KR101599532B1 - Method and apparatus for generating a mimo (multiple input multiple output) codebook - Google Patents

Abstract

The present invention relates to a codebook-based precoding technique, wherein specific codebook entries used in a lower dimension (i.e., having a relatively small number of antennas) MIMO systems are higher dimensionally (i. e., with a relatively large number of antennas) in a MIMO system. That is, it aims at providing a method of reducing the required memory space of a mobile station (UE or MS) by enabling a codebook design which is easier by generating a high standard codebook based on a low standard codebook.

Description

[0001] METHOD AND APPARATUS FOR GENERATING A MIMO [0002] MULTIPLE INPUT MULTIPLE OUTPUT [0003] CODEBOOK [

The present invention relates to a codebook-based precoding technique, wherein specific codebook entries used in a lower dimension (i.e., having a relatively small number of antennas) MIMO systems are higher (i. e., having a relatively large number of antennas). < / RTI >

The multiple-access (or multiple-access) or MIMO (Multiple Input Multiple Output) technique is known in the art. That is, a technique for improving the wireless communication performance by using a plurality of (multiple) antennas in both the transmitting end and the receiving end. That is, using multiple antennas can increase link range and data throughput without additional bandwidth or transmit power.

In the prior art, multiple input multiple output (MIMO) processing procedures have been problematic due to complex computation and overall processing burden, so the prior art does not adequately address these issues and does not provide a suitable solution.

The present inventors have recognized the problems of the prior art and invented various features to be described later to enable MIMO processing procedures to be performed more efficiently and effectively. a codebook based precoding technique can be adapted to situations requiring a larger number of transmit antennas in a MIMO system. As a result, complex computation and processing burden can be minimized.

In order to solve the above problem, a method of constructing a codebook related to a plurality of antennas on a multiple input multiple output (MIMO) system, includes the steps of: connecting a codebook related to a MIMO system having four transmission antennas wherein the codebook is stored in a storage medium and the codebook has entries for rank 1 to rank 4 and each rank is assigned a specific spatial multiplexing order ); Using the entries in rank 1 of the codebook associated with the MIMO system with the four transmit antennas to rank 8 to rank 8 of the two base matrices for the MIMO elementary codebook associated with the eight antennas Constructing new entries for the new entry; And transmitting feedback indicating an entry index of the MIMO basic codebook related to the eight transmit antennas to a base station. .

Advantageously, the two basis matrices are defined as a specific matrix.

Advantageously, the base station is characterized by providing signaling to the mobile station to indicate what type of antenna configuration is used.

Advantageously, the eight transmit antennas are cross-polarized antennas.

Advantageously, the MIMO-based codebook associated with the eight transmit antennas supports a technique related to IEEE 802.16 or a technology related to 3GPP LTE.

In addition, in order to solve the above-mentioned problems, there is provided an apparatus related to a Multiple Input Multiple Output (MIMO) system, comprising: a transceiver for transmitting and receiving radio signals carrying data and information to or from a base station; A storage medium for storing data and information comprising at least one codebook associated with a MIMO system having four transmit antennas; And accessing a codebook in combination with the storage medium and having entries for rank 4 in rank 1, each rank representing a specific spatial multiplexing order, Using the entries in rank 1 of the codebook associated with the MIMO system to construct the entries for rank 8 from rank 1 to rank 2 of the two base matrices for the MIMO elementary codebook associated with the 8 antennas And transmitting feedback to the base station indicating an entry index of the MIMO-based codebook associated with the eight transmit antennas. ≪ RTI ID = 0.0 >

In accordance with the present invention, certain codebook entries used in a lower dimension (i.e., having a relatively small number of antennas) MIMO systems have higher dimensions (i.e., Number of antennas) in a MIMO system. Characteristically, using the entries in rank 1 of the codebook associated with the MIMO system with four transmit antennas, the entries of rank 8 to rank 2 in the two basic matrices for the MIMO elementary codebook associated with the eight transmit antennas Has been proposed by the present invention, and this proposal of the present invention has a great effect on generating a high-standard codebook based on a low-standard codebook, enabling an easier codebook design and reducing a required memory space of a mobile station (UE or MS) have.

Figure 1 illustrates exemplary concepts of Orthogonal Frequency Division Multiplexing (OFDM) technology.
Figure 2 illustrates exemplary concepts of Multiple Input Multiple Output (MIMO) techniques.
Figure 3 illustrates exemplary concepts of MIMO precoding techniques.
Figure 4 illustrates exemplary link throughput gains of a newly proposed codebook in contrast to using an AWD (Amendment Working Document) codebook in uncorrelated cases.
FIG. 5 illustrates exemplary link throughput gains of a newly proposed codebook in contrast to using an AWD (Amendment Working Document) codebook when correlated.
Figure 6 illustrates exemplary experimental results for spectral efficiency (bits / tone) versus signal-to-noise ratio (SNR) using 4-bit, 6-bit, 8-bit and AWD codebooks.
FIG. 7 illustrates exemplary performance comparison results generated by applying the base matrices B1 and B2 of the present invention in the case of a cross-polarized antenna.
Figure 8 shows that there is very little performance loss due to the base matrices B1 and B2 on the highly correlated ULA.
Figure 9 shows an exemplary codebook matrix selection ratio when using an AWD codebook for rank 2 on ULA and XPOL.
Figure 10 shows two exemplary antenna indexing formats for XPOL type 1 and XPOL format 2.
Figure 11 shows exemplary experimental results for the selectivity of the codebook matrices over several antenna configurations.
Figure 12 shows an exemplary table containing codebook entries for the two newly generated matrices realized in the SU-MIMO basis codebook according to the present invention.
13 is an exemplary flowchart for explaining a basic method procedure using specific codebook entries in which a newly generated codebook is used for a higher dimension MIMO system in a lower dimension MIMO system, And FIG. 2 is a block diagram of a conventional MIMO apparatus.
FIG. 14 is a flowchart illustrating an exemplary concept of protocol layers (MAC-PHY) and various entities supporting MIMO-OFDM at a transmission end in downlink data transmission.
15 is a flowchart illustrating an exemplary concept of protocol layers (MAC-PHY) and various entities supporting MIMO-OFDM at a receiving end in uplink data transmission.

A basic concept of the present invention is a method for constructing a codebook related to a plurality of antennas on a Multiple Input Multiple Output (MIMO) system, comprising: accessing a codebook associated with a MIMO system having four transmit antennas Wherein the codebook is stored in a storage medium and the codebook has entries for rank 1 to rank 4, each rank representing a specific spatial multiplexing order; Using the entries in rank 1 of the codebook associated with the MIMO system with the four transmit antennas to rank 8 to rank 8 of the two base matrices for the MIMO elementary codebook associated with the eight antennas Constructing new entries for the new entry; And transmitting feedback indicating an entry index of the MIMO basic codebook related to the eight transmit antennas to a base station. And proposes a wireless mobile communication terminal capable of performing such a method.

Hereinafter, configurations and operations of embodiments according to the present invention will be described with reference to the accompanying drawings.

The inventive concepts and features herein are described primarily on the basis of OFDM-MIMO technology. However, these details are not intended to limit the features described herein and are applicable to other types of mobile and / or wireless communication systems and methods.

For downlink and uplink transmissions, several types of multiplexing and multiple access techniques have to be considered. Downward, various multiplexing techniques may be used such as time division multiplexing, frequency division multiplexing, and code division multiplexing. Upwards, various multiple access techniques may be used such as time division multiple access, frequency division multiple access, and code division multiple access.

Referring to FIG. 1, OFDM is one of frequency division multiplexing techniques. In particular, OFDM is a bandwidth efficient technique for multi-carrier modulation. That is, OFDM is used as a digital multi-carrier modulation method as a frequency division multiplexing (FDM) scheme. A large amount of adjacent orthogonal sub-carriers are used to carry the data, the data being divided into a number of parallel data streams or channels, one for each sub-carrier. Each sub-carrier is modulated at a low symbol rate using a particular modulation scheme, such as quadrature amplitude modulation (QAM) or phase shift keying (PSK), which is a single carrier modulation scheme of the same bandwidth Similar overall data rates can be maintained.

Fundamental features and operating principles of OFDM include using inter-symbol interference (ISI) minimization using orthogonality, using a Fast Fourier Transform (FFT) algorithm (and / or an inverse FFT (Or forward error correction (FEC), frequency and / or time interleaving, use of adaptive transmissions, etc.), adaptation of guard intervals, equalization of frequency response characteristics, Spatial diversity, linear transmitter power amplification, etc. In the present invention, for the sake of brevity, the above-mentioned features and principles will not be described in detail. However, It will be appreciated by those skilled in the art that the disclosed features and principles may be applied in whole or in part to the present invention.

In addition, the present invention can be applied to various OFDM systems including OFDMA, IEEE, 802.11a / g / n, 802.16, 802.20, 3GPP LTE, 3GPP LTE-A, Evolved UMTS Terrestrial Radio Access (E-UTRA) Applicable to technical standards.

Referring to FIG. 2, a multiple access (or multiple access) technique may be referred to as a Multiple Input Multiple Output (MIMO) technique. Basically, the MIMO is related to using a plurality of (multiple) antennas in both the transmitting end and the receiving end to improve wireless communication performance. The MIMO may be considered as one of the technologies called smart antenna technology.

The MIMO technique is very attractive in wireless communication in order to dramatically increase the link range and data throughput without additional bandwidth or transmission power. Such an increase may be due to high reliability or split due to high spectral efficiency and reduced fading which allows more bits of hertz or more bits per second to be transmitted and received. It is achievable.

The MIMO may be divided into a single-user (SU) MIMO, a multi-user (MU) MIMO, a cooperative MIMO, and a networked MIMO. The single-user (SU) MIMO is concerned with using multiple (multiple) antennas at both the transmitter and the receiver (point-to-point). The multi-user (MU) MIMO is concerned with communicating concurrently with a plurality of independent wireless terminals using the same frequency (point-to-multipoint or multipoint-to-point) . The cooperative MIMO is concerned with the use of a combination of distributed antennas belonging to several different nodes in the transmitter or receiver to achieve the benefits of MIMO. The networked MIMO is concerned with using coupling between network nodes to achieve the benefits of MIMO. Each of the MIMOs thus divided represents special features and certain characteristics in IEEE 802.11, 802.16, 3GPP LTE (LTE-A), etc., and all the MIMOs are applicable to the present invention.

The MIMO may be divided into three major categories such as spatial multiplexing, diversity coding, and precoding.

The spatial multiplexing is related to splitting the high rate signal into a plurality of lower rate streams, each stream being transmitted in different transmission antennas in the same frequency channel. If such signals arrive at sufficiently different spatial signatures and receive antenna arrays, the receiving end can distinguish such streams to produce parallel channels or streams. The spatial multiplexing can increase the channel capacity at a high signal-to-noise ratio (SNR). The maximum number of spatial streams is limited to be smaller than the number of antennas at the transmitting end or the receiving end. The spatial multiplexing may be used with or without transmission channel information.

The diversity coding is a technique used when there is no transmission channel information at the transmitting end. In the diversity method, only one stream (unlike a plurality of streams in spatial multiplexing) is transmitted, the signals being coded using space-time or space-time coding techniques. The signals are either completely orthogonal coded or emit from respective transmit antennas using specific principles that are close to orthogonal coding. The diversity uses independent fading in a plurality of antenna links to enhance signal splitting. In the diversity coding, there is no beamforming or array gain from the diversity coding because there is no channel information.

Precoding relates to the concept of beamforming in which the same signal is radiated from each transmit antenna with an appropriate phase weighting such that the signal power is at a maximum at the receive end input. The beamforming increases signal gain from structural combining and multipath fading effects. When the receiving end includes a plurality of antennas, such beamforming can not simultaneously maximize the signal level to all of the receiving antennas, and therefore the precoding is used. In general, the precoding requires channel state information (CSI) at the transmitter.

Referring to FIG. 3, in the MIMO precoding scheme, the transmitted data is divided into a plurality of independent transport streams, each stream being precoded separately and all precoded streams being superimposed before transmission superimposed. The number of transport streams is represented or represented by a transmission rank. Here, the transmission rank may be defined as the spatial multiplexing order. The transmission rank is optimally selected for a given channel taking into account the transmission power and various channel characteristics. For each transport stream, a precoder is a beamforming vector consisting of a set of weights for each antenna multiplied by the transmitted symbols before transmission.

The precoder of the rank r is a matrix constituting the r Beamforming vectors as columns. The average transmit power may be equally divided among all transport streams so that the norm of all beamforming vectors is the same and normalized to one. Also, because of the overlap of the transport streams, the beamforming vectors are selected orthogonal for the sake of simplicity of decoding. As a result, the precoder is generally a semi-unitary matrix.

In a technique called codebook based precoding, a predefined codebook is provided to the base station (BS) and all the receiving stations (MS) (Or other value / indicator) related to the selected precoder. Selection of Precoder Rank [0064] The selection of the precoder rank can also be performed by selecting the precoder from the codebook that maximizes its performance, .

In addition, the MIMO may use a closed-loop operation or an open-loop operation.

In the closed loop operation, feedback from the receiving side is required. Here, a packet for generating channel state information (CSI) may be used at the transmitting end. In addition, explicit feedback such as quantized CSI utilization may be used. Additionally, a pointer may be used in the quantized codebook of precoding matrices. Also, the concepts of beamforming and precoded spatial division multiplexing (SDM) may be applied.

In the open loop operation, feedback from the receiving side is not required. Also, transmit diversity, switched beams, spatial division multiplexing (SDM), and hybrid methods (e.g. combining transmission and spatial division multiplexing) may be used. Here, in the SDM, the precoding matrix is an identity matrix or a matrix arbitrarily selected from the codebook.

The inventors of the present invention recognize that various operations related to codebook-based precoding techniques can be improved from the problems of the prior art described above. These prior art problems have been recognized as being very important based on focused investigations, rigorous simulations, and experimental tests conducted by the inventors. As a result, the present inventors have found that certain codebook entries used in a lower dimension (i.e., having a relatively small number of antennas) MIMO systems have a higher dimension , With a relatively large number of antennas) in a MIMO system. Characteristically, using the entries in rank 1 of the codebook associated with the MIMO system with four transmit antennas, the entries of rank 8 to rank 2 in the two basic matrices for the MIMO elementary codebook associated with the eight transmit antennas Is not provided or suggested in any known codebook-based precoding method prior to being performed by the present inventor described in this specification and in the priority document publication.

Several exemplary embodiments will be described in more detail below.

First, the generation of the codebook using both the DFT basic matrix and the block diagonal matrix will be described.

a) a base codebook matrix using an N-point DFT matrix

In the MIMO method, an N-point DFT matrix WN (or a transform matrix using both column permutation, phase shift, or the like, or both the matrix and transform matrix) is a base matrix, Can be used to generate transmit antenna codebooks.

b) a basic codebook matrix using a block diagonal matrix

(Or both the matrix and the transformation matrix using two N / 2 point DFT matrices (or column permutation, phase shift, etc.) or two N / 2 times N / 2 zero A block diagonal matrix W _B consisting of N / 2-by-N / 2 zero matrices W _B can be used as a base matrix to generate N-transmit antenna codebooks.

c) Generating a codebook for rank R

The codebook in the other ranks consists of a subset of the basis codebook vectors. The codebook for the rank R is constructed by using R column vectors of the base matrix W. [ If w (i) represents the i-th column vector of the base matrix W, then the codebook C for rank R is:

d) Generate N-point DFT basis matrix and N-bit codebook

The n-bit codebook is constructed from the basis matrix by using 2n combinations of codebook column vectors.

Example: Generating a 2-bit codebook Ci (where i = 0, 1, 2, 3) for rank 2 by using a 4-point DFT basis matrix, where index i represents a codebook index:

e) Generating a codebook using multiple basis matrices

Assuming that there are a plurality of base matrices, a rank R codebook configuration is possible by using a column vector from the plurality of base matrices or a column vector from one base matrix.

f) generate a codebook using both the DFT basis matrix and the block diagonal matrix

W _N The square of the column element is always 1 / N. W _B , the square of the column element for the half of the WB is 2 / N and the square of the column element for the other half is 0 due to N / 2-by-N / 2 zero matrices in the matrix. Here, since there is a difference in column elements between W _N and W _B , the following codebook configuration restriction is required.

For rank R codebook generation (when R = n + m), n represents the number of column vectors from W _N and m represents the number of column vectors from W _B. Where m must be an even number, the m / 2 column vectors of the codebook occupy one of the first half columns of W _B , and the other m / 2 column vectors occupy the remaining half columns of WB.

If n is not zero and R is odd, the n column vectors of the codebook are taken from W _N. If n is 0 and R is even, then the codebook can be constructed using only W _B. If m is 0, then the codebook can be constructed using only W _N , irrespective of the value of R. Where R is n + m and m is an even number so that all transmit antennas have the same power.

In the above description of the proposed codebook generation method, a DFT (or FFT) matrix is used. However, the use of an IDFT (or IFFT) matrix instead of the DFT matrix in the proposed method is also applicable. Also, a transformation matrix (i.e., a transformation matrix using both column permutation, phase shift, or the like, or both the matrix and transformation matrix) of the IDFT can be used.

For W _N , any single unitary matrix without zero elements can be placed. For W _B , all block diagonal matrices or modified versions of such matrices can be considered.

The following exemplary procedures (1) through (5) illustrate how a higher dimensional codebook is constructed, created, or applied using a lower dimensional codebook recognized and developed by the present inventors.

(1) 4G MIMO configuration of 8 transmit codebooks using 4 codebooks

In 3GPP Rel-8 (i.e., LTE system), four transmit antenna MIMO transmissions are supported in the downlink. However, the LTE-A system can support MIMO transmission up to 8 transmit antennas to improve the throughput of the comparison system with the LTE system. For the purpose of achieving remarkable performance in terms of system throughput, the LTE-A system supporting eight transmit antenna systems must provide a CL-MIMO method based on the use of a codebook or CSI direct feedback. Here, IEEE 802.16 may also consider an 8 Tx CL-MIMO system for the same purpose of the LTE-A system.

For the code book design for the CL-MIMO, there are a few points to be considered and necessary when used in a specific system such as 3GPP LTE-A or IEEE 802.16. It is generally not difficult to design a codebook up to 4Tx CL-MIMO while satisfying the above-mentioned points to be considered and the points required. However, the codebook design for the 8TX CL-MIMO is a challenging task. Even if a new codebook for the 8TX CL-MIMO is provided, it is necessary to use three different codebooks each for the UE or MS (mobile station) 2Tx, 4Tx, and 8TX CL-MIMO operations, It is not desirable due to the need for additional space in the memory required to store all of the other codebooks. Therefore, the present inventors have developed an 8Tx CL-MIMO operation using a codebook for 4Tx CL-MIMO, regardless of the type of system (e.g. 3GPP LTE-A, IEEE 802.16, etc.) without any additional memory requirement for 8Tx codebook storage A simple method of constructing a codebook for the above-mentioned codebook is proposed and the above-mentioned problem is addressed.

In the LTE-A system, the codebook for 8Tx CL-MIMO is recommended to have the following characteristics.

(1) Unitary; (2) Constant modulus; (3) Nested structure; (4) Constrained alphabet, e.g., 8PSK.

However, the codebook configuration method used for 4Tx CL-MIMO in LTE does not suggest that all of the above-mentioned four characteristics are satisfied in 8Tx CL-MIMO. Therefore, several proposals have been made to generate an 8Tx codebook satisfying the above-mentioned four characteristics.

In the general IEEE 802.16m, the proposed codebooks for 4Tx and 8Tx CL-MIMO do not have any relation to each other, and the mobile station (MS) The codebooks that are not related to each other must be stored.

In order to provide an easy configuration of the 8Tx codebook for reducing the memory space in the UE (or MS) and / or the eNode-B (or BS), reuse of the existing 4Tx codebook in the configuration of the new 8Tx codebook is highly desirable Do.

That is, if a specific 4Tx codebook is specified, the 8Tx codebook can be configured more easily by reusing the designated 4Tx codebook. For example, the following methods (Equation 1) can be used to design an 8Tx codebook for rank 1.

 ≪ Formula 1-1 >

는 랭크 R 에 대한 N_TXTx 코드북의 k 번째 엔트리를 의미하고, L_R 은 랭크 R 에 대한 8Tx 코드북의 엔트리 수를 나타낸다. 또한, α_R 및 β_R 는 스케일링 인자(scaling factors)들이다.In the above equation

Denotes the kth entry of the N _TX TX codebook for rank R, and L _R represents the number of entries in the 8Tx codebook for rank R. < RTI ID = 0.0 > Also,? _R and? _R are scaling factors.

을 만들기 위해 설정될 수 있다. 상기 IEEE 802.16m에서, α_R and β_R 의 값들은

을 만들기 위해 설정될 수 있다. 상기 LTE, LTE-A 시스템과 IEEE 802.16m 시스템에서,

가 상기 값의 하나의 예로 사용될 수 있다.In the LTE and the LTE-A system, the value of α and _R β _R are

As shown in FIG. In the IEEE 802.16m, the value of α and _R β _R are

As shown in FIG. In the LTE, LTE-A system and the IEEE 802.16m system,

Can be used as an example of the above value.

The simplest method of constructing a codebook for rank 1 is to set m, n, k equal (m = n = k). Another alternative is to use n = k and m = mod (k + k1, L1) settings. From these settings, the L1 entries of the 8Tx codebook for rank 1 can be obtained.

Hereinafter, a mathematical expression applying the following extension rule may be used to acquire another entry of the 8Tx codebook for the rank 1.

 ≪ EMI ID =

Equation (1-2) may also be expressed by the following equation.

≪ Equation (1-3)

For example, B may be set to 4 for a stroke of 4 additional entries.

Based on a method similar to the above, an 8Tx codebook for rank 2, 3, and 4 can be constructed as follows.

≪ Equation (1-4)

In LTE, LTE-A and IEEE 802.16m systems,

및,

And

Can be applied as an example.

However, it is meaningless to generate the 8Tx codebook for rank 4 in the 4Tx codebook for rank 4. Because the Chordal distance between the entries is all zero. Therefore, in the present invention, the 4Tx codebook for rank 3 and rank 2 is reused as shown in the following equation to construct an 8Tx codebook for rank 4.

≪

In the LTE and LTE-A systems, it can be set as follows.

는 다음 예시처럼 설정될 수 있다.In an IEEE 802.16m system,

Can be set as shown in the following example.

As another method of setting the 8Tx codebook for rank 4, the following equation can be used.

≪ Equation (1-6) >

는 다음 예시처럼 설정될 수 있다.In LTE and LTE-A systems,

Can be set as shown in the following example.

는 다음 예시처럼 설정될 수 있다.In an IEEE 802.16m system,

Can be set as shown in the following example.

Further, an 8Tx codebook for any rank higher than rank 4 can be constructed in a manner similar to the above. The formula for this is as follows.

≪ Equation (1-7) >

는 다음 예시처럼 설정될 수 있다.In LTE and LTE-A systems,

Can be set as shown in the following example.

다음 예시처럼 설정될 수 있다.In an IEEE 802.16m system,

It can be set as shown in the following example.

The simplest way of constructing an 8Tx codebook for rank 5-8 is to set m, n, k equal (m = n = k). The above Equations 1-2 to 1-3 in which the extension rule is used are applicable to Equations 1-4 to 1-7.

The portions of the 8Tx codebook per rank may comprise at least a portion of the 8Tx codebook for each rank using the entries generated by the methods mentioned above.

In addition, the features, equations, and procedures described above are applicable to any number of dimension codebook configurations, such as a codebook configuration for 16Tx.

In other words, a high specification codebook may be constructed using a relatively low specification codebook. As an example, an 8Tx codebook may be constructed using the 4Tx codebook as described above. Here, an 8Tx codebook for a rank lower than rank 4 can be configured with a corresponding rank of a 4Tx codebook. On the other hand, the 8Tx codebook for ranks greater than or equal to rank 4 can be configured using entries in the 4Tx codebook of ranks less than rank 4 of the 4Tx codebook. For example, as described above, an 8Tx codebook for rank 4 can be constructed using 4Tx codebooks for rank 3 and rank 1. That is, the concept of reducing the required memory space of the mobile station (UE or MS) and enabling a codebook design that is easier by generating a high-standard codebook based on the low-specification codebook is a feature of the present invention proposed by the present inventor.

A method for constructing an 8Tx codebook using a 4Tx codebook for rank 1 will be described below.

An 8Tx codebook can be easily constructed by reusing a given 4Tx codebook. For example, the following methods (Equation 2) may be used in constructing the 8Tx codebook for rank 1.

Equation (2-1)

If n = m = k,

For the configuration of the 8Tx codebook for rank 2, the following equation can be used.

Equation (2-2)

The simplest example of creating a codebook for rank 2 is as follows.

For the configuration of the 8Tx codebook for rank 3, the following equation can be used.

≪ EMI ID =

The simplest example of creating a codebook for rank 3 is as follows.

In a manner similar to the above, an 8Tx codebook for rank 8 in rank 4 can be made as a simple example as follows.

(Equation 2-4)

(2-5)

(Equation 2-6)

(Equation 2-7)

(Equation 2-8)

 Another example of using phase rotation in a row or column is as follows.

In the above-mentioned method, a high-rank 8Tx codebook generation process from a low rank to a 4Tx codebook for rank 1 is described. However, conversely, it is possible to generate an 8Tx codebook from a high rank. That is, an 8Tx codebook for rank 8 may be generated according to Equation 2-8, and an appropriate row or column corresponding to each rank may be extracted. For example, based on the 8Tx codebook for rank 8 (using equations 2-8), an 8Tx codebook for rank 1 is configurable using the first row of each matrix of 8Tx codebook for rank 8 Do. Similarly, an 8Tx codebook for rank 2 is also configurable using the first and second rows of each matrix of the 8Tx codebook for rank 8.

The particular size of the codebook for each rank may be different. For example, the codebooks for rank 1 and rank 2 may have 64 and 32 entries (or bits) respectively and the codebooks for rank 8 in rank 3 may have only 4 entries (or bits) have. In order to increase the size of the codebook for rank 1, the following exemplary method can be applied.

As for the rank 1, the codebook size for rank 2 can be increased in the same way as follows.

The portions of the 8Tx codebook per rank may comprise at least a portion of the 8Tx codebook for each rank using the entries generated by the methods mentioned above.

In addition, the features, equations, and procedures described above are applicable to any number of dimension codebook configurations, such as a codebook configuration for 16Tx.

The following describes the IEEE 802.16m 8Tx codebook.

The codebooks for the 2Tx, 4Tx, and 8Tx CL-MIMO schemes usually require a larger codebook size for each rank as the number of transmit antennas increases. For example, if a 4Tx rank 1 codebook of 6 bits size is given, a 2Tx rank 1 codebook of 3 bits size is adopted. If the 8Tx codebook has only 4 bits per rank, it will not be enough to support high-specification transmit antennas such as 8Tx.

In the present invention, a new 8Tx codebook having a larger size than the 8Tx codebook obtained by the 4Tx codebook reuse is provided through the following process.

In order to construct a larger-sized 8Tx codebook, it is most simply applied to reuse a 4Tx codebook having a 6-bit size. The reuse of the 4Tx codebook has the following two advantages. The first advantage is that the MS does not require additional memory for the 8Tx codebook since it already stores the 4Tx codebook itself. A second advantage is that it can be easily extended to larger-sized codebooks. For example, an 8Tx codebook configuration of 8 bits can be deduced from a 4Tx codebook subset as follows.

The codebook results represent required properties such as having unitary matrices and nested structures. A codebook of 6 bits can be made in the same manner as that of constructing an 8-bit codebook. The configuration of the 6-bit codebook for rank 2 at rank 1 can be constructed using a subset of 4Tx codebook as follows.

Figure 4 shows an exemplary link throughput increase over simulation using the codebook proposed here rather than an amendment working document (AWD) codebook already known in the uncorrelated case.

In the simulation, an 8Tx-2Rx CL-MIMO basis codebook under Ped-B 3km / h was evaluated. The time periods of PMI feedback, CQI feedback, band selection, and rank adaptation are 5 ms, 5 ms, 40 ms, and 40 ms, respectively. According to the above results, a codebook size of at least 8 bits is required for achieving a comparable precoding increase according to the SVD-based approach.

FIG. 5 shows exemplary link throughput gains of a newly proposed codebook in contrast to using an AWD (Amendment Working Document) codebook when correlated. That is, FIG. 5 shows an increase in the throughput when the proposed codebook of 8 bits and 6 bits is used in comparison with the AWD codebook in the mutual relation.

Figure 6 shows exemplary experimental results for spectral efficiency (bits / tone) versus signal-to-noise ratio (SNR) when using 4-bit, 6-bit, 8-bit and AWD codebooks. As shown in FIG. 6, the use of a 6-bit codebook is not so much improved as compared with the use of a 4-bit codebook. However, the newly proposed 8-bit codebook has an average performance of 5% .

If the uplink (UL) feedback overhead for 8Tx CL-MIMO is simply less than or equal to the uplink feedback overhead of 4Tx CL-MIMO, then a 6-bit codebook is preferable to an 8-bit codebook. However, if the uplink (UL) feedback overhead for the 8Tx CL-MIMO is greater than the uplink feedback overhead of 4Tx CL-MIMO, then the downlink throughput is significantly increased Therefore, an 8-bit codebook is preferable for CL-MIMO based on 8Tx codebook since it provides a great efficiency throughout the system.

The following codebook entries may be added to the 4-bit 8Tx codebook proposed in the present invention.

The above-mentioned processes can be summarized as being related to a base codebook (especially SU-MIMO basis codebook) for 8 transmit antennas. That is, the basic codebook for the 8 transmit antennas can be constructed as follows from the basic codebook for the 4 transmit antennas.

As described above, the above-described features are applicable to IEEE 802.16m, 3GPP LTE, and 3GPP LTE-Advance.

The downlink (DL) codebook for 8Tx precoding will be described below.

In the LTE-Advance (LTE-A) system, up to 8Tx antenna techniques should be supported in the downlink for achieving remarkable spectral efficiency. In particular, since most throughput on the downlink results from CL-MIMO operation, it is natural to apply CL-MIMO in the 8Tx antenna technique. There are two approaches to achieving CL-MIMO for the 8Tx antenna case in FDD. The first is based on the codebook in the LTE system and the second is based on the CSI direct feedback.

The process described below will focus on a codebook-based scheme and propose a new codebook for the 8Tx antenna scheme.

As a guideline for such a codebook design, it is necessary to consider various conditions based on the release 8 (Rel-8) system design, and the 8Tx codebook should satisfy at least the following characteristics.

1) constant modulus properties 2) finite alphabet 3) highly correlated, uncorrelated, cross polarized antenna configurations, and the like. 4) moderate codebook size, and 5) nested properties.

The following procedure proposes to construct a new codebook for the 8Tx antenna system from the codebook currently used in the Rel-8 4Tx transmission scheme. First, two basis matrices using the Rel-8 4Tx codebook are used for Rank 1 transmission, and the two matrices are as follows.

 The above-described characteristics 1) and 2) are satisfied before the construction of the 8Tx codebook from the basic matrices using the 4Tx codebook. For the above characteristic 4), a codebook having a size of 4 bits will be considered for the first time. All of the entries of the codebook are extracted from the basic matrices so that the characteristic 5) is satisfied.

The following Table 1 shows exemplary entries in a 4-bit codebook used for 8Tx precoding. In Table 1 above, the indices of the rows of the base matrices are shown in parentheses.

In addition, the following Table 2 shows yet another exemplary entries in the 4-bit codebook used for 8Tx precoding. In Table 2 above, the indices of the rows of the base matrices are shown in parentheses.

However, in order to satisfy the above characteristic 3), it is also necessary to introduce other basic matrices such as B1 and B2 shown below. The basic matrices shown below are more suitable for a cross polarized antenna configuration than the basic matrices A1 and A2 shown above.

및

는 k 번째 값(element)에 0이 아닌 값을 구비하는unit vector임을 의미한다.here

And

Is a unit vector having a non-zero value for the k-th element (element).

Although the above characteristic 5) is partially unsatisfactory, the increase in performance obtained by applying the basic matrices B1 and B2 is clear, as can be seen in Fig. 7, which shows a performance comparison in the case of cross-polarized antennas. On the other hand, FIG. 8 shows that there is almost no performance loss due to the base matrices B1 and B2 on the correlated ULA.

The basic matrices B1 and / or B2 are interchangeable with any block diagonal matrix or a modified version thereof. The matrices A1 and / or A2 may be interchangeable with any single unitary matrix containing non-zero values. In designing the codebook using these basic matrices A1 to B2, the codebook entries for the odd rank (e.g., rank 1, 3, 5, 7) are configurable by using the A1 and A2, The codebook entries for rank 2, 4, 6, and 8 are configurable by using A1, B1, and B2 as shown in Table 2 above. As another example, the codebook entries for the odd ranks may be configured by mixing the even numbers in the B1 and / or B2 rows with the odd numbers in the A1 and / or A2 rows.

The above procedure can be applied to a system using any number of transmit antennas.

In this procedure, two optional downlink (DL) codebooks were provided for 8Tx precoded transmissions. One is to maintain the nested property strictly, and the other one is focused on achieving good performance regardless of the configuration of the antenna, such as cross polarization and high mutual ULA. At first, the size of the codebook is considered as 4 bits, and the actual codebook size can be transformed according to a specific uncorrelated case and an MU-MIMO technique.

The following describes an 8Tx codebook that is improved over IEEE 802.16m.

The current AWD has an 8Tx codebook of 4 bits in size per rank for sufficient beamforming increase in a highly correlated case. Applying such an 8Tx codebook to a MIMO system of ULA and XPOL antennas, multiple entries of the codebook for rank 2 do not perform well, as shown in FIG. The last 8 entries of the codebook for rank 2 are identified through the SLS as non-significant entries. In order to improve the performance of SU-MIMO, new entries of the codebook for Rank 2 are proposed in the present invention.

Figure 9 shows an exemplary codebook matrix selection ratio when using an AWD codebook for rank 2 on ULA and XPOL. In FIG. 9, XPOL types 1 and 2 mean two types of antenna indexing in the XPOL shown in FIG.

In the above-mentioned processes, the features of the newly proposed 8Tx codebook can be summarized as follows. That is, the 8-15 entries of the codebook for rank 2 can be replaced by the following matrices.

For performance comparison, in XPOL type 2, the base matrices can be changed as follows.

Figure 11 shows exemplary experimental results for the selectivity of the codebook matrices over several antenna configurations. That is, FIG. 11 shows a comparison result of selection ratios for the codebook matrices through SLS based on the following simulation assumptions.

11, the last 8 entries of the current codebook for rank 2 do not play a significant role in the ULA and XPOL antenna configurations. On the other hand, the 8 entries proposed in the present invention play a notable role in the case of XPOL. FIG. 11 illustrates that, for the XPOL configuration, block diagonal matrices or modified versions thereof are more appropriate base matrices.

The following table 3 demonstrates this implication by sector power. As shown in Table 3, the codebook proposed in the present invention exhibits remarkable increase in the amount of power without increasing the feedback overhead in the case of the XPOL setting, and has the same performance in the case of ULA.

Thus, in the IEEE 802.16m technology standard and / or other related technical standards, the following two matrices may be used for the SU-MIMO basis codebook.

Also, in the IEEE 802.16m technical standard and / or other related technical standards, the table shown in Fig. 12 may be defined.

Various inventive concepts and features of the present invention can be summarized as follows.

The present invention provides an improved method of codebook-based precoding technology. That is, certain codebook entries used in a MIMO system of a lower dimension (i.e., having a relatively small number of antennas) are of a higher dimension (i.e., a relatively large number Antennas) in a MIMO system.

In particular, referring to FIG. 13, in accessing a codebook associated with a MIMO system with four transmit antennas, the codebook is stored in memory and has entries for rank 8 in rank 1, Represents a specific spatial multiplexing order and uses the entries in rank 1 of the codebook associated with the MIMO system with four transmit antennas to generate two basis matrices < RTI ID = 0.0 > < / RTI > constructing new entries from rank 1 to rank 8 of base matrices; And transmits to the base station feedback indicating the entry index of the MIMO basic codebook associated with the eight transmit antennas.

The processes proposed and proposed in the present invention may lead to a method of selecting specific codebook entries to be used in another additional or new MIMO system. For example, in selecting a particular codebook entry for use in a low-specification MIMO system to generate a codebook for a MIMO system of a high-specification MIMO system, one or more particular odd-numbered rows or even- The rows may be appropriately selected and the particular odd numbered rows or even numbered rows thus selected may be used to construct the codebook entries used in the high specification MIMO system.

13, the present invention includes an apparatus capable of performing and implementing the methods described above, such as a processing unit, a controller, a CPI, a microprocessor, a memory or a storage device, etc. Lt; / RTI >

FIG. 14 is a flowchart illustrating an exemplary concept of protocol layers (MAC-PHY) and various entities supporting MIMO-OFDM at a transmission end in downlink data transmission.

First, transport blocks are transferred from the MAC layer to the PHY layer through one or more transport channels. (1402). After performing the CRC insertion 1404 and the process of code block segmentation 1406, turbo encoding 1408, rate matching 1410, and code block concatenation, (1402) and the like are performed. Where the HARQ entity 1411 is associated with performing the rate matching 1410. In addition, after performing the CRC insertion (1404), a convolutional encoding 1414, a rate matching 1416, and a channel interleaving 1418, May be performed.

Thereafter, a scrambling process 1420, a constellation mapping 1422, a layer mapping 1424, a spatial precoding 1426, and an OFDM signal generation OFDM signal generation 1428 and the like are performed. Here, the MAC scheduler 1421 is related to the constellation mapping 1422, the spatial precoding 1426, and the OFDM signal generation 1428. In addition, the number of layers between the layer mapping 1424 and the spatial precoding 1426 processes may be less than or equal to the rank of the channel.

Various processes related to a method for newly generating a codebook for a high-specification MIMO system using specific codebook entries for a low-specification MIMO system may be performed in connection with the spatial precoding 1426. [ That is, during the spatial precoding 1426, the entries in rank 1 of the codebook associated with the MIMO system with four transmit antennas are ranked at the rank 1 of the two foundations matrices for the MIMO based codebook associated with the eight transmit antennas. Newly used to configure up to 8 entries.

15 is a flowchart illustrating an exemplary concept of protocol layers (MAC-PHY) and various entities supporting MIMO-OFDM at a receiving end in uplink data transmission.

14, the processes corresponding to FIG. 14 are performed inversely until a constellation mapping 1522 when signals are received through multiple antennas, and scrambling is performed (FIG. 1520 are slightly different from scrambling in FIG. In other words, guard intervals 1530, an N-point inverse FFT 1528, a sub-carrier mapping 1526, and an M-point DFT (M-point) DFT 1524 and the like are performed and the remaining processes 1512-1510-1508-1506- (1518-1516-1514) -1504 transfer the transmission blocks 1502 from the PHY layer to the MAC layer The same processes are similar to the processes described in FIG. 14 and are performed only in the reverse direction.

Hereinafter, a terminal according to the present invention will be described.

A terminal according to the present invention includes all types of terminals capable of using a service for exchanging data with each other over wireless. That is, the terminal according to the present invention may be a mobile communication terminal (e.g., a UE, a mobile phone, a cellular phone, a DMB phone, a DVB-H phone, a PDA phone, a PTT phone, And laptops, laptop computers, digital TVs, GPS navigation, portable gaming devices, MP3 and other consumer electronics, and the like.

A terminal according to the present invention may include a basic hardware configuration (a transceiver, a processing unit or a control unit, a storage unit, and the like) necessary for performing a function and an operation for generating a code book of a high standard on the MIMO system exemplified by the present invention .

The method according to the present invention described so far can be implemented in software, hardware, or a combination thereof. For example, the method according to the present invention can be stored in a storage medium (e.g., a mobile terminal or a base station's internal memory, flash memory, hard disk, etc.) Lt; / RTI > may be implemented as codes or instructions within a software program that may be executed by a processor (e.g., an internal microprocessor).

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it will be understood by those of ordinary skill in the art that various changes in form and details may be made therein without departing from the scope of the present invention . Accordingly, the true scope of the present invention should be determined by the technical idea of the appended claims.

Claims (10)

CLAIMS What is claimed is: 1. A method for transmitting data on a Multiple Input Multiple Output (MIMO) system supporting eight transmit antennas,
The method comprising: performing precoding on data to be transmitted, the precoding being performed using a precoding matrix selected from a codebook associated with the eight transmit antennas; ; And
And transmitting the precoded data to a receiver through the eight transmit antennas,
The codebook includes two base matrices, and the two base matrices V8 (:,:, 4) and V8 (:,:, 5)

or

≪ / RTI > The method according to claim 1,
Further comprising receiving from the receiving end signaling indicating a type of antenna configuration to be used. The method according to claim 1,
Wherein the eight transmit antennas are cross-polarized antennas. The method according to claim 1,
Wherein the precoding matrix is selected by the receiving end. 5. The method of claim 4,
Further comprising receiving from the receiving end information associated with an index indicating the precoding matrix. As an apparatus for transmitting data on a Multiple Input Multiple Output (MIMO) system supporting eight transmit antennas,
A precoder for performing precoding on data to be transmitted, wherein the precoding is performed using a precoding matrix selected from a codebook associated with the eight antennas; And
And a transmitter for transmitting precoded data by the precoder to a receiving end,
The codebook includes two base matrices, and the two base matrices V8 (:,:, 4) and V8 (:,:, 5)

or

≪ / RTI > The method according to claim 6,
Wherein the eight transmit antennas are cross-polarized antennas. The method according to claim 6,
Wherein the transmitter is further configured to receive from the receiver a signaling indicating a type of antenna configuration to be used, The method according to claim 6,
And the precoding matrix is selected by the receiving end. 10. The method of claim 9,
Wherein the transmitter is further configured to receive, from the receiving end, information associated with an index indicating the precoding matrix.

Images