‾}{h}}_i\right)}_{(m-1)N+1}]$ 【4】

$a_i^{\left(h\right)}=[{\left({\stackrel{\‾}{h}}_i\right)}_1,{\left({\stackrel{\‾}{h}}_i\right)}_2,{\left({\stackrel{\‾}{h}}_i\right)}_3,...,{\left({\stackrel{\‾}{h}}_i\right)}_N],$ 【5】

Calculate Λ _V,i and Λ _H,i respectively:

Among them, F _M and F _N are M×M and N×N dimensional DFT matrices respectively, and the elements of the m row and nth column are respectively ${\left[f_m\right]}_{m,\mathrm{no}}=\frac{1}{\sqrt{m}}{e}^{j2\mathrm{\π}(m-1)(\mathrm{no}-1)/m}$ and ${\left[f_N\right]}_{m,\mathrm{no}}=\frac{1}{\sqrt{N}}{e}^{j2\mathrm{\π}(m-1)(\mathrm{no}-1)/N}.$ Finally, find out the vertical main mode and the horizontal main mode of the channel mean value of user i respectively, that is, the largest element on the diagonal elements of Λ _V,i and Λ _H,i and The corresponding indices l _i and j _i : in and are the l _i -th and j _i -th diagonal elements of Λ _V,i and Λ _H,i respectively. For the FDD system, user i uses the result of its channel estimation to calculate K _i , Λ _V,i and Λ _H,i according to the formula [2] - formula [7], and find out its channel mean value in the vertical direction main mode and the horizontal direction respectively main mode l _i and j _i , and feed back K _i , l _i and j _i to the base station.

2. User Grouping

The base station divides users into M×N categories according to l _i and j _i of each user, i=1,...,L, and the classification criterion is: if l _i =m and j _i =n, user i is classified into the first (m,n) class.

3. User Scheduling

The base station selects the user with the largest Rice factor from each type of users, that is, the user with the largest _Ki , and if the number of selected users is greater than the maximum number of users U in the system, then select Lai from the previously selected users. The U users with the largest Si factor are service users, denoted as user 1, user 2, ..., user U.

4. Send precoding vector calculation

The precoding vector w _i of user i scheduled in this scheme is calculated by the following formula:

Among them, the total transmit power of P, for the matrix column l _i of , for the matrix j _i column of .

The specific embodiment of the present invention is as follows:

1) If it is a TDD system, skip to step 2); if it is an FDD system, skip to step 4).

2) The base station uses the results of uplink channel estimation and channel reciprocity to obtain the downlink normalized channel row vector h _i of user i, and use the formulas [2] and [3] to calculate the channel mean value of user i respectively and the Rice factor K _i .

3) Calculate Λ _V,i and Λ _H,i using the formulas [4]-[7], and find out the vertical main mode and the horizontal main mode of the channel mean value of user i respectively, that is, the pair of Λ _V,i and Λ _H,i largest element on corner element and For the corresponding indexes l _i and j _i , go to step 7).

4) User i uses the result of its channel estimation to calculate the channel mean value of user i respectively according to the formulas [2] and [3] and the Rice factor K _i .

5) User i calculates its Λ _V,i and Λ _H, i using the formulas [4]-[7], and finds out the main mode in the vertical direction and the main mode in the horizontal direction of the channel mean value of user i respectively, that is, Λ _V,i and Λ _{H , the largest element on the i} diagonal element and The corresponding indices l _i and j _i .

6) User i feeds back its K _i , l _i and j _i to the base station.

7) The base station divides users into M×N categories according to l _i and j _i of each user, i=1,...,L, and the classification criterion is: if l _i =m and j _i =n, then user i is classified into into the (m, n)th category.

8) The base station selects the user with the largest Rice factor from each type of users, that is, the user with the largest K _i . If the number of selected users is greater than the maximum number of users U in the system, select the user among the previously selected users. The U users with the largest Rice factor are service users, denoted as user 1, user 2, ..., user U.

9) For the user i selected in step 8), use the formula [8] to calculate the precoding vector w _i .

10) Using w _i calculated in step 9), perform precoding transmission on the U users selected in step 8) according to the formula [1].

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 are also possible. It should be regarded as the protection scope of the present invention.

Claims (1)

1. utilize a multiuser downstream three-dimensional statistics beam forming adaptive transmission method for channel mean information, it is characterized in that: comprise the steps:

(1) acquisition of BTS channel mean information: base station adopts uniform planar antenna array, this antenna array vertical direction has M capable, horizontal direction often row has N number of bay, M × N number of bay altogether, often row and every array antenna array element all form uniform linear antenna array, adjacent antenna array element distance is the half of carrier wavelength in the horizontal and vertical directions, and the reception antenna number of each user is 1, and total number of users is L;

Normalization channel row vector between base station and user i is h _i, meet h _i(m-1) N+n element (h _i) _{(m-1) N+n}channel coefficients between the bay arranged for base station m capable n-th and user i, subscript represent conjugate transpose, E{} representative is averaged, (α) _ii-th element of representation vector α;

For TDD system, first, base station utilizes the result of uplink channel estimation and the reciprocity of channel, obtains the channel h of base station to user i _i; Then, base station calculates the channel average of user i and Rice factor and then obtain channel average vertical component $a_i^{\left(v\right)}=[{\left({\stackrel{\‾}{h}}_i\right)}_1,{\left({\stackrel{\‾}{h}}_i\right)}_{N+1},{\left({\stackrel{\‾}{h}}_i\right)}_{2N+1},...,{\left({\stackrel{\‾}{h}}_i\right)}_{(M-1)N+1}]$ And horizontal component $a_i^{\left(h\right)}=[{\left({\stackrel{\‾}{h}}_i\right)}_1,{\left({\stackrel{\‾}{h}}_i\right)}_2,{\left({\stackrel{\‾}{h}}_i\right)}_3,...,{\left({\stackrel{\‾}{h}}_i\right)}_N],$ And calculate respectively with wherein F _mand F _nbe respectively the DFT matrix of M × M and N × N, F _mand F _nthe element of capable n-th row of m is respectively ${\left[F_M\right]}_{m,n}=\frac{1}{\sqrt{M}}{e}^{j2\mathrm{\π}(m-1)(n-1)/M}$ With ${\left[F_N\right]}_{m,n}=\frac{1}{\sqrt{N}}{e}^{j2\mathrm{\π}(m-1)(n-1)/N};$ Finally, base station finds out its vertical direction holotype and horizontal direction holotype respectively, namely finds out Λ respectively _v,iand Λ _h,ielement maximum on diagonal element with corresponding index l _iand j _i, namely wherein with be respectively Λ _v,iand Λ _h,il _iand jth _iindividual diagonal element;

For FDD system, first, the result that user i utilizes down channel to estimate, obtains the normalization channel vector h of base station to user i _i; Then, user i calculates its channel average and Rice factor and then obtain its channel average vertical component $a_i^{\left(v\right)}=[{\left({\stackrel{\‾}{h}}_i\right)}_1,{\left({\stackrel{\‾}{h}}_i\right)}_{N+1},{\left({\stackrel{\‾}{h}}_i\right)}_{2N+1},...,{\left({\stackrel{\‾}{h}}_i\right)}_{(M-1)N+1}]$ And horizontal component $a_i^{\left(h\right)}=[{\left({\stackrel{\‾}{h}}_i\right)}_1,{\left({\stackrel{\‾}{h}}_i\right)}_2,{\left({\stackrel{\‾}{h}}_i\right)}_3,...,{\left({\stackrel{\‾}{h}}_i\right)}_N],$ And calculate respectively with wherein F _mand F _nbe respectively the DFT matrix of M × M and N × N, F _mand F _nthe element of capable n-th row of m is respectively ${\left[F_M\right]}_{m,n}=\frac{1}{\sqrt{M}}{e}^{j2\mathrm{\π}(m-1)(n-1)/M}$ With ${\left[F_N\right]}_{m,n}=\frac{1}{\sqrt{N}}{e}^{j2\mathrm{\π}(m-1)(n-1)/N};$ Finally, user i finds out its vertical direction holotype and horizontal direction holotype respectively, namely finds out Λ respectively _v,iand Λ _h,ielement maximum on diagonal element with corresponding index l _iand j _i, and by K _i, l _iand j _ifeed back to base station;

(2) users classification: base station is according to the l of each user _iand j _i, i=1 ..., L, is divided into M × N class by user, sorting criterion is: if l _i=m and j _i=n, be then included into (m, n) class by user i;

(3) user scheduling: the maximum user of its Rice factor is selected respectively in base station from every class user, i.e. K _imaximum user, if selected number of users is greater than system maximum number of user U, then in previously selected user, select the maximum U of a Rice factor user for service-user, be designated as user 1, user 2 ..., user U;

(4) to the user that step (3) is selected, base station calculates the precoding vectors of each user: the precoding vectors of user i is wherein, P total emission power, for matrix l _irow, for matrix jth _irow;

(5) precoding vectors calculated in step (4) is utilized to carry out precoding transmissions to user.