KR20130118111A - Method for assigning index of codebook in multiple antenna system, and recording medium thereof - Google Patents

Abstract

The present invention relates to a method for arranging codebook indexes in a multi-antenna system in which a transmitter having a plurality of transmit antennas and a receiver having a plurality of receive antennas communicate on a predetermined codebook in an incorrect feedback channel. Specifying a size (M) of a reference codebook, p is an integer, and when a total number of candidate reference codebooks that can be generated is N, a method of deleting one beamforming matrix from a child codebook of size N Generating N candidate reference codebooks having a size of N-1, measuring a minimum relative distance between the N candidate reference codebooks, and referring to a candidate reference codebook having the largest value among the minimum relative distances as an optimal candidate. Designating as a codebook, calculating a size (N _ref ) of the best candidate reference codebook and the best candidate If the size (N _ref ) of the reference codebook is the same as the size (M) of the reference codebook to be generated, designating the optimal candidate reference codebook as a reference codebook. According to the present invention, it shows an adaptive characteristic free from the size or selective constraints of the existing parent codebook, and has an effect of obtaining an improved performance gain compared to the existing scheme in an incorrect feedback channel.

Description

Method for assigning index of codebook in multiple antenna system and recording medium

The present invention uses a chordal distance between beamforming matrices in a codebook to have enhanced characteristics in an incorrect feedback channel for a line coded transmission method using a codebook in a closed circuit-based mobile communication system using multiple transmit antennas. An improved grouping-based codebook index placement technique having adaptive characteristics by measuring relative distances.

The precoding method using multiple antennas has the advantage of maximizing spatial diversity gain and reducing interference of other signals. However, since the pre-coding method requires feedback of channel state information (CSI) from the receiving end to the transmitting end, it causes an overhead of infinite precision.

In order to solve this problem, a pre-coding transmission method using a set of quantized channel matrices defined as a codebook is mainly used, and a very efficient and optimized random channel state based on a system measure such as information capacity of a communication system is used. An information precoding matrix is mapped to a plurality of indices constituting the codebook.

On the other hand, the effect of feedback error in an incorrect feedback channel causes a problem of system performance degradation and an outage area increase due to incorrect index delivery in a codebook.

To solve this problem, the index assignment (IA) technique is very simple and robust to feedback errors.

The index placement technique has the advantage of not requiring additional feedback bits or signal processing compared to other feedback error prevention techniques such as error control coding.

The index placement technique that can be applied according to the codebook size can be classified into the group-based IA (GIA) technique using global search and grouping. To apply the technique, a specific codebook named 'parent codebook' of size M (M <N) and having the same number of transmit antennas is necessary.

Existing researches used codebooks defined in standardization documents such as WiMAX or codebooks based on Grassmannian beamforming-based generation as parent codebooks, but codebooks that can be selected as parent codebooks are limited depending on the number of transmit antennas and codebook size M. There is this.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems. In consideration of an inaccurate feedback channel necessarily occurring in an actual wireless channel, noses between matrices in a codebook to extend a range of codebook selection in an indexing scheme using grouping (chordal) Produces a codebook of desired size using the optimal codebook design technique according to distance, provides an improved grouping-based codebook index placement method using the index relocation method using grouping, and provides a new line coding method using the same Its purpose is to.

The objects of the present invention are not limited to the above-mentioned objects, and other objects not mentioned can be clearly understood by those skilled in the art from the following description.

In order to achieve the above object, the present invention provides a method for arranging codebooks in a multi-antenna system in which a transmitter having a plurality of transmit antennas and a receiver having a plurality of receive antennas communicate on a predetermined codebook in an incorrect feedback channel. In the step of specifying the size (M) of the reference codebook to be generated, p is an integer, when the total number of candidate reference codebooks that can be generated is N, one of the child codebook of size N Generating N candidate reference codebooks having a size of N-1 by deleting a beamforming matrix, measuring a minimum relative distance between the N candidate reference codebooks, and having the largest value among the minimum relative distances designating the candidate reference codebook optimally reference candidate codebook, the (N _ref) the size of the reference candidate optimal codebook When under the same phase and magnitude (M) of the reference codebook to be generated with the (N _ref) The size of the reference candidate optimal codebook, and a step of specifying the reference candidate optimal codebook by the reference codebook.

If the size (N _ref ) of the best candidate reference codebook and the size (M) of the reference codebook to be generated are not the same, the size (N _ref ) of the best candidate reference codebook is the total number of candidate reference codebooks (N). The method may further include designating a new child codebook, and generating the candidate reference codebook to calculating the size of the reference codebook until N _ref = M.

The minimum relative distance between the N candidate reference codebooks may be determined by using a chordal distance representing a relative distance of a subspace belonging to a Grassmannian manifold. have.

When k and l are integers satisfying k ≠ l, 1≤k, l≤N-1, the nose distance is

It can be calculated using the following equation.

Given the minimum nose distance δ _min , the minimum nose distance

It can be calculated using the following equation.

In the step of designating a candidate reference codebook having the largest value among the minimum relative distances as an optimal candidate reference codebook, the p _opt- th candidate codebook having the largest δ _min among the minimum chordal distances δ ^p _min is the best candidate reference. Decide on a codebook, May be applied.

According to the present invention, the relative distance using the chordal distance between beamforming matrices in the codebook is compared with the indexing scheme using the conventional grouping, which has limited requirements for determining the parent codebook including the number of transmit antennas and the codebook size. By measuring and generating the reference codebook of the desired size from the child codebook, it shows the adaptive characteristics free from the size or the selective constraints of the existing parent codebook, and has the effect of improving the performance gain compared to the conventional technique in the inaccurate feedback channel. have.

1 is an exemplary view showing a closed loop multiple antenna system using a codebook supporting multiple users according to an embodiment of the present invention.
2 is an explanatory diagram schematically illustrating a generation principle of a reference codebook according to an embodiment of the present invention.
3 is a conceptual diagram illustrating a block diagram of a reference codebook generation principle according to an embodiment of the present invention.
4 is an exemplary diagram comparing a spatial representation of a codebook designed according to a minimum chordal distance according to an embodiment of the present invention.
FIG. 5 is a graph showing QPSK modulation and demodulation performance by applying an improved grouping-based index arrangement scheme proposed in a closed loop multi-antenna system using a codebook supporting multiple users according to an embodiment of the present invention.
6 is a graph comparing QPSK modulation and demodulation performance by applying an index placement technique using a conventional grouping and an improved grouping-based index arrangement technique proposed by the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the drawings, the same reference numerals are used for the same reference numerals even though they are shown in different drawings. In the following description of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear. Also, throughout this specification, when a component is referred to as "comprising ", it means that it can include other components, aside from other components, .

In the following description, specific details of an improved grouping-based codebook index placement technique in a multi-antenna system using the codebook of the present invention are presented to provide a more comprehensive understanding of the present invention, without these specific details and also in their variations. It will be apparent to those skilled in the art that the present invention can be easily implemented by the present invention.

In the following description, the general process of mapping codebook indices in a transmitter according to feedback information and a process of index arrangement using global search and grouping in a multi-antenna system supporting a closed loop scheme are described below. We will look at it.

In addition, an improved grouping-based codebook index with more adaptive characteristics is applied by applying the optimal codebook design method according to the chordal distance between vectors in the codebook to the index placement technique that is robust to inaccurate feedback channel. Let's take a look at the process of placement techniques.

First, a general process of mapping a codebook index in a transmitter according to feedback information in a multi-antenna system supporting a closed loop scheme will be described.

The following notation is used to facilitate easy understanding of the present invention. u (m, n) is a set of m × n matrices of orthogonal columns, I _M is an identity matrix of size M × M, bold identifiable capital letter A is a matrix, bold spermatic small letter a is Vector, A ^H is the conjugate transposition of matrix A,

Is a complex matrix with m rows and n columns, and

Means Frobenius norm.

In a simple feedback system based on a multi-antenna with a number of receiving antennas M _R and a number of transmitting antennas M _T , when a layer number _R ≤ min ( M _T , M _R ), which means a virtual layer transmitting independent data in a spatial domain, Transmission signal

Is a normalized precoding matrix

Transmitted after being multiplied by and having an iid (independent and identically distributed) characteristic

The received baseband signal y is given by Equation 1 below.

At this time,

Is an additive white gaussian noise (AWGN) vector component. The amount of feedback information for transmit beamforming increases linearly as the number of taps of the channel, the number of transmit antennas, and the number of users increases, which acts as an overhead for the overall system. To solve this problem, a codebook, which is a set of quantized precoding matrices, is used to compress and transmit feedback information to a minimum within a range where performance degradation does not occur. Codebook W is W =

,

Where N is the total number of indexes in the codebook.

1 is an exemplary view showing a closed loop multiple antenna system using a codebook supporting multiple users according to an embodiment of the present invention.

In FIG. 1, one transmitter 110 and a plurality of receivers 120-1 and 120 -M are included. In the following description, it is based on one receiver 120-1. The following description is regarded to apply equally to the remaining receiver.

The receiver 120-1 generates feedback information based on the channel state information corresponding to each data stream. The generation method of the feedback information may be determined in consideration of a signal detection method, whether the codebook 116 is used at the transmitting side, and the like.

The receiver 120-1 transmits the generated feedback information. In this case, the feedback information is a channel quality indicator (CQI) indicating the status information of the channel, the spatial multiplexing method using the channel matrix is applicable, and how many rank (rank) can be sent if possible? A rank indicator (RI) for informing, and a precoding matrix indicator (PMI), that is, an index, which helps the transmitter's feedback information processing unit 114 to set the precoding matrix well, may be included.

The transmitter 110 receives feedback information from all receivers 120-1 and 120 -M. 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 single user mode, and multiple users are selected in 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, the method of detecting a signal at the receiving side, an operation mode, and the number of data streams supporting the selected user may be further considered.

The transmitter 110 designates 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 -M. 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 #M _R ), a channel estimator 122-1, and a feedback information generator 124-1. It consists of. The transmitter 110 is composed of a plurality of transmit antennas (Ant_tx # 1, Ant_tx # 2, ..., Ant_tx #M _T ), a feedback information processor 114 and a signal transmitter 112.

Referring to the receiving apparatus 120-1, a signal received from at least one receiving antenna Ant_rx # 1, Ant_rx # 2,..., Ant_rx #M _R is input to the channel estimating unit 122-1. The channel estimator 122-1 estimates channel characteristics corresponding to each data stream from the received signal by a preset signal detection method. In this case, it is assumed that the channel characteristics are quasi-static flat fading with CN (0,1) distribution, and that the channel characteristics are constant in symbol transmission. That is, it is assumed that fading in the radio channel model occurs independently for each symbol.

As a signal detection method, a linear detection method and a nonlinear detection method exist. A representative example of the linear detection method is the LMMSE method, and a representative example of the nonlinear detection method is the SIC method. The signal detection method 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 method 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 the feedback information by the channel characteristic matrix provided from the channel estimator 122-1. When generating the feedback information, the feedback information generator 124-1 considers the signal detection method 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 the codebooks 116 and 126-1 are used. . If the transmitter 110 uses the codebook 116, the feedback information includes a precoding matrix indicator, that is, an index.

The receiver 120-1 transmits the feedback information generated by the feedback information generator 124-1 to the transmitter 110. The feedback information is preferably transmitted periodically by the receiver 120-1. However, if the elements for determining the transmission time point are promised in advance between the transmitter 110 and the receiver 120-1, it is possible to transmit the feedback information aperiodically.

Referring to the transmitter 110, feedback information received from all receivers 120-1 and 120 -M 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. In addition, an encoding method and an MCS level for transmitting a data stream of at least one selected user are determined.

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

In order to determine the K _opt (0 # k N-1) th index in the optimal codebook fed back using the channel information H estimated at the receiver to maximize the overall system performance, the feedback information processor 114 selects the optimal codebook index. Decisions are made using techniques. A representative optimal codebook index selection technique uses the received signal-to-noise ratio (SNR) or channel capacity.

The selection technique using the received SNR is determined by Equation 2. At this time, k is an integer.

The selection technique using the channel capacity is determined by Equation 3. Is the estimated received SNR.

Typically, feedback is achieved by selecting the appropriate size codebooks 116 and 126-1 that are optimized for generation efficiency and storage memory.

The following describes the general index placement method through the index placement technique using global search and grouping.

In order to solve this problem, the feedback error is caused by incorrect index propagation, resulting in system performance degradation and increased communication outage area. Representative researches include the FSMC-based search method using the characteristics of the finite state markov channel (FSMC) model, the Bayesian search method using the characteristics of the average signal-to-noise ratio, and the adaptive diversity method.

In addition to the above research method, the index placement technique is very simple and robust to feedback errors. Index placement techniques have been developed from conventional vector quantization designs and have the advantage of not requiring additional feedback bits or signal processing compared to other feedback error prevention techniques, such as error control coding. The index placement technique that can be applied according to the codebook size can be classified into the index placement technique using global search and grouping. First, the index placement technique through global search will be described.

If the size of the index constituting the codebook is less than 3 bits, the index placement technique using the global search technique is applied. In an actual system, since the probability of an error occurring more than 2 bits is very small, the present invention assumes that no error exceeding a maximum of 1 bit per symbol is generated. For a codebook consisting of L-bit indexes, index placement through global search is applied through the following steps.

Step 1) Generate the number of all cases for the codebook that can be generated in consideration of the arrangement order of the indices. At this time, the number of codebooks that can be generated is 2 ^L !.

Step 2) Through Equation 4, the cost function C ^l , l = 1, ..., 2 ^L ! According to the hamming distance of the bits constituting the index in each codebook generated in Step 1) is obtained. Calculate each.

In this case, W ^l _i and W ^l _j are precoding matrices corresponding to the i and j th indexes in the l th codebook. Hamming distance between i th index i ^l _i and j th index I ^l _j in the l th codebook

According to I (

) Is determined by Equation 5. In other words, if the Hamming distance between two indices is 1, I (

) Is 1; otherwise, I (

) Is defined as 0.

Step 3) Through Equation 4, the l-th codebook having the largest cost function is determined as the codebook to which the optimal index placement technique is applied through global search.

In order to apply the global search-based index placement technique for codebooks consisting of L-bit indexes, 2 ^L ! You need two codebooks. If the codebook size is 16 or more, 2 ⁴ !> 10 ¹³ codebooks have to be generated in step 1), which is not systematically possible. For this, we use index placement using grouping, which is one of the semi-optimal techniques when the codebook size is large.

The following describes the process of index placement using grouping.

Index placement using grouping uses a child codebook of size N C _ch = [ W ₀ , W ₁ , ..., W _{N -1} ] and a global search of size M (M <N). It is constructed using the following steps using a parent codebook C _p = [ C ₀ , C ₁ , ..., C _{M −1} ] to which the index placement technique is applied.

Step 1) Create M empty groups.

Step 2) Specify m = 0.

Step 3) Perform l- [m] _M. At this time,

Is a modulo-M operation.

Step 4) Finding the inner product of the first coding matrix in all the child codebooks with respect to the l th beamforming matrix C _l in the parent codebook, and then applying the beamforming matrix in the child codebook having the largest dot product to the first group. Include it.

Step 5) The remaining beamforming vectors except for the beamforming matrix in the child codebook selected in step 4) are designated as the new child codebook.

Step 6) Specify m = m + 1. Steps 3) to 6) are repeated until m <M is satisfied.

Step 7) Precoding matrices in the N / M child codebooks are arranged in each of the M groups. Map the indexes in order according to the group order. At this time, the order of arranging the precoding matrices in each group is arbitrarily determined because the influence on the overall performance is very small. Through this, the codebook to which the optimal index placement technique is applied is determined.

Parent codebook designation is necessary to apply the index placement technique using grouping. In the previous research, codebooks defined in standardized documents such as WiMAX or grassmannian beamforming-based generation were used, but the number of transmit antennas and codebook size M The choice of codebooks was limited. Codebooks defined in recent standardized documents have only 2, 4, and 8 transmit antennas, and their sizes are limited to 8, 16, 32, etc., depending on the number of transmit antennas. If a codebook is not specified in the standardized document, a codebook through generation of Grassmannian beamforming-based method should be used. Therefore, the requirements for determining the parent codebook are limited.

The improved grouping-based index placement technique (enhanced GIA, EGIA) proposed by the present invention determines the parent codebook in the child codebook, and can be simply configured without limitations according to the number of transmit antennas and size M. The proposed grouping-based index placement scheme newly defines a reference codebook instead of a parent codebook.

2 is an explanatory diagram schematically illustrating a generation principle of a reference codebook according to an embodiment of the present invention.

Referring to FIG. 2, selecting precoding matrices that minimize the impact on overall performance while continuing to remove the precoding matrix corresponding to one index from the child codebook of size N to the reference codebook of size M. Indicates.

3 is a conceptual diagram illustrating a block diagram of a reference codebook generation principle according to an embodiment of the present invention.

Referring to FIG. 3, if a size of a reference codebook to be generated is specified (S310), a reference codebook is generated through the following steps.

Step 1) By generating one beamforming matrix from the child codebook of size N, candidate codebooks F ^' _p and p (1≤p≤N) having a size of N-1 are generated (S312). At this time, p is an integer, and the total number of candidate reference codebooks that can be generated is N.

Step 2) In order to select an optimal candidate reference codebook among N candidate reference codebooks, a chordal distance closely related to the performance of the codebook is used in a Grassmannian subspace packing related study. The chordal distance, which represents the relative distance of the subspace belonging to the Grassmannian manifold, is calculated using Equation 6. At this time, k and l are integers satisfying k ≠ l, 1≤k, l≤N-1.

When the minimum chordal distance δ _min is defined through Equation 7, an optimal codebook in the Grassmannian subspace packing technique is designed to maximize the minimum chordal distance (S314).

A method of arranging the codebook index of the present invention described above with reference to FIG. 3 is as follows.

In a method for arranging codebook indexes in a multi-antenna system in which a transmitter having a plurality of transmit antennas and a receiver having a plurality of receive antennas communicate in an incorrect feedback channel based on a predetermined codebook, a reference codebook to be generated codebook) size M is specified (S310).

When p is an integer and the total number of candidate reference codebooks that can be generated is N, N candidate reference codebooks having a size of N-1 are generated by deleting one beamforming matrix from a child codebook of size N. (S312).

The minimum relative distance between the N candidate reference codebooks is measured using the nose length (S314).

A candidate reference codebook having the largest value among the minimum relative distances is designated as an optimal candidate reference codebook (S316).

The size N _ref of the best candidate reference codebook is calculated (S318).

The size N _ref of the best candidate reference codebook is compared with the size M of the reference codebook to be generated (S330).

If the size N _ref of the best candidate reference codebook and the size M of the reference codebook to be generated are the same, the best candidate reference codebook is designated as the reference codebook (S322).

If the size of the best candidate reference codebook (N _ref ) and the size of the reference codebook (M) to be generated are not the same, the new child codebook using the size of the best candidate reference codebook (N _ref ) as the total number of candidate reference codebooks (N). It is specified as (S320).

Then, steps S312 to S330 are repeated until N _ref = M.

4 is an exemplary diagram comparing a spatial representation of a codebook designed according to a minimum chordal distance according to an embodiment of the present invention.

FIG. 4 illustrates a Voronoi diagram according to a minimum chordal distance when codewords corresponding to an index in an arbitrary codebook of size 6 in a two-dimensional Grassmannian manifold, designed according to a minimum chordal distance. An example diagram comparing the spatial representation of codebooks.

Referring to FIG. 4, the Voronoi diagram shown in FIG. 4- (a) shows that the codewords are distributed evenly in all subspaces in the two-dimensional Grassmannian manifold in order to successfully operate on various channels, but smaller than in FIG. 4- (a). It can be seen that in Fig. 4- (b) with the minimum chordal distance, there is an uneven spatial distribution. Therefore, among the N candidate reference codebooks generated in step 1), an optimal codebook in which quantization distortion for CSI can be minimized is the minimum chordal distance δ ^p _min of each candidate reference codebook by Equation (8). The p _opt th candidate codebook having the largest δ _min among the files is determined as the best candidate reference codebook (S316).

Step 3) When the size of the best candidate reference codebook selected in step 2) is N _ref (S318), when N _ref ≠ M, the selected best candidate reference codebook is designated as a new child codebook having a size of N = N _ref (S320). ).

Step 4) Repeat steps 1) to 3) until N _ref = M.

Step 5) When M _ref = M, the codebook finally selected in Equation 8 is designated as the final reference codebook (S322).

The above-described global search based index placement scheme is applied to the determined reference codebook. The subsequent process is the same as the existing index placement technique using grouping, except that the reference codebook is used instead of the parent codebook.

The improved grouping-based index placement scheme proposed in the present invention has an extended adaptive characteristic than the existing parent codebook selection scheme in generating a reference codebook. That is, since a reference codebook of a desired size can be generated from the child codebook, it has an advantage of deviating from the size or selective constraint of the parent codebook, which was limited in the existing research.

In order to accurately measure the performance of the improved grouping-based index placement scheme proposed by the present invention, a bit error rate (BER) is measured and measured in a system environment with feedback errors through Monte-Carlo-based simulations. Analyzed. In the long term evolution (LTE) based system model, the radio channel model has a semi-static flat fading with CN (0,1) distribution and assumes that channel characteristics are constant in symbol transmission. That is, fading in the radio channel model occurs independently for each symbol. The power delay profile (PDP) for the multipath channel is Pedestrian A of the ITU-R model, and the codebook specified in the LTE and WiMAX standard documents is used. The channel estimation and synchronization and the correlation between the antennas are all assumed to be ideal, and the noise uses additive white Gaussian noise with a distribution of CN (0, N ₀ ).

FIG. 5 is a graph showing QPSK modulation and demodulation performance by applying an improved grouping-based index arrangement scheme proposed in a closed loop multi-antenna system using a codebook supporting multiple users according to an embodiment of the present invention.

Referring to FIG. 5, when a 4-bit 3GPP LTE 4Tx rank 1 codebook (hereinafter, 4-bit LTE codebook) is applied to a multi-antenna based orthogonal frequency division multiplexing (OFDM) system having 4 transmit antennas and 2 receive antennas, feedback error probability (Feedback BER) of 10 ^- if the ^second BER is to apply an improved group-based index of the arrangement scheme ( 'LTE using CB proposed GIA' notation) to offer to the codebook in LTE ^10-3 index placement method 0.5 dB gain is obtained when is not applied ('LTE CB without IA' notation). In addition, the feedback error probability ^{10- 3} when the BER is ^10-4 obtained in a 0.5 dB gain. That is, when the feedback error occurs, it can be confirmed that the application of the technique proposed in the present invention provides a significant gain.

6 is a graph comparing QPSK modulation and demodulation performance by applying an index placement technique using a conventional grouping and an improved grouping-based index arrangement technique proposed by the present invention.

Referring to FIG. 6, in order to apply the existing GIA technique ('Conventional GIA'), a parent codebook should be defined. In the LTE codebook, since there are no codebooks consisting of 3 bits when the number of transmitting antennas is 4, the table in the WiMAX standard document The 3-bit WiMAX codebook V (4,1,3) described in 298o was defined as the parent codebook. The proposed technique always generates 1-bit feedback error or 10 ^-2 feedback error probability, and gains 0.1 to 0.2 dB over the conventional indexing scheme. This suggests that the proposed scheme, which can determine the reference codebook within the child codebook, improves the performance gain, compared to the indexing scheme using the existing grouping, in which the selectable parent codebook is limited according to the number of transmit antennas and the codebook size. Can be.

On the other hand, the codebook generation method according to an embodiment of the present invention can also be implemented as computer-readable code on a computer-readable recording medium. A computer-readable recording medium includes all kinds of recording apparatuses in which data that can be read by a computer system is stored.

For example, the computer-readable recording medium may be a ROM, a RAM, a CD-ROM, a magnetic tape, a hard disk, a floppy disk, a removable storage device, a nonvolatile memory, , Optical data storage devices, and the like, as well as carrier waves (for example, transmission over the Internet).

In addition, the computer readable recording medium may be distributed and executed in a computer system connected to a computer communication network, and may be stored and executed as a code readable in a distributed manner.

While the present invention has been described with reference to several preferred embodiments, these embodiments are illustrative and not restrictive. It will be understood by those skilled in the art that various changes and modifications may be made therein without departing from the spirit of the invention and the scope of the appended claims.

110 Transmitter 120-1, 120-M Receiver
112 Signal transmitter 114 Feedback information processor
116, 126-1 Codebook 122-1 Channel Estimator
124-1 Feedback Information Generator

Claims (7)

A method for arranging codebook indexes in a multi-antenna system in which a transmitter having a plurality of transmit antennas and a receiver having a plurality of receive antennas communicate in an incorrect feedback channel based on a predetermined codebook,
Designating a size M of a reference codebook to be generated;
When p is an integer and the total number of candidate reference codebooks that can be generated is N, N candidate reference codebooks having a size of N-1 are generated by deleting one beamforming matrix from a child codebook of size N. Making;
Measuring a minimum relative distance between the N candidate reference codebooks;
Designating a candidate reference codebook having the largest value among the minimum relative distances as an optimal candidate reference codebook;
Calculating a size (N _ref ) of the best candidate reference codebook; And
And designating the best candidate reference codebook as a reference codebook if the size (N _ref ) of the best candidate reference codebook and the size (M) of the reference codebook to be generated are the same.
The method of claim 1,
If the size (N _ref ) of the best candidate reference codebook and the size (M) of the reference codebook to be generated are not the same, the size (N _ref ) of the best candidate reference codebook is the total number of candidate reference codebooks (N). To designate a new child codebook,
Repeating the step of generating the candidate reference codebook to calculating the size of the reference codebook until N _ref = M.
The method of claim 1,
Measuring the minimum relative distance between the N candidate reference codebook,
And a minimum relative distance between the N candidate reference codebooks using a chordal distance representing a relative distance of a subspace belonging to a Grassmannian manifold.
The method of claim 3,
When k and l are integers satisfying k ≠ l, 1≤k, l≤N-1, the nose distance is

Codebook index arrangement method, characterized in that calculated using the equation.
5. The method of claim 4,
Given the minimum nose distance δ _min , the minimum nose distance

Codebook index arrangement method, characterized in that calculated using the equation.
The method of claim 5,
Designating a candidate reference codebook having the largest value among the minimum relative distances as an optimal candidate reference codebook,
And determine the most p _opt second candidate codebooks having a larger δ from the minimum _min kodeul (chordal) distance δ ^p _min to see optimal candidate codebook,

Codebook index placement method, characterized in that the equation is applied.
A computer-readable recording medium having recorded thereon a program capable of executing the method of any one of claims 1 to 6.

Images