WO2015037936A1

WO2015037936A1 - Method and apparatus for grouping antennas in multiple-input multiple-output antenna system

Info

Publication number: WO2015037936A1
Application number: PCT/KR2014/008513
Authority: WO
Inventors: 심병효; 이병주; 김태영; 설지윤
Original assignee: 삼성전자 주식회사; 고려대학교 산학협력단
Priority date: 2013-09-13
Filing date: 2014-09-12
Publication date: 2015-03-19

Abstract

Disclosed are a method and an apparatus for grouping antennas in a multiple-input multiple-output antenna system. The method of the present invention comprises the steps of: measuring a channel vector for a plurality of antennas of a base station; grouping channel coefficients of the channel vector in accordance with a plurality of antenna grouping patterns and determining grouped codebook vectors corresponding to the grouped channel coefficients; selecting one of the antenna grouping patterns using the grouped codebook vectors for the antenna grouping patterns; and feeding back, to the base station, a pattern index indicating the selected antenna grouping pattern and a codebook index indicating the grouped codebook vector corresponding to the selected antenna grouping pattern.

Description

Method and device for antenna grouping in multi-input / output antenna system

The present invention relates to a method and apparatus for grouping antennas in a communication system having multiple input and output antennas.

A full dimension multiple input multiple output (FD-MIMO) system may be implemented in a frequency division duplexing (FDD) system and a time division duplexing (TDD) system. Since the TDD type system utilizes channel reciprocity of uplink and downlink, the burden on channel feedback is not large. However, when a large capacity MIMO technology is applied to an FDD-based system used in many cellular networks, feedback of channel state information for supporting beamforming of a transmitter is required.

Channel feedback is typically represented by a finite number of feedback bits. If the feedback information is incomplete, beamforming at the transmitter may also be inaccurate. In fact, if beamforming is not performed correctly in a multiuser MIMO system, the interference between users is not completely eliminated, and as a result, the signal to interference and noise ratio (SINR) of the receiver is degraded. The accuracy of channel state information directly affects the downlink multiplexing gain in a multiuser MIMO system. In particular, the number of feedback bits for quantizing a channel should be set in proportion to the number of base station antennas and the signal-to-noise ratio in order to maintain a difference in sum rate between perfect channel state information and inaccurate channel state information within a certain range. The FD-MIMO system is a technique for obtaining a high data rate by utilizing a very large number of antennas from dozens to hundreds at a base station, and the feedback load is inevitably increased in proportion to the number of antennas.

The performance of finite channel feedback depends on the accuracy of the channel vector quantization. When beamforming using an inaccurate quantized channel vector, it is difficult to expect an increase in total capacity because it becomes an interference limiting system in which interference between users is not completely eliminated. In particular, in the case of a large capacity MIMO communication system, as the number of transmit antennas increases, the dimension of the channel vector increases linearly. Therefore, as the number of transmit antennas increases, the amount of feedback for channel vector quantization increases. However, in order not to increase the overhead of the uplink, the channel feedback should be expressed with as few bits as possible, which may cause a problem that the accuracy of channel vector quantization is reduced and it is difficult to obtain an increase in performance through beamforming.

The present invention provides a method and apparatus for transmitting and receiving feedback information on a channel state in a communication system.

The present invention provides a method and apparatus for performing antenna grouping for uplink feedback.

The present invention provides a method and apparatus for feeding back channel state information at a receiving end to perform beamforming at a transmitting end of a large-capacity multiple input / output antenna system.

The present invention provides a method and apparatus for reducing the dimension of a vector for channel quantization and reducing system feedback load by performing antenna grouping.

The present invention provides a method and apparatus for performing antenna grouping to reduce feedback load on the uplink of a large capacity multiple input and output antenna (FD-MIMO) system.

Method according to a preferred embodiment of the present invention; An antenna grouping method of a multiple input / output antenna system, comprising: measuring channel vectors for a plurality of antennas of a base station, grouping channel coefficients of the channel vector according to a plurality of antenna grouping patterns, and grouping the channel coefficients Determining grouped codebook vectors corresponding to the plurality of antennas; selecting an antenna grouping pattern of one of the antenna grouping patterns by using grouped codebook vectors for the antenna grouping patterns; and selecting the selected antenna grouping pattern And feeding back a pattern index indicating a and a codebook index indicating a grouped codebook vector corresponding to the selected antenna grouping pattern to the base station.

Method according to an embodiment of the present invention; An antenna grouping method of a multiple input / output antenna system, the method comprising: receiving a pattern index indicating one of a plurality of antenna grouping patterns, a codebook index indicating a grouped codebook vector from a user terminal, and receiving the grouped codebook vector. Determining a codebook vector for the plurality of antennas by expanding according to the antenna grouping pattern indicated by the pattern index, constructing a beamforming matrix using the determined codebook vector, and forming the beamforming matrix Precoding the data using the data and transmitting the precoded data to the user terminal.

Apparatus according to an embodiment of the present invention; A terminal device supporting antenna grouping in a multiple input / output antenna system, comprising: a channel measuring unit measuring a channel vector of a plurality of antennas of a base station, and grouping channel coefficients of the channel vector according to a plurality of antenna grouping patterns A control unit for determining grouped codebook vectors corresponding to the grouped channel coefficients and selecting one antenna grouping pattern of the antenna grouping patterns using the grouped codebook vectors for the antenna grouping patterns; And a transmitter for feeding back a pattern index indicating an antenna grouping pattern and a codebook index indicating a grouped codebook vector corresponding to the selected antenna grouping pattern to the base station.

An apparatus according to an embodiment of the present invention is a base station apparatus supporting antenna grouping in a multiple input / output antenna system, comprising: a pattern index indicating one of a plurality of antenna grouping patterns and a codebook index indicating a grouped codebook vector And a receiver for receiving from the user terminal, and extending the grouped codebook vector according to the antenna grouping pattern indicated by the pattern index to determine codebook vectors for the plurality of antennas, and using the determined codebook vector to beam And a transmitter configured to preform data using the beamforming matrix, and to transmit data precoded to the user terminal.

1 illustrates a schematic structure of a large capacity multiple input / output communication (FD-MIMO) system including beamforming.

2 is a block diagram illustrating a structure of a beamforming system based on antenna grouping according to an embodiment of the present invention.

3 is a diagram illustrating an antenna group pattern extension procedure according to an embodiment of the present invention.

4 is a block diagram showing the configuration of a pattern selection unit 228 according to an embodiment of the present invention.

5 illustrates antenna grouping patterns according to an embodiment of the present invention.

6 is a flowchart illustrating an operation of a user terminal according to an exemplary embodiment of the present invention.

7 is a flowchart illustrating a procedure of determining a pattern index in a user terminal according to an embodiment of the present invention.

8 is a flowchart illustrating the operation of a base station according to an embodiment of the present invention.

9 illustrates an antenna array according to an embodiment of the present invention.

10, 11, 12 and 13 are graphs showing the total capacity performance of antenna grouping according to embodiments of the present invention.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. In describing the present invention, when it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be omitted. Terms to be described later are terms defined in consideration of functions in the present invention, and may be changed according to intentions or customs of users or operators. Therefore, the definition should be made based on the contents throughout the specification.

Referring to FIG. 1, the transmitter 110 has an antenna array 120 composed of N _t antennas, and generates signals for the

K receiver devices

130, 132, and 134 to transmit through the antennas 120. It is provided with a beam forming unit 115 that can be. Herein, a case in which each of the

receiving devices

130, 132, and 134 has one antenna is illustrated, but the

receiving devices

130, 132, and 134 may include an antenna array including one or more antennas, and the antennas of the transmitting end 110 ( A signal transmitted from 120 is received through the channel 100.

Each

receiver

130, 132, 134 performs channel vector quantization to select a codebook index that maximizes SINR. In general, the received signal of the k-th receiving device may be represented by Equation 1 below.

Equation 1

Where h _k denotes a channel vector from the antenna array 120 of the transmitter 110 to the k-th receiver, x denotes a transmission signal vector having an average power limit of ρ, and z _k denotes k Gaussian noise in the first receiver. Here, w _i is a beamforming vector applied by the transmitting end 110 for the i-th receiving device, and s _i represents a data symbol to be transmitted to the i-th receiving device.

When linear beamforming is applied at the transmitter 110, the received signal at the k-th receiver may be represented again as in Equation 2 below.

Equation 2

When the interference by other receiving devices other than the k-th receiving device, that is, the j-th receiving device, is regarded as noise, the sum rate of the receiving devices may be expressed by Equation 3 below.

Equation 3

Here, p denotes a sum transmission power predetermined at the transmitter, and the transmitter may uniformly or non-uniformly distribute the power to the receiving devices within the predetermined sum transmission power.

In the finite channel feedback system, the receiving apparatus calculates channel direction information (CDI) as shown in Equation 4 below and quantizes the calculated channel direction information.

Equation 4

The receiving device is a codebook according to a predetermined number of feedback bits B.

The codebook index j for maximizing the received signal-to-noise ratio is selected based on the channel direction information. Here, the codebook is a set containing 2 ^B N _t × 1 codebook vectors c _i . That is, the index j corresponding to the codebook vector c _j most similar to the original channel among 1 to 2 ^B is determined. Equation 5 below shows an example of a formula for determining the codebook index j.

Equation 5

The selected codebook index is composed of B bits and is fed back from each receiving apparatus to the transmitting end 110.

The transmitter 110 receives a codebook index from the

K receivers

130, 132, and 134, and configures a composite channel matrix as shown in Equation 6 below.

Equation 6

That is, the transmitter 110 determines the codebook vector c _j from the codebook C given in advance based on the codebook index j fed back from the k-th receiving device, and selects the codebook vector c _j .

Map to. That is, the channel state information of the kth receiver is

Becomes The channel matrix information is obtained based on the K codebook indices fed back for the K receivers.

It is configured by connecting.

The zero forcing (ZF) beamforming matrix obtained with the composite channel matrix may be expressed as Equation 7 below.

Equation 7

The beamforming vector for the k-th receiving device may be obtained as shown in Equation 8 by normalizing the k-th column vector of the zero forcing beamforming matrix to satisfy the power limitation of the transmission signal.

Equation 8

In a FD-MIMO system supporting beamforming, downlink performance can be increased through feedback of channel state information and proper transmission beamforming using the fed back information. The amount of channel state information fed back increases in proportion to the number of antennas of the transmitting end. In an actual system, the amount of channel state information fed back is limited.

In particular, in a multi-user MIMO system, if beamforming is not performed correctly, inter-user interference is not completely eliminated and is much more affected by a finite number of feedback bits than a single-user MIMO system. For single-user MIMO systems, the accuracy of channel state information affects performance but does not change the multiplexing gain. On the contrary, since the multi-user MIMO system no longer improves performance due to interference under finite feedback conditions, the accuracy of the channel state information directly affects the multiplexing gain of the multi-user MIMO downlink.

In order to obtain a total capacity similar to that in which perfect channel state information is fed back, the amount of feedback for quantizing the channel should be set in proportion to the number of antennas and the signal to noise ratio of the transmitter. As an example, the number of feedback bits required when using the random vector quantization technique may be expressed as in Equation 9 below.

Equation 9

Where ρ _dB is the sum of the predetermined transmit powers in dB at the transmitter (ρ _dB = 10log ₁₀ ρ). In a large-scale MIMO system, since the transmitter operates a large number of antennas, it is impossible to simply extend the number of feedback bits. In the following embodiment, a feedback scheme capable of reducing a feedback load while minimizing performance degradation is proposed.

2 is a block diagram illustrating a structure of a beamforming system based on antenna grouping according to an embodiment of the present invention. Although only one receiving device 220 is shown here, it will be apparent that there may be one or more receiving devices that operate similarly in an actual communication environment.

Referring to FIG. 2, the transmitter 210 includes an antenna group pattern extension unit 212 that receives a data symbol s to be transmitted, a beamformer 214 having a beamforming matrix W, and N _t antennas. It is configured to include a configured antenna array 216. Although the antenna group pattern extension unit 212 and the beamforming unit 214 are shown as separate entities, they may be implemented as one or more controllers or processors, depending on the implementation.

The receiving device 220 includes one or more receiving antennas 232, a channel measuring unit 222, an antenna grouping unit 224, a vector quantization and group pattern extension unit 226, and a pattern selecting unit 228. ) And a feedback transmitter 230. Although the antenna grouping unit 224, the vector quantization and group pattern extension unit 226, and the pattern selection unit 228 are illustrated as separate entities, they may be implemented as one or more controllers or processors, depending on the implementation. have.

The antenna group pattern extension unit 212 at the transmitting end 210 groups antennas grouped according to the antenna grouping pattern selected by the receiving device 220 among the predetermined antenna grouping patterns based on a pattern index fed back from each receiving device. Determine channel state information corresponding to the channel information, that is, the channel vector h. The beamforming unit 214 determines a beamforming matrix W based on the determined channel vector, precodes the input data symbol stream s according to the beamforming matrix W, and transmits the same through the plurality of antennas 216. do.

The channel measuring unit 222 of the receiving device 220 is a channel vector for the antennas 216 of the transmitting end 210 based on the signal received by the receiving antenna 232.

The antenna grouping mapping unit 224 maps the channel vector to antenna groups according to predetermined antenna grouping patterns, thereby channeling the channel vectors grouped by antenna grouping pattern.

Calculate The grouped channel vector includes channel coefficients corresponding to the antenna groups, each channel coefficient being calculated using the channel coefficients measured for the antennas belonging to one antenna group. The vector quantization and group pattern extension unit 226 quantizes the grouped channel vectors to group the codebook vector for each antenna group according to the antenna grouping pattern.

And expand the grouped codebook indices according to the corresponding antenna grouping pattern to expand the codebook vector for each antenna 216.

Calculate Here, the codebook coefficients of the grouped codebook vector are determined from channel coefficients corresponding to the antenna groups, and the codebook coefficients of the demapped codebook vector are calculated by extending the grouped codebook coefficients belonging to the same antenna group according to the antenna group size. .

The pattern selector 228 obtains extended codebook vectors corresponding to the plurality of antenna grouping patterns, and determines one antenna grouping pattern that maximizes a received signal-to-noise ratio using the original channel vector and the extended codebook vectors. The pattern index indicating the determined antenna grouping pattern is transmitted to the vector quantization and group pattern extension unit 226.

The vector quantization and group pattern extension unit 226 transmits a pattern index for identifying the determined antenna grouping pattern and a codebook index indicating a grouped codebook vector corresponding to the determined antenna grouping pattern to the feedback transmitter 230. The feedback transmitter 230 configures the pattern index and the codebook index into a message or a packet according to a predetermined signaling method, and feeds back the transmitter 210.

As described above, the system according to the embodiment of the present invention groups the antennas according to a predetermined antenna grouping pattern, quantizes the grouped channel vectors, expands according to the antenna grouping pattern, and maximizes the received signal-to-noise ratio. Select the indexes at the same time.

3 is a diagram illustrating an antenna group pattern extension procedure according to an embodiment of the present invention. Herein, an extension of one specific antenna grouping pattern is illustrated. In the following description, the index k of the receiving device will be omitted.

Referring to FIG. 3, the receiving apparatus performs antenna grouping 315, that is, antenna group mapping on the channel vector 310, using a specific antenna grouping pattern known to both the transmitting end and the receiving apparatus. In the illustrated example, the channel coefficients measured for the eight antennas of the transmitting end are h ₁ , h ₂ ,..., H ₈ , and a specific antenna grouping pattern groups each of two adjacent antennas. That is,

antennas

1 and 2 become antenna group 1, antennas 3 and 4 become antenna group 2, antennas 5 and 6 become antenna group 3, and antennas 7, 8 become antenna group 4. Antenna group mapping 315 generates grouped channel coefficients g ₁ , g ₂ , g ₃ , g ₄ corresponding to four antenna groups, respectively.

The receiving device quantizes the grouped channel vector 325 to determine grouped codebook coefficients v ₁ , v ₂ , v ₃ , v ₄ 330 corresponding to four antenna groups. In particular, the codebook vector v _i is selected from a codebook given in advance using the grouped channel vector g _i . In one embodiment, the codebook is configured to be common to all antenna grouping patterns. In another embodiment, the codebook may be configured for each antenna grouping pattern or may be configured for each antenna group size or antenna group number.

The grouped channel vectors are extended codebook coefficients corresponding to eight antennas through antenna group demapping 335 according to the corresponding antenna grouping pattern v ₁ , v ₁ , v ₂ , v ₂ , v ₃ , v ₃ is extended to an extended codebook vector 340 comprising v ₄ , v ₄ .

The receiving apparatus obtains extended codebook vectors corresponding to the plurality of antenna grouping patterns, respectively, and obtains a difference from coefficients h ₁ , h ₂ ,..., H ₈ of the original channel vector among the extended codebook vectors. An extended codebook vector to minimize and an antenna grouping pattern corresponding thereto are determined, and a pattern index indicating the determined antenna grouping pattern is determined to be fed back.

Hereinafter, a procedure of determining an antenna group pattern extension and a pattern index performed in a receiving apparatus will be described in more detail.

Equation 10 below represents a grouped channel vector generated through the antenna group mapping 315.

Equation 10

Where g ⁱ is Is the vector grouped by the i th antenna grouping pattern, and G ⁱ is the i th antenna grouping pattern

And group mapping matrix, _P N is the number of antenna grouping pattern.

The group mapping matrix G ⁱ determines a representative value of channel coefficients to be grouped by antenna group. Representative values may be determined by linear combining and averaging. In the example of FIG. 3, the antenna group mapping by averaging is illustrated. For example, g ₁ corresponding to antenna group 1 including antennas 1 and ₂ becomes (h ₁ + h ₂ ) / 2.

The dimension of the grouped channel vector g ⁱ is N _g , which is smaller than the dimension N _t of the original channel vector h. Here, N _g may be _obtained as N _t / K, where K is the antenna group size, that is, the number of antennas included in one antenna group. In the example of FIG. 3, the antenna group size is 2, and thus the dimension of the grouped channel vectors is 4.

When the number of feedback bits required for channel vector quantization is B, the number of feedback bits in the antenna grouping technique is divided into the number of bits B _p for selecting an antenna grouping pattern and the number of bits BB _p for quantizing each grouped channel vector. Lose. That is, the feedback bit number B _{p, which} is part of the overall feedback load, is used to obtain the pattern diversification gain. B _p is the number of bits needed to quantize the pattern index, and is related to the number of antenna grouping patterns given.

The feedback amount per channel entry of the antenna grouping technique and the conventional beamforming technique according to the embodiment of the present invention is represented by (BB _p ) / N _g and B / N _t , respectively. The feedback amount per channel entry in the antenna grouping scheme is larger than the feedback amount per channel entry in the conventional beamforming scheme when the number of bits B _p for the selection of the antenna grouping pattern satisfies the following Equation (11).

Equation 11

That is, B _p may be determined to satisfy the condition of Equation 11 above.

In general, the transmit antennas of a large-capacity multiple input / output communication system have a high correlation, and among them, the correlation between adjacent antennas is higher. Therefore, B _p is relatively smaller than B because adjacent antennas are mainly grouped, and the feedback amount per channel entry of the antenna grouping technique is increased, thereby increasing the quantization accuracy of the grouped channel vectors.

As described above, the reception apparatus quantizes the grouped channel vector to determine the grouped codebook vector, and extends the antenna grouping pattern. The receiving device then selects a pattern index that maximizes the received signal-to-interference noise ratio based on the grouped codebook vector for each antenna grouping pattern.

A formula for selecting a pattern index and a codebook index at the receiving device may be expressed as Equation 12 below.

Equation 12

Here p ^* means a pattern index and c ^* means a codebook index. In addition, E is an extension matrix, and there is an inverse relationship (E ⁱ = (G ⁱ ) ^T ) with the group mapping matrix G ⁱ of Equation 10, and c is a codebook of BB _p .

Is the codebook vector of.

The procedure of selecting a pattern index will be described in detail below.

The receiving device firstly codes a code vector of the channel vector g ⁱ grouped by the i th antenna grouping pattern and a given number of feedback bits BB _p .

Then, as shown in Equation 13 below, a grouped codebook vector that minimizes quantization error with g ⁱ is determined.

Equation 13

here

Denotes channel direction information of the grouped channel vector according to the i th antenna grouping pattern. The grouped codebook vectors v ⁱ are extended as shown in Equation 14 according to the corresponding antenna grouping pattern.

Equation 14

here

Is an extended codebook vector based on the i th antenna grouping pattern, E ⁱ is an extension matrix for the i th antenna grouping pattern, and N _p is the number of antenna grouping patterns. As an example, the extension matrix may perform an operation of copying one beamforming coefficient to beamforming coefficients of the same antenna group according to the grouping characteristic.

The extended codebook vector has the same dimension as the original channel vector, so that the receiver uses the original channel vector and the extended codebook vector to maximize the signal-to-noise ratio of the receiver, as shown in Equation 15 below. Determine the pattern index p ^* .

Equation 15

Referring to FIG. 4, the pattern selector 228 includes a multiplier 410 and a comparator 420. The multiplier 410 is a channel vector h measured for the signal received from the base station and an extended codebook vector for each antenna grouping pattern.

With h as

Calculate the product of. The comparator 420 is used for h and antenna grouping patterns.

Comparing the products of with each other to determine the antenna grouping pattern corresponding to the maximum product, and outputs a pattern index p ^* identifying the antenna grouping pattern.

The receiving device transmits the determined pattern index p ^* and the corresponding codebook index c ^* to the transmitting end. Where codebook index c ^* is the codeword

Determined by the index of.

The antenna grouping technique can maintain the channel vector quantization performance while reducing the amount of feedback per channel entry because the grouping can reduce the dimension of the actual beamforming vector and obtain the diversification gain due to the antenna grouping patterns.

Referring to FIG. 5, the channel vector h _i denotes a channel coefficient corresponding to the i th antenna, and in the illustrated example, the antenna array indicates a configuration of antennas arranged in a grid shape of 4 × 4. Antenna grouping pattern 1 groups two adjacent antennas in a horizontal direction into one antenna group, and the first two antennas become one antenna group. Antenna grouping pattern 2 groups two adjacent antennas in a vertical direction into one antenna group. In the antenna grouping pattern 3, an antenna that is diagonal to the first antenna becomes one antenna group, and two adjacent antennas among the other antennas are grouped into one antenna group. The remaining antennas are grouped in a similar manner.

Referring to the antenna grouping technique according to an embodiment of the present invention. For example, given the antenna grouping pattern P ₁ = ({1,2}, {3,4}), the antenna group mapping matrix corresponding to the antenna grouping pattern is given by Equation 16 below.

Equation 16

Here, the antenna grouping pattern P ₁ = ({1,2}, {3,4}) means that the first antenna and the second antenna are grouped into one antenna group, and the third and fourth antennas are grouped into one antenna group. . The grouped vector by the antenna group mapping matrix of Equation 16 is

Appears. Vectors obtained by performing vector quantization of grouped vectors

Is expanded through Equation 17 below.

Equation 17

here

Is an extension matrix according to the antenna grouping pattern P ₁ = ({1,2}, {3,4}) and is inversely related to the antenna group mapping matrix (

).

On the other hand, when the antenna grouping pattern P ₂ = ({1,3}, {2,4}), the antenna group mapping matrix according to this is given by Equation 18 below.

Equation 18

Similar to the case of antenna grouping pattern P _1, the grouped vector obtained using antenna grouping pattern P ₂ = ({1,3}, {2,4})

Appears. The quantized vector for antenna grouping pattern P ₂ is

Antenna pattern expansion matrix

The extended codebook vector can be obtained as shown in the following equation.

In high-capacity multi-input and output communication systems, the correlation between the antennas of the transmitting end is high, so it is effective to group the adjacent antennas as much as possible. In general, the closer the antennas are physically, the higher the correlation. In addition, the beamforming vector should be configured in consideration of the phase of the base station and the user. If the amount of uplink feedback per user is B bits in a large-capacity multiple input / output communication system, in the antenna grouping scheme, B _p bits among the B bits are used to obtain diversified gains of the antenna grouping patterns, and BB _p bits are used for vector quantization. Used.

In a finite channel feedback system, B _{p is} also finite. The closer the antenna is, the higher the correlation and thus the higher the correlation coefficient. Therefore, B _p may be determined to obtain diversification gains of the antenna grouping patterns according to the correlation coefficient. If B _p is determined, the shifting property is used to shift one or two squares from adjacent antennas.

Can be grouped according to finite antenna grouping patterns.

As one embodiment, the scenario of the feedback operation aspect in the receiving device, i.e., the user terminal may appear as follows.

First, B _p and BB _p may be determined at each user terminal. Each user terminal selects a codebook index of a codebook vector closest to the original channel and a pattern index corresponding thereto through allocation of feedback resources. To this end, a maximum B _p is predetermined, and antenna grouping patterns accordingly are also known to the base station and the user terminals.

Second, the antenna group size may be determined in addition to the diversity gain of the antenna grouping patterns and resource allocation for vector quantization in the user terminal. In this case, a diversification gain for the antenna grouping patterns can be obtained, but a memory space for storing a large number of antenna grouping patterns according to the antenna group size is needed at the user terminal.

Referring to FIG. 6, in step 610, a user terminal performs channel measurement using a pilot signal transmitted from a base station to obtain channel vectors for antennas of the base station. In step 615, the user terminal performs antenna group mapping according to each of the given antenna grouping patterns, and determines the grouped codebook vector. In operation 620, the user terminal demaps the codebook vectors grouped according to the plurality of antenna grouping patterns and selects a pattern index indicating an antenna grouping pattern having optimal performance. In step 625, the user terminal determines a grouped codebook vector corresponding to the antenna grouping pattern of the selected pattern index, and feeds back the codebook index and the pattern index of the grouped codebook vector to the base station. In step 630, the user terminal demodulates the data received through the precoded signal of the base station according to the scheduling of the base station.

Referring to FIG. 7, in step 710, the user terminal sets the pattern index i to 1 and determines the grouped channel vector by performing antenna group mapping on the channel vector according to the i-th antenna grouping pattern in step 715. In step 720, the user terminal performs quantization on the grouped channel vector to select a grouped codebook vector, and determines a codebook index for identifying the grouped codebook vector. In step 725, the user terminal determines the extended codebook vector by extending the grouped codebook vector according to the i th antenna grouping pattern.

In step 730, the user terminal increases the pattern index i by 1, and then in step 735, the user terminal determines whether i is greater than the number N _p of the given antenna grouping patterns. If i is not greater than N _p , the user terminal returns to step 715 and repeats operations 715 to 735 for the next antenna grouping pattern.

In step 740, the user terminal selects the codebook vector most similar to the original channel vector and the pattern index of the corresponding antenna grouping pattern using the codebook vectors determined for all the antenna grouping patterns.

Referring to FIG. 8, in step 810, the base station transmits a pilot signal using a predetermined resource. In step 815, the base station receives feedback information including a pattern index and a codebook index from each user terminal through a predetermined resource. In step 820, the base station determines a grouped codebook vector corresponding to the codebook index, an extended codebook vector according to the antenna grouping pattern of the pattern index, and connects extended codebook vectors for a plurality of user terminals to form a beamforming matrix. Configure W. In step 825, the base station precodes data using the beamforming matrix and transmits the data to user terminals in step 830.

In one embodiment of the present invention, the antenna grouping patterns may be defined in consideration of the correlation between the antennas. Constructing the antenna grouping patterns is the same as constructing the group mapping matrix G ⁱ or inversely related group pattern extension matrix E ⁱ . One easy way to construct a pattern set is to group highly correlated antennas. In general, close antennas can be grouped because they have a high correlation. An embodiment is shown in which antennas close to FIG. 5 are grouped.

As another embodiment, a pattern set of antenna grouping patterns may be configured by using information of a transmitter correlation matrix and a subspace packing technique. Since the number of possible patterns also increases as the number of transmitting antennas increases, patterns that represent subspaces of all possible patterns need to be selected. To this end, the pattern set S may be obtained by applying the modified subspace packing scheme through the following procedure.

In step 1, when the number of transmit antennas and the number of antenna groups are given, the total number of possible patterns is defined. When the first total set of possible patterns is represented by T = {1, 2, ..., N _possible }, the total number of possible patterns is represented by Equation 19 below.

Equation 19

Where K is N _t / N _g and represents the number of antennas per antenna group. If N _possible becomes quite large, the high computational cost will limit the hardware implementation. For example, when N _t = 32 and N _g = 16, the total number of possible patterns is approximately

In this case, the number of feedback bits of B _p = 58 is required for pattern selection.

In the second step, an effective transmitter correlation matrix is defined and Frobenius norm values are calculated for each pattern candidate. An effective sender correlation matrix

If defined as, each pattern candidate

Effective Source Correlation Matrix for

Frobenius norm is calculated by Equation 20.

Equation 20

In step 3, subsets of L = 2 ^{B + 1} patterns are constructed in ascending Frobenius norm values. The subsets can be expressed as in Equation 21 below.

Equation 21

Where r ^{L + 1} represents the L + 1 th largest Frobenius norm value. In the last step, the final pattern set S is obtained by using a modified subspace packing technique. In order to apply the subspace packing scheme, a relative distance value between antenna group patterns is required. The relative distance value is calculated through the correlation matrix distance, which indicates the degree of orthogonality of the correlation matrices. When there are two matrices X and Y, the correlation matrix distance can be calculated as shown in Equation 22 below.

Equation 22

Where tr (.) Is the trace function to represent the sum of the diagonal elements. d _corr is a value representing the similarity between two matrices. The smaller the d _corr value, the more similar the two matrices and the larger the two matrices.

To find the final pattern set S

Candidate sets

Is obtained. And each candidate set

The minimum correlation matrix distance of is obtained by the following Equation 23.

Equation 23

here

Is the i th effective correlation matrix of the k th candidate set. As a result, the final pattern set S may be determined as in Equation 24 below.

Equation 24

That is, a set of maximized minimum correlation matrix distances is selected as the final pattern set S as in the subspace packing technique.

The channel vector of the k-th receiving device considering the transmitter correlation matrix is represented by Equation 25 below.

Equation 25

In Equation 25 above

Is the source correlation matrix,

Each element of is a channel vector following iid (independent and identical distribution). If a general exponential model is applied to determine the effect of the correlation coefficient on the transmitter correlation matrix R _{t, k} , the correlation matrix R _{t, k} can be expressed as Equation 26.

Equation 26

here

, α is the correlation coefficient, and θ _k is the phase of the k-th receiving device, the phase between the receiving devices is independent.

The antenna grouping scheme proposed in the embodiments of the present invention can be modified in a way of effectively allocating feedback resources. As mentioned above, the total number of possible patterns N _possible increases exponentially with the number of transmit antennas and the number of antenna groups. To control the pattern complexity, the modified embodiment divides the antenna array into a predetermined number of zones and applies the antenna grouping technique to only one or more selected zones. In this case, zone information indicating the selected one or more zones may be additionally fed back. As another example, when the selection criteria of the zone is previously promised between the transmitting device and the receiving device, the feedback of the zone information may be omitted.

9 illustrates zone separation of a two-dimensional antenna array in accordance with an embodiment of the present invention.

Referring to FIG. 9, the original channel vector h is partial vectors corresponding to N _R regions.

Divided into Where r _i is

The vector represents the channel vector representing the i th region and N _R represents the number of divided regions. And, beamforming gain value for each zone

This is calculated.

Next, the receiving apparatus assumes that the beamforming gain values for each zone are arranged in ascending order (

Apply the antenna grouping technique to a specific zone. In one embodiment, an area having a low beamforming gain value, as an example

An antenna grouping technique may be applied to the in-zone. In this case, the length N _g of the grouped channel vectors becomes as follows.

Equation 27

Where N ' _{R, g} represents the number of antenna groups in the selected zone.

An advantage of the embodiment shown in FIG. 9 is that by applying antenna grouping techniques to specific areas, the total number of possible patterns can be significantly reduced. The number of possible patterns in each zone is shown in Equation 28 below.

Equation 28

here

Represents the number of antennas in each zone. For example, given N _t = 32, N _R = 4, and N ' _{R, g} = 4, N' _possible = 105 and the number of pattern feedback bits to represent pattern candidates in this region is B ' _p. == 7 Since the N ' _R zone is selected, the final number of pattern feedback bits is B _p = N' _R B ' _p . That is, it can be seen that the feedback complexity is significantly reduced compared to other embodiments.

In addition, in the embodiment of FIG. 9, the antenna grouping technique is applied to a region having a low beamforming gain value, so that a larger number of feedback bits is allocated to a region of N _R -N ' _R. For example, when B _q is the number of feedback bits for vector quantization, the number of feedback bits B _e per antenna is

Is given by In this case, N ' _{R, g} B _e bits are allocated to the zone to which antenna grouping is applied and to the remaining zone.

The number of bits is allocated.

Because of this, more feedback resources are allocated to the areas with high beamforming gains and less feedback resources are allocated to the areas with low beamforming gains.

In FIG. 10, when N _t antennas are used (Conventional beamforming) 1010 and N _g antennas to be randomly selected (Random selection with N _g ) 1020 according to the number N _g of antenna groups, antennas In the case of applying antenna grouping patterns determined in consideration of the correlation between (Anntenna grouping with N _g ) 1030, the total capacity according to the number of bitback bits B is shown. As shown, it can be seen that the total capacitance 1030 in the case of applying the antenna grouping patterns according to the embodiment of the present invention is superior to

other cases

1010 and 1020.

In FIG. 11, when all N _t antennas are used and more feedback bits are used (Conventional beamforming (B = 32)) according to the correlation coefficient α, 1110, N _t antennas are used and the same feedback bit number is used. When using (Conventional beamforming (B = 16)) (1140), when randomly selected N _g antennas (Random selection with N _g ) (1130), when the antenna grouping patterns determined in consideration of the correlation between the antennas are applied The total dose for (Antenna grouping with N _g ) (1120) is shown. As shown, the total capacitance 1120 in the case of applying the antenna grouping patterns according to the exemplary embodiment of the present invention is superior to that in the other cases (1130 and 1140), and the performance in the case of using more feedback bits (1110) considerably. It can be seen that the proximity.

In FIG. 12, when N _t ∈ {32,64} antennas are used, the antenna grouping method according to the correlation coefficient α is not applied (Conventional beamforming) 1220 and 1240 and the antenna grouping method according to the embodiment of FIG. 9 is applied. The combined dose for the case (Antenna grouping) (1210, 1230) is shown. As shown, it can be seen that the higher the total capacitance when the antenna grouping technique according to the embodiment of the present invention is applied, the higher the correlation is. In particular, when the correlation coefficient is high, it can be seen that the case where the antenna grouping method having a small number of transmitting antennas is applied is higher than the case where the antenna grouping method having a large number of transmitting antennas is not applied.

In FIG. 13, when the correlation coefficient α ∈ {0.6, 0.9}, when the antenna grouping technique according to the feedback bit number B is not applied (Conventional beamforming) 1330 and 1340 and the antenna grouping technique according to the embodiment of FIG. For the applied group (Antenna grouping) (1310, 1320), the total capacity according to the number of bits of bitback B is shown. As shown, it can be seen that the difference between the total capacitance when the antenna grouping technique according to the embodiment of the present invention is applied and the total capacitance when different is higher as the correlation coefficient increases. In particular, it can be seen that the case where the antenna grouping technique having a small correlation coefficient is applied is higher than the case where the antenna grouping technique having a high correlation coefficient is not applied.

Embodiments of the present invention that operate as described above by reducing the amount of feedback for channel vector quantization at the receiving end of a large-capacity multiple input and output antenna communication system, by grouping the antennas using a high correlation between the antennas of the large-capacity MIMO system, By increasing the channel quantization performance, the same performance as the prior art can be obtained with a smaller number of feedback bits.

Meanwhile, in the detailed description of the present invention, specific embodiments have been described, but various modifications may be made without departing from the scope of the present invention. Therefore, the scope of the present invention should not be limited to the described embodiments, but should be determined not only by the scope of the following claims, but also by those equivalent to the scope of the claims.

Claims

In the antenna grouping method of a multiple input and output antenna system,

Measuring a channel vector for a plurality of antennas of the base station;

Grouping channel coefficients of the channel vector according to a plurality of antenna grouping patterns, and determining grouped codebook vectors corresponding to the grouped channel coefficients;

Selecting an antenna grouping pattern of one of the antenna grouping patterns by using grouped codebook vectors for the antenna grouping patterns;

And feeding back a pattern index indicating the selected antenna grouping pattern and a codebook index indicating a grouped codebook vector corresponding to the selected antenna grouping pattern to the base station.
The method of claim 1, wherein each of the grouped codebook vectors is:

And codebook vector coefficients corresponding to antenna groups according to the corresponding antenna grouping pattern, wherein each of the codebook vector coefficients is calculated as a linear combination or average of channel coefficients for the antennas of the corresponding antenna group. Grouping method.
The method of claim 1, wherein the selecting of the antenna grouping pattern comprises:

Expanding the grouped codebook vectors according to the plurality of antenna grouping patterns to determine extended codebook vectors corresponding to the antenna grouping patterns, wherein each of the extended codebook vectors corresponds to the plurality of antennas Codebook vector coefficients,

Selecting an extended codebook vector closest to the channel vector among the extended codebook vectors;

And selecting an antenna grouping pattern corresponding to the selected extended codebook vector.
In the antenna grouping method of a multiple input and output antenna system,

Receiving a pattern index indicating one of a plurality of antenna grouping patterns and a codebook index indicating a grouped codebook vector from a user terminal;

Determining the codebook vectors for the plurality of antennas by extending the grouped codebook vectors according to the antenna grouping pattern indicated by the pattern index;

Constructing a beamforming matrix using the determined codebook vector;

And precoding data using the beamforming matrix and transmitting the precoded data to the user terminal.
The method of claim 4, wherein the grouped codebook vector,

Codebook vector coefficients corresponding to antenna groups according to the antenna grouping pattern indicated by the pattern index, wherein each of the codebook vector coefficients is calculated as a linear combination or average of channel coefficients for the antennas of the corresponding antenna group. And antenna grouping method.
The method of claim 4, wherein the antenna grouping pattern indicated by the pattern index,

Determining extended codebook vectors corresponding to the antenna grouping patterns by extending grouped codebook vectors corresponding to the plurality of antenna grouping patterns according to the plurality of antenna grouping patterns, wherein the extended codebook vectors Each of which includes codebook vector coefficients corresponding to the plurality of antennas,

Selecting an extended codebook vector closest to the channel vector among the extended codebook vectors;

And selecting the antenna grouping pattern corresponding to the selected extended codebook vector.
The method of claim 1 or 6, wherein the selecting of the extended codebook vector comprises:

Calculating products of the extended codebook vectors and the channel vector;

And comparing the products with each other to select an extended codebook vector corresponding to the maximum product.
The method of claim 1, wherein the plurality of antenna grouping patterns include:

And a plurality of antennas having a relatively high correlation with each other are grouped into the same antenna group.
The method of claim 1 or 4, wherein the number of bits of the pattern index is

And at least one of a total number of bits for feedback, a number of the plurality of antennas, and a number of antenna groups of each antenna grouping pattern.
The method of claim 1 or 4, wherein the number of antenna grouping patterns is:

And the number of bits of the pattern index, the number of the plurality of antennas, and the number of antenna groups of each antenna grouping pattern.
The method of claim 1 or 4, wherein each antenna grouping pattern,

Configured to group adjacent antennas in one of the plurality of antennas in any one of a horizontal direction, a vertical direction, and a diagonal direction into one antenna group, or

Is configured to group a plurality of antennas to which the transmit end correlation matrix and the subspace packing scheme are applied to the same antenna group, or

The plurality of antennas are divided into a predetermined number of zones, and configured to group antennas included in one or more selected ones of the zones into one antenna group,

The antennas are divided into a predetermined number of zones, and the antennas included in one or more selected zones whose beamforming gain values are below a predetermined reference value are configured to group into one antenna group. Antenna grouping method.
The method of claim 11,

And information on zone information indicating the selected one or more zones is transmitted from the user terminal.
A terminal device supporting antenna grouping in a multiple input / output antenna system,

A channel measuring unit measuring a channel vector of the plurality of antennas of the base station;

Group channel coefficients of the channel vector according to a plurality of antenna grouping patterns to determine grouped codebook vectors corresponding to the grouped channel coefficients, and use the grouped codebook vectors for the antenna grouping patterns to perform the antenna A controller for selecting one antenna grouping pattern among the grouping patterns;

And a transmitter for feeding back a pattern index indicating the selected antenna grouping pattern and a codebook index indicating a grouped codebook vector corresponding to the selected antenna grouping pattern to the base station.
A base station apparatus supporting antenna grouping in a multiple input / output antenna system,

A receiver configured to receive a pattern index indicating one of a plurality of antenna grouping patterns and a codebook index indicating a grouped codebook vector from a user terminal;

A control unit configured to extend the grouped codebook vectors according to the antenna grouping pattern indicated by the pattern index to determine codebook vectors for the plurality of antennas, and to configure a beamforming matrix using the determined codebook vectors;

And a transmitter configured to precode data using the beamforming matrix and to transmit the precoded data to the user terminal.