KR101043808B1 - Beamforming apparatus and method for mimo system - Google Patents

Abstract

를 추출하는 선처리 행렬 추출부, 그리고 선처리 행렬과 심벌을 이용하여 연산을 수행하여 송신 신호를 생성하는 선처리부를 포함한다. 이때 선처리 행렬

의

번째 열벡터

는 전송 최대비 결합 행렬

과 가중치 벡터

의 선형 조합으로 이루어진다. 본 발명에 의하면, 낮은 연산 복잡도를 가지면서도 SVD 방식과 동일한 성능을 제공할 수 있다.The present invention relates to a beamforming apparatus and method for a multiple input / output system, wherein the apparatus comprises a preprocessing matrix using channel state information from a receiving apparatus.

A preprocessing matrix extracting unit for extracting a, and a preprocessing unit for generating a transmission signal by performing an operation using the preprocessing matrix and symbols. Preprocessing matrix

of

Vector

Is the transmission maximum ratio combining matrix

And weights vector

Consists of a linear combination of. According to the present invention, it is possible to provide the same performance as the SVD method while having low computational complexity.

Multiple input / output system, beamforming, singular value decomposition, channel matrix, preprocessing matrix

Description

Beamforming Apparatus and Method for Multiple Input / Output System {BEAMFORMING APPARATUS AND METHOD FOR MIMO SYSTEM}

The present invention relates to a beamforming apparatus and method for a multiple input / output system.

Recently, many 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. As part of this research, research on multiple-input multiple-output (MIMO) systems using channels in the spatial domain has been actively conducted. MIMO technology can provide a high data rate by increasing the channel capacity within a limited frequency resource by using multiple antennas at both ends of the transmission and reception.

As a technique of using multiple transmit / receive antennas, radar technology using an array antenna, sonar technology in an underwater space, beamforming technology of a wireless communication system, and the like have been studied for a long time. Recently, as the demand for increasing capacity in a wireless communication system increases, Attention has been focused on such MIMO technology research. MIMO technology, also known as space-time technology, can theoretically achieve channel capacity that is proportional to the smaller number of transmit and receive antennas by using multiple transmit / receive antennas in a scatter-rich channel environment. Indeed, many studies, especially the Bell Labs, show that high system capacity can be achieved by using multiple antennas at the transceiver and using appropriate signal processing techniques.

In the MIMO environment, the closed-loop transmission technique is a technique in which a transmitter receives feedback channel information estimated by a receiver and transmits data using the information, thereby improving system capacity and average error rate by using channel information. In this case, the transmitter can be classified into a total channel feedback and a limited channel feedback scheme according to characteristics of the channel information received from the receiver, and various transmission schemes can be applied from the channel information. However, in this case, it has a problem of occupying many frequency resources of the uplink in order to feed back channel information, and shows a disadvantage in that the performance deteriorates rapidly when an error occurs in the feedback channel. In the case of full channel feedback, all channel information of the MIMO system estimated by the receiver is fed back to the transmitter, and the transmitter can apply a transmission scheme for improving system performance using the channel information.

If the transceiver knows all the channel information, it can separate the channel into virtual parallel subchannels without interference and use the eigenvalues for each subchannel from Singular Value Decomposition (SVD) of the channel matrix. Signal processing to optimize channel capacity and average error rate. When the total transmit power is limited, an optimal solution for channel capacity can be obtained by allocating transmit power for each antenna using a water-filling technique. This channel capacity is larger than the method of allocating the same power for each transmit antenna in the open loop scheme. In particular, the smaller the signal-to-noise ratio and the greater the correlation between channels, the greater the difference. However, since the optimal SVD-based transmission method requires the SVD operation of the channel matrix, the computational complexity increases greatly as the number of antennas increases, which makes it difficult to actually implement.

An object of the present invention is to provide a beamforming apparatus and method of a closed loop MIMO system having substantially the same performance as that of an SVD-based transmission scheme, which is an optimal beamforming scheme, and having low computational complexity.

를 추출하는 선처리 행렬 추출부, 그리고 상기 선처리 행렬과 심벌을 이용하여 연산을 수행하여 송신 신호를 생성하는 선처리부를 포함하며, 상기 선처리 행렬

의

번째 열벡터

는 전송 최대비 결합 행렬

과 가중치 벡터

의 선형 조합으로 이루어지고,

이고, 상기

은 상기 송신 신호가 동시에 독립적으로 전송되는 데이터 스트림의 개수이다.In order to solve this technical problem, a beamforming apparatus of a multiple input / output system according to an exemplary embodiment of the present invention uses a preprocessing matrix by using channel state information from a receiving apparatus.

A preprocessing matrix extracting unit for extracting a preprocessing matrix, and a preprocessing unit generating a transmission signal by performing an operation using the preprocessing matrix and the symbol;

of

Vector

Is the transmission maximum ratio combining matrix

And weights vector

Consists of a linear combination of

And

Is the number of data streams in which the transmission signal is independently transmitted simultaneously.

인 경우 상기 열벡터

이고, 상기 전송 최대비 결합 행렬

이며, 상기 가중치 벡터

이고, 상기

는 상기 채널 정보로부터 연산된 채널 행렬

의 에르미트 전치 행렬이며, 상기

는 첫 번째 위치가 1이고 다른 위치가 0으로 이루어진 단위 벡터이고, 상기

은 상기 전송 최대비 결합 행렬

의 첫 번째 열벡터이다.remind

The column vector if

And the transmission maximum ratio coupling matrix

And the weight vector

And

Is a channel matrix computed from the channel information

Hermitian transpose of

Is a unit vector of which the first position is 1 and the other position is 0,

Is the transmission maximum ratio coupling matrix

Is the first column vector of.

인 경우 상기 열벡터

이고, 상기 전송 최대비 결합 행렬

이며, 상기 가중치 벡터

이고, 상기

는 상기 채널 정보로부터 연산된 채널 행렬

의 에르미트 전치 행렬이며, 상기

는 첫 번째 위치가 1이고 다른 위치가 0으로 이루어진 단위 벡터이고, 상기

은 상기 전송 최대비 결합 행렬

의 첫 번째 열벡터이며, 상기

는 양의 실수이다.remind

The column vector if

And the transmission maximum ratio coupling matrix

And the weight vector

And

Is a channel matrix computed from the channel information

Hermitian transpose of

Is a unit vector of which the first position is 1 and the other position is 0,

Is the transmission maximum ratio coupling matrix

Is the first column vector of

Is a positive real number.

인 경우 상기 열벡터

이고, 상기 전송 최대비 결합 행렬

(

)이며, 상기 가중치 벡터

이고, 상기

는 상기 채널 정보로부터 연산된 채널 행렬

의 에르미트 전치 행렬이며, 상기

는 m번째 위치가 1이고 다른 위치가 0으로 이루어진 단위 벡터이고, 상기

은 상기 전송 최대비 결합 행렬

의 m번째 열벡터이다.remind

The column vector if

And the transmission maximum ratio coupling matrix

(

) And the weight vector

And

Is a channel matrix computed from the channel information

Hermitian transpose of

Is a unit vector of m-position 1 and other positions 0,

Is the transmission maximum ratio coupling matrix

M-th column vector of.

인 경우 상기 열벡터

이고, 상기 전송 최대비 결합 행렬

(

)이며, 상기 가중치 벡터

이고, 상기

는 상기 채널 정보로부터 연산된 채널 행렬

의 에르미트 전치 행렬이며, 상기

는 m번째 위치가 1이고 다른 위 치가 0으로 이루어진 단위 벡터이고, 상기

은 상기 전송 최대비 결합 행렬

의 m번째 열벡터이며, 상기

는 양의 실수이다.remind

The column vector if

And the transmission maximum ratio coupling matrix

(

) And the weight vector

And

Is a channel matrix computed from the channel information

Hermitian transpose of

Is a unit vector of m-position 1 and other positions 0,

Is the transmission maximum ratio coupling matrix

M-th column vector of

Is a positive real number.

를 추출하는 단계, 그리고 상기 선처리 행렬과 심벌을 이용하여 연산을 수행하여 송신 신호를 생성하는 단계를 포함하며, 상기 선처리 행렬

의

번째 열벡터

는 전송 최대비 결합 행렬

과 가중치 벡터

의 선형 조합으로 이루어지고,

이며, 상기

은 상기 송신 신호가 동시에 독립적으로 전송되는 데이터 스트림의 개수이다.According to another aspect of the present invention, a beamforming method of a multiple input / output system includes a preprocessing matrix using channel state information from a receiving device.

Extracting and generating a transmission signal by performing an operation using the preprocessing matrix and the symbol;

of

Vector

Is the transmission maximum ratio combining matrix

And weights vector

Consists of a linear combination of

And said

Is the number of data streams in which the transmission signal is independently transmitted simultaneously.

As described above, according to the beamforming apparatus and method of the MIMO system according to the present invention, it is possible to obtain substantially the same system performance while having a lower computational complexity than the SVD-based transmission scheme.

DETAILED DESCRIPTION Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art may easily implement the present invention.

First, a transmission apparatus and a reception apparatus including a beamforming apparatus of a closed loop MIMO system according to an embodiment of the present invention will be described in detail with reference to FIGS. 1 and 2.

1 is a block diagram illustrating a transmission apparatus of a MIMO system according to an embodiment of the present invention, and FIG. 2 is a block diagram illustrating a receiving apparatus of a MIMO system according to an embodiment of the present invention.

A transmission apparatus according to an embodiment of the present invention includes a channel coder 110, a bit interleaver 120, a serial-to-parallel converter 130, a plurality of mapping units 140 and 145, a preprocessor 150, and a plurality of transmissions. Antennas 160 and 165, and a preprocessing matrix calculator 170, and the receiver includes a channel decoder 210, a bit deinterleaver 220, a parallel-to-serial converter 230, and a plurality of demapping units 240. , 245, a post processor 250, and a plurality of receive antennas 260 and 265. Here, the number of mapping units 140 and 145 and the number of transmitting antennas 160 and 165 are the same, and the number of demapping units 240 and 245 and the number of receiving antennas 260 and 265 are the same. The beamforming apparatus according to the embodiment of the present invention includes the preprocessor 150 and the preprocessing matrix extractor 170 in a narrow sense, but in a broad sense, the channel coder 110, the bit interleaver 120, and the serial and parallel converter. 130 and the mapping units 140 and 145 may be further included.

는 송신 안테나(160, 165)의 개수를,

은 수신 안테나(260, 265)의 개수를,

은 동시에 전송되는 독립적인 데이터 스트림의 개수를 나타내는 것으로 한다. 이때 식

이 성립한다. 그리고 보통 문자는 스칼라를, 볼드체 소문자는 벡터를, 볼드체 대문자는 행렬을 나타내고,

는 공액 복소수(complex conjugate)를,

는 전치(transpose)를,

는 에르미트 전치(Hermitian transpose)를 나타내는 것으로 한다.Below

Is the number of transmit antennas 160, 165,

Denotes the number of receive antennas 260 and 265,

Denotes the number of independent data streams transmitted simultaneously. Where expression

This holds true. And the usual characters represent scalars, the lowercase bold letters to vectors, the bold uppercase letters to matrices,

Is a complex conjugate,

Is a transpose,

Denotes Hermitian transpose.

The channel coder 110 encodes information bits to be transmitted to the receiving device. That is, the channel coder 110 may include a cyclic redundancy check code (CRC) code, a convolutional code, and the like in the information bits to correct a random error that may occur during radio wave transmission at a receiving device. Insert

The bit interleaver 120 stores an array of bit streams provided from the channel encoder 110 so that the receiving device can correct burst errors or fading errors that may occur during radio wave transmission. Change it. The bit interleaver 120 may change the arrangement of the bit stream using methods such as block interleaving, helical interleaving, random interleaving, and the like.

The serial-to-parallel converter 130 converts the bit streams serially input from the bit interleaver 120 into a plurality of parallel bit streams according to the number of the mapping units 140 and 145, and sends them to the corresponding mapping units 140 and 145. . Each parallel stream is generated by cutting the serial bit stream by a predetermined number of bits corresponding to the modulation levels of the mapping units 140 and 145. In this case, the modulation level is determined according to the modulation scheme used in the mapping units 140 and 145. For example, if the number of mapping units 140 and 145 is two and the modulation scheme is 16-QAM, two parallel bit streams may be converted using a serially input bit stream in units of 4 bits.

차원의 심벌 벡터

로 변환할 수 있다. 여기서 심벌 벡터

는

로 나타낼 수 있으며,

는

번째 타임 슬롯(time slot)를 나타낸다.The mapping units 140 and 145 are parallel bits provided from the parallel-parallel conversion unit 130 using modulation methods such as quadrature amplitude modulation (QAM) and quadrature phase shift keying (QPSK). Convert the stream to a symbol. That is, the mapping units 140 and 145 perform bit streams.

Dimensional symbol vector

Can be converted to Where symbol vector

Is

Can be represented by

Is

Represents the first time slot.

와 선처리 행렬(Precoding Matrix)

을 이용하여 아래의 [수학식 1]과 같은 연산을 수행하고 선처리된 신호 벡터

를 생성한다. 선처리 행렬

는 빔형성 행렬(Beamforming Matrix)이라고도 하며 이하 이 둘을 혼용하여 사용하기로 한다.The preprocessor 150 is a symbol

And precoding matrix

Perform the operation as shown in [Equation 1] below using the preprocessed signal vector

. Preprocessing matrix

Also known as a beamforming matrix, the two will be used interchangeably hereinafter.

은 행렬

의

번째 열벡터를 나타낸다.here

Silver matrix

of

The second column vector.

를 추출한다. 선처리 행렬 추출부(170)는 채널 상태 정보로서 채널 행렬(channel matrix) 자체를 수신하고 이를 이용하여 연산 처리를 수행함으로써 선처리 행렬

를 결정할 수도 있지만, 수신 장치로부터 채널 행렬과는 직접 관련이 없는 데이터 스트림을 받아 이로부터 채널 행렬을 추출하고 채널 행렬을 이용하여 연산 처리를 수행함으로써 선처리 행렬

를 결정할 수도 있다.The preprocessing matrix extractor 170 receives channel state information (CSI) from the receiving apparatus and preprocesses the matrix.

Extract The preprocessing matrix extractor 170 receives the channel matrix itself as channel state information and performs arithmetic processing using the preprocessing matrix.

It is also possible to determine, but by receiving the data stream not directly related to the channel matrix from the receiving device, extracting the channel matrix from it and performing arithmetic processing using the channel matrix

May be determined.

에 대응하는 전파를 내보낸다.The transmit antennas 160 and 165 are signal vectors provided from the preprocessor 150.

Send out the corresponding radio wave.

를 후처리부(250)로 보낸다. 벡터

는 아래의 [수학식 2]와 같다.Receive antennas 260 and 265 are signal vectors received from a transmitting device

Send to the post-processing unit (250). vector

Is shown in Equation 2 below.

는 복소 가우시안 노이즈(Gaussian noise) 벡터를 나타내며, 채널 행렬

는

와 같다. 여기서

는 j번째 송신 안테나와 i번째 수신 안테나 사이의 채널 응답을 나타내고,

는 채널 행렬

의 i번째 행벡터(row vector)를 나타낸다. 그리고 아래의 [수학식 3]이 성립한다고 가정한다.here

Denotes a complex Gaussian noise vector, and the channel matrix

Is

Same as here

Denotes a channel response between the j th transmit antenna and the i th receive antenna,

Is the channel matrix

Represents the i th row vector of the. And it is assumed that Equation 3 below holds true.

는 유클리디안 놈(Euclidean norm)을 나타낸다.here

Represents the Euclidean norm.

에 대하여 필터링을 수행하여 심벌

을 생성한다. 즉, 후처리부(250)는 수신 안테나(250, 265)로부터 제공받은 신호 벡터

에 수신 행렬(receive matrix)

를 곱하여 심벌

를 생성한다.Post-processing unit 250 is a signal provided from the receiving antenna (260, 265) as shown in Equation 4 below

Filter on the symbol

. That is, the post processor 250 receives the signal vectors provided from the receiving antennas 250 and 265.

Receive matrix

Multiply the symbol

.

는 필터 출력 노이즈(filter output noise)를 나타내며, 심벌 벡터

는

로 나타낼 수 있다.here

Denotes the filter output noise, and the symbol vector

Is

It can be represented as.

를 비트 스트림으로 변환한다. 즉, 디매핑부(240, 245)는 후처리부(250)로부터 제공받은 심벌 벡터

를 복조하여 비트 스트림으로 변환한다.The demapping units 240 and 245 are symbols in a demodulation scheme corresponding to the modulation scheme used in the mapping units 140 and 145.

To a bit stream. That is, the demapping units 240 and 245 are symbol vectors provided from the post processing unit 250.

Demodulate to convert to a bit stream.

The parallel converter 230 converts a bit stream input in parallel from the demapping units 240 and 245 into a serial bit stream.

The bit deinterleaver 220 changes the arrangement of the bit stream provided from the parallel-to-serial converting unit 230 according to the interleaving method used by the transmitting apparatus, and converts the bit stream to the original bit stream.

The channel decoder 210 checks for an error using a CRC code, a convolutional code, etc. inserted in the bit stream provided from the bit deinterleaver 220, and corrects the error if there is an error. Generate an estimated bit that is expected to have been sent by the transmitting device.

를 결정하는 방법에 대해 설명한다. 먼저 동시에 전송되는 독립적인 데이터 스트림의 개수가 1개인 경우(단일 빔형성)부터 설명하고, 이후 데이터 스트림의 개수가 복수 개인 경우(복수의 빔형성)에 대해 설명한다.Then, the preprocessing matrix in the preprocessing matrix extractor 170 of the transmitting apparatus according to the embodiment of the present invention.

It explains how to determine. First, the case where the number of independent data streams transmitted at the same time is one (single beamforming) will be described first, and the case where the number of the data streams is plural (multiple beamforming) will be described.

) Single beamforming (

)

(위에서 설명한 빔형성 행렬

의 첫 번째 열벡터)를 전송 최대비 결합(Transmit Maximal Ratio Combining, TMRC) 벡터

의 조합으로 나타내면 아래의 [수학식 5]와 같다.Beamforming vector

(Beamforming matrix described above

Transmit Maximal Ratio Combining (TMRC) Vector

When expressed as a combination of Equation 5 below.

는

로 정의된다. [수학식 5]는 다음과 같은 행렬 형태로 쓸 수 있다.

, 여기서 TMRC 행렬

는

로 정의되고, 가중치 벡터

는

로 정의된다. 단일 빔형성의 경우 식

가 성립된다.Where TMRC vector

Is

Is defined as Equation 5 can be written in the following matrix form.

, Where TMRC matrix

Is

Defined by the weight vector

Is

Is defined as For single beamforming

Is established.

과 같다. 여기서

는 평균 성좌점 에너지(average constellation energy)를 의미한다. 따라서 이러한 신호대 잡음비(SNR)를 최대화하는 것은

을 최대화하는 것과 동일하다. 그러면 문제는 아래의 [수학식 6]과 같이 수식화된다.Assuming that the maximum ratio combining (MRC) technique is applied at the receiving device, the signal-to-noise ratio (SNR) of the maximum ratio combining (MRC) combiner is

Is the same as here

Is the average constellation energy. Therefore, maximizing this signal-to-noise ratio (SNR)

Is the same as maximizing. The problem is then formulated as in Equation 6 below.

를 미리 결정된 벡터로서 아래의 [수학식 7]과 같이 선택할 수 있다.Simply to solve [Equation 6]

May be selected as Equation 7 below as a predetermined vector.

는 i번째 위치는 1이고 다른 위치는 0으로 이루어진

단위 벡터(unit vector)를 의미한다. [수학식 7]과 같이 선택하는 것은 관측에 기초 한 것으로서, 하나의 행벡터

에 매치되는 TMRC 벡터를 이용하면 적절한

를 선택함으로써 SVD 방식의 성능과 거의 근접한 성능을 제공할 수 있다. [수학식 7]에서

이라고 선택한 것은 [수학식 3]에서의 가정으로 인하여 빔형성 벡터

이 최대 에너지를 가지는 채널 행벡터에 매치된다는 것을 의미한다. 이러한 에너지 기반 선택은

일 때 거의 최적이 되나,

인 경우라도 성능 손실은 큰 차이가 없다(도 3 참조). 지금까지 설명한 방식을 일반화된 송신 최대비 결합 타입 Ⅰ(Generalized TMRC Type Ⅰ, G-TMRC Type-Ⅰ)이라 명명한다.here

Is the i position is 1 and the other is 0

It means a unit vector. Equation (7) is based on observation, one row vector

Using TMRC vectors that match

By selecting, we can provide a performance close to that of the SVD method. In [Equation 7]

The beamforming vector is chosen due to the assumptions in [Equation 3].

This means that a match is made to the channel row vector having the maximum energy. This energy-based choice

Is almost optimal when

Even if the performance loss is not a big difference (see Fig. 3). The method described so far is referred to as Generalized TMRC Type I (G-TMRC Type-I).

On the other hand, how to apply the gradient ascent algorithm to improve the performance more than G-TMRC Type-I.

가

에서

방향으로 진행한다면 함수

가 빠르게 증가한다는 관측 결과에 기초한다. 여기서

는

가

일 때 함수

의 기울기를 나타낸 것으로서

로 나타낼 수 있다.The slope increasing algorithm is a function of Equation 6

end

in

Function to proceed in the direction of

Is based on the observation that increases rapidly. here

Is

end

Function when

As the slope of

It can be represented as.

라고 정의되면

가 성립하며 충분히 작은 양의 실수

는 [수학식 9]와 같이 모델링 할 수 있다.therefore

Is defined as

A small enough mistake

Can be modeled as shown in [Equation 9].

은 양의 실수이다.here

Is a positive real number.

를

으로 선택하면 최적의 방식에 거의 근사한 성능을 제공할 수 있다. 이러한 방법을 일반화된 송신 최대비 결합 타입 Ⅱ(Generalized TMRC Type Ⅱ, G-TMRC Type-Ⅱ)라 명명한다. G-TMRC Type-Ⅱ에 의하면 G-TMRC Type-Ⅰ의 성능보다 다소 향상된 성능을 제공할 수 있다(도 3 참조).like this

To

In this case, we can provide a performance that approximates the optimal way. This method is called Generalized TMRC Type II (G-TMRC Type-II). According to the G-TMRC Type-II, a performance slightly improved than that of the G-TMRC Type-I may be provided (see FIG. 3).

G-TMRC Type-I and G-TMRC Type-II are summarized as follows.

인 2Ⅹ2, 4Ⅹ4, 6Ⅹ6 시스템의 신호대 잡음비의 누적 분포 함수를 보여주고 있다. G-TMRC Type-Ⅱ의 경우 단순히 계산 복잡도를 최소화하기 위하여

를 150으로 하였다. SVD 기반 단일 선처리 방식(SVD-based Unitary Precoding, SVD-UP)은 최적의 성능을 보여주지만 본래의 반 복 연산으로 인하여 계산 복잡도와 수치 감도가 매우 높다. G-TMRC Type-Ⅰ은 SVD-UP와 비교하여 1dB 미만의 성능 저하가 있지만 계산 복잡도는 가장 낮다. G-TMRC Type-Ⅱ는 SVD-UP에 비하여 계산 복잡도가 낮을 뿐만 아니라 매우 근접한 성능을 제공한다.3 is a graph showing simulation results in order to compare the performance of the G-TMRC method and the SVD method.

The cumulative distribution function of the signal-to-noise ratios of 2Ⅹ2, 4Ⅹ4, and 6Ⅹ6 systems is shown. In the case of G-TMRC Type-II, simply to minimize the computational complexity

Was set to 150. SVD-based unitary precoding (SVD-UP) shows optimal performance, but due to the inherent repetitive operations, the computational complexity and numerical sensitivity are very high. G-TMRC Type-I has a performance degradation of less than 1dB compared to SVD-UP, but has the lowest computational complexity. G-TMRC Type-II is not only low computational complexity, but also very close to the SVD-UP.

) Multiple beamforming (

)

First, the SVD scheme will be briefly described, and then the case where there are a plurality of independent data streams transmitted simultaneously will be described.

를

로 두면, 우특이 벡터(right singular vectors)

는 [수학식 10]과 같이 나타낼 수 있다.Preprocessing Matrix in SVD-UP Method

To

If left, right singular vectors

Can be expressed as shown in [Equation 10].

은 길이가

인 모든 단위 벡터의 집합이고,

은 벡터

와 직교하는 모든 단위 벡터의 집합을 나타낸다. 이것은 모든

차원의 단위 벡터 중에서

가

일 때

이 최대화되는 반면, 벡터

와 직교하는 벡터 공간 내에서는

가

일 때

이 최대화된다는 것을 의미한다.here

Is the length

Is a set of all unit vectors,

Silver vector

Represents the set of all unit vectors orthogonal to. This is all

Out of the unit vectors of the dimension

end

when

While this is maximized, the vector

In a vector space orthogonal to

end

when

This means that it is maximized.

은 TMRC 벡터의 선형 조합으로서 [수학식 11]과 같이 표현할 수 있다.Returning to the G-TMRC method according to the embodiment of the present invention, considering Equation 10, the beamforming vector

May be expressed as Equation 11 as a linear combination of TMRC vectors.

은 복수의 빔형성 벡터와 구별하기 위한 것이다.Superscript here

Is for distinguishing from the plurality of beamforming vectors.

을 초기값으로 하면, 첫 번째 빔형성 벡터

은

이며,

은 앞서 설명한 단일 빔형성인 경우의

를 선택하는 방법을 이용하여 결정된다. 즉,

은 [수학식 12]와 같이 결정된다.

Is the initial value, the first beamforming vector

silver

Is,

Is the case of the single beamforming described above.

Is determined using the method of selecting. In other words,

Is determined as shown in Equation 12.

은 식

을 충족하도록 선택된다.here

Silver expression

Is chosen to meet.

에 직교하는 빔형성 벡터

를 구할 수 있다. 이 때문에

를 계산할 수 있으며, 이것은 다음 [수학식 13]과 동등하다.And using orthogonal projection, beamforming vector

Beamforming Vector Orthogonal to

Can be obtained. Because of this

Can be calculated, which is equivalent to Equation 13 below.

은

의 i번째 열을 나타낸다.here

silver

It represents the i th column of.

를 계산하기 위해 가중치 벡터

를 결정해야 한다. G-TMRC Type-Ⅰ의 경우에

는 [수학식 14]와 같이 미리 결정될 수 있다.

Weight vector to calculate

Must be determined. In the case of G-TMRC Type-Ⅰ

May be predetermined as shown in [Equation 14].

는 가장 큰 행벡터

가 아니라 두 번째로 큰 행벡터

에 매칭된다.

Is the largest row vector

Not the second largest row vector

Is matched to

을 [수학식 15]와 같이 선택할 수 있다.As in the case of single beamforming, the gradient vector increases the weight vector in the case of G-TMRC Type-II.

Can be selected as shown in [Equation 15].

은 식

이 성립하도록 결정된다.here

Silver expression

It is determined to hold.

도 결정할 수 있다. 복수의 빔형성인 경우의 G-TMRC 방식을 정리하면 다음과 같다.In this way

Can also be determined. The G-TMRC scheme in the case of a plurality of beamforming is summarized as follows.

보다 작을 때

선처리 행렬의 계산을 위하여 단지

개의 행벡터가 요구되므로 피드백 오버헤드(feedback overhead)를 줄일 수 있다. 이와 반대로 G-TMRC Type-Ⅱ 방식이면 SVD 방식과 마찬가지로 전체 채널 상태 정보(Full CSI)가 요구된다. 그럼에도 불구하고 G-TMRC Type-Ⅱ 방식의 연산 복잡도는 SVD 방식에 비하여 확실히 낮다. 한편 시분할 듀플렉스(Time Division Duplex, TDD) 시스템에서, G-TMRC Type-Ⅰ 방식의 수신 장치는 채널 상태 정보를 전송하기 위한 업링크(uplink)용으로

개의 안테나가 아니라

개의 안테나가 요구되므로 시스템 비용을 더욱 줄일 수 있다.When comparing the amount of channel state information (CSI) required by the transmitting apparatus, in the frequency division duplex (FDD) system, if the G-TMRC Type-I scheme, the number of transmission data streams is

Less than

Only for the calculation of the preprocessing matrix

Row vectors are required, which reduces the feedback overhead. In contrast, in the G-TMRC Type-II method, full channel state information (Full CSI) is required as in the SVD method. Nevertheless, the computational complexity of the G-TMRC Type-II scheme is clearly lower than that of the SVD scheme. Meanwhile, in a time division duplex (TDD) system, a receiving device of a G-TMRC Type-I type is used for uplink for transmitting channel state information.

Not antennas

Two antennas are required, further reducing system cost.

Next, the beamforming method (G-TMRC) and the SVD method of the MIMO system according to the embodiment of the present invention will be described with reference to FIGS. 4 to 6 in terms of computational complexity and performance.

4 is a graph for comparing the computational complexity of the system and the SVD system according to an embodiment of the present invention, Figure 5 is a graph for comparing the performance of the system and SVD system according to an embodiment of the present invention in the case of single beam forming 6 is a graph for comparing the performance of the system and the SVD system according to an embodiment of the present invention in the case of a plurality of beamforming.

)의 함수로서 G-TMRC 방식과 SVD 방식에서 요구되는 부동 소수점 곱셈의 횟수를 도시한 그래프이다. SVD 방식의 연산을 위하여 Power Method와 Jacobi Method라는 두 개의 SVD 연산 알고리즘을 사용하였다. 도 4에 보이는 것처럼 G-TMRC Type-Ⅰ 방식과 G-TMRC Type-Ⅱ 방식의 연산 횟수가 SVD 방식에 비하여 현저하게 낮다는 것을 알 수 있으며, Power Method의 연산 횟수에 비하여 각각 90% 및 50%가 줄어드는 것을 알 수 있다.The graph shown in FIG. 4 shows the number of transmit / receive antennas (

Is a graph showing the number of floating point multiplications required by the G-TMRC and SVD methods. Two SVD algorithms, Power Method and Jacobi Method, are used for the SVD operation. As shown in FIG. 4, it can be seen that the number of operations of the G-TMRC Type-I method and the G-TMRC Type-II method is significantly lower than that of the SVD method, and is 90% and 50%, respectively, compared to the number of calculations of the Power Method. It can be seen that the decrease.

이고 4-QAM 변조 방식의 복수의 빔형성 시스템의 성능을 나타내는 그래프로서, 가로축은 신호대 잡음비(Signal-to-Noise Ratio, SNR)을 나타내고 세로축은 프레임 에러율(Frame Error Rate, FER)을 나타낸다. 도 5 및 도 6에 보이는 것처럼 G-TMRC Type-Ⅰ 방식은 SVD-UP 방식에 비하여 1dB 미만의 성능 손실을 보이지만 G-TMRC Type-Ⅱ 방식은 모든 안테나 구성에 대하여 SVD-UP 시스템과 거의 동일한 성능을 제공하는 것을 알 수 있다.5 is a graph showing the performance of a 16-QAM modulation single beamforming system, the graph shown in FIG.

And a horizontal axis indicates a signal-to-noise ratio (SNR) and a vertical axis indicates a frame error rate (FER). As shown in FIGS. 5 and 6, the G-TMRC Type-I method exhibits a performance loss of less than 1 dB compared to the SVD-UP method, but the G-TMRC Type-II method has almost the same performance as the SVD-UP system for all antenna configurations. It can be seen that it provides.

Next, a beamforming method of the MIMO system according to an embodiment of the present invention will be described with reference to FIG. 7.

7 is a flowchart illustrating a beamforming method of a MIMO system according to an embodiment of the present invention.

First, a transmitter of a MIMO system according to an embodiment of the present invention performs channel encoding for inserting a CRC code, a convolutional code, and the like into an information bit so that a receiver may correct a random error that may occur during radio wave transmission. In operation S720, interleaving is performed to change the arrangement of the bit stream using a method such as block interleaving, helical interleaving, random interleaving, and the like so that the error caused by the burst error or fading may be corrected in the reception apparatus (S720).

로 변환한다(S740).Thereafter, the transmitting apparatus converts the serially changed bit stream into at least one parallel bit stream based on the modulation level (S730), and uses modulation schemes such as quadrature amplitude modulation (QAM) and quadrature phase shift modulation (QPSK). Symbolize a parallel bit stream

Convert to (S740).

를 추출한다(S750). 그런 후 송신 장치는 [수학식 1]과 같이 선처리 행렬

과 심벌

를 곱하여 선처리된 신호

를 생성(S760)해 내고 이를 송신 안테나를 통하여 전송한다(S770).The transmitting apparatus receives the channel state information from the receiving apparatus and uses the equations (5) to (15) to preprocess the matrix.

To extract (S750). Then, the transmitting device performs a preprocessing matrix as shown in [Equation 1].

And the symbol

Preprocessed signal by multiplying

It generates (S760) and transmits it through the transmit antenna (S770).

이 추출되는 것으로 도 7에 도시하였지만 단계(S750)는 단계(S710) 내지 단계(S740)와 별도로 수행될 수 있으며 단계(S760) 이전에 수행되면 된다.For convenience of explanation, the preprocessing matrix after step S740.

Although this is illustrated in FIG. 7 as being extracted, step S750 may be performed separately from steps S710 to S740, and may be performed before step S760.

Next, a method of processing a signal received from a transmitting device by the receiving device will be described.

에 수신 행렬

를 곱하여 심벌

를 생성한다. 수신 장치는 송신 장치에서 이용한 변조 방식에 대응하는 복조 방식으로 심벌

를 복조하여 비트 스트림으로 변환한다. 그런 후 수신 장치는 변환된 병렬 비트 스트림을 직렬 비트 스트림으로 변환한다. 수신 장치는 송신 장치에서 이용한 인터리빙 방법에 따라 직렬 비트 스트림의 배열 등을 변경하고, 비트 스트림 내에 삽입되어 있는 CRC 코드, 컨볼루셔널 코드 등을 이용하여 에러 여부를 체크하고, 에러가 존재하는 경우 에러를 정정하여 추정 비트를 생성한다. 수신 장치는 추정 비트를 생성하는 과정과는 별도로 송신 장치에 채널 상태 정보를 전송하여 송신 장치로 하여금 선처리 행렬을 추출할 수 있도록 한다.The receiving device receives the signal received through the receiving antenna

Receive matrix

Multiply the symbol

. The receiving device uses the demodulation method corresponding to the modulation method used in the transmitting device.

Demodulate to convert to a bit stream. The receiving device then converts the converted parallel bit stream into a serial bit stream. The receiving apparatus changes the arrangement of the serial bit stream according to the interleaving method used by the transmitting apparatus, checks whether there is an error using the CRC code, convolutional code, etc. inserted in the bit stream, and if there is an error, Is corrected to generate an estimated bit. The receiving apparatus transmits channel state information to the transmitting apparatus separately from the process of generating the estimated bits so that the transmitting apparatus can extract the preprocessing matrix.

Although the preferred embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements of those skilled in the art using the basic concepts of the present invention defined in the following claims are also provided. It belongs to the scope of rights.

1 is a block diagram illustrating a transmission apparatus of a MIMO system according to an embodiment of the present invention.

2 is a block diagram illustrating a receiving apparatus of a MIMO system according to an embodiment of the present invention.

3 is a graph showing a simulation result in order to compare the performance of the G-TMRC method and the SVD method.

4 is a graph for comparing the computational complexity of the system and the SVD system according to an embodiment of the present invention.

5 is a graph for comparing the performance of the system and the SVD system according to an embodiment of the present invention in the case of single beamforming.

6 is a graph for comparing the performance of the system and the SVD system according to an embodiment of the present invention in the case of a plurality of beamforming.

7 is a flowchart illustrating a beamforming method of a MIMO system according to an embodiment of the present invention.

<Explanation of symbols for main parts of the drawings>

110: channel coder, 120: bit input,

130: serial and parallel conversion unit 140, 145: mapping unit,

150: preprocessor, 160, 165: transmit antenna,

170: preprocessing matrix extractor, 210: channel decoder,

220: bit de-inverter, 230: parallel-to-serial conversion unit,

240, 245: demapping portion, 250: post-processing portion,

260, 265: receiving antenna

Claims (10)

A beam forming apparatus of a multiple input / output system, Preprocessing Matrix Using Channel State Information from the Receiver

A preprocessing matrix extracting unit for extracting A preprocessor configured to generate a transmission signal by performing an operation using the preprocessing matrix and the symbol; The preprocessing matrix

of

Vector

Is the transmission maximum ratio combining matrix

And weights vector

Consists of a linear combination of

And

Is the number of data streams to which the transmission signal is independently transmitted simultaneously. Beamforming apparatus. delete delete delete In claim 1, remind

The column vector if

And the transmission maximum ratio coupling matrix

(

) And the weight vector

And

Is a channel matrix computed from the channel state information.

Hermitian transpose of

Is a unit vector of m-position 1 and other positions 0,

Is the transmission maximum ratio coupling matrix

M-th column vector of

Is a positive real number. As a beamforming method of a multiple input / output system, Preprocessing Matrix Using Channel State Information from the Receiver

Extracting, and Generating a transmission signal by performing an operation using the preprocessing matrix and the symbol; The preprocessing matrix

of

Vector

Is the transmission maximum ratio combining matrix

And weights vector

Consists of a linear combination of

And said

Is the number of data streams to which the transmission signal is independently transmitted simultaneously. Beamforming method. delete delete delete In claim 6, remind

The column vector if

And the transmission maximum ratio coupling matrix

(

) And the weight vector

And

Is a channel matrix computed from the channel state information.

Hermitian transpose of

Is a unit vector of m-position 1 and other positions 0,

Is the transmission maximum ratio coupling matrix

M-th column vector of

Is a positive real number.

Images