CN102185643B - Cooperative communication multi-resolution self-adapting wave beam forming method - Google Patents

Abstract

The invention discloses a cooperative communication multi-resolution self-adapting wave beam forming method which solves the codebook mismatching phenomenon and the problem of low backward channel frequency spectrum utilization ratio in the prior art. The method comprises the following concrete steps of: (1) building a plurality of codebooks of different resolutions according to performance requirements of a system, and storing at a transmitting end and a receiving end; (2) estimating an equivalent channel matrix from the transmitting end to the receiving end, and carrying out singular value decomposition (SVD); (3) searching an optimal wave beam forming matrix from the codebook set according to the performance requirements of the system, and feeding back index numbers to the transmitting end; and (4) finding out an optimal wave beam forming matrix from a local codebook set according to the received index numbers by the transmitting end, thus forming a wave beam. The cooperative communication multi-resolution self-adapting wave beam forming method has the advantages of stronger self adaptivity and smaller average feedback data amount on different channel scenes, realizes the channel adaptation as highest as possible, and improves the frequency spectrum utilization ratio of the backward channel of a cooperative communication system.

Description

Collaboration communication multi-resolution adaptive beam-forming method

Technical field

The invention belongs to communication technical field, further relate to collaboration communication transmitting terminal adaptive beamforming method in wireless communication field, have the robustness of raising collaboration communication and the characteristics of down channel frequency spectrum resource utilization rate.

Background technology

When carrying out beam forming in collaboration communication, acceptance point is to launch point, acceptance point is to via node and via node to the feedback that all needs to carry out channel condition information (Channel State Information (CSI)) between launch point, therefore need the CSI data volume of feedback larger, can reduce number of bits of feedback with Limited Feedback, improve the backward channel utilance.The effect quality of Limited Feedback adopts beam forming matrix code book closely related with transmitting-receiving two-end, so the design problem of beam forming matrix code book is very crucial.

The people such as Behrouz Khoshnevis are at article " Grassmannian Beamforming for MIMOAmplify-and-Forward Relaying " (IEEE Journal on Selected Areas in Communications, Volume:26, Issue:8pp.1397-1407, October 2008) in the beam forming that a kind of code book method for designing based on Jim Glassman space vanning (Grassmanian Space Packing) realizes collaboration communication is proposed.The implementation step of the method is: first, suppose that receiving terminal has CSI, that is: via node knows that transmitting terminal is to the CSI between via node, receiving terminal knows that transmitting terminal is to the CSI between receiving terminal and via node to the CSI between receiving terminal, communication system is stored in respectively transmitting terminal and receiving terminal with the Jim Glassman code book that chooses, and the Jim Glassman code books that another is different are stored in respectively via node and receiving terminal; The second, transmitting terminal is selected transmitting terminal beam forming vector and is carried out beam forming in code book according to channel information, and via node and receiving terminal carry out respectively matching treatment to received signal; The 3rd, via node is selected junction waves beam shaping vector and is carried out beam forming in code book according to channel information, and receiving terminal carries out matching treatment to the received signal.The method has proved that the code book design problem of optimal beam shaping matrix can be converted into optimum Jim Glassman space bin packing, adopting Jim Glassman space vanning code book can make the minimization of loss of system signal noise ratio, is the optimal selection that quantizes the beam forming matrix based on the code book of Jim Glassman space vanning.

The weak point of said method is: when the many antenna configurations scene that relates to, use scenes, when the combination of channel scene is changeable, comparatively serious code book mismatch phenomenon can occur, cause the decline of the performance such as beam forming adaptivity and the availability of frequency spectrum in collaboration communication.

Summary of the invention

The object of the invention is to overcome the deficiency of above-mentioned prior art, provide a kind of beam-forming method based on " the Jim Glassman space vanning of multiresolution " quantization method for collaboration communication beam forming problem, with the adaptivity of beam forming and the availability of frequency spectrum of system in the lifting collaboration communication.

The present invention realizes that the concrete steps of above-mentioned purpose are as follows:

(1) storage codebook set

1a) determine the code book number: the communication system data center processor is determined the maximum code book number that allows use of Jim Glassman space packing method of multiresolution according to the restriction of the average number of bits of feedback of communication system;

1b) select code book: the communication system data center processor selects a plurality of resolution by the codebook set of the code book that increases progressively from low to high as communication system in Jim Glassman space vanning result, comprise a plurality of code words in each code book, code word is the matrix that carries out beam forming;

1c) storage codebook set: communication system is stored in the codebook set that chooses respectively in the register of transmitting terminal and receiving terminal.

(2) channel decomposing

2a) transmitting training sequence: transmitting terminal is to via node and receiving terminal transmitting training sequence, and the via node signal processor adopts the mode of amplification forwarding to be transmitted to receiving terminal the training sequence that receives;

2b) channel estimating: the channel estimator of via node estimates transmitting terminal to the channel matrix of via node according to the training sequence that receives, the information of channel matrix is forwarded to receiving terminal, the receiving terminal channel estimator estimates respectively the channel matrix of transmitting terminal to the channel matrix of receiving terminal and via node to receiving terminal according to the training sequence that receives, and the receiver signal processor merges processing to three channel matrixes and obtains transmitting terminal to the equivalent channel matrix of receiving terminal;

2c) channel decomposing: the receiver signal processor is combined according to following formula the equivalent channel matrix H that obtains and carries out singular value decomposition:

H＝UDV ^H

Wherein, H is the equivalent channel matrix that the receiver estimation obtains, and U and V are that signal processor carries out the unitary matrice that singular value decomposition obtains, and D is that signal processor decomposes the diagonal matrix that obtains, [] ^HThe conjugate transpose of representing matrix.

(3) select optimal beam shaping matrix

3a) select optimal beam shaping matrix in the minimum code book of resolution: receiver processor uses step 2c according to minimum mean square error criterion) in the unitary matrice that obtains of singular value decomposition at step 1a) in search optimal beam shaping matrix in the minimum code book of the code book resolution that obtains;

3b) judge whether selected matrix meets the demands: if the mean square error that selected optimal beam shaping matrix produces is less than or equal to the mismatch error of communication system, receiver register records the code book index number of this beam forming matrix and codewords indexes number, completes the selection of optimal beam shaping matrix; If the mean square error that selected optimal beam shaping matrix produces is greater than the mismatch error of communication system, execution in step 3c);

3c) judge that code book resolution is whether maximum: if current to use the resolution of code book be not that the codebook set intermediate-resolution is the highest, receiver processor switches to the higher leveled code book of resolution with current code book, use step 2c according to minimum mean square error criterion) in the unitary matrice search optimal beam shaping matrix that obtains of singular value decomposition, return to execution in step 3b); If current to use the resolution of code book be that the codebook set intermediate-resolution is the highest, receiver register records the code book index number of current selected optimal beam shaping matrix and codewords indexes number, completes the selection of optimal beam shaping matrix.

(4) feedback channel information: receiver number feeds back to transmitting terminal with the code book index of the optimal beam shaping matrix selected number and codewords indexes.

(5) beam forming

5a) search beam forming matrix: the transmitter signal processor is according to the code book index of receiving terminal feedback number and codewords indexes number, search out corresponding code book and corresponding code word in the codebook set of transmitter register, with selected code word as the beam forming matrix;

5b) beam forming: transmitter uses 5a) in the beam forming matrix premultiplication that the searches out acquisition weighted signal that transmits, beam forming is completed in transmitting after the antenna transmission weighting.

The present invention compared with prior art has the following advantages:

First, the codebook set that uses due to the present invention is comprised of by the code book that increases progressively from low to high resolution, overcome that the Jim Glassman space packing method code book capacity of fixedly low resolution that prior art exists is fixing can't tackle the deficiency of channel scene changes, have advantages of different channel scene adaptives stronger, the realization of maximum possible channel adaptation.

Second, because at first the present invention searches for optimal beam shaping matrix in the lower code book of resolution, secondly search for optimal beam shaping matrix in the higher code book of resolution, overcome the excessive lower deficiency of cooperation communication system backward channel utilance that causes of fixing high-resolution Jim Glassman space packing method feedback data amount that prior art exists, have advantages of that average feedback data amount reduces greatly, has improved the availability of frequency spectrum of cooperation communication system backward channel.

Description of drawings

Fig. 1 is flow chart of the present invention.

Fig. 2 is collaboration communication beam forming scene schematic diagram of the present invention.

Embodiment

The present invention will be further described referring to accompanying drawing.

In collaboration communication beam forming scene shown in Figure 2, the process that the present invention completes the collaboration communication beam forming as shown in Figure 1, the step of its realization is as follows:

Step (1) storage codebook set

1a) determine the code book number.The communication system data center processor calculates the maximum code book number that allows use of Jim Glassman space packing method of multiresolution according to the restriction of the average number of bits of feedback of communication system according to following formula:

${\log }_{2}L+\underset{k=0}{\overset{L-1}{\mathrm{\Σ}}}{\log }_2\left(2^{k}m\right)\≤B$

Wherein, L is the number of code book in codebook set, ∑ () expression cumulative and, k for the counting factor (0≤k≤L-1), m is the element number in the minimum code book of resolution,

Expression rounds up, N _TBe the number of transmit antennas sum of transmitting terminal number of transmit antennas and via node, B is the maximum average number of bits of feedback of communication system restriction;

1b) select code book.The communication system data center processor is G (N in dimension _T, N _T) Jim Glassman space vanning result in select L resolution by the code book that increases progressively from low to high, the case dimension of result of G () expression Jim Glassman space, N _TBe the number of transmit antennas sum of transmitting terminal number of transmit antennas and via node, the L that selects code book forms the codebook set of communication system, and the code word number that comprises in code book is respectively 2 from low to high by resolution ^kM (k=0,1 ... L-1), m is the element number in the minimum code book of resolution, and code word is matrix form, and in order to carry out beam forming, the dimension of code word is N _T* N _T

1c) storage code book: communication system is stored in the codebook set that chooses respectively in the register of transmitting terminal and receiving terminal.

Step (2) channel decomposing

2a) transmitting training sequence.Transmitting terminal is to via node and receiving terminal transmitting training sequence, and the via node signal processor adopts the mode of amplification forwarding to be transmitted to receiving terminal the training sequence that receives, and the gain amplifier of via node amplifier is A;

2b) channel estimating.Embodiments of the invention have adopted the channel estimating based on training sequence, be applicable to the system of burst transfer mode based on the channel estimating of training sequence, by sending known training sequence, carry out initial channel estimating at receiving terminal, when sending the Useful Information data, utilize initial channel estimation results to carry out a judgement and upgrade, complete real-time channel estimating.The embodiment of the present invention is due to the channel estimation technique that adopts based on training sequence, although the availability of frequency spectrum is lower, delay performance is better.

The channel estimator of via node estimates transmitting terminal to the channel matrix of via node according to the training sequence that receives, the information of channel matrix is forwarded to receiving terminal, the receiving terminal channel estimator estimates respectively the channel matrix of transmitting terminal to the channel matrix of receiving terminal and via node to receiving terminal according to the training sequence that receives, three channel matrixes is merged processing according to following formula obtain transmitting terminal to the equivalent channel matrix of receiving terminal:

H＝[AH ₂H ₁?H ₀]

Wherein, H be transmitting terminal to the equivalent channel matrix of receiving terminal, A is the gain amplifier of via node amplifier, H ₂Be the channel matrix of via node to receiving terminal, H ₁Be the channel matrix of transmitting terminal to via node, H ₀Be the direct channel matrix of transmitting terminal to receiving terminal.

2c) channel decomposing.The receiver signal processor carries out singular value decomposition according to following formula to estimating the equivalent channel matrix H that obtains:

H＝UDV ^H

Wherein, H is the equivalent channel matrix that the receiver estimation obtains, and U and V are that signal processor carries out the unitary matrice that singular value decomposition obtains, and D is that signal processor decomposes the diagonal matrix that obtains, [] ^HThe conjugate transpose of representing matrix.

Step (3) is selected optimal beam shaping matrix

3a) select optimal beam shaping matrix in the minimum code book of resolution.Receiver processor uses step 2c) in the unitary matrice that obtains of singular value decomposition, at step 1a) in the minimum code book F of code book resolution that obtains ₁according to following formula search optimal beam shaping matrix W _opt:

$W_{\mathrm{opt}}=\arg \underset{W\∈F_1}{\min }{\left|\right|U^H\mathrm{HV}-U^H\mathrm{HW}\left|\right|}^2$

Wherein, W _optBe the optimal beam shaping matrix that search obtains, arg min () expression is got minimum value according to the variation of independent variable, F ₁Be the minimum code book of resolution, W is that optimal beam shaping matrix search is brought in funds, and U and V are step 2c) in the unitary matrice that obtains, H is step 2b) in the equivalent channel matrix that obtains, || || ²Expression is asked square value to matrix norm.

3b) judge whether selected matrix meets the demands.The receiver signal processor judges that according to following formula mean square error that current selected optimal beam shaping matrix produces is whether less than or equal to the mismatch error of communication system:

$\frac{{\left|\right|U^H\mathrm{HV}-U^{H}H{W}_{\mathrm{opt}}\left|\right|}^2}{{\left|\right|H\left|\right|}^2}\≤\mathrm{\δ}$

Wherein, U and V are step 2c) in the unitary matrice that obtains, H is step 2b) in the equivalent channel matrix that obtains, W _optThe optimal beam shaping matrix that obtains for search, || || ²Expression is asked square value to matrix norm, and δ is the patient maximum mismatch error of communication system.

If W _optThe mean square error that produces is less than or equal to the maximum mismatch error of communication system, and receiver register records W _optCode book index number and codewords indexes number, complete the selection of optimal beam shaping matrix; If W _optThe mean square error that produces is greater than the maximum mismatch error of communication system, execution in step 3c);

3c) judge whether code book resolution is maximum.The receiving end signal processor compares all code books in current use code book and codebook set, and the resolution that judges current use code book is that the codebook set intermediate-resolution is the highest.If current to use the resolution of code book be not that the codebook set intermediate-resolution is the highest, receiver processor switches to the higher leveled code book of resolution with current code book, according to step 3a) in searching method search optimal beam shaping matrix, return to execution in step 3b); If current to use the resolution of code book be that the codebook set intermediate-resolution is the highest, receiver register records the code book index number of current selected optimal beam shaping matrix and codewords indexes number, completes the selection of optimal beam shaping matrix.

Step (4) feedback channel information.Receiver is with the optimal beam shaping matrix W of selecting _optCode book index number and codewords indexes number feed back to transmitting terminal.

Step (5) beam forming

5a) search beam forming matrix.The transmitter signal processor searches out corresponding code book and corresponding code word according to the code book index of receiving terminal feedback number and codewords indexes number in the codebook set of transmitter register, with selected code word as the beam forming matrix.

5b) beam forming.Transmitter uses 5a) in the beam forming matrix premultiplication that the searches out acquisition weighted signal that transmits, beam forming is completed in transmitting after the antenna transmission weighting.

Claims (4)

1. a collaboration communication multi-resolution adaptive beam-forming method, comprise the steps:

(1) storage codebook set

1a) determine the code book number: the communication system data center processor is determined the maximum code book number that allows use of Jim Glassman space packing method of multiresolution according to the restriction of the average number of bits of feedback of communication system;

1b) select code book: the communication system data center processor selects a plurality of resolution by the codebook set of the code book that increases progressively from low to high as communication system in Jim Glassman space vanning result, comprise a plurality of code words in each code book, code word is the matrix that carries out beam forming;

1c) storage codebook set: communication system is stored in the codebook set that chooses respectively in the register of transmitter and receiver;

(2) channel decomposing

2a) transmitting training sequence: transmitter is to via node and receiver transmitting training sequence, and the via node signal processor adopts the mode of amplification forwarding to be transmitted to receiver the training sequence that receives;

2b) channel estimating: the channel estimator of via node estimates transmitter to the channel matrix of via node according to the training sequence that receives, the information of channel matrix is forwarded to receiver, the receiver channel estimator estimates respectively the channel matrix of transmitted from transmitter to receiver and via node to the channel matrix of receiver according to the training sequence that receives, and the receiver signal processor merges to three channel matrixes the equivalent channel matrix that processing obtains transmitted from transmitter to receiver;

2c) channel decomposing: the receiver signal processor is combined according to following formula the equivalent channel matrix H that obtains and carries out singular value decomposition:

H＝UDV ^H

Wherein, H is the equivalent channel matrix that the receiver estimation obtains, and U and V are that signal processor carries out the unitary matrice that singular value decomposition obtains, and D is that signal processor decomposes the diagonal matrix that obtains, [] ^HThe conjugate transpose of representing matrix;

(3) select optimal beam shaping matrix

3a) select optimal beam shaping matrix in the minimum code book of resolution: the receiver signal processor uses step 2c according to minimum mean square error criterion) in the unitary matrice that obtains of singular value decomposition at step 1a) in search optimal beam shaping matrix in the minimum code book of the code book resolution that obtains;

3b) judge whether selected matrix meets the demands: if the mean square error that selected optimal beam shaping matrix produces is less than or equal to the mismatch error of communication system, receiver register records the code book index number of this beam forming matrix and codewords indexes number, completes the selection of optimal beam shaping matrix; If the mean square error that selected optimal beam shaping matrix produces is greater than the mismatch error of communication system, execution in step 3c);

3c) judge that code book resolution is whether maximum: if current to use the resolution of code book be not that the codebook set intermediate-resolution is the highest, the receiver signal processor switches to the higher leveled code book of resolution with current code book, use step 2c according to minimum mean square error criterion) in the unitary matrice search optimal beam shaping matrix that obtains of singular value decomposition, return to execution in step 3b); If current to use the resolution of code book be that the codebook set intermediate-resolution is the highest, receiver register records the code book index number of current selected optimal beam shaping matrix and codewords indexes number, completes the selection of optimal beam shaping matrix;

(4) feedback channel information: receiver number feeds back to transmitter with the code book index of the optimal beam shaping matrix selected number and codewords indexes;

(5) beam forming

5a) search beam forming matrix: the transmitter signal processor is according to the code book index of receiver feedback number and codewords indexes number, search out corresponding code book and corresponding code word in the codebook set of transmitter register, with selected code word as the beam forming matrix;

5b) beam forming: transmitter uses 5a) in the beam forming matrix premultiplication that the searches out acquisition weighted signal that transmits, beam forming is completed in transmitting after the antenna transmission weighting.

2. collaboration communication multi-resolution adaptive beam-forming method according to claim 1, it is characterized in that, described step 1a) in, the communication system data center processor is according to the restriction of the average number of bits of feedback of communication system, determine the maximum code book number that allows use of Jim Glassman space packing method of multiresolution, carry out according to following formula:

${\log }_{2}L+\underset{k=0}{\overset{L-1}{\mathrm{\Σ}}}{\log }_2\left(2^{k}m\right)\≤B$

Wherein, L is the quantity of code book in codebook set, ∑ () expression cumulative and, k is the counting factor 0≤k≤L-1, m is the element number in the minimum code book of resolution,

Expression rounds up, N _TBe the transmitter antenna number, B is the maximum average number of bits of feedback of communication system restriction.

3. collaboration communication multi-resolution adaptive beam-forming method according to claim 1, is characterized in that described step 3a) in minimum mean square error criterion determined by following formula:

$W_{\mathrm{opt}}=\arg \underset{W\∈F_1}{\min }\left|\right|U^H\mathrm{HV}-U^H\mathrm{HW}|{|}^2$

Wherein, W _optBe the optimal beam shaping matrix that search obtains, argmin () expression is got minimum value according to the variation of independent variable, F ₁Be the minimum code book of resolution, W is that optimal beam shaping matrix search is brought in funds, and U and V are step 2c) in the unitary matrice that obtains, H is step 2b) in the equivalent channel matrix that obtains, || || ²Expression is asked square value to matrix norm.

4. collaboration communication multi-resolution adaptive beam-forming method according to claim 1, it is characterized in that, described step 3b) judge that the mean square error of current selected optimal beam shaping matrix generation whether less than or equal to the mismatch error of communication system, carries out according to following formula in:

$\frac{\left|\right|U^H\mathrm{HV}-U^{H}H{W}_{\mathrm{opt}}|{|}^2}{\left|\right|H|{|}^2}\≤\mathrm{\δ}$

Wherein, U and V are step 2c) in the unitary matrice that obtains, H is step 2b) in the equivalent channel matrix that obtains, W _optThe optimal beam shaping matrix that obtains for search, || || ²Expression is asked square value to matrix norm, and δ is the patient maximum mismatch error of communication system.

Images