KR20110025058A - Codebook for multiple-user multiple input multiple output communication and communication device of using the codebook - Google Patents

Abstract

PURPOSE: A code book and a communication device using the same are provided to determine beam forming vectors by using a beam forming codebook in which a transmitter copes with each of the receivers. CONSTITUTION: A memory stores a first channel codebook of a first receiver(220) and a second channel codebook of a second receiver. A codeword included in the first channel codebook is received from the first receiver. A first beam forming vector for the first receiver and a second beam forming vector for the second receiver are determined based on a first code word and a second code word.

Description

CODEBOOK FOR MULTIPLE-USER MULTIPLE INPUT MULTIPLE OUTPUT COMMUNICATION AND COMMUNICATION DEVICE OF USING THE CODEBOOK}

The following embodiments relate to a codebook for multi-user multi-input / output communication and a communication device using the codebook.

Recently, various studies have been conducted to provide various multimedia services including voice services in a wireless communication environment and to support high quality and high speed data transmission. In particular, technology for a multi input multi output (MIMO) communication system using a plurality of channels in a spatial domain is rapidly developing.

In a MIMO communication system, base stations and terminals use codebooks to properly cope with the channel environment. The specific space may be quantized into a plurality of codewords according to a predetermined criterion, and the generated plurality of codewords may be stored in a base station and terminals as a codebook. This codebook can be used by base stations and terminals to share channel information, and can also be used to determine beamforming vectors.

A communication method of a transmitter for multi-user multi-input / output communication according to an embodiment of the present invention includes maintaining a memory in which a first channel codebook corresponding to a first receiver and a second channel codebook corresponding to a second receiver are stored; Information about a first preferred codeword among codewords included in the first channel codebook from the first receiver and information about a second preferred codeword among codewords included in the second channel codebook from the second receiver. Receiving information; And determining a first beamforming vector for the first receiver and a second beamforming vector for the second receiver based on the first preferred codeword and the second preferred codeword. Here, the first channel codebook and the second channel codebook are different from each other.

The information about the first preferred codeword is related to the covariance of the first channel between the transmitter and the first receiver, and the information about the second preferred codeword is between the transmitter and the second receiver. It may be related to the covariance of the second channel.

Codewords included in the first channel codebook are defined based on vectors in a grasmanian line packing codebook of M _t M _r1 x 1 dimension, and code words included in the second channel codebook are M _t M _r2 x 1 M _r1 is the number of receive antennas of the first receiver, M _r2 is the number of receive antennas of the second receiver, and M _t is the number of receive antennas of the transmitter. It may be the number of transmit antennas.

Vectors in the Grasmanian line packing codebook of M _t M _r1 x 1 dimension are converted into m _t x M _r1 dimension matrices, and each of the codewords included in the first channel codebook is M _t x M _r1 dimension Vectors defined in each of the m _t M _r2 x 1-dimensional Grasmanian line packing codebooks are converted into m _t x M _r2 -dimensional matrices, and included in the second channel codebook. Each of these may be defined by each of the m _t x M _r2 dimensions of the matrix.

Determining a first beamforming vector for the first receiver and a second beamforming vector for the second receiver may include determining the first beamforming vector using a first beamforming codebook for the first receiver. And determining the second beamforming vector by using a second beamforming codebook for the second receiver different from the first beamforming codebook.

A minimum distance between first beamforming codewords included in the first beamforming codebook and a second distance between second beamforming codewords included in the second beamforming codebook may be the same.

Determining a first beamforming vector for the first receiver and a second beamforming vector for the second receiver is based on the assumption that each of the first and second receivers is a zero forcing linear receiver. Calculating a sum rate for a receiver and the second receiver; And selecting a first beamforming vector for the first receiver and a second beamforming vector for the second receiver based on the sum of rates for the first receiver and the second receiver.

Each of the first beamforming codebook and the second beamforming codebook may be designed such that a correlation between the first beamforming codewords and the second beamforming codewords is minimized.

The first beamforming codebook is generated by inner product of each of the codewords included in a first unitary rotation matrix and a predefined basic codebook, and the second beamforming codebook is the first unitary rotation matrix. And a second unitary rotation matrix different from and each of the codewords included in the predefined basic codebook.

The method of communication of the transmitter may further include providing information about the first unitary rotation matrix to the first receiver and providing information about the second unitary rotation matrix to the second receiver. .

The method of communication of the transmitter is based on the dot product between the first beamforming vector and the hermitian of the first beamforming vector, and the interference statistics based on the dot product between the second beamforming vector and the hermitian of the second beamforming vector. calculating (statistics); And providing information regarding the interference statistics to the first receiver and the second receiver.

A communication method of a second receiver for a multi-user multi-input / output communication system including a transmitter, a first receiver, and a second receiver according to an embodiment of the present invention may include a second channel distinguished from a first channel codebook corresponding to the first receiver. Maintaining a memory in which a second channel codebook corresponding to a receiver is stored; Calculating a covariance of a second channel between the transmitter and the second receiver; Selecting a second preferred codeword among codewords included in the second channel codebook based on the covariance of the second channel; Providing information regarding the second preferred codeword to the transmitter; Identifying the second beamforming vector when the transmitter uses a second beamforming vector based on the second preferred codeword; And calculating a reception combining vector according to the second beamforming vector and the second channel.

The method of communication of the receiver further comprises calculating an effective channel between the transmitter and the second receiver using a dedicated pilot for the second receiver transmitted from the transmitter, wherein the second beamforming Identifying the vector may be determining the second beamforming vector based on the effective channel.

In the communication method of the receiver, the transmitter determines the first beamforming vector using a first beamforming codebook for the first receiver, and a second beamforming for the second receiver that is different from the first beamforming codebook. The second beamforming vector is determined using a codebook, and the second beamforming codebook is generated by inner product of each of the codewords included in a unitary rotation matrix and a predefined basic codebook. And receiving information about a unitary rotation matrix from the transmitter, and determining the second beamforming vector comprises: identifying the second beamforming vector using the information about the unitary rotation matrix. It may be a step.

The method of communication of the receiver includes interference statistics from the transmitter, wherein the interference statistics are a dot product between the first beamforming vector for the first receiver and the hermitian of the first beamforming vector and the second beamforming vector. And calculating information based on the dot product between the second beamforming vector and the hermian of the second beamforming vector, wherein calculating the reception combining vector comprises: the reception combiner based on the interference statistics. Calculating an inning vector.

1 is a diagram illustrating a closed loop multi-user multi-input / output communication system according to an embodiment of the present invention.
2 is a diagram illustrating a transmitter provided with M _t transmit antennas and a receiver provided with M _r receive antennas.
FIG. 3 is a diagram illustrating a vector in a Grasmanian line packing codebook of M _t M _r x 1 and a matrix of M _t x M _r dimensions generated based on the vector.
4 illustrates an example in which codewords included in one codebook and codewords included in each of the other codebooks have little correlation.
5 is a flowchart illustrating a communication method of a transmitter and a receiver i in a multi-user multi-input / output communication system according to an exemplary embodiment of the present invention.
6 is a block diagram showing a transmitter according to an embodiment of the present invention.
7 is a block diagram illustrating a receiver according to an embodiment of the present invention.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram illustrating a closed loop multi-user multi-input / output communication system according to an embodiment of the present invention.

Referring to FIG. 1, a closed loop multi-user multi-input / output communication system includes a base station 110 and terminals 120, 130, and 140 of users. Although embodiments of the present invention may be applied to the uplink, the following description will mainly describe a downlink case for convenience of description. 'Closed loop' means that the terminals 120, 130, and 140 feed back feedback information including channel information to the base station 110, and the base station 110 performs beamforming using the feedback information. do.

The base station 110 may be provided with a plurality of antennas, and one or more antennas may be installed in each of the terminals 120, 130, and 140 of the users. In addition, channels are formed between the base station 110 and the terminals 120, 130, and 140 of the users, and a signal is transmitted / received through each of the channels.

When the channel between the base station 110 and the user k terminal is referred to as H _k , the base station 110 transmits a common pilot so that the user k terminal can estimate the channel H _k . At this time, the base station 110, information relating to the terminal of the user k determines the preferred code words of the code words included in the after estimating the channel H _k, the codebook based on the channel H _k, the preferred code words Feedback. The base station 110 determines a beamforming vector for the terminal of user k used to beamform the data streams S ₁ , S _N based on the information about the preferred codeword.

As will be described in detail below, according to the embodiments of the present invention, the base station may transmit a dedicated pilot for the user k terminal, and the user k terminal may transmit the effective channel H _k w _k using the dedicated pilot. It can be estimated. Here, w _k is a beamforming vector for the user k terminal. In addition, the base station 110 according to the embodiments of the present invention may not only use a codebook to grasp information (channel information) regarding the channel H _k , but also use a codebook in which candidates of the beamforming vector are stored. . When a codebook in which candidates of the beamforming vector are stored is used, the base station 110 may select any one of the candidates of the beamforming vector as the beamforming vector.

2 is a diagram illustrating a transmitter provided with M _t transmit antennas and a receiver provided with M _r receive antennas.

Referring to FIG. 2, transmitter 1 210 includes M _t transmit antennas, and each of receivers 220, 230, and 240 includes M _r receive antennas. Here, it is assumed that the number of receivers 220, 230, and 240 is K. In addition, the transmitter 1 210 may be a base station, a repeater, etc. in the downlink, and may be a terminal in the uplink. Similarly, each of the receivers 220, 230, 240 may be a terminal in the downlink and may be a base station or a repeater in the uplink.

In the following, when M _r is greater than or equal to M _t, and M _r if M is less than _t will be described embodiments of the present invention for each.

M _r M _t Greater than or equal to

If each of the transmitters and receivers uses a channel codebook with a limited size to share channel information, the transmitter may call partial channel information (channel state information) from each of the receivers. Can be obtained). And, if it is assumed that the transmission power for each of the receivers is the same, each of the receivers is assigned a transmission power of P / K. P represents the total transmission power. At this time, the received signal vector in a receiver i y _i can be expressed as Equation (1).

[Equation 1]

이고, C^{Mr x Mt}는 M_r x M_t 차원의 컴플렉스 공간을 나타낸다.

는 수신기 i를 위한 빔포밍 벡터를 나타내며,

는 규준화된 노이즈 벡터를 나타낸다.

의 엔트리들은 표준 정규 분포

에 따라 identically independent distributed하다. 그리고,

는 전송 심볼을 나타내며, 전송 심볼의 에너지는

이며,

는 신호 대 잡음 비(SNR)이다.here,

And C ^{Mr x Mt} represents a complex space of M _r x M _t dimension.

Denotes a beamforming vector for receiver i,

Denotes a normalized noise vector.

Entries in the standard normal distribution

According to identically independent distributed. And,

Represents a transmission symbol, and the energy of the transmission symbol is

,

Is the signal-to-noise ratio (SNR).

를 계산할 수 있다. 여기서, H_i ^*는 H_i의 허미시안을 나타내며,

는 H_i의 프로비니어스 놈의 제곱(square)을 의미한다.Receiver i may estimate the channel H _i between receiver i and the transmitter using the common pilot transmitted from the transmitter. When modeling the received signal vector y _i at the receiver i as shown in Equation 1, the receiver i may calculate H _i ^* H _i , which is a covariance of the channel H _i . As well as the receiver i

Can be calculated. Here, H _i ^* represents the hermitian of H _i ,

Is the square of the Provenom norm of H _i .

를 이용하여

에 대응하는 선호 코드워드를 선택할 수 있다. 여기서,

는 하기 수학식 2와 같이 N1 개의 코드워드들을 포함한다.
Receiver i is a channel codebook of size N1

Using

A preferred codeword corresponding to may be selected. here,

Includes N1 codewords as shown in Equation 2 below.

[Equation 2]

에 포함되는 코드워드들 중 선호 코드워드는 하기 수학식 3을 통하여 선택될 수 있다.
here,

Among the codewords included in the preferred codeword may be selected through Equation 3 below.

&Quot; (3) "

이고, 수신기들 각각은 서로 다른 채널 코드북

를 갖는다. 예를 들어, 수신기 1은 채널 코드북 g⁽¹⁾, 수신기 2는 채널 코드북 g⁽²⁾를 갖는다. 그리고, 코드워드들 각각

는 하기 수학식 4를 통해 알 수 있는 바와 같이, M_tM_r x 1 차원의 그래스매니안 라인 팩킹 코드북에 있는 벡터들을 기초로 생성될 수 있다.
here,

And each of the receivers has a different channel codebook

Has For example, receiver 1 has channel codebook g ⁽¹⁾ and receiver 2 has channel codebook g ⁽²⁾ . And each of the codewords

As can be seen through Equation 4, can be generated based on the vectors in the grasmanian line packing codebook of M _t M _r x 1 dimension.

&Quot; (4) "

는 그래스매니안 라인 팩킹 코드북

에 있는 벡터이다. 함수

는 벡터

를

차원의 매트릭스로 변환한다. 특히, 함수

는 벡터

의 원소들을 컬럼별로(columnwisely) 쌓음(stacking)으로써

차원의 매트릭스를 생성할 수 있다.

는 A x B 차원의 오쏘노말(orthonormal) 매트릭스 스페이스를 나타낸다.
here,

Grasmania Line Packing Codebook

Is a vector. function

Vector

To

Convert to a matrix of dimensions. Specifically, the function

Vector

By stacking the elements of column by column

You can create a matrix of dimensions.

Denotes an orthonormal matrix space of A x B dimension.

FIG. 3 is a diagram illustrating a vector in a Grasmanian line packing codebook of M _t M _r x 1 and a matrix of M _t x M _r dimensions generated based on the vector.

Referring to FIG. 3, the dimension of the vector 310 is M _t M _r x 1 and the dimension of the matrix 320 is M _t x M _r . Assume that M _t is 2 and M _r is 4.

의 원소이며, 매트릭스(320)은 벡터(310)의 원소들을 컬럼별로(columnwisely) 쌓음(stacking)으로써 생성된다.
Vector 310 is a Grasmanian line packing codebook in M _t M _r x 1

Element 320, and the matrix 320 is generated by stacking the elements of the vector 310 columnwise.

를 추정하고, 복조 시간에서 전용 파일럿을 이용하여 에러 없이

의 실효 채널

를 추정한다고 가정한다. 여기서,

는 수신기 i를 위해 송신기에 의해 사용되는 빔포밍 벡터를 나타내며,

는 N2 사이즈의 빔포밍 코드북으로부터 선택된다. 이 때, 제로포싱 선형 컴바이너는 하기 수학식 5와 같이 나타낼 수 있다.
Suppose receivers use a zero forcing linear combining technique, and receiver I uses a common pilot at measurement time without error

Is estimated and error-free using a dedicated pilot at the demodulation time.

Effective channel of

Suppose we estimate. here,

Denotes the beamforming vector used by the transmitter for receiver i,

Is selected from the N2 size beamforming codebook. In this case, the zero-forcing linear combiner may be represented by Equation 5 below.

[Equation 5]

라고 가정한다. 여기서, n_i는 수신기 i를 위한 빔포밍 코드북에서 수신기 i를 위한 빔포밍 벡터의 인덱스를 나타낸다. 전용 파일럿이 사용되는 경우, 수신기 i는 실효 채널을 추정함으로써,

를 파악할 수 있지만, 전용 파일럿이 사용되지 않는 경우, 송신기는 다운링크에서

를 수신기 i로 알려줄 수 있다.Here, the beamforming vector for receiver i can be distinguished from other receivers so that the beamforming vector for receiver i can be distinguished.

Assume that Where n _i represents the index of the beamforming vector for receiver i in the beamforming codebook for receiver i. If a dedicated pilot is used, receiver i estimates the effective channel,

, But if a dedicated pilot is not used, the transmitter

Can be informed to receiver i.

라고 한다면, 수신기들 각각은

를 송신기로 피드백한다. 여기서,

은 채널 코드북

에서 l 번째 코드워드를 의미한다. 송신기가 적절한 스케쥴링을 수행할 수 있도록 송신기들 각각은

를 피드백한다.Preferred codeword of each receiver

, Each of the receivers

Feeds back to the transmitter. here,

Silver channel codebook

Means the lth codeword in. Each of the transmitters must be configured so that the transmitter can perform proper scheduling.

Feedback.

"로 나타낼 수 있다. 그리고, 수신기들 각각에 대응하는 빔포밍 코드북은 그래스매니안 라인 팩킹 코드북일 수 있다. 즉, 수신기 1 내지 수신기 K의 빔포밍 코드북은 하기 수학식 6과 같이 나타낼 수 있다.
The transmitter generates K beamforming vectors for the receivers based on the preferred codeword of each of the receivers. In this case, each of the receivers may have a different beamforming codebook, and the transmitter selects K beamforming vectors from the beamforming codebook corresponding to each of the receivers. The beamforming vectors selected by the transmitter are "

In addition, the beamforming codebook corresponding to each of the receivers may be a Grassmanian line packing codebook. That is, the beamforming codebooks of the receivers 1 to K may be represented by Equation 6 below.

&Quot; (6) "

는 수신기 i를 위한 빔포밍 코드북을 나타내며, N2는

의 사이즈이다. 이러한 빔포밍 코드북들은 수신기들마다 서로 다르게 디자인되지만, 거의 그래스매니안 매니폴드(manifold) 내에서 동일한 '최소 거리(minimum distance)' 특성을 가질 수 있다. '최소 거리'는

에 포함된 코드워드들 사이의 거리들 중에서 가장 작은 거리를 의미한다.here,

Denotes a beamforming codebook for receiver i, where N2

Is the size of. These beamforming codebooks are designed differently for each receiver, but may have the same 'minimum distance' characteristics in a nearly grasmanian manifold. "Minimum distance"

It means the smallest distance among the distance between the codewords included in.

When each of the receivers performs zero forcing linear combining, the transmitter may select K beamforming vectors using Equation 7 below.

[Equation 7]

Referring to Equation 7, the transmitter selects K beamforming vectors such that the sum rate to the receivers is maximized.

The sum of average transmission rates of the transmitters may be expressed by Equation 8 below.

[Equation 8]

In a broadcast channel, the sum of rate is limited due to the high interferences between receivers in high SNR region. In this high SNR region, it may be more efficient to suppress the interference power than to simply increase the desired signal power of each of the receivers. An efficient beamforming codebook design method requires increasing the sum of rate by reducing the interference power between receivers. As one possible method for efficient beamforming codebook design, an interference limited beamforming codebook is described below.

A rate agreement formula with complete channel information may be represented by Equation 9 below.

[Equation 9]

The sum of transmission rates in Equation 9 is lower bounded by Equation 10 below.

[Equation 10]

에 대하여

가 취해진다.

에 대하여

가 취해지면, 하기 수학식 11을 나타낼 수 있다.
Assuming interference is limited,

about

Is taken.

about

Is taken, the following equation (11) can be represented.

[Equation 11]

Referring to Equation 11, the codebook design criterion in the high SNR region means that the sum of data rates is dominant by a separation characteristic between codebooks of other receivers in a single sphere. . In other words, the sum of rates in the high SNR region is greatly influenced by the separation characteristic between codebooks of other receivers.

Each of the K receivers has one preferred codeword. All possible permutations of the index of the preferred codewords may be defined as in Equation 12 below.

[Equation 12]

는 K 튜플(tuple) 벡터이며, K 튜플 벡터의 엘리먼트들은 선호 코드워드들의 인덱스들이다. 빔포밍 코드북 설계 목표 함수는 하기 수학식 13과 같이 나타낼 수 있다.
here,

Is a K tuple vector, and the elements of the K tuple vector are indices of preferred codewords. The beamforming codebook design target function may be represented by Equation 13 below.

[Equation 13]

이고, 다른 빔포밍 코드북들이

(

)인 경우,

는

의 기울어진(tilted) 버전이거나,

의 회전된 버전일 수 있다. 그리고,

및

는 단위 구 내에 위치하며, 동일한 최소 거리 특성을 최대한 갖는 것이 효율적일 수 있다.
The objective function of Equation 13 means that the beamforming codebook should be designed so that correlation between codewords of the beamforming codebook for other receivers is minimized. Intuitively given beamforming codebook

And other beamforming codebooks

(

),

Is

Is a tilted version of

It can be a rotated version of. And,

And

Is located in the unit sphere, and it may be efficient to have the same minimum distance characteristics to the maximum.

4 illustrates an example in which codewords included in one codebook and codewords included in each of the other codebooks have little correlation.

Each of the spheres 410, 420, and 430 of FIG. 4 means a space including codewords. The codewords included in the phrase 410, the codewords included in the phrase 420, and the codewords included in the phrase 430 are orthogonal and thus have zero correlation.

As shown in FIG. 4, when codewords of a codebook corresponding to a specific receiver and codewords of codebooks corresponding to other receivers are orthogonal, optimal codebooks of Equation 13 may be designed.

를 가정하는 경우, K 개의 수신기들을 위한 코드북은 하기 수학식 14와 같이 설계될 수 있다.
It can be very difficult to optimize equation (13) for any beamforming codebooks. In this case, an embodiment of the present invention may provide a systematic beamforming codebook design approach. Mt's Grasmanian Line Packing Codebook

In the following case, a codebook for K receivers may be designed as shown in Equation 14 below.

[Equation 14]

는 수신기 i를 위한 Mt 차원의 유니터리 회전 매트릭스를 의미한다. 주어진 베이스 코드북

에 대하여, 유니터리 회전 매트릭스

는 상기 수학식 13을 통하여 최적화될 수 있다.here,

Denotes the unitary rotation matrix of the Mt dimension for the receiver i. Given base codebook

Relative rotation matrix

May be optimized through Equation 13.

과 회전 매트릭스의 후보들을 포함하는

를 미리 저장할 수 있고, 수신기들이 송신기에 의해 서빙되기 위하여 셀로 접속할 때마다, 송신기는 회전 매트릭스의 Ui의 인덱스를 수신기들로 피드포워드할 수 있다. 이러한 경우, 수신기들은 회전 매트릭스 Ui를 이용하여 기본 코드북

을 회전함으로써 코드북 구조(configuration)를 쉽게 변경할 수 있다. 특히, 전용 파일럿이 사용될 수 있다면, 수신기들은 Ui 및

를 알 필요가 없다.The above systematic beamforming codebook structure may be an efficient structure. In particular, the receivers are base codebooks

And candidates for the rotation matrix

Can be stored in advance, and whenever the receivers connect to the cell to be served by the transmitter, the transmitter can feed forward the index of Ui of the rotation matrix to the receivers. In this case, the receivers use the rotation matrix Ui to code the basic codebook.

You can easily change the codebook configuration by rotating. In particular, if a dedicated pilot can be used, the receivers are

There is no need to know.

To improve overall performance, assume that a feedforward link from the transmitter to each of the receivers is available. Through this feedforward link, the transmitter can broadcast information about interference statistics to all receivers.

In this case, the interference statistics may be expressed by Equation 15 below.

[Equation 15]

를 설계하는 데에 도움을 준다. 최적의 컴바이닝 벡터

는 하기 수학식 16과 같이 나타낼 수 있다.
Assuming that an interference limited beamforming codebook is used, as shown in Equation 14, the interference statistic of Equation 15 indicates that the receiver i is an optimal combining vector.

It helps to design. Optimal Combination Vectors

Can be expressed as in Equation 16 below.

[Equation 16]

는 하기 수학식 17을 기초로 계산될 수 있다.
here,

May be calculated based on Equation 17 below.

[Equation 17]

Each of the receivers can design the optimal combiner described in Equation 16 above. In addition, based on the partial channel information fed back from each of the receivers, it is possible to obtain K beamforming vectors maximizing the sum of the rates.

Equation 18

와 같이 특이값 분해가 수행되는 경우,

는 우측 특이 매트릭스(right singular 매트릭스)이고,

는 대각 특이 값(diagonal singular value)이다.here,

If singular value decomposition is performed,

Is the right singular matrix,

Is a diagonal singular value.

(

)을 기초로, 송신기는 수신기들로 데이터 스트림을 송신할 수 있다. 이 때, 실제의 전송률 합은 하기 수학식 19와 같이 나타낼 수 있다.
Beamforming vectors

(

The transmitter may transmit the data stream to the receivers. In this case, the actual sum of transmission rates may be expressed by Equation 19 below.

[Equation 19]

In the above, the case where M _r is greater than or equal to M _t has been described. The following describes the case where M _r is smaller than M _t .

M _r This M _t Less than

When M _t is greater than M _r , the received signal vector of the receiver i may be expressed by Equation 20 below.

[Equation 20]

이고, C^{Mr x Mt}는 M_r x M_t 차원의 컴플렉스 공간을 나타낸다.

는 수신기 i를 위한 빔포밍 벡터를 나타내며,

는 규준화된 노이즈 벡터를 나타낸다.

의 엔트리들은 표준 정규 분포

에 따라 identically independent distributed하다. 그리고,

는 전송 심볼을 나타내며, 전송 심볼의 에너지는

이며,

는 신호 대 잡음 비(SNR)이다.here,

And C ^{Mr x Mt} represents a complex space of M _r x M _t dimension.

Denotes a beamforming vector for receiver i,

Denotes a normalized noise vector.

Entries in the standard normal distribution

According to identically independent distributed. And,

Represents a transmission symbol, and the energy of the transmission symbol is

,

Is the signal-to-noise ratio (SNR).

는 사이즈 N2의 빔포밍 코드북으로부터 선택된 빔포밍 벡터라고 가정한다. 이 때, 최소 자승 결합(least square combining) 기법에 따른 컴바이너는 하기 수학식 21과 같이 나타낼 수 있다.
Assume that receivers use a least square combining technique. And,

Is assumed to be a beamforming vector selected from a beamforming codebook of size N2. In this case, the combiner according to the least square combining technique may be represented by Equation 21 below.

[Equation 21]

라고 가정한다. 여기서, n_i는 수신기 i를 위한 빔포밍 코드북에서 수신기 i를 위한 빔포밍 벡터의 인덱스를 나타낸다. 전용 파일럿이 사용되는 경우, 수신기 i는 실효 채널을 추정함으로써,

를 파악할 수 있지만, 전용 파일럿이 사용되지 않는 경우, 송신기는 다운링크에서

를 수신기 i로 알려줄 수 있다.The beamforming vector for receiver i can be distinguished from other receivers so that the beamforming vector for receiver i can be distinguished.

Assume that Where n _i represents the index of the beamforming vector for receiver i in the beamforming codebook for receiver i. If a dedicated pilot is used, receiver i estimates the effective channel,

, But if a dedicated pilot is not used, the transmitter

Can be informed to receiver i.

A rate agreement formula with complete channel information may be represented by Equation 22 below.

[Equation 22]

는

의 row 서브스페이스를 나타내며, U_j는

이고,

이며,

이다. 상기 수학식 22의 전송률 합은 하기 수학식 23에 의해 하한된다(lower bounded).
here,

Is

Represents the row subspace of, where U _j is

ego,

,

to be. The sum of transmission rates in Equation 22 is lower bounded by Equation 23 below.

&Quot; (23) "

에 대하여

가 취해진다.

에 대하여

가 취해지면, 하기 수학식 24를 나타낼 수 있다.
Reconsidering interference suppression,

about

Is taken.

about

If is taken, the following equation (24) can be represented.

&Quot; (24) "

를 양자화한다.Interpreting Equation 24, the following embodiments may be proposed. That is, the receiver i uses a channel codebook of N1 size.

Quantize

In this case, the channel codebook for the receiver i may be represented by Equation 25 below.

[Equation 25]

Here, codewords included in the channel codebook for the receiver i may be generated based on Equation 26 below.

[Equation 26]

를 갖고, 각각의 채널 코드북은 그래스매니안 라인 팩킹 코드북이라고 가정한다. 수신기들 각각의 선호 코드워드를

라고 한다면, 수신기들 각각은

를 송신기로 피드백한다.

은 채널 코드북

에서 l 번째 코드워드를 의미한다. 송신기가 적절한 스케쥴링을 수행할 수 있도록 송신기들 각각은

를 피드백한다.In addition, each channel has a different channel codebook

And assume that each channel codebook is a Grasmanian line packing codebook. Preferred codeword of each receiver

, Each of the receivers

Feeds back to the transmitter.

Silver channel codebook

Means the lth codeword in. Each of the transmitters must be configured so that the transmitter can perform proper scheduling.

Feedback.

로 나타낼 수 있다. 그리고, 수신기들 각각에 대응하는 빔포밍 코드북은 그래스매니안 라인 팩킹 코드북일 수 있다. 즉, 수신기 1 내지 수신기 K의 빔포밍 코드북은 하기 수학식 27과 같이 나타낼 수 있다.
The transmitter generates K beamforming vectors for the receivers based on the preferred codeword of each of the receivers. In this case, each of the receivers may have a different beamforming codebook, and the transmitter selects K beamforming vectors from the beamforming codebook corresponding to each of the receivers. The beamforming vectors selected by the transmitter

It can be represented as. The beamforming codebook corresponding to each of the receivers may be a grassmanian line packing codebook. That is, the beamforming codebooks of the receivers 1 to K may be expressed by Equation 27 below.

[Equation 27]

은 수신기 i를 위한 빔포밍 코드북을 나타내며, N2는

의 사이즈이다. LSC를 사용하는 수신기 i의 신호 파워를 최대화하는 빔포밍 방향은 서브스페이스

위에 있어야 한다. 따라서,

는 서브스페이스

위에 있어야 하며, 송신기에서의 K 개의 빔포밍 벡터들

는

와 함께 계산되는 전송률 합을 최대화하도록 결정되어야 한다는 두 가지 제한(constrain)을 제안할 수 있다.Where the beamforming codebook

Denotes a beamforming codebook for receiver i, where N2

Is the size of. The beamforming direction for maximizing the signal power of receiver i using LSC is subspace

Should be above therefore,

Is a subspace

K beamforming vectors at the transmitter

Is

We can propose two constraints that must be determined to maximize the sum of the rates calculated with.

와 관련하여,

이 N1 개 존재하며,

는

(j=1, 2, 3, . . , N1)과 관련된다. 송신기가 수신기들 각각으로부터

를 얻은 이후에, 송신기는 K 개의 빔포밍 벡터들을 선택한다. 특히, 송신기는 하기 수학식 28을 이용하여 K 개의 빔포밍 벡터들을 선택할 수 있다.
Subspace codebook allocated to receiver i

In connection with

There are N1,

Is

(j = 1, 2, 3,..., N1). From each of the receivers

After obtaining, the transmitter selects K beamforming vectors. In particular, the transmitter may select K beamforming vectors using Equation 28 below.

[Equation 28]

는

안에 있어야 하는 제한, 채널 서브스페이스 양자화를 가지고, 아래에서 설명하는 것과 같이 코드북 디자인 제한을 얻을 수 있다.
The following describes how the beamforming codebook is designed. Equation (24) provides a design criterion for the beamforming codebook in the interference limited region.

Is

With the constraints that need to be in, channel subspace quantization, we can get the codebook design constraints as described below.

Possible permutations of indices of preferred codewords from the K receivers can be represented by Equation 29 below.

[Equation 29]

는 K 튜플 벡터이며, K 튜플 벡터의 엘리먼트들은 선호 코드워드들의 인덱스들이다. 빔포밍 코드북 설계 목표 함수는 하기 수학식 30과 같이 나타낼 수 있다.
here,

Is a K tuple vector, and the elements of the K tuple vector are indices of preferred codewords. The beamforming codebook design target function may be represented by Equation 30 below.

Equation 30

The beamforming codebook design objective function means that the beamforming codebook associated with the quantized channel subspace of receiver i should be designed such that the correlation between codewords from other receivers is minimized.

To improve overall performance, assume that a feedforward link from the transmitter to each of the receivers is available. Through this feedforward link, the transmitter can broadcast information about interference statistics to all receivers.

In this case, the interference statistics may be expressed by Equation 31 below.

Equation 31

를 설계하는 데에 도움을 준다. 최적의 컴바이닝 벡터

는 하기 수학식 32과 같이 나타낼 수 있다.
The interference statistic of Equation 31 indicates that the receiver i is the optimal combining vector.

It helps to design. Optimal Combination Vectors

Can be expressed as in Equation 32 below.

Equation 32

는 하기 수학식 33을 기초로 계산될 수 있다.
here,

May be calculated based on Equation 33 below.

[Equation 33]

The transmitter may select K beamforming vectors capable of maximizing the sum of rates based on the quantized channel matrix from each of the receivers using Equations 32 and 33. In particular, the transmitter may use Equation 34 below.

[Equation 34]

5 is a flowchart illustrating a communication method of a transmitter and a receiver i in a multi-user multi-input / output communication system according to an exemplary embodiment of the present invention.

를 추정한다.First, embodiments of the present invention will be described with reference to FIG. 5 when M _r is greater than or equal to M _t . Referring to FIG. 5, the transmitter transmits a common pilot. Each of the plurality of receivers estimates a channel, and receiver i, which is one of the plurality of receivers, is also a channel.

Estimate

를 이용하여 추정된 채널

에 대응하는 선호 코드워드

를 선택한다. 보다 구체적으로, 수신기 i는 상기 수학식 3을 이용하여

를 선택할 수 있다.Each of the receivers has a different channel codebook. Receiver i channel codebook

Estimated Channel Using

Preferred codeword for

Select. More specifically, the receiver i using the equation (3)

Can be selected.

가 선택되면, 수신기 i는

의 인덱스를 송신기로 피드백한다. 마찬가지로, 송신기에 의해 서빙되기 위하여 셀에 접속한 모든 수신기들 각각은 해당 선호 코드워드의 인덱스를 송신기로 피드백한다. 또한, 위에서 설명한 바와 같이,

도 피드백된다.

Is selected, the receiver i

Feeds back the index to the transmitter. Similarly, each of the receivers connected to the cell to be served by the transmitter feeds back the index of the preferred codeword to the transmitter. Also, as described above,

Is also fed back.

The transmitter selects K beamforming vectors based on the index fed back from each of the receivers. That is, the transmitter selects K beamforming vectors from beamforming codebooks corresponding to each of the receivers described in Equation 6 above.

의 인덱스를 수신함으로써,

를 파악할 수 있다. 뿐만 아니라, 송신기가 전용 파일럿을 사용할 수 있는 경우, 수신기들은 해당 실효 채널을 파악함으로써, 해당 빔포밍 벡터를 파악할 수 있다.If the transmitter can use the feedforward link, information about the selected beamforming vector can be feedforwarded. For example, receiver i

By receiving the index of

Can be identified. In addition, when the transmitter can use the dedicated pilot, the receiver can determine the corresponding beamforming vector by identifying the corresponding effective channel.

If the transmitter uses beamforming codebooks such as Equation 14, the transmitter may feedforward information about the rotation matrix and may feedforward information about interference statistics.

Since the above-described steps may be well applied even when M _r is smaller than M _t , detailed description thereof will be omitted.

6 is a block diagram showing a transmitter according to an embodiment of the present invention.

Referring to FIG. 6, a transmitter for multi-user multi-input / output communication includes a memory 610, a receiver 720, a processor 730, and a transmitter 740.

The memory 610 stores a channel codebook and a beamforming codebook corresponding to each of the receivers, and additionally stores candidates of a rotation matrix and a base codebook.

Under the control of the processor 630, the transmitter 640 transmits a common pilot, and after each of the receivers estimates a channel, the receiver 640 feeds back partial channel information (information on a preferred codeword) and channel quality information. do. In this case, since the channel codebooks corresponding to each of the receivers are different from each other, the transmitter should refer to the corresponding channel codebook to determine partial channel information of each of the receivers.

That is, the receiver 620 receives information about a preferred codeword from each of the receivers, and the processor 630 refers to the channel codebook of each of the receivers pre-stored in the memory 610 to obtain the preferred codeword of each of the receivers. Figure out.

Once the preferred codeword of each of the receivers is known, the processor 630 may determine K (the number of receivers) from the beamforming codebook of each of the receivers previously stored in the memory 610 based on the preferred codeword of each of the receivers. Generate beamforming vectors.

Once the beamforming vectors are determined, the processor 630 may generate information about the beamforming vectors, information about interference statistics, information about the rotation matrix corresponding to each of the receivers, etc. when the feedforward link is available. have.

The transmitter 640 transmits the precoded data stream to the receivers using the beamforming vectors. In this case, the transmitter 640 may selectively transmit information about beamforming vectors, information about interference statistics, information about a rotation matrix corresponding to each of the receivers, and the like. In addition, when the dedicated pilot can be used, the transmitter 640 may transmit the dedicated pilot.

Since all of the descriptions of FIGS. 1 to 5 may be applied to the transmitter illustrated in FIG. 6, detailed descriptions thereof will be omitted.

7 is a block diagram illustrating a receiver according to an embodiment of the present invention.

Referring to FIG. 7, a receiver for multi-user multi-input / output communication includes a memory 710, a channel estimator 720, a receiver 730, a processor 740, and a transmitter 740.

The memory 610 stores a channel codebook and a beamforming codebook corresponding to the corresponding receiver, and additionally stores candidates of a rotation matrix and a base codebook.

When the common pilot is transmitted from the transmitter, the received common pilot is provided from the receiver 730 to the channel estimator 720. Channel estimator 720 estimates the channel between the receiver and the transmitter.

The processor 740 selects a preferred codeword from the channel codebook previously stored in the memory 610 based on the estimated channel. Information about the preferred codeword is fed back through the transmitter 740.

After the transmitter determines the beamforming vector corresponding to the receiver, when the data stream is precoded using the beamforming vector, the received data stream is processed by the processor 740. At this time, the processor 740 is optimized based on the information about the beamforming vectors fed forward from the transmitter, the information on the interference statistics, the information on the rotation matrix, or the effective channel estimated based on a dedicated pilot from the transmitter. Combiners can be designed. The received data stream is processed according to the optimal combiner.

Since all of the descriptions of FIGS. 1 to 5 may be applied to the receiver illustrated in FIG. 7, a detailed description thereof will be omitted.

The methods described above may be embodied in the form of program instructions that may be executed by various computer means and may be recorded in a computer readable medium. The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. Program instructions recorded on the media may be those specially designed and constructed for the purposes of the present invention, or they may be of the kind well-known and available to those having skill in the computer software arts. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tape, optical media such as CD-ROMs, DVDs, and magnetic disks, such as floppy disks. Magneto-optical media, and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine code generated by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like. The hardware device described above may be configured to operate as one or more software modules to perform the operations of the present invention, and vice versa.

As described above, the present invention has been described by way of limited embodiments and drawings, but the present invention is not limited to the above embodiments, and those skilled in the art to which the present invention pertains various modifications and variations from such descriptions. This is possible.

Therefore, the scope of the present invention should not be limited to the described embodiments, but should be determined not only by the claims below but also by the equivalents of the claims.

110: base station

Claims (15)

In the communication method of the transmitter for multi-user multi-input and output communication,
Maintaining a memory in which a first channel codebook corresponding to the first receiver and a second channel codebook corresponding to the second receiver are stored;
Information about a first preferred codeword among codewords included in the first channel codebook from the first receiver and information about a second preferred codeword among codewords included in the second channel codebook from the second receiver. Receiving information; And
Determining a first beamforming vector for the first receiver and a second beamforming vector for the second receiver based on the first preferred codeword and the second preferred codeword.
Including,
And the first channel codebook and the second channel codebook are different from each other.

The method of claim 1,
The information about the first preferred codeword is related to the covariance of the first channel between the transmitter and the first receiver, and the information about the second preferred codeword is between the transmitter and the second receiver. A method of communication of a transmitter for multi-user multi-input / output communication relating to covariance of a second channel.

The method of claim 1,
Codewords included in the first channel codebook are defined based on vectors in a grasmanian line packing codebook of M _t M _r1 x 1 dimension,
Codewords included in the second channel codebook are defined based on vectors in a grasmanian line packing code of M _t M _r2 x 1 dimension,
M _r1 is the number of receive antennas of the first receiver, M _r2 is the number of receive antennas of the second receiver, and M _t is the number of transmit antennas of the transmitter. .

The method of claim 3,
Vectors in the Grasmanian line packing codebook of M _t M _r1 x 1 dimension are converted into m _t x M _r1 dimension matrices, and each of the codewords included in the first channel codebook is M _t x M _r1 dimension Defined by each of the matrices of
The vectors in the Grasmanian line packing codebook of M _t M _r2 x 1 dimension are converted into m _t x M _{r 2} dimensional matrices, and each of the codewords included in the second channel codebook is M _t x M _{r 2} dimensional. A method of communication for a transmitter for multi-user multi-input / output communication defined by each of the matrices of.

The method of claim 1,
Determining a first beamforming vector for the first receiver and a second beamforming vector for the second receiver
Determine the first beamforming vector using a first beamforming codebook for the first receiver, and use the second beamforming codebook for the second receiver that is different from the first beamforming codebook. A method of communication of a transmitter for multi-user multi-input / output communication, comprising determining a vector.

The method of claim 5,
The minimum distance between the first beamforming codewords included in the first beamforming codebook and the minimum distance between the second beamforming codewords included in the second beamforming codebook are equal to each other. Method of communication of the transmitter.

The method of claim 1,
Determining a first beamforming vector for the first receiver and a second beamforming vector for the second receiver
Calculating a sum rate for the first receiver and the second receiver assuming that each of the first receiver and the second receiver is a zero forcing linear receiver; And
Selecting a first beamforming vector for the first receiver and a second beamforming vector for the second receiver based on the sum of rates for the first receiver and the second receiver;
Communication method of the transmitter for multi-user multi input and output communication comprising a.

The method of claim 5,
Each of the first beamforming codebook and the second beamforming codebook
And a method of communication between the first beamforming codewords and the second beamforming codewords to minimize the correlation.

The method of claim 5,
The first beamforming codebook
Generated by inner product of each of the codewords included in the first unitary rotation matrix and the predefined basic codebook,
The second beamforming codebook
And a second unitary rotation matrix different from the first unitary rotation matrix and an inner product of each of the codewords included in a predefined basic codebook.

10. The method of claim 9,
Providing information about the first unitary rotation matrix to the first receiver and providing information about the second unitary rotation matrix to the second receiver
Communication method of the transmitter for multi-user multi input and output communication further comprising.

The method of claim 1,
Calculating interference statistics based on a dot product between the first beamforming vector and a hermithian of the first beamforming vector, and a dot product between the second beamforming vector and the hermitian of the second beamforming vector; step; And
1 ′ providing information about the interference statistics to the first receiver and the second receiver
Communication method of the transmitter for multi-user multi input and output communication further comprising.

A communication method of a second receiver for a multi-user multi-input / output communication system including a transmitter, a first receiver, and a second receiver,
Maintaining a memory in which a second channel codebook corresponding to a second receiver distinct from the first channel codebook corresponding to the first receiver is stored;
Calculating a covariance of a second channel between the transmitter and the second receiver;
Selecting a second preferred codeword among codewords included in the second channel codebook based on the covariance of the second channel;
Providing information regarding the second preferred codeword to the transmitter;
Identifying the second beamforming vector when the transmitter uses a second beamforming vector based on the second preferred codeword; And
Calculating a reception combining vector according to the second beamforming vector and the second channel
Communication method of the second receiver for multi-user multi-input and output communication comprising a.

The method of claim 12,
Calculating an effective channel between the transmitter and the second receiver using a dedicated pilot for the second receiver sent from the transmitter
Further comprising:
Identifying the second beamforming vector
And identifying the second beamforming vector based on the effective channel.

The method of claim 12,
The transmitter determines the first beamforming vector using a first beamforming codebook for the first receiver, and uses the second beamforming codebook for the second receiver that is different from the first beamforming codebook If the second beamforming vector is determined, and the second beamforming codebook is generated by inner product of each of the codewords included in the unitary rotation matrix and the predefined basic codebook,
Receiving information about the unitary rotation matrix from the transmitter
Further comprising:
Identifying the second beamforming vector
And identifying the second beamforming vector using the information about the unitary rotation matrix.

The method of claim 12,
Interference statistics from the transmitter-the interference statistics are a dot product between the first beamforming vector for the first receiver and the hermitian of the first beamforming vector, and the second beamforming vector and the second beamforming. Receiving information about the product of the product between the hermitians of the vectors
Further comprising:
Computing the reception combining vector
And calculating the received combining vector based on the interference statistics.

Images