KR101042170B1 - Method for searching code book index in a multi antenna system of closed loop scheme thereof receiver - Google Patents

Abstract

The present invention proposes a codebook index search scheme that is faster than a conventional search scheme in a quantized dynamic gain transmission multiple input multiple output system. The proposed fast index search technique quickly selects a few candidate index vectors using a newly defined simple metric, and determines the optimal codebook index by applying existing search techniques to the selected vectors. According to the present invention, the proposed symbol error probability is almost the same as that of the conventional search method, but the computation amount and the search time are reduced. In addition, even if the size of the codebook stored in the transceiver is increased, the search time is reduced compared to the conventional search method.

Multiple Input Multiple Output (MIMO), Closed Loop, Multiple User, Receive Beamforming Matrix, CQI, Diversity, Codebook, Dynamic Gain Transmission

Description

Codebook index search method and receiver supporting it in closed loop multi-antenna system {METHOD FOR SEARCHING CODE BOOK INDEX IN A MULTI ANTENNA SYSTEM OF CLOSED LOOP SCHEME THEREOF RECEIVER}

The present invention relates to a closed loop multi-antenna system, and more particularly, to a method for searching codebook indexes in a closed loop multi-antenna system and a receiving apparatus therefor.

Recently, the speed of development of mobile communication system is being accelerated by the endless demands of users and the efforts of system developers to secure new markets while satisfying the demands of users. For example, code division multiple access (CDMA), which was a second-generation (2G) mobile communication technology, is now being used as a wideband CDMA (WCDMA), a third-generation (3G) mobile communication technology currently widely used.

In addition, four-generation (TG) mobile communication technologies, such as the upcoming Long Term Evolution (LTE) and world interoperability for microwave access (WiMAX), are now reaching the standardization and commercialization stages with the attention of users and system developers.

These next-generation mobile communication technologies must satisfy efficient use of bandwidth, fast data rates, and low error rates.

However, unlike the wired channel, the wireless channel environment has low reliability due to multipath interference, shadowing, attenuation, time varying noise, and interference. This is a representative cause of the failure to increase the data transmission speed in a mobile communication system.

Therefore, many techniques have been proposed to overcome this. Representative examples are error control coding and diversity techniques. The error control coding technique suppresses the influence of the signal distortion and noise. The diversity scheme copes with fading by receiving a plurality of signals that have undergone independent fading.

The diversity scheme includes time, frequency, multipath, and spatial diversity. The time diversity combines channel coding and interleaving to obtain a diversity effect in time. The frequency diversity enables diversity effects by allowing signals transmitted at different frequencies to be received through different multipaths. The multipath diversity divides the multipath signal using different fading information to obtain a diversity effect. The spatial diversity uses multiple antennas at the transmitter, the receiver, or both to achieve diversity effects by fading signals independent of each other. Typically, the spatial diversity scheme uses an antenna array.

The system using the antenna array (hereinafter referred to as a "multi-antenna system") is a system having multiple antennas in a transmitter / receiver, which uses a spatial domain to increase frequency efficiency. The multi-antenna system is easy to obtain a higher transmission rate in an environment where the time domain and the frequency domain are limited by utilizing the spatial domain.

Since the multi-antenna system transmits independent information for each antenna, it corresponds to a multi-input multi-output (MIMO) system. In the MIMO system, an efficient signal processing algorithm is required for data service with high transmission rate as well as high quality.

An example of the signal processing algorithm is a resource allocation algorithm. The resource allocation algorithm aims to allocate a resource, that is, a data rate for each antenna, so as to obtain a target error rate while consuming a minimum amount of resources. The resource allocation algorithm may be classified into a signal processing algorithm at a transmitting side and a signal processing algorithm at a receiving side. In this case, when a data rate for each antenna is given, a transmitter allocates resources to reach a target error rate while consuming the smallest energy.

Existing resource allocation algorithms are classified into a uniform allocation method, a fixed allocation method, and a full-search allocation method.

The uniform allocation scheme allocates the same data rate for each antenna. This is the simplest resource allocation method that does not require the transmission of feedback information. However, not only the linear detection technique but also the successive interference cancellation (SIC) technique has a disadvantage in that the performance is not good due to the high error rate.

The fixed allocation scheme determines an optimal allocation and uses the determined allocation for all channels. The optimal allocation is determined according to the statistics of the channel. This shows better performance than single allocation if the channel statistics are correct. However, not only the error rate performance is limited by the constraint of fixing the assignment, but also the fixed assignment does not work reliably on channel changes.

Finally, the full-search allocation scheme uses all available combinations as candidates for data rate allocation, and uses the candidates requiring the smallest power as allocations for the current channel. This shows the best performance because all cases are considered. However, the complexity and feedback information has the disadvantage that increases. In the full-search allocation scheme, an iterative algorithm has been proposed to reduce the complexity.

The signal processing algorithm at the receiving side searches for a channel state for each transmission channel and provides a method for feeding back the search result to the transmitting side. The signal processing algorithm at the transmitting side provides a method for allocating resources for each transmitting antenna based on a search result fed back from the receiving side.

As a signal processing algorithm on the transmitting side, BLAST (Bell Labs Layered Space Time) technology is typically present. The BLAST technology can use multiple antennas to increase the amount of data transmission without increasing the frequency domain used by the system.

This BLAST technology is divided into D-BLAST (Diagonal-BLAST) and V-BLAST (Vertical-BLAST). The D-BLAST uses a specific block coding between data transmitted from each transmit antenna for diagonal transmission, and thus has high frequency efficiency. However, there is a disadvantage in that the complexity is high in the implementation. In contrast, V-BLAST can reduce the complexity by transmitting data independently from each antenna.

On the other hand, as a signal processing algorithm on the receiving side, an algorithm used to detect signals transmitted from each transmitting antenna by using a received signal can be classified into a linear detection technique and a non-linear detection technique. .

The linear detection technique includes a zero forcing (ZF) technique and a minimum mean square error (MMSE) technique.

The ZF technique is to remove the interference between symbols by calculating the reference for each column vector of the channel matrix, detecting the symbols corresponding to the column vectors having the largest vector size, and then removing the detected signal components from the received signal. . The MMSE technique minimizes a mean square error between originally transmitted symbols and signals estimated at a receiver.

The nonlinear technique includes a maximum likelihood detection (ML) technique and a successive interference cancellation (SIC) technique.

The ML technique can greatly improve performance by substituting all transmittable symbols from all transmit antennas and selecting an input having a least squared Euclidean distance. However, the complexity increases exponentially with the number of transmit antennas and the modulation order. Therefore, the ML technique shows the best performance, but has a disadvantage in that it is complicated to implement with an increase in the amount of calculation.

The SIC technique is an interference cancellation technique that improves performance by first detecting and removing a channel having a large signal-to-interference plus noise ratio (SINR). The SIC technique requires separate ordering to obtain the best performance.

As described above, in order to improve the performance of the MIMO system, not only the transmission signal can be detected more accurately from the received signal but also a good signal detection technique is urgently required from the viewpoint of the calculation amount.

Accordingly, in order to solve the foregoing, the present invention provides a method for reducing the search time of a code book index in a multi-antenna system supporting a closed loop scheme using a code book, and a receiver for supporting the same.

The present invention also provides a method for searching a codebook index using a proximity between a received signal and a codebook in a multi-antenna system supporting a closed loop scheme using a codebook, and a receiver for supporting the same.

In addition, the present invention provides a codebook index search method for performing a search only for the index of the codebook satisfying a predefined proximity reference value in a multi-antenna system supporting a closed loop method using a codebook and a receiving apparatus supporting the same to provide.

In a first aspect for achieving the above, the present invention supports a closed loop scheme, and a multiplexing device in which a transmitter having a plurality of transmit antennas and a receiver having a plurality of receive antennas communicate based on a predetermined codebook. In the receiving apparatus of the antenna system, a process of generating a candidate index vector group by selecting a candidate index for the codebook index search among all the indexes of the codebook, and searching the codebook index by the generated candidate index vector group Including the process of performing

In this case, the candidate index proposes a codebook index search method, wherein the proximity corresponding to a predefined random proximity reference value among all the indexes is measured.

In a second aspect to achieve the above, the present invention supports a closed loop scheme, and the multiplexing device having a plurality of transmitting antennas and a receiving device having a plurality of receiving antennas communicate based on a predetermined codebook. In the antenna system, feedback information for generating a candidate index vector group by selecting a candidate index for searching a codebook index among all indexes of the codebook and performing a codebook index search by the generated candidate index vector group Including a generator,

Here, the feedback information generator proposes a receiver for searching for a codebook index, wherein the index that measures a degree of proximity satisfying a predefined proximity reference value among all the indexes is selected as a candidate index.

According to the new fast index search scheme in the dynamic gain transmission system of the present invention as described above, using the specific phase of the weight vector calculated from the channel state information, the index vector phase of each of the QEGT codebooks is compared and the proximity of the two phases is compared. Since the system designer selects only the vector index of the QEGT codebook which has a value smaller than any predefined reference value, the feedback process is performed. Therefore, the index search time is reduced and the performance difference is hardly observed. have.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings so that the present invention can be more easily implemented. It should be noted that the terminology used in the embodiments of the present invention should be defined by the general description in the detailed description.

A detailed description of an embodiment of the present invention, which will be described later, will be described with respect to a general process of a receiver transmitting feedback information to a transmitter in a multiple antenna system supporting a closed loop scheme. We will also look at codebook-based feedback schemes in multi-antenna systems that support closed loop schemes.

Then, as a preferred embodiment of the present invention, in suggesting a feedback scheme based on a codebook in a multi-antenna system supporting a closed loop scheme, a method for reducing a time for searching a codebook index and a receiver accordingly Let's take a look at the specific operation of.

First, a description will be given of a general scheme for a receiver to transmit feedback information to a transmitter in a multi-antenna system supporting a closed loop scheme.

가

번째 (

) 송신 안테나의 가중치

과의 곱을 통해 전송될 때,

번째 (

) 수신 안테나에서의 출력

은 하기 <수학식 1>로 표현할 수 있다.Source information signal

end

th (

) Weight of the transmitting antenna

When sent through a product of,

th (

Output at the receiving antenna

Can be expressed by Equation 1 below.

은 무기억성 페이딩 채널로써

의 분포를 가진다. 이 때,

은 복소 정규 분포를 의미하며,

은

번째 수신 안테나에서의

의 분포를 가지는 잡음이다. 만약

,

일 때,

은 h_{j ,i}에 독립이며, n_m 또한 n_j와 독립이다. here

Is a noninvasive fading channel

Has a distribution of. At this time,

Means a complex normal distribution,

silver

At the first receive antenna

Noise with a distribution of. if

,

when,

Is independent of h _{j , i} and n _{m is} also independent of n _j .

번째 수신 안테나에 의해 수신된 신호

에 수신측 가중치 값인

(

은 복소수 값으로

은 공액(conjugation)을 의미)을 곱한다. 이러한 과정을 모든 수신 안테나

에 적용하고, 각 수신 안테나

의 적용에 의해 출력되는 수신신호들을 결합하면 복조된 신호

을 생성할 수 있다. 이러한 시스템 모델을 매트릭스 형태로 나타내면 하기 <수학식 2>와 같다.The channel assumes constant channel characteristics for some symbol transmissions. Receiver

Received by the first receive antenna

On the receiving side weight value

(

Is a complex value

Multiply by conjugation). This process all receive antenna

Applied to each receiving antenna

Demodulated signal by combining received signals output by

Can be generated. This system model is represented by the following Equation (2).

,

,

그리고

는

크기를 갖는 페이딩 채널로써

은

의

번째 요소이다. At this time,

,

,

And

Is

As a fading channel with magnitude

silver

of

Second element.

와

는 각각 송신 및 수신 다이버시티에서의 가중치 벡터를 나타내는데, 최적의 성능을 위해 수신 신호 대 잡음비를 최대화하는 채널의 함수로써 정의된다.Also

Wow

Represent the weight vectors in transmit and receive diversity, respectively, defined as a function of the channel that maximizes the received signal-to-noise ratio for optimal performance.

When the MIMO channel by the same gain combining (EGT) and equal gain combining (EGC) is configured, the optimum channel gain is expressed by Equation 3 below.

와 동 이득 조합 가중치 벡터

는 하기 <수학식 4>와 <수학식 5>를 통해 채널의 함수로써 명확히 정의한다.Dynamic gain transmission weight vector to obtain this optimal channel gain

Vortex Gain Combination Weight Vector

Is clearly defined as a function of the channel through Equations 4 and 5 below.

,

,

,

,

,

,

,

,

에 대해

즉,

형태의 단위 이득(unit gain)을 곱해주어도 전체 이득에는 변함이 없기 때문에 상기 <수학식 4>에서는 유일한 값, 즉 근사해(closed form solution)를 갖지 않는다. 또한 상기 <수학식 5>에서

또한 같은 이유로 근사 해를 갖지 않는다.Arbitrary phase angle

About

In other words,

Multiplying the unit gain of the form does not change the overall gain, so Equation 4 does not have a unique value, that is, a closed form solution. Also in Equation 5 above

It also does not have an approximate solution for the same reason.

는 하기 <수학식 6>을 통해 계산된다.As described above, in order to obtain an optimal channel gain through Equation 3 without having an approximate solution, the phase vector of Equation 4 is obtained.

Is calculated by the following Equation 6.

Frequency division duplex (FDD) in which control data overhead is limited in order to increase a data rate of a feedback channel, that is, a user of a predetermined bandwidth, for feeding back such channel state information to a transmitter. Many difficulties arise in closed-loop MIMO systems. Also, since the gain gain / combination MIMO system does not have an approximate solution for any of the transmit / receive antennas in Equations 4 and 5, it is necessary to perform repetition through many clock cycles. The complexity is increased.

To solve this problem, five quantized channel state information transmission schemes have been proposed in IEEE 802.16e. The most popular technique is the feedback method based on the codebook.

비트로 이루어져 있어 작은 상향 링크 대역폭 을 통해서도 충분한 채널 용량 획득 및 다이버시티 성능을 얻을 수 있는 장점이 있다. 이 때, 코드 북은 생성 효율성 및 저장 메모리에 최적화되어 있으며, 송/수신단 모두 동일한 코드 북을 가지고 있다.Using the codebook, instead of feeding back the entire channel state information, various numbers of channel state information are quantized in advance into an arbitrary vector, and then only specific index information corresponding to the quantized channel state information vector is transmitted through the feedback channel. Can be sent. This causes the index size to

It consists of bits, so that even with a small uplink bandwidth, sufficient channel capacity acquisition and diversity performance can be obtained. At this time, the codebook is optimized for generation efficiency and storage memory, and both transmitters and receivers have the same codebook.

를 양자화한 특정 크기의 코드 북을 사용함으로써 상기에서 설명한 동 이득 전송에서의 문제점을 해결하였다Among the various codebooks, the quantized dynamic gain transfer (QEGT) technique proposed by David J. Love is based on Equation (6).

The problem of the above described dynamic gain transmission is solved by using a codebook of a specific size quantized.

로 정의한다. 예를 들어

코드 북에서 전체 인덱스 수는

이다.In the following description, for the convenience of explanation, the notation for codebooks is used.

. E.g

The total number of indexes in the codebook is

to be.

Table 1 below shows the sizes of various QEGT codebooks.

는 각각의 위상 각도를 코드화시키기 위해 사용된 비트 수이며, 코드 북 크기

에서

은 코드 북 인덱스 수,

은 코드 북 인덱스를 구성하기 위해 필요한 비트 수이다.Referring to <Table 1> above,

Is the number of bits used to code each phase angle, and codebook size

in

Is the number of codebook indexes,

Is the number of bits needed to construct the codebook index.

Meanwhile, in the general QEGT overall codebook, the size of each weight vector is the same, and the phase angles of all weight vectors for the first transmitting antenna are equal to zero. However, the phase angles of the weight vectors for the second transmit antennas are different from each other.

Hereinafter, a closed loop multi-antenna system proposed by the present invention will be described with reference to the accompanying drawings.

1 illustrates a closed loop multiple antenna system supporting multiple users according to an exemplary embodiment of the present invention. In FIG. 1, one transmitter 110 and a plurality of receivers 120-1 and 120 -N are configured. The transmitter 110 may be assumed to be a base station, and the plurality of receivers 120-1 and 120 -N may be assumed to be mobile terminals. In the following description, it is based on one receiver 120-1. The following description is regarded to apply equally to the remaining receiver.

Referring to FIG. 1, the receiver 120-1 performs channel estimation by receiving a received signal as an input. Channel quality information (CQI) corresponding to each data stream is obtained by the channel estimation. The CQI information may be represented by a CQI value. In the following description, the CQI information is collectively used for convenience. The data stream may be defined differently depending on whether a precoder is used at the transmitting side.

The receiver 120-1 generates feedback information based on channel quality information corresponding to each data stream. The method of generating the feedback information may be determined in consideration of a signal detection technique, whether a precoder is used at the transmitter, or the like. The receiving device 120-1 transmits the generated feedback information.

The transmitter 110 receives feedback information from all receivers 120-1 and 120 -N. The transmitter 110 selects at least one user (ie, receiver) based on the feedback information. The number of users selected here may be determined by the operation mode. The operation mode consists of a single-user mode and a multi-user mode. One user is selected in the single user mode, and a plurality of users are selected in the multi-user mode. The operation mode is determined based on the feedback information.

The transmitter 110 determines a modulation & coding selection level (MCS level) based on the feedback information. In determining the MCS level, a signal detection technique, an operation mode, and the number of data streams supporting the selected user may be further considered.

The transmitter 110 specifies an encoding and modulation scheme based on the determined MCS level, and transmits a data stream corresponding to at least one user selected using the designated encoding and modulation scheme.

As described above, according to the present invention, the transmitter 110 determines an operation mode based on feedback information provided from each receiver 120-1 or 120 -N. That is, the transmitter 110 may selectively use a single user mode and a multi-user mode.

To this end, the receiver 120-1 includes at least one reception antenna (Ant_rx # 1, Ant_rx # 2,..., Ant_rx #N _r ), a channel estimator 122-1, and a feedback information generator 124-. It consists of 1). The transmitter 110 is composed of a plurality of transmit antennas (Ant_tx # 1, Ant_tx # 2, ..., Ant_tx #N _r ), a feedback information processor 114 and a signal transmitter 112.

Looking at the receiver 120-1, a signal received from at least one receive antenna Ant # 1, Ant # 2, ..., Ant #N _r is input to the channel estimator 122-1. . The channel estimator 122-1 estimates channel characteristics corresponding to each data stream from the received signal by a preset signal detection technique. The signal detection technique includes a linear detection technique and a nonlinear detection technique. A representative example of the linear detection technique is the LMMSE technique, and a representative example of the nonlinear detection technique is the SIC technique. A signal detection scheme in the channel estimator 122-1 is specified in advance. However, in the case of supporting the dual mode, the channel estimator 122-1 may selectively use an optimal signal detection technique in consideration of channel conditions.

The channel characteristic matrix of the received signal estimated by the channel estimator 122-1 is provided to the feedback information generator 124-1.

The feedback information generator 124-1 generates feedback information by a channel characteristic matrix provided from the channel estimator 122-1. When generating the feedback information, the feedback information generator 124-1 considers a signal detection technique used by the channel estimator 122-1.

Meanwhile, the feedback information generator 124-1 may generate feedback information by further considering the number of data streams provided to the user from the transmitter 110 and whether a precoder (code book) is used. If the transmitter 110 uses a precoder, a precoder index (Precoder_index) is added to the feedback information. The precoder index (Precoder_index) is an index that specifies a precoding matrix capable of obtaining maximum single user performance in a codebook.

The receiving device 120-1 transmits the feedback information generated by the feedback information generating unit 124-1 to the transmitting device 110. The transmission of the feedback information is preferably periodically transmitted by the receiving device 120-1. However, if elements for determining a transmission time point are previously agreed between the transmitter 110 and the receiver 120-1, it is possible to transmit feedback information aperiodically.

Looking at the transmitter 110, the feedback information received from all of the receivers (120-1, 120-N) is provided to the feedback information processor 114. The feedback information processor 114 selects at least one user based on the feedback information received from each receiver. A coding scheme and an MCS level for transmitting a data stream of the at least one user selected are determined.

In order to select at least one user, the feedback information processor 114 considers the operation mode of the transmitter. Therefore, the feedback information processor 110 determines an operation mode before selecting a user. As mentioned above, the operation mode is divided into a single user mode and a multi-user mode.

A detailed operation for searching the codebook index by the method proposed by the embodiment of the present invention by the feedback information processor 114 will be described later.

과

를 복소평면 상에서 보이고 있다.2 is a QEGT codebook

and

Is shown on the complex plane.

Referring to FIG. 2, the magnitude of each weight vector is the same in the entire QEGT codebook, and the phase angles of all the weight vectors for the first transmitting antenna are equal to zero. However, the phase angles of the weight vectors for the second transmit antennas are different from each other.

The receiver can obtain an optimal codebook index from the generated QEGT codebook by Equation 7 below.

는 코드 북의 인덱스이며,

은 1-norm을 의미한다. here

Is the index of the codebook,

Means 1-norm.

Typically, feedback is achieved by selecting a codebook of the appropriate size that is optimized for generation efficiency and storage memory. However, as the number of transmitting antennas increases, the QEGT codebook needs to use a codebook having a larger number of indexes than a codebook having a smaller number of indexes. This is to get closer to the ideal performance. However, when the total number of indexes increases, the time for searching for the optimal index increases linearly.

Hereinafter, a detailed description according to an embodiment of the present invention will be described in detail for a method for fast index search in the same gain transmission system.

In the codebook index search method according to an embodiment of the present invention to be described later, a candidate index group is generated by extracting a candidate index among all indexes constituting the codebook, and a code is generated by the candidate index constituting the generated candidate index group. Suggest ways to search for books.

Meanwhile, in the embodiment of the present invention, a procedure of generating a candidate index group is divided into four steps.

The first step is to generate a new channel characteristic matrix by compensating the phase angle of the channel characteristic matrix of the received signal. That is, a function of the new channel characteristic matrix is defined by compensating the phase angles of the remaining columns except for the first column corresponding to the first transmitting antenna in the channel characteristic matrix estimated from the signals received through the plurality of receiving antennas.

The second step compensates for the phase angle of the codebook to generate an index vector matrix of the codebook. That is, a function of the index vector matrix of the codebook is calculated by compensating the phase angles of the vector values of the indices corresponding to the remaining transmit antennas except the first transmit antenna of the codebook.

The third step is to obtain a proximity for each index in the codebook by using the function of the newly defined channel characteristic matrix and the index vector matrix of the codebook. That is, the absolute value of the difference between the newly defined function of the channel characteristic matrix and the function of the index vector matrix of the codebook corresponds to each of the remaining transmit antennas except the first transmit antenna among the entire indexes of the codebook. Compute the proximity of an index.

The fourth step generates a candidate index group by a candidate index that satisfies any predefined proximity reference value among the calculated proximity.

3 illustrates a control flow for fast index search according to an embodiment of the present invention. As described above, the fast index search may be performed by the feedback information generating unit constituting the receiver by the control flow shown in FIG. 3. However, in the following description, the reception device will be described as a subject of operation for convenience of description.

Referring to FIG. 3, in step 310, the receiving apparatus extracts a candidate index for fast searching from all indexes (N = 2 ^L ) in the codebook. In other words, fast searching is performed in four steps to select candidate index vectors. After receiving the candidate index in step 320, the receiver performs detailed search for selecting an optimal weight vector index among the determined candidate index vectors in step 330.

In the foregoing description, FIG. 2 defines that the weight vector has the same magnitude in the first antenna of the QEGT codebook and that the phase angles of all the weights for the first transmit antenna are equal to zero.

이 아닌 특정 인덱스만을 탐색함으로써 기존의 탐색 기법보다 탐색 시간이 감소한다. 단지 본 발명에서 상기 <수학식 7>을 이용하여 최적의 코드 북 인덱스를 선택할 시에는 k의 범위는 N이 아닌 후보 인덱스로 한정되어야 할 것이다.In other words, the phase angle of all weights in the first antenna of every QEGT codebook has a real vector value of zero, not a complex number. The present invention utilizes the characteristics of the QEGT codebook and converts the vector value of the first transmit antenna, which is complex in the channel state information estimated at the receiver, into a real number having the phase angle equal to 0, the same as the codebook. At this time, the phase angle required for the conversion is compensated to the transmit antennas other than the first transmit antenna to measure the proximity to the phase angle of the codebook except the first antenna. By selecting only the specific index of the codebook in comparison with the value, and using the equation (7), by selecting the optimal codebook index that maximizes the diversity gain in these indexes, the total number of indexes

The search time is reduced by searching only a specific index rather than the conventional search method. In the present invention, when selecting an optimal codebook index using Equation 7, the range of k should be limited to candidate indexes rather than N.

At this time, the number of candidate index vectors is defined as NCI (Number of Candidate Indices). Since the NCI value is always smaller than the total number N of indexes, the index search time is reduced compared to the conventional search technique.

을 고려하는데, 이는 하기 <수학식 8>과 <수학식 9>에 의해 정의된다.As a group of candidate index vectors

Considering this, it is defined by the following equation (8) and (9).

는 빠른 탐색을 위한 새로운 행렬의 함수이며,

은 시스템 설계자에 의해 시스템 성능과 탐색 복잡도 간의 유연한 관계(tradeoff)에 의해 사전에 결정된 최적화 위상 경계(threshold) 각도이다.

은 새로운 채널 행렬이며,

은

번째(

) 송신 안테나,

은

번째(

) 수신 안테나이며,

은

의 위상 각도를 의미한다. 그리고

는 송신 안테나의 총 개수를 의미한다.

Is a function of the new matrix for fast navigation,

Is the optimization phase threshold angle predetermined by the system designer by a flexible tradeoff between system performance and search complexity.

Is the new channel matrix,

silver

th(

) Transmitting antenna,

silver

th(

) Receiving antenna,

silver

Means the phase angle. And

Denotes the total number of transmit antennas.

는 하기 <수학식 10>으로 표현될 수 있다. Complex MIMO Channels

May be expressed by Equation 10 below.

를 상기 <수학식 10>으로 정의할 때, 상기 <수학식 9>를 이용한 새로운 채널 행렬인

을 계산하는 것이다. 상기 새로운 채널 행렬의 함수

는 하기 <수학식 11>에 의해 획득할 수 있다. 하기 <수학식 11>에서 사용하고 있는

는 하기 <수학식 12>에 의해 정의되며,

는 하기 <수학식 13>에 의해 정의된다.The first step of fast searching consisting of 1 to 4 steps according to the embodiment of the present invention described above

When is defined as Equation 10, it is a new channel matrix using Equation 9

Will be calculated. Function of the new channel matrix

Can be obtained by Equation 11 below. In Equation 11 below

Is defined by Equation 12 below,

Is defined by Equation 13 below.

은 송/수신단 페이딩 채널 요소들의 위상 각도를 나타낸다. 그리고

을 정의하고 있는 상기 <수학식 13>은

로 일반화시킬 수 있을 것이다.Meanwhile, Equations 11 to 13 may be used.

Denotes the phase angle of the transmit / receive end fading channel elements. And

Equation 13, which defines

Can be generalized to

In Equation 10, the complex value of the first column (column corresponding to the first transmitting antenna) is converted into a real number. Converting to the real number means converting the phase angle of the complex value of the first column to 0 as a vector value for the first transmit antenna. At this time, the phase angle required for the conversion is compensated to the other antenna. In Equation 12, the channel characteristics of the compensation for the phase angle are shown.

After the compensation is completed, as shown in Equation 13, the first step is completed by adding and taking the absolute values of the compensated phase angles for all the transmitting antennas.

을 계산하는 것이다. QEGT 코드 북 인덱스의 첫 번째 송신 안테나에 대한 벡터 값은 모두 실수이기 때문에 상기 <수학식 9>는 하기 <수학식 14>와 같이 간단히 나타낼 수 있다. The second step is in Equation 9

Will be calculated. Since the vector values for the first transmitting antenna of the QEGT codebook index are all real numbers, Equation 9 may be simply expressed as Equation 14 below.

는 하기 <수학식 15>에 의해 정의된다.In Equation 14 above

Is defined by Equation 15 below.

은 인덱스가

일 때, 각 인덱스의 위상 각도를 의미한다. 한편 상기 <수학식 15>는

로 일반화시킬 수 있다.here

Has an index

When, it means the phase angle of each index. Meanwhile, Equation 15 is

Can be generalized to

를 계산하는 것이다.The third step is the proximity between the newly defined channel matrix and the index vector matrix of the codebook using Equation 16 below.

Will be calculated.

The final step is to select candidate indices of the codebook using Equation 17 below.

4 shows on the complex plane any candidate indices selected by fast searching in the QEGT codebook.

인 코드 북에서 두 번째 안테나에 대한 벡터 값은 모두 순환적 대칭을 이루기 때문에 가장 작은

값을 가지는 코드 북 인덱스를 후보로 선택 시 후보 인덱스 수는 항상 2가 된다.

인 경우, 시스템 설계자가 정한 적절한 근접 기준 값인 임계 각도

보다 작은

를 선택하는데, 이는

인 QEGT 코드 북들은 첫 번째 송신 안테나를 제외한 나머지 송신 안테나들이 최소 2개 이상의 복소평면에서 표현되기 때문이다. 이때, 임계 (threshold) 각도

를 작은 값으로 선택하면 인덱스 탐색 시간은 감소하지만, 후보 인덱스 수 감소로 인해 성능이 저하된다.As shown in Figure 4 is represented on one complex plane

In the codebook, the vector values for the second antenna are the smallest because they all have circular symmetry.

When selecting a codebook index having a value as a candidate, the number of candidate indexes is always two.

, The critical angle, which is the appropriate proximity reference value determined by the system designer.

lesser

, Which means

This is because the QEGT codebooks are expressed in at least two complex planes except for the first transmit antenna. At this time, the threshold angle

Selecting a small value reduces the index search time, but decreases the performance due to the decrease in the number of candidate indexes.

Meanwhile, when the candidate index is selected by the fast search as described above, the optimal codebook index is selected by using Equation 7, which is an Accurate Searching technique. As described above, when applying the codebook index search excluded in the present invention, NCI obtained by the fast search is used instead of using N, which is the total number of indexes in the codebook as k in Equation (7). Used. Through the proposed scheme, NCI can always be searched faster because its number is always smaller than N.

를 1.5 라디안으로 지정하였고, 전체 시뮬레이션 반복 횟수는 200만 회이다.Diversity performance for fast index search in the new gain gain system is measured using narrow-band BPSK transmission / reception system using Monte Carlo computer simulation. At this time, carrier frequency is 2 GHz, data rate is 9.6 kbps, threshold angle

Is specified in 1.5 radians and the total number of simulation iterations is 2 million.

의 분포로써 준정적 플랫 페이딩을 가지며, 심벌 전송에 있어 채널 특성이 일정하다고 가정한다. 즉, 무선 채널 모델에서의 페이딩 현상은 각 심벌마다 독립적으로 발생한다. 또한 잡음은

의 분포를 가지는 가산성 백색 가우시안 잡음(AWGN)이며, 이 때

은 복소 정규 분포를 의미한다.The wireless channel model used in the computer simulation

It is assumed that the distribution has a semi-static flat fading and that channel characteristics are constant in symbol transmission. That is, fading in the radio channel model occurs independently for each symbol. Also the noise

Additive white Gaussian noise (AWGN) with a distribution of

Denotes a complex normal distribution.

It is assumed that channel estimation and synchronization and correlations between antennas are all ideal.

FIG. 5 shows the performance of codebook C (2,3) by applying a fast index search scheme when the number of transmit antennas is 1 in the gain transmission system.

FIG. 6 shows the performance of codebook C (3,6) by applying a fast index search scheme when the number of transmit antennas is 1 in the gain transmission system.

5 and 6, it can be seen that the transmission diversity gains of applying the fast index search scheme are the same as compared with the ideal dynamic gain transmission scheme and the existing index search scheme.

FIG. 7 shows the performance of codebook C (2,3) by applying a fast index search scheme when the number of transmit antennas is 2 in the same gain transmission system.

FIG. 8 shows the performance of codebook C (3,6) by applying a fast index search scheme when the number of transmit antennas is 2 in the same gain transmission system.

Referring to FIGS. 7 and 8, it can be seen that the transmission diversity gain is almost the same when the fast index search scheme is applied compared to the ideal dynamic gain transmission scheme and the conventional index search scheme.

The performances shown in FIGS. 5 to 8 are results obtained on the assumption that the user's moving speed is infinite. However, the closed loop mobile communication system using the actual codebook is used when the user moves at a low speed of 1 to 60 km / h rather than a high speed, so the user needs to know the performance of 3 km / h or 60 km. In the case of moving at a speed of / h, it is confirmed how much diversity gain can be obtained in a dynamic gain transmission MISO channel system having 2 or 3 transmit antennas and 1 receive antenna.

FIG. 9 shows the performance of the codebook C (2,3) by applying a fast index search scheme when the number of transmit antennas is 2 and the number of receive antennas is the same in the gain-transmitting system considering the moving speed.

FIG. 10 shows the performance of the codebook C (3,6) by applying a fast index search scheme when the number of transmit antennas and the number of receive antennas are 3 in the same gain gain system considering the moving speed.

9 and 10, when the number of transmit antennas is 2 according to the user's moving speed, the same diversity gain is obtained as compared with the conventional index search scheme, and when the number of transmit antennas is 3, the performance is higher than that of the conventional index scheme. It can be seen that this is about 0.1 to 0.7 dB lower.

이 증가할수록 QEGT 코드 북의 탐색 시간이 기존의 인덱스 탐색 시간에 비해 현저히 감소함을 알 수 있다.11 is a comparison of the operation time of the conventional index vector search method and the fast index search method proposed by the present invention for various QEGT code books. As the code book size increases, that is, the number of code book index bits

As the number increases, the search time of the QEGT codebook is significantly reduced compared to the existing index search time.

Table 2 below shows the average candidate index number NCI for each codebook when the QEGT codebook is used in a dynamic gain transmission MISO channel having a receiving antenna of 1 using the fast index search scheme proposed by the present invention. This indicates that NCI is smaller than the number of indexes N in the conventional index search technique.

에 대해 송신 안테나가 3 이상인 경우, 그 값이 커지면 NCI는 증가하는 대신 SRI는 감소하고, 그 값이 작아지면 NCI는 감소하는 대신 SRI가 증가함을 보인다. Referring to Table 2, it can be seen that when the fast index search method is used through the SRI (search time reduction indicator), the overall index search time is greatly reduced. In addition, the optimization phase boundary angle determined in advance by the flexible relationship between system performance and search complexity

If the transmit antenna is greater than or equal to 3, as the value increases, the SRI decreases instead of increasing the NCI, and when the value decreases, the SRI increases instead of decreasing the NCI.

값을 1 또는 1.5로 설정한 경우의 후보 인덱스 수와 탐색 시간 감소율을 보인다. 만약 시스템 설계자가 최적화 위상 경계 각도를 적절히 선택한다면 보다 유연한 성능을 얻을 수 있다.<Table 2> is

When the value is set to 1 or 1.5, the number of candidate indexes and the decrease in search time are shown. If the system designer chooses the optimal phase boundary angle properly, more flexibility is achieved.

On the other hand, but shown and described with respect to the preferred embodiment of the present invention, the present invention is not limited to the specific embodiments described above, in the technical field to which the invention belongs without departing from the spirit of the invention claimed in the claims Various modifications can be made by those skilled in the art, and these modifications should not be individually understood from the technical spirit or the prospect of the present invention.

1 is a block diagram of a closed loop mobile communication system for performing a fast index search according to an embodiment of the present invention;

과

을 복소평면 상에 보이고 있는 도면;2 is a QEGT codebook

and

Showing on a complex plane;

3 is a diagram illustrating a control flow for fast index search according to an embodiment of the present invention;

4 shows on the complex plane any candidate indices selected by fast searching in a QEGT codebook;

FIG. 5 is a diagram showing the performance of codebook C (2,3) by applying a fast index search scheme when the number of transmitting antennas is 1 in the same gain transmission system; FIG.

FIG. 6 is a diagram showing the performance of codebook C (3,6) by applying a fast index search scheme when the number of transmitting antennas is 1 in the same gain transmission system; FIG.

7 is a diagram illustrating performance of codebook C (2,3) by applying a fast index search scheme when the number of transmit antennas is 2 in the same gain transmission system;

8 is a diagram showing the performance of codebook C (3,6) by applying a fast index search scheme when the number of transmit antennas is 2 in the same gain transmission system;

9 is a diagram illustrating the performance of codebook C (2,3) by applying a fast index search scheme when the number of transmit antennas is 2 and the number of receive antennas in the same gain-transmitting system considering the moving speed;

FIG. 10 is a diagram illustrating performance of codebook C (3,6) by applying a fast index search scheme when the number of transmit antennas is 3 and the number of receive antennas is 1 in the gain gain system considering the moving speed; FIG.

11 is a view showing a comparison of the operation time of the conventional index vector search method and the fast index search method proposed in this paper for various QEGT codebooks.

Claims (18)

A codebook index search method by the receiver in a multi-antenna system supporting a closed loop scheme and a transmitter having a plurality of transmit antennas and a receiver having a plurality of receive antennas communicate based on a predetermined codebook. , Generating a candidate index vector group by selecting candidate indexes for searching a codebook index among all indexes of the codebook; Performing a codebook index search by the generated candidate index vector group; The generating of the candidate index vector group may include compensating phase angles of the remaining columns except for the first column corresponding to the first transmitting antenna in the channel characteristic matrix estimated from the signals received through the plurality of receiving antennas. Defining a function of the feature matrix, Calculating a function of the index vector matrix of the codebook by compensating the phase angles of the vector values of the indices corresponding to the remaining transmit antennas except the first transmit antenna of the codebook; The absolute value of the difference between the function of the newly defined channel characteristic matrix and the function of the index vector matrix of the codebook corresponds to each of the remaining transmit antennas except the first transmit antenna among all indexes of the codebook. Calculating the proximity of the index, Generating a candidate index group by a candidate index that satisfies any predefined proximity reference value among the calculated proximity, wherein the candidate index is any predefined proximity among all the indexes. Codebook index search method characterized in that the proximity satisfying the value corresponds to the measured index. The method of claim 1, wherein the generating of the candidate index vector group comprises: A function of a new matrix is generated according to Equation 1 below, and the candidate index vector group is selected by a candidate index selected from all indices of the codebook by applying the generated new matrix function to Equation 2 below.

Codebook index search method. [Equation 1]

[Equation 2]

here

Is a function of the new matrix,

Is any predefined proximity criterion value,

Is the new channel matrix,

(

) Is the index of the transmitting antenna,

(

) Is the index of the receiving antenna,

silver

The phase angle of. delete The method of claim 1, wherein the defining of the function of the new channel characteristic matrix comprises: Acquiring a phase error angle required to convert a complex vector value corresponding to the first transmit antenna into a real number having a phase angle of 0 in a channel characteristic matrix estimated from signals received through the plurality of receive antennas; , A function of the new channel characteristic matrix is obtained by a new channel characteristic matrix which compensates for the phase angle of each of the vector values corresponding to the remaining transmit antennas except for the first transmit antenna among the estimated channel characteristic matrix based on the acquired phase error angle. Codebook index search method comprising the step of defining. The method of claim 1, Function of the new channel characteristic matrix

silver

Is defined as where

silver

ego,

Is the phase angle of the transmit / receive fading channel elements,

Is a new channel characteristic matrix which compensates for the phase angle. 6. The method according to claim 1 or 5, The index vector matrix of the codebook

silver

Is calculated by where

silver

ego,

Is a phase angle of each index in the codebook. The method of claim 6, The candidate index group is

Is generated by Where PRX is the calculated proximity;

Is any predefined proximity criterion value,

Is a total number of transmitting antennas. The method according to claim 1 or 2, The codebook index search by the generated candidate index vector group

Is made by here

Is the index of the codebook, NCI is the total number of candidate indices in the generated candidate index vector group,

Codebook index search method, characterized in that 1-norm (norm). The method of claim 7, wherein The codebook index search by the generated candidate index vector group

Is made by here

Is the index of the codebook, NCI is the total number of candidate indices in the generated candidate index vector group,

Codebook index search method, characterized in that 1-norm (norm). In the receiver for supporting a closed loop scheme, the transmitter having a plurality of transmit antennas and the receiver having a plurality of receive antennas to search the codebook index in a multi-antenna system to communicate based on a predetermined codebook, A candidate information vector group is generated by selecting candidate indices for searching a codebook index among all indexes of the codebook, and includes a feedback information generator for performing a codebook index search by the generated candidate index vector group; , The feedback information generator compensates the phase angles of the remaining columns except for the first column corresponding to the first transmitting antenna in the channel characteristic matrix estimated from the signals received through the plurality of receiving antennas, thereby calculating a function of the new channel characteristic matrix. Define, Compute the function of the index vector matrix of the codebook by compensating the phase angle for the vector values of the index corresponding to the other transmit antennas except the first transmit antenna of the codebook, The absolute value of the difference between the function of the newly defined channel characteristic matrix and the function of the index vector matrix of the codebook corresponds to each of the remaining transmit antennas except the first transmit antenna among all indexes of the codebook. Calculate the proximity of the index, A candidate index group is generated by a candidate index that satisfies any predefined proximity reference value among the calculated proximity, wherein the feedback information generator satisfies any predefined proximity reference value among all the indexes And selecting an index of which proximity is measured as a candidate index. The method of claim 10, wherein the feedback information generation unit, A function of a new matrix is generated by Equation 3 below, and the candidate index vector group is selected by a candidate index selected from all indices of the codebook by applying the generated new matrix function to Equation 4 below.

Receiver) characterized in that for generating a). &Quot; (3) &quot;

&Quot; (4) &quot;

here

Is a function of the new matrix,

Is any predefined proximity criterion value,

Is the new channel matrix,

(

) Is the index of the transmitting antenna,

(

) Is the index of the receiving antenna,

silver

The phase angle of. delete The method of claim 10, wherein the feedback information generation unit, Obtaining a phase error angle necessary for converting a vector value of a complex number corresponding to the first transmitting antenna into a real number having a phase angle of 0 in a channel characteristic matrix estimated from signals received through the plurality of receiving antennas, A function of the new channel characteristic matrix is obtained by a new channel characteristic matrix which compensates for the phase angle of each of the vector values corresponding to the remaining transmit antennas except for the first transmit antenna among the estimated channel characteristic matrix based on the acquired phase error angle. Receiving device, characterized in that. The method of claim 10, wherein the feedback information generation unit,

Function of the new channel characteristic matrix by

Defines the Where above

silver

ego,

Is the phase angle of the transmit / receive fading channel elements,

Is a new channel characteristic matrix which compensates for the phase angle. The method of claim 10 or 14, wherein the feedback information generating unit,

The index vector matrix of the codebook by

, Where above

silver

ego,

Is a phase angle of each index in the codebook. The method of claim 15, wherein the feedback information generation unit,

The candidate index group is generated by Where PRX is the calculated proximity;

Is any predefined proximity criterion value,

Is a total number of transmitting antennas. The method of claim 10 or 11, wherein the feedback information generating unit,

Perform a codebook index search by the generated candidate index vector group by here

Is the index of the codebook, NCI is the total number of candidate indices in the generated candidate index vector group,

Receiver is characterized in that 1-norm (norm). The method of claim 16, wherein the feedback information generation unit,

Perform a codebook index search by the generated candidate index vector group by here

Is the index of the codebook, NCI is the total number of candidate indices in the generated candidate index vector group,

Receiver is characterized in that 1-norm (norm).

Images