JP4594976B2 - Codebook generator, codebook, and method for obtaining update matrix used in precoding scheme using MIMO transmission - Google Patents

Description

  The present invention relates to the field of multiple-input-multiple-output (MIMO) transmission. MIMO transmission is increasingly being used in mobile communication scenarios and is considered a future mobile communication system.

A MIMO system can generally significantly improve the system capacity of a mobile communication system. This is achieved, for example, because a MIMO system can simultaneously transmit and receive multiple data streams using multiple transmit and receive antennas. FIG. 17 shows a general MIMO system 1700. In this scenario, a transmission unit 1710 using N _t transmit antennas 1712 is shown. A signal transmitted from the transmission antenna 1712 is received by a reception unit 1720 using N _r reception antennas 1722. There is a mobile radio channel between the transmit antenna 1712 and the receive antenna 1722, which is shown as H _{k in} FIG. The signal transmitted by the transmission unit 1710 is denoted as x _k and is precoded by the precoding unit 1730. Precoding unit 1730 applies a precoding matrix F _k to the actual data stream s _k to be transmitted.

The precoding matrix F _k is based on feedback received from the receiving device. Preferably, the receiving apparatus feeds back the index m _opt using a code book. In the scenario shown in FIG. 17, these indices are determined based on a codebook 1740 available at the transmitting and receiving devices. The received signal is detected by the detection unit 1750 of the receiving apparatus that performs channel estimation. That is, the channel H _k is estimated, precoding is evaluated based on the channel estimate, coded based on the available codebook 1740, and an index for each codebook is provided to the transmitter. Accordingly, FIG. 17 shows a MIMO system using a codebook based precoding scheme. The MIMO precoding scheme can transmit multiple data streams with multiple beams 1760, thus increasing system capacity.

  When complete channel information is fed back from the receiving device, the transmitting device can perform full precoding, but this requires a large amount of feedback information, i.e. a channel to signal from the receiving device to the transmitting device These channels will use high data rates because the feedback information is detailed. Therefore, reducing the amount of feedback is a major issue in the field of MIMO systems using codebook-based precoding. In codebook-based linear precoding, it is only necessary to transmit a preferable index of a predetermined codebook set as feedback information, and the necessary feedback information is also reduced.

ここで、プリコーディング行列Ｆ_kは、コードブックの要素から計算される。式１において、ＭＩＭＯチャネルＨ_kは、

と定めることができる。 Refer to FIG. When the number of transmission antennas is N _t , the number of reception antennas is N _r , and the number of streams transmitted simultaneously is N _s , the transmission signal vector at time k can be expressed as follows.

Here, the precoding matrix F _k is calculated from the elements of the codebook. In Equation 1, the MIMO channel H _k is

Can be determined.

  Codebook design concepts can be classified as either “one-shot codebooks” or “tracking codebooks”. Table 1 shows the main differences between the two codebook design concepts.

  In the case of a one-shot codebook, each element of the codebook can be a matrix or a vector. The former is called a matrix codebook, and the latter is called a vector codebook. For tracking codebooks, conventional systems use vector codebooks primarily with recursive transmission algorithms, as will be described in more detail below.

  FIG. 18 shows a graph of channel coefficients of the mobile radio channel with respect to the time axis. Suppose that feedback is to be provided and that channel samples are taken at equally spaced periods indicated by black dots in FIG. It can be seen from the sampling points in FIG. 18 that the channel changes to some extent between the two channel estimates and the feedback period. This degree depends on, for example, mobility in a communication scenario. Based on this change, a codebook design concept is selected. If the channel change is small, a tracking codebook should be preferred because the channel can be tracked. If the change is somewhat large, a one-shot codebook is preferred. In an actual wireless communication scenario, a channel between two sampling points can be associated, for example, by slow fading or a fast sampling frequency corresponding to a short feedback period. In this case, the tracking codebook can provide a better tradeoff between the required feedback bits, ie the required data rate and performance in the feedback channel.

に基づいて特異値分解（singular value decomposition）を行い、ユニタリ行列Ｖ_kを求める。 Hereinafter, an example of a system in which the number of transmission antennas is N _t = 4, the number of reception antennas is N _r = 2, and the number of data streams is N _s = 2 will be described. Further, throughout this description, k is a time index. FIG. 19 shows the structure of a conventional receiving apparatus 1900. A conventional receiving apparatus 1900 includes a channel estimation unit 1910 for providing a received signal. The channel estimator 1910 estimates a radio channel based on a reference symbol and provides a channel estimate H _k to the processing unit 1920. The processing unit 1920

A singular value decomposition is performed based on the above to obtain a unitary matrix V _k .

The receiving apparatus 1900 further includes a memory 1930 that stores the previous precoding matrix F _k−1 .

に基づいて、以前のプリコーディング行列Ｆ_k-1に列を追加して、（Ｆ’_k-1）^Hを求めるカラムアダー１９４０をさらに備えている。 The receiving device 1900

Is further provided with a column adder 1940 for adding (F ′ _k−1 ) ^H to the previous precoding matrix F _k−1 .

に基づいて、カラムアダー１９４０の出力と処理ユニット１９２０の出力とが乗算部１９５０において掛け合わされ、乗算部１９５０がデルタ行列Ｄ_kを出力し、さらにデルタ行列Ｄ_kがコードブックサーチャ１９６０の入力とされる。 Equation 4 shows that the column adder 1940 adds a dummy column to the previous precoding matrix F _k−1 . To find the delta matrix

Based on the output of the output processing unit 1920 of Karamuada 1940 is multiplied in the multiplication unit 1950, the multiplication unit 1950 outputs a delta matrix D _k, is further delta matrix D _k is an input of the codebook searcher 1960 .

に基づいて、コードブック１９７０を用いてデルタ行列Ｄ_kをエンコードし、ベクトルコードブック１９７０のインデックスとしてインデックス

を出力する。 Codebook Searcher 1960

Based on, it encodes the delta matrices D _k using a code book 1970, the index as an index of the vector codebook 1970

Is output.

As can be seen from Equation 6, the delta matrix D _k is encoded column by column. For this reason, the first column is taken out and a suitable matching vector is searched in the code book 1970, and based on this, the dimension of the matrix is lowered by one dimension using the householder function Q _n shown in Equation 6. In a matrix with lower dimensions, the next column is one dimension lower than the first column and therefore requires less feedback. The details of the structure of the receiving apparatus and the mathematical background are described in Non-Patent Document 1.

を受信する。行列コンストラクタ２０１０は、受信装置で使用されるのと同じコードブックであるコードブック２０２０へ接続されている。再帰的なアルゴリズムを用いて、行列コンストラクタは、

に基づき、コードブックインデックスからそれぞれのデルタ行列Ｄ_kの推定値

を求める。 FIG. 20 is a block diagram of each transmitting apparatus 2000. The transmission apparatus 2000 includes a matrix constructor 2010. The matrix constructor 2010 is the codebook index

Receive. The matrix constructor 2010 is connected to a code book 2020 which is the same code book used in the receiving device. Using a recursive algorithm, the matrix constructor

Based on the codebook index to estimate each delta matrix D _k

Ask for.

に基づいて、カラムアダー２０４０がダミー列を追加するための以前のプリコーディング行列Ｆ_k-1を記憶するメモリ２０３０をさらに備えている。 The transmission device 2000

, The column adder 2040 further includes a memory 2030 for storing the previous precoding matrix F _k−1 for adding a dummy column.

に基づいて、列が追加された以前のプリコーディング行列と推定されたデルタ行列とを掛け合わせて、新たなプリコーディング行列を求める。次いでこの新たなプリコーディング行列をデータストリームｓに適用して、送信信号ｘを求める。 Next, the multiplication unit 2050

Based on the above, a new precoding matrix is obtained by multiplying the previous precoding matrix with the added column and the estimated delta matrix. This new precoding matrix is then applied to the data stream s to determine the transmission signal x.

Instead, replace the above equation with x = F _k Ps
It can also be modified. In this equation, P is an Ns × Ns power allocation matrix.

The conventional transmitting apparatus 2000 and the conventional receiving apparatus 1900 perform a recursive delta matrix coding algorithm using the same codebook 2020, 1970, each composed of vectors of different lengths. Equation 10 describes the codebook CB in more detail.

  Details of the structure of the transmitter and the mathematical background are described in Non-Patent Document 2.

により、以前のプリコーディング行列を左側から掛けることによってプリコーディング行列が更新されることによる。 19 and 20 show a receiving apparatus 1900 and a transmitting apparatus 2000 in the case of a conventional precoding scheme with a tracking function in which a current precoding matrix is updated using a previous precoding matrix. Yes. The conventional scheme has a disadvantage in that it requires a column adder for adjusting the matrix size. this is,

Thus, the precoding matrix is updated by multiplying the previous precoding matrix from the left side.

を求めるために、送信装置が行列コンストラクタ２０１０を必要とするという短所がある。これは、従来のスキームではベクトルコードブックを使用するために複数のインデックスがフィードバックされるという事実によるものであり、その場合、複数のインデックスは、より多くのフィードバックビットまたはより高いフィードバックデータレートを必要とするという短所ももたらす。しかも、従来のスキームは、特異値分解を行うために処理ユニット１９２０を必要とするという短所ももたらす。特異値分解は、複雑な計算であり、受信装置はさらなる処理能力を必要とする。概して、従来のシステムには、プリコーディングスキームを実施するために、ハードウェア及びソフトウェアの高い複雑度が求められるという短所がある。 The number of columns or dummy columns added depends on the number of data streams transmitted. In addition, the traditional scheme has a feedback index

Therefore, the transmission apparatus needs a matrix constructor 2010 to obtain the above. This is due to the fact that the traditional scheme feeds back multiple indexes to use a vector codebook, where multiple indexes require more feedback bits or higher feedback data rate. It also brings the disadvantage of. In addition, the conventional scheme also has the disadvantage of requiring a processing unit 1920 to perform singular value decomposition. Singular value decomposition is a complex calculation and the receiving device requires more processing power. In general, conventional systems have the disadvantage of requiring high hardware and software complexity to implement the precoding scheme.

R1-060868, Intel, "Performance and complexity of scaleable Precoded MIMO (LLS & SLS)," 3GPP RAN1 LTE, Mar 2006 R1-060868, Intel, "Performance and complexity of scaleable precoded MIMO (LLS & SLS)", 3GPP RAN1 LTE, Mar 2006

  It is an object of the present invention to provide a more efficient concept for generating a tracking matrix codebook for a precoding scheme in a transmission system that allows higher system capacity and link capacity.

  The object is to generate a codebook according to claim 1, a method for obtaining a plurality of marching matrices according to claim 21, a codebook according to claim 23, and a codebook according to claim 24. Can be achieved by a storage medium for storing.

  The object is to determine one or more update matrices for use in a precoding scheme in a transmission scheme using a first number of transmit antennas and a second number of receive antennas, and including the update matrix A code book generating device for generating a code book, wherein the first number is greater than 1, the second number is 1 or more, a first matrix is obtained, and the first matrix is obtained according to a certain degree of correlation And a combiner (410) that combines the first matrix and the second matrix to obtain one update matrix. The

  The object is to obtain one or more update matrices used in a precoding scheme in a transmission scheme using a first number of transmitting antennas and a second number of receiving antennas, A sub-step for obtaining a first matrix, wherein the second number is greater than 1, and a second step for obtaining a second matrix associated with the first matrix by a certain degree of correlation; Respectively including the steps of obtaining respective update matrices. The method further includes the step of obtaining one update matrix by combining the first matrix and the second matrix.

  The above object is a codebook including at least two update matrices used in precoding, each of which is associated with an index of an update matrix, wherein the update matrix is a first A matrix, a step of obtaining a second matrix associated with the first matrix with a certain degree of correlation, and combining the first matrix and the second matrix to obtain one update matrix. And the step of determining is achieved by a codebook in which the same degree of correlation as the different first matrix or a different degree of correlation is used to calculate a different update matrix.

  The object is to provide that the codebook includes at least two update matrices used in precoding, each update matrix being associated with an index of an update matrix, wherein the update matrix determines a first matrix. A step of obtaining a second matrix associated with the first matrix with a certain degree of correlation, and obtaining one update matrix by combining the first matrix and the second matrix. This is achieved by a storage medium storing the codebook, wherein the same correlation degree as the different first matrix or a different correlation degree is used to calculate the different update matrix.

  The present invention is based on the knowledge that a tracking codebook including an updating matrix that enables a precoding scheme in MIMO transmission is efficiently obtained based on the synthesis of correlated matrices. The matrix can be obtained so as to be unitary and random. For example, a unitary random matrix is obtained from an arbitrary random matrix so that the updating matrix codebook has a tracking function by using singular value decomposition. be able to. The code book can be efficiently generated using the code book generating device according to the present invention or the method according to the present invention for obtaining a plurality of update matrices included in the code book, It can be stored in a storage medium. The first matrix and the second matrix that are correlated use a time shift matrix of a random generator having a coherence time, or first obtain an independent matrix, such as a forgetting filter. By associating the second matrix and the first matrix using a filter, it is obtained efficiently. The present invention further provides the advantage that a codebook including a precoding scheme and an update matrix for MIMO transmission is efficiently sought while simultaneously providing a tracking function.

  Embodiments of the present invention will be described in detail with reference to the accompanying drawings.

  FIG. 1 shows an embodiment of a receiving device 100. The receiving apparatus 100 includes a channel estimation unit 110, and the channel estimation unit 110 is connected to the processing unit 120. The processing unit 120 is connected to a code book 122, a memory 124 for storing a precoding matrix, and a searcher 130. The searcher 130 is connected to the transmission unit 140. Receiving apparatus 100 controls precoding in the transmitting apparatus in a transmission scheme using the first number of transmitting antennas and the second number of receiving antennas. Here, the first number is greater than 1 and the second number is 1 or more. The receiving device 100 includes a channel estimation unit 110 that estimates a wireless channel between the transmitting device and the receiving device 100 based on the received signal and obtains a wireless channel estimation value.

  The receiving apparatus 100 provides a new matrix based on the update matrix included in the update matrix codebook 122 and the previous precoding matrix stored in the memory 124 that stores the precoding matrix. A processing unit 120 is further provided. The update matrix may have at least two rows and at least two columns. In general, the processing unit provides a new precoding matrix based on the update information items included in the update codebook and the previous precoding matrix. Preferably, the update codebook includes an update matrix as an update information item, wherein the update matrix has at least two rows and at least two columns.

  The receiving apparatus 100 searches for a new precoding matrix that satisfies the optimization criterion based on the radio channel estimation value among the new precoding matrices, and a new precoding found when searching for the new precoding matrix. The transmission unit 140 further transmits an index of the update matrix associated with the coding matrix.

  In one embodiment, the second number, the number of receive antennas, is greater than 1 and the precoding matrix has at least two rows and at least two columns. The processing unit can provide a new precoding matrix by multiplying the update matrix and the previous processing matrix. Furthermore, the receiving apparatus 100 can receive a multicarrier signal and control precoding in each carrier or a series of carriers of the multicarrier signal.

  In one embodiment, the processing unit 120 is implemented as a multiplier and multiplies the previous precoding matrix stored in the memory 124 storing the precoding matrix with the update matrix contained in the codebook 122. . Once a new precoding matrix is obtained, the processing unit can update the precoding matrix for the next update.

ここで、サーチされる新たな一連のプリコーディング行列を決定するために、右側から以前のプリコーディング行列Ｆ_k-1が掛け合わされている。式１２では、（Ｂより）小さい数を個数とする更新されたプリコーディング行列を計算することもできる。一般に、ｌはＢ以下の整数であり、ｐは１以上の下位乗算インデックス（lower multiplication index）であり、ｑはＢ以下の上位乗算インデックス（higher multiplication index）である。よって、１からＢまでのインデックスのサブセット、例えば１、２、４、Ｂなどを計算することもできる。この場合、ｐが１であり、ｑがＢであり、その間には２、４の２つの数しかないことになる。 FIG. 2 shows another embodiment of the receiving device 150. The receiving device 150 includes a channel estimation unit 155. The channel estimation unit 155 evaluates the channel estimation value H _k based on the received signal and provides the channel estimation value H _k to the codebook searcher 160. The receiving apparatus further includes a memory 165 that stores the previous precoding matrix F _k−1 . The previous precoding matrix F _k−1 is multiplied by the update matrix C _m of the code book 175 by the multiplication unit 170 to obtain a new series of precoding matrix C ′ _m , which is also the code book searcher. 160. The updated precoding matrix is given by the following equation.

Here, in order to determine a new series of precoding matrices to be searched, the previous precoding matrix F _k-1 is multiplied from the right side. Equation 12 can also calculate an updated precoding matrix with a number smaller than (B). In general, l is an integer equal to or less than B, p is a lower multiplication index of 1 or more, and q is a higher multiplication index of B or less. Thus, a subset of indices from 1 to B, such as 1, 2, 4, B, etc., can also be calculated. In this case, p is 1, q is B, and there are only two numbers of 2, 4 between them.

を見つけるために、更新されたすべてのプリコーディング行列Ｃ’_mをサーチする。最適化されたプリコーディング行列は、例えば一実施形態では、所与のチャネル推定値Ｈ_kについてチャネル容量を最大にするものであり、以下のように表すことができる。

ここで、σ²はノイズ分散またはノイズ電力である。別の実施形態では、最適化基準が異なっていてもよく、例えば、信号対雑音比または通信リンクの品質に関連する他のメトリックを最大にするものでもよい。コードブックサーチャ１６０は、最適なプリコーディング行列

を求める。次いで、この最適なプリコーディング行列が、プリコーディング更新の次の反復におけるＦ_kとして、以前のプリコーディング行列を記憶するメモリ１６５へ提供される。さらに、コードブックサーチャ１６０は、見つけられた最適化されている更新用行列

のインデックスｍ_optを送信装置へ提供する。 In another embodiment, another matrix is multiplied to perform any other precoding or adaptation task. That is, C ′ _k = C _k A _k F _k−1 . Codebook searcher 160 is an optimized precoding matrix

To find all the updated precoding matrices C ′ _m . The optimized precoding matrix, for example, in one embodiment, maximizes the channel capacity for a given channel estimate H _k and can be expressed as:

Here, σ ² is noise variance or noise power. In another embodiment, the optimization criteria may be different, for example to maximize other metrics related to signal to noise ratio or communication link quality. Codebook searcher 160 is an optimal precoding matrix

Ask for. This optimal precoding matrix is then provided to the memory 165 that stores the previous precoding matrix as F _k in the next iteration of the precoding update. In addition, the codebook searcher 160 may find the optimized update matrix found.

Index m _opt is provided to the transmitter.

  In another embodiment, codebook searcher 160 determines a maximum capacity based on detection of minimum mean square error successive interference cancellation (MMSE-SIC). In another embodiment, the processing unit 120 shown in FIG. 1 includes storage means for storing a code book based on the code book 175.

  FIG. 3 shows an embodiment of the transmission device 200. The transmitting apparatus 200 includes a receiving unit 210, storage means 220, an updating matrix generator 230, a combiner 240, and a memory 245 for storing a precoding matrix.

  Transmitting apparatus 200 transmits a precoded signal. This signal is pre-predicted using an updated precoding matrix stored in a memory 245 that stores the precoding matrix in a transmission scheme using a first number of transmit antennas and a second number of receive antennas. It is coded. Here, the first number is greater than 1 and the second number is 1 or more. The transmitting device 200 includes a receiving unit 210 that receives an update index, and a storage unit 220 that stores a codebook including different update information items each associated with the update index. The transmission apparatus 200 uses the received update index and the update information item associated in the codebook stored in the storage unit 220 to use the update matrix having at least two rows and at least two columns. Further comprising: an update matrix generator 230 for obtaining an update matrix; and a combiner 240 for combining the previously used precoding matrix 245 and the update matrix provided by the generator 230 to obtain an updated precoding matrix. Yes.

In one embodiment, the receiving unit 210 can receive the index (m _opt ) of the received update matrix. The storage means 220 stores a codebook that includes different update matrices (C _m ) each associated with an update matrix index (m) and having at least two rows and at least two columns.

とを掛けることにより行われる。ここで、以前に使用されたプリコーディング行列（Ｆ_k-1）が右側から掛けられる。受信アンテナの数である第２の数は１より大きくてもよく、プリコーディング行列（Ｆ_k-1、Ｆ_k）は正方行列とすることができる。さらに、更新用行列（Ｃ_m）も正方行列とすることができる。送信装置２００は、マルチキャリア信号を送信して、マルチキャリア信号内の各搬送波または一連の搬送波におけるプリコーディングを制御することもできる。 In another embodiment, the combiner 240 determines a precoding matrix (F _k ) updated by multiplication. This multiplication is performed by using the previously used precoding matrix (F _k-1 ) and the updating matrix.

This is done by multiplying Here, the previously used precoding matrix (F _k-1 ) is multiplied from the right side. The second number, which is the number of receiving antennas, may be larger than 1, and the precoding matrix (F _k−1 , F _k ) may be a square matrix. Further, the update matrix (C _m ) can also be a square matrix. The transmission apparatus 200 can also transmit a multicarrier signal to control precoding on each carrier or a series of carriers in the multicarrier signal.

を提供することができる。送信装置２５０は、以前のプリコーディング行列Ｆ_k-1を記憶するメモリ２６０をさらに備えている。さらに、送信装置２５０は乗算部２６５を備えている。乗算部２６５は、

に基づいて、更新用行列と以前のプリコーディング行列とを掛け合わせる。 FIG. 4 shows another embodiment of the transmission device 250. The transmission device 250 includes a code book 255, which includes the index m _opt of the optimal update matrix received from the reception device. Using the received update matrix index, the codebook uses the associated update matrix

Can be provided. The transmitting apparatus 250 further includes a memory 260 that stores the previous precoding matrix F _k−1 . Further, the transmission device 250 includes a multiplication unit 265. The multiplier 265

Based on, the update matrix is multiplied by the previous precoding matrix.

Then, the multiplier 265 can provide an updated precoding matrix F _k . According to the embodiment shown in FIG. 4, the combiner 240 can be implemented as the multiplication unit 265. In another embodiment, each of the transmission devices 200 and 250 includes a precoder that applies a precoding matrix F _k to a transmission signal.

Furthermore, in another embodiment, the precoding matrix F _k can be initialized. Initialization helps reduce tracking errors in some situations. Initialization can be performed periodically or can be triggered by a control signal or feedback from a receiving device. The matrix for initialization can be one element of the codebook. In one embodiment, it can be composed of unit vectors extracted from the unit matrix. For example, when the number of transmission antennas is N _t = 4 and the number of data streams is N _s = 2, the operation is as follows.

に右側から以前のプリコーディング行列を掛け合わせることにより、プリコーディング行列が更新されることであるという点に留意されたい。これにより、行列のサイズを調整するためのカラムアダーと、特異値分解とが不要になる。さらに、従来のスキームにおけるベクトルコードブックの代わりに行列コードブックを使用することにより、行列コンストラクタが不要となり、加えて、必要なフィードバック情報も低減できる。したがって、実施形態は、送信装置及び受信装置のハードウェア及びソフトウェアの複雑度を低減するという利点を提供する。さらに、必要なフィードバック情報が従来のスキームよりも少なく、したがって、フィードバックビット数、データレート、及びフィードバックチャネルの伝送リソースもより少なくて済む。 The essential difference from the conventional scheme is

Note that the precoding matrix is updated by multiplying by the previous precoding matrix from the right. Thereby, the column adder for adjusting the size of the matrix and the singular value decomposition become unnecessary. Furthermore, by using a matrix codebook instead of the vector codebook in the conventional scheme, a matrix constructor becomes unnecessary, and in addition, necessary feedback information can be reduced. Thus, embodiments provide the advantage of reducing the hardware and software complexity of the transmitter and receiver. Furthermore, less feedback information is required than in conventional schemes, and therefore fewer feedback bits, data rates, and transmission resources for the feedback channel are required.

FIG. 5 shows a MIMO transmission system 300 with a receiving device 310 according to one embodiment. The reception device 310 includes N _r reception antennas 312 connected to the reception unit 314. The reception unit 314 provides a reception signal to the channel estimation unit 316 and the MMSE-SIC detection unit / decoder 318. Next, the channel estimation unit 316 provides channel estimation values to the MMSE-SIC detection and decoder 318 and the codebook searcher 320. The code book searcher 320 is connected to the multiplication unit 322, and performs multiplication of the previous precoding matrix F _k−1 stored in the storage means 324 similar to that described above and the update matrix of the code book 326. The result is received from the multiplication unit 322. When finding the optimal update matrix, the codebook searcher 320 provides the optimal update matrix index m _opt to the transmitter 350. In this MIMO scheme, the first number and the second number are preferably chosen so that the number of rows of the precoding matrix F _k or F _k−1 is greater than the number of columns. However, the number of receiving antennas can be one.

The transmission device 350 includes a code book 352, and the code book 352 receives an optimal update matrix index m _opt from the reception device 310. The code book 352 can then provide the optimal update matrix to the multiplier 354. The multiplier 354 further receives the previous precoding matrix F _k−1 from the memory 356 that stores the previous precoding matrix. The encoder and symbol mapper 358 provides the transmission signal to the multiplication unit 354, and the multiplication unit 354 multiplies the transmission signal and the updated precoding matrix and provides the result to the transmission unit 360. The transmission unit 360 uses the N _r transmission antennas 362 and transmits a signal to the reception device 310 via the radio channel.

In the embodiment shown in FIG. 5, the codebook searcher 320 can evaluate the optimal update matrix, for example, by optimizing the link capacity or system capacity that can be calculated on the assumption of MMSE-SIC detection. . Therefore, in one embodiment, the codebook searcher 320 and the MMSE-SIC detector and decoder 318 can partially share circuitry. This leads to a further reduction in hardware and software complexity. Furthermore, in another embodiment, each of the codebook matrices is a unitary matrix, which will be described in detail below. In an embodiment of a codebook based precoding system, the transmitting device and the receiving device need to share the same codebook. The codebook can be generated off-line and stored in a storage medium or in the memory of the transmitter and receiver. In general, codebook generation is extremely important because codebook generation is related to system performance. According to the embodiment described above, each element C _{k of the} codebook is an N _t × N _t matrix. Therefore, a tracking precoding scheme is envisaged as described above, so the codebook must be generated so that it can be tracked.

  FIG. 6 shows an embodiment of a codebook generator 400 that includes a generator 405 and a combiner 410. According to this embodiment, the codebook generator 400 can use one or more updates used in a precoding scheme with a transmission scheme using a first number of transmit antennas and a second number of receive antennas. Each matrix is calculated, and a codebook including these update matrices is generated. Here, the first number is greater than 1 and the second number is 1 or more. The code book generating apparatus 400 generates a first matrix and obtains a second matrix associated with the first matrix with a certain degree of correlation (405), a first matrix and a first matrix And a combiner (410) for obtaining one update matrix by synthesizing two matrices.

  The code book generating apparatus 400 generates a code book including at least two update matrices used in precoding. Each update matrix is associated with an index of the update matrix. The update matrix is obtained by combining the step of obtaining a first matrix, the step of obtaining a second matrix associated with the first matrix by a certain degree of correlation, and the first matrix and the second matrix. And calculating one update matrix. In order to calculate different update matrices, the same correlation degree as that of the different first matrix or a different correlation degree is used.

  In one embodiment, the generator 405 has a magnitude in the range of 0.5 to 1.5, and generates a column vector whose angle between any two column vectors is in the range of 45 ° to 145 °. Find the matrix you have. In another embodiment, the generator 405 can determine a unitary matrix that is a matrix with orthogonal columns or a substantially orthogonal matrix, for example, based on a random or pseudo-random matrix. In another embodiment, each update matrix has an associated unique update matrix index. In another embodiment, the synthesized update matrix has as its diagonal elements elements that are at least twice the size of any off-diagonal elements. The degree of correlation between the first matrix and the second matrix can be realized by a generator having a coherence time, for example. That is, the degree of correlation corresponds to a time shift between the two generated matrices. The degree of correlation can also be related to the fading characteristics of the radio channel. In one embodiment of a larger codebook, a matrix with a plurality of different correlations is used to determine the update matrix. For example, the first matrix and the second matrix can be generated by the Jakes model.

  In another embodiment of the codebook generator 400, the generator 405 obtains a random matrix for applying singular value decomposition to obtain a unitary matrix, or triangulation of any matrix and the first matrix and A second matrix can be determined. In this case, the unitary matrices are multiplied together to obtain an update matrix that is also a unitary matrix having a tracking function. In another embodiment, the code book described above is stored on a storage medium. When using a matrix codebook, preferably each element matrix of the codebook is a unitary matrix, which has orthogonal columns that provide better performance and lower complexity.

による行列Ｇであり、式１６に示したように１つの更新用行列Ｃ_mが得られる。図７の実施形態によれば、コードブックの各要素は、２つのランダムユニタリ行列を掛け合わせることによって生成される。したがって、結果として生じる行列もユニタリ行列である。加えて、２つのユニタリ行列は、互いに時間シフトしたバージョンである。すなわち、ジェネレータがコヒーレンス時間を有する限り、これらの２つの行列は関連付けられており、そのため、結果として得られるコードブックの要素はトラッキング機能を有する。より具体的には、再び式１６を参照すると、２つのサンプル周期の間、すなわち２つのフィードバック周期の間のチャネルの変動が小さいとき、Ｃ_mの非対角要素ε_ijは、対角要素ｇ_iiに比べて小さい。かかる行列は、より細かいプリコーディング行列の調整のためのものである。２つのサンプル周期の間のチャネル変動が、例えば、無線チャネルの高速フェージングなどにより大きい場合、非対角要素ε_ijは、より動的なプリコーディング行列の調整を可能とするために、対角要素ｇ_iiと同程度の大きさとする必要がある。したがって、各要素行列が様々なΔｔから生成されてコードブックが求められる場合、更新用行列は、様々なチャネル環境に対して生成できる。したがって、各実施形態は、コードブックが、それぞれ相関度に基づいて、時間シフトΔｔに基づく様々な環境に適応できるという利点を提供する。 FIG. 7 shows an embodiment of a codebook generation method. A first unitary matrix V (t) is generated by a random unitary matrix generator based on a given time shift Δt. The two matrices are then multiplied at step 440 and the result is

A matrix G is obtained, and one update matrix C _m is obtained as shown in Expression 16. According to the embodiment of FIG. 7, each element of the codebook is generated by multiplying two random unitary matrices. Accordingly, the resulting matrix is also a unitary matrix. In addition, the two unitary matrices are time-shifted versions of each other. That is, as long as the generator has coherence time, these two matrices are related so that the resulting codebook elements have a tracking function. More specifically, referring again to Equation 16, when the channel variation between two sample periods, ie, two feedback periods, is small, the non-diagonal element ε _ij of C _m is the diagonal element g Smaller than _ii . Such a matrix is for finer adjustment of the precoding matrix. If the channel variation between two sample periods is larger, eg due to fast fading of the radio channel, the off-diagonal element ε _ij is a diagonal element to allow for more dynamic precoding matrix adjustments. It needs to be as large as g _ii . Therefore, when each element matrix is generated from various Δt and a code book is obtained, an update matrix can be generated for various channel environments. Thus, each embodiment provides the advantage that the codebook can be adapted to different environments based on the time shift Δt, each based on the degree of correlation.

FIG. 8a illustrates pseudo code for one embodiment of a method for generating a codebook. In the pseudo code shown in FIG. 8a, B is the size of the code book. Thus, the loop is repeated B times, as indicated by the “for” loop of step 450 and step 460. In step 451, two matrices of size N _t × N _t that are time shifted by Δt are generated. In step 452, the singular value decomposition of both matrices is calculated, and in step 453, the two unitary matrices obtained in step 452 are multiplied to obtain one update matrix C _m . In this way, the loop is performed B times. The random matrix generated in step 451 can also be obtained based on the Jakes model, which is also widely used for obtaining fading coefficients. Singular value decomposition is used to find a unitary matrix. In step 453, each element of the codebook is determined. Furthermore, the process of FIG. 8a can be repeated a sufficient number of times to obtain an optimal set of codebooks or an optimal set of update matrices in terms of channel capacity that can be achieved depending on the environment.

として定めることができる。すなわち、Ｖ（ｔ）Ｖ（ｔ＋Δｔ）^Hのサブセット、例えば行ベクトルまたは列ベクトルなどを使って、コードブックの更新用行列が求められる。一実施形態では、コンバイナ（４１０）は、第１の行列と第２の行列との合成のサブセットのみを使って更新用行列を求めることができ、別の実施形態では、この部分が行ベクトルまたは列ベクトルに対応する。 In another embodiment, a subset of matrix G is

Can be determined as That is, a codebook update matrix is obtained using a subset of V (t) V (t + Δt) ^H , for example, a row vector or a column vector. In one embodiment, the combiner (410) may determine the update matrix using only a subset of the first and second matrix synthesis, and in another embodiment, this portion may be a row vector or Corresponds to a column vector.

  The codebook generator (400) can comprise a generator for determining a first matrix and a second matrix in order to optimize the channel capacity by several trials. The code book generating apparatus (400) may further include means for quantizing the update matrix. The update matrix may be a square matrix in another embodiment.

であってもよい。アンテナ間の相関や、アンテナ間または伝搬経路間の平均電力の不均衡など、実際の伝搬チャネルの特性を考慮してコードブックを生成する場合、リンクまたはシステム容量をさらに一層増加させることができる。さらに、コードブックは、単位行列に近い更新用行列、または単位行列と同一でさえある更新用行列を含むこともできる。これは、場合によってはプリコーディング行列を更新しないというオプションを提供するものである。この場合、プリコーディング行列を更新しない方がより有利となることもある。 The matrix determined as part of the pseudo code shown in FIG. 8a has propagation channel characteristics. In one embodiment, these matrices can be determined as mutually random independent identically distributed (iid) variables. That is, the degree of correlation can be zero. In another embodiment, these matrices can be determined as random variables that are correlated with each other. Here, the correlation value or the correlation degree can be adjusted. In another embodiment, a range of different correlation values are covered. There may also be an imbalance between the average power of each element. That is,

It may be. When the codebook is generated in consideration of characteristics of actual propagation channels such as correlation between antennas and average power imbalance between antennas or propagation paths, the link or system capacity can be further increased. In addition, the codebook may include an update matrix that is close to the unit matrix, or an update matrix that is even identical to the unit matrix. This provides the option of not updating the precoding matrix in some cases. In this case, it may be more advantageous not to update the precoding matrix.

FIG. 8b is a graph depicting the improvement in capacity that can be achieved using one embodiment. In FIG. 8b, two graphs of link capacity versus average signal-to-noise ratio (SNR) are shown. The upper graph is achieved using this embodiment, and the lower graph is achieved using a conventional scheme. Consider the example scenario where the number of transmit antennas is N _t = 4, the number of receive antennas is N _r = 2 and the number of data streams is N _s = 2. In this embodiment, a capacity improvement of about 0.3 b / (s Hz) can be realized.

  Codebook-based precoding with tracking uses channel coherence between two sample points, i.e., one feedback period of coherence, and provides more capacity than one-shot codebook-based precoding. However, reliable precoding matrix tracking can be difficult if feedback errors occur or if there is a sudden change in the propagation channel. Therefore, instead of providing a difference of update information as in codebook-based precoding having a tracking function, in some cases, providing complete precoding information as in one-shot codebook-based precoding. Is more advantageous. FIG. 9 shows a receiving apparatus 700 including a detection unit 710 and a transmission unit 720. The receiving apparatus 700 provides precoding information, and includes a detection unit 710 that detects an event, and transmits complete precoding information at a first time in response to detection of the occurrence of the event. A transmission unit 720 that transmits the differential precoding information at a second time different from the time.

  In one embodiment, the detection unit 710 detects an elapse of a predetermined time from the previous time when the complete precoding information is sent as an event, or detects a complete precoding more complete than the differential precoding information for the communication link. A situation where coding information is advantageous is detected as an event.

  FIG. 10 shows another embodiment of the receiving device 750. The receiving device 750 includes a channel estimation unit 755 that provides a channel estimation value to a codebook searcher 760 that searches a one-shot codebook. Channel estimator 755 also provides channel estimates to another codebook searcher 765 that searches in a trackable codebook. In the embodiment of receiving device 750 shown in FIG. 10, codebook searcher 765 and codebook searcher 760 access the same codebook 770. In another embodiment, these codebook searchers may have separate codebooks. The receiving device 750 further includes a memory unit 775. This memory unit 775 provides the previous precoding matrix to the codebook searcher 765 having a tracking function to enable the search of the aforementioned codebook. Furthermore, the embodiment shown in FIG. 10 includes a switch 780. The switch 780 can transmit precoding information based on a one-shot codebook, that is, complete precoding information, while the trackable update information based on a codebook having a tracking function, that is, differential precoding information. It shows that coding information can be transmitted.

  In another embodiment of the receiving device 700, each of the detectors 710 may benefit from complete precoding information based on a range of complete precoding information and various differential precoding information. Is detected. In this embodiment, differential precoding information is provided to track slight changes in the radio channel. However, the codebook cannot track faster changes in the radio channel, e.g. induced by very fast fading, high mobility or Doppler scenarios. Whenever the channel change is too great, the detector 710 can decide to transmit complete precoding information. In another embodiment, the sending unit 720 sends an indicator that indicates whether the currently transmitted precoding information is complete or differential. This indicator can be implemented by a single bit transmitted over a feedback or control channel that indicates how the contained control information should be interpreted.

  In another embodiment, a receiving device can receive a multicarrier signal and provide precoding information for each carrier or series of carriers in the multicarrier signal. The receiving apparatus (700) can provide precoding information as an index of a matrix codebook, for example. In another embodiment, the receiving device (700) can provide differential precoding information based on a matrix codebook, where the matrix codebook contains a special codebook index for complete precoding information. Complete precoding information is provided based on the column vector of the complete precoding matrix.

  FIG. 11 shows an embodiment of the transmission device 800. Transmitting apparatus 800 transmits a signal using precoding updated based on precoding information, and includes a detection unit 810 that detects an event, and a first time according to detection of the occurrence of the event. Means 820 for performing update using complete precoding information and performing update using differential precoding information at a second time different from the first time. In one embodiment, the detection unit 810 detects that a predetermined time has elapsed from the time when the complete precoding information is received as an event, or is more complete than the differential precoding information for the communication link. A situation where precoding information is advantageous is detected as an event. In another embodiment, the transmitting apparatus 800 further comprises a receiver that receives complete precoding information or differential precoding information. In another embodiment, the receiver provides an indicator of whether a situation has occurred where complete precoding information is advantageous over differential precoding information. In one embodiment, the transmitting device can determine the situation by estimating the radio channel itself.

FIG. 12 shows another embodiment of the transmitter 850. The transmitter 850 includes a memory 855 that stores the previous precoding matrix F _k−1 . The previous precoding matrix F _k-1 is provided to a precoding matrix codebook searcher 860 having a tracking function. The transmission apparatus 850 includes a one-shot mode precoding matrix searcher 865, a tracking mode precoding matrix codebook searcher 860, and a common codebook 870. The transmission device 850 further includes a switch 875. The switch 875 can switch between tracking, that is, precoding based on differential precoding information, and precoding based on one-shot update, that is, precoding based on complete precoding information. In another embodiment of the transmitting device 800, separate codebooks can be used for complete precoding information and differential precoding information.

  The transmitter (800), in one embodiment, transmits a multicarrier signal and precodes each carrier or series of carriers in the multicarrier signal using differential precoding information or complete precoding information. The transmitting apparatus (800) can pre-code based on the matrix codebook index. The matrix codebook may have a special codebook index for complete precoding information, and the complete precoding information is provided based on the column vector of the complete precoding matrix.

  Each of the embodiments described above provides a mechanism for switching between tracking codebook mode and one-shot codebook mode in order to cope with situations where abrupt channel changes or feedback errors occur. According to one embodiment, the generation of the precoding matrix and the codebook search in the transmission device are performed in a tracking codebook mode or a one-shot codebook mode. In one embodiment, one codebook is shared between the two modes, while in another embodiment, separate codebooks are used. For example, in one embodiment, switching between the two modes is periodically performed based on a predetermined time recognized by the transmission device and the reception device. In another embodiment, the switching is performed based on feedback from the receiving device, eg, by an indicator used in the feedback.

  FIG. 13 shows a table with four cells, the left cell representing the transmitting device, the right cell representing the receiving device, the upper cell representing the one-shot mode, and the lower cell representing the tracking mode. Yes. In one-shot mode, as described in detail in the document “R1-050903, Qualcomm, Europe,“ Description and link simulations of MIMO schemes for OFDMA based E-UTRA downlink evaluation ””, the column vector index is precoded vector. It is preferable to use as.

As can be seen from FIG. 13, in the one-shot mode transmission apparatus, a column vector is selected from a codebook and used as a precoding matrix. In the example shown in FIG. 13, four transmit antennas are used, ie N _t = 4. In the one-shot mode receiver, for example, as shown in FIG. 13, column vectors are appropriately selected from the same codebook, that is, the same matrix, in order to maximize the channel capacity. The preferred column portion is then notified to the transmitting device. In the tracking mode, the transmission apparatus applies the update matrix notified as described above to the previous precoding matrix. The receiving apparatus evaluates the index of the update matrix optimized as described above, and feeds it back to the transmitting apparatus. In the tracking mode, an initial precoding matrix determined in the one-shot mode is set using the one-shot mode. The initial precoding matrix can then be updated using the tracking mode.

FIG. 14 illustrates one embodiment of a signaling scheme. It is assumed that the signaling word has a plurality of digits based on the signaling word 1400. In the embodiment shown in FIG. 14, signaling word 1400 includes an indicator in its first digit indicating whether tracking mode or one-shot mode is used. The other digits represent the codebook index in the tracking mode and the column portion of the matrix in the one-shot mode, respectively. Referring again to the example of four transmit antennas, ie N _t = 4, two sample codewords 1410 and 1420 are obtained. The sample signaling word 1410 includes 1 indicating the one-shot mode in its first digit. Therefore, the transmitting device, potentially recognize that the remaining digits represents the index C ₁ column portion. Upon receiving the signaling word 1420, the transmitting device recognizes that the tracking mode is used by the first digit being zero. In this case, the remaining 4 digits of the signaling word, for example, represents the codebook index of the codebook from C ₂ to C _17. In FIG. 14, the code book describing the tracking mode and the one-shot mode includes a matrix having B = 17 as the number, and the first one C ₁ is dedicated to the one-shot mode, and C ₂ to C _17. Up to is for tracking mode only.

  The first bit of the 5-bit index is a bit indicating whether the tracking mode or the one-shot mode is used. Each of the other bits is an index that is a subscript of the column in the case of one shot, and is a codebook index in the case of tracking. Thus, the same feedback bit is used in both modes.

  If the switching between the two modes is performed periodically according to a predetermined time scheme, the first bit x in FIG. 14 is clearly unnecessary and can be omitted. Each of the above-described embodiments can be directly combined with, for example, an OFDM (Orthogonal Wave Frequency Division Multiplexing) system. In this case, precoding needs to be performed on a subcarrier or a subcarrier chunk. In general, these embodiments can be used with any number of connection schemes.

FIG. 15 is a graph depicting the improvement in capacity that can be achieved using one embodiment. FIG. 15 shows a diagram of two capacities with respect to the average SNR in the unit of b / (s Hz) in the case of this embodiment and the conventional scheme of only the one-shot mode. In the illustrated scenario, the number of transmit antennas is N _t = 4, the number of receive antennas is N _r = 2 and the number of data streams is N _s = 2. When severe radio channel changes are expected in certain environments, conventional schemes use only the one-shot mode, while this embodiment can utilize the switching mechanism described above. The capacity achieved using this embodiment is shown above, and the capacity achieved using the conventional concept is shown below in FIG. It can be seen that this embodiment can achieve a capacity gain of about 0.5 b / (s Hz) using one spare bit as the mode indicator.

FIG. 16a is a graph evaluating the link capacity in units of b / (s Hz) with respect to the number of feedback bits. In the simulation of the capacity graph shown in FIG. 16a, the number of transmitting antennas is N _t = 4, the number of receiving antennas is N _r = 2 and the number of data streams is N _s = 2. In this graph, conventional schemes are indicated by circles, and this embodiment is indicated by crosses. It can be seen that this embodiment can achieve a higher capacity regardless of the number of feedback bits used. Thus, the advantage of this embodiment is that it provides higher link capacity and system capacity.

FIG. 16b is another graph showing the link capacity in terms of b / (s Hz) with respect to the average SNR. In the simulation shown in FIG. 16b, the number of transmitting antennas is N _t = 4, the number of receiving antennas is N _r = 2, and the number of data streams is N _s = 2. The graph includes three capacity graphs: the method according to the present invention indicated by a cross, the conventional method indicated by a circle, and the theoretical upper bound (Shannon limit). ). The graph shows that the capacity of the method according to the present invention is closer to the upper bound than the capacity of the conventional method. Thus, embodiments provide the advantage that the capacity of the link or system is greater than that of conventional approaches.

  In summary, the embodiments provide the advantage that greater capacity can be achieved in a MIMO transmission system using precoding. In addition, the embodiment requires less feedback than conventional techniques. This improves the spectral efficiency of the MIMO system and can utilize additional capacity with less feedback.

  Depending on some implementation requirements of the inventive method, the inventive method can be implemented as hardware or software. The implementation may be performed using a digital storage medium, in particular a disk, DVD or CD, on which electronically readable control signals are stored that perform the method of the invention in cooperation with a programmable computer system. it can. Thus, in general, the present invention is a machine-readable medium and program code causes a computer to perform the method according to the present invention. Therefore, in other words, the method according to the present invention is a computer program including program code that causes a computer to execute at least one of the methods according to the present invention.

It is a block diagram which shows one Embodiment of a receiver. It is a block diagram which shows another embodiment of a receiver. It is a block diagram which shows one Embodiment of a transmitter. It is a block diagram which shows another embodiment of a transmitter. 1 is a block diagram illustrating an embodiment of a MIMO transmission system. FIG. It is a block diagram which shows one Embodiment of a code book production | generation apparatus. It is explanatory drawing which shows one Embodiment of the method of calculating | requiring several update matrix. It is explanatory drawing which shows the pseudo code of one Embodiment of the method of calculating | requiring several update matrixes. 6 is a graph showing the improvement in capacity achievable with each embodiment. It is a block diagram which shows another embodiment of a receiver. It is a block diagram which shows another embodiment of a receiver. It is a block diagram which shows another embodiment of a transmitter. It is a block diagram which shows another embodiment of a transmitter. It is explanatory drawing which shows one Embodiment of the method of updating precoding information using complete precoding update information and differential precoding update information. FIG. 6 is an explanatory diagram illustrating an embodiment of a method for providing precoding information. 6 is a graph showing the improvement in capacity achievable with each embodiment. It is a graph of the simulation result showing the improvement of the capacity | capacitance by each embodiment. 6 is a graph of simulation results representing capacity improvement relative to the average signal to noise ratio achievable by each embodiment. It is explanatory drawing showing a general MIMO system. It is a graph of the channel coefficient of the mobile radio channel with respect to a time axis. It is a block diagram showing the structure of the conventional receiver. It is a block diagram of a transmitter.

Claims (31)

Determining one or more update matrices to be used in a precoding scheme in a transmission scheme using a first number of transmit antennas and a second number of receive antennas, and generating a codebook including the matrices A code book generation device (400), wherein the first number is greater than 1 and the second number is 1 or more;
A generator (405) for determining a first matrix and determining a second matrix associated with the first matrix with a certain degree of correlation;
A codebook generation device (400) comprising: a combiner (410) that combines the first matrix and the second matrix to obtain one update matrix.   The codebook generation according to claim 1, wherein a plurality of update matrices are determined, wherein the same correlation or different correlations with different first matrices are used to calculate different update matrices. Device (400).   The codebook generation device (400) according to claim 1 or 2, wherein the combiner (410) obtains the update matrix using only a subset of the combination of the first matrix and the second matrix. .   The codebook generator (400) of claim 3, wherein the subset is a row vector or a column vector.   The generator (405) determines a matrix having a column vector with a normalized magnitude of 0.5 to 1.5, or the angle between any two column vectors is 45 ° to 135 ° The code book generation device (400) according to any one of claims 1 to 4.   The codebook generation device (400) according to any one of claims 1 to 4, wherein the update matrix is a symmetric matrix or a complex symmetric matrix.   The codebook generation device (400) according to any one of claims 1 to 6, wherein the generator (405) determines a matrix having orthogonal columns or substantially orthogonal columns. The generator (405) obtains a unitary matrix or a substantially unitary matrix;
The code according to any one of claims 1 to 7, wherein the unitary matrix or the substantially unitary matrix has orthogonal columns or substantially orthogonal columns, or a column vector whose norm is 1 or approximately 1. Book generator (400).   The codebook generation device (400) according to any one of claims 1 to 8, wherein the generator (405) calculates a random matrix or a pseudo-random matrix.   The codebook generation device (400) according to any one of claims 1 to 9, wherein each of the update matrices is associated with a unique index of the update matrix.   The generator (405) includes the first matrix and the matrix so that the average size of diagonal elements of the obtained update matrix is at least twice the average size of all off-diagonal elements. The codebook generation device (400) according to any one of claims 1 to 10, wherein the codebook generation device (400) is for obtaining a second matrix.   The codebook generation device (400) according to any one of claims 1 to 11, wherein the updating matrix is a square matrix.   The code according to any one of the preceding claims, wherein the generator (405) determines the first matrix and the second matrix using a time shift related to the degree of correlation. Book generator (400).   The codebook generation apparatus (400) according to any one of claims 1 to 13, wherein the correlation is related to a fading characteristic of a radio channel.   15. The codebook generation device (400) according to any one of claims 1 to 14, wherein the generator (405) obtains a matrix using a plurality of different correlation degrees.   The code book generation device (400) according to any one of claims 1 to 15, wherein the generator (405) is configured to obtain the first matrix and the second matrix based on a Jakes model. .   The generator (405) obtains the first matrix and the second matrix by obtaining a random matrix to which matrix triangulation is applied, or the elements of the first matrix and the second matrix The code book generation device (400) according to any one of claims 1 to 16, wherein the elements of the matrix follow an independent same distribution.   18. The generator (405) according to any one of claims 1 to 17, wherein the generator (405) determines the first matrix and the second matrix in order to optimize channel capacity by several trials. Code book generation apparatus (400).   19. The codebook generator (1) according to any one of claims 1 to 18, further comprising means for quantizing the update matrix, or wherein the generator (405) obtains a quantized matrix. 400).   The code book generation device (400) according to any one of claims 1 to 19, wherein the combiner (410) combines the first matrix and the second matrix by multiplication. A method for obtaining one or more update matrices used in a precoding scheme in a transmission scheme using a first number of transmitting antennas and a second number of receiving antennas, wherein the first number is 1 And the second number is greater than or equal to 1,
Obtaining each of the update matrices by a sub-step of obtaining a first matrix and a sub-step of obtaining a second matrix associated with the first matrix by a certain degree of correlation;
Combining the first matrix and the second matrix to obtain one update matrix. A computer program comprising program code for causing a computer to execute the method of claim 21 . A codebook including at least two update matrices used in precoding,
Each of the update matrices is associated with an index of the update matrix;
The update matrix combines the step of obtaining a first matrix, the step of obtaining a second matrix associated with the first matrix by a certain degree of correlation, and the first matrix and the second matrix And calculating one update matrix, and
A codebook in which the same degree of correlation as the different first matrix or a different degree of correlation is used to calculate different update matrices. The codebook includes at least two update matrices used in precoding;
Each of the update matrices is associated with an index of the update matrix;
The update matrix combines the step of obtaining a first matrix, the step of obtaining a second matrix associated with the first matrix by a certain degree of correlation, and the first matrix and the second matrix And calculating one update matrix, and
A storage medium storing the codebook, wherein the same correlation or different correlation with a different first matrix is used to calculate a different update matrix.   The storage medium according to claim 24, wherein the codebook is quantized and stored. In a transmission scheme using a first number of transmitting antennas and a second number of receiving antennas, a storage medium provided in a receiving apparatus for controlling precoding in a transmitting apparatus,
The first number is greater than 1 and the second number is 1 or greater;
The receiving device is
A channel estimation unit that estimates a radio channel between the transmission device and the reception device based on a received signal and obtains a radio channel estimation value;
A processing unit for providing a new precoding matrix based on the update matrix included in the codebook of the update matrix and a previous precoding matrix, wherein the update matrix includes at least two A processing unit having a row and at least two columns;
A searcher for searching for a new precoding matrix satisfying an optimization criterion based on a radio channel estimation value among the new precoding matrices;
A transmission unit that transmits an index of an update matrix associated with the new precoding matrix,
The storage medium according to claim 24 or 25. A storage medium provided in a transmission device for transmitting a precoded signal,
The signal is precoded with an updated precoding matrix in a transmission scheme using a first number of transmit antennas and a second number of receive antennas;
The first number is greater than 1 and the second number is 1 or greater;
The transmitting device is
A receiving unit for receiving the update index; and
Storage means for storing a codebook including different update information items, each of the update information items including an update index associated with itself;
An update matrix generator for determining an update matrix having at least two rows and at least two columns using the received update index and the update information item associated in the codebook;
Combining a precoding matrix used before and the update matrix to obtain an updated precoding matrix,
The storage medium according to any one of claims 24 to 26. A storage medium provided in a receiving device that provides precoding information,
The receiving device is
A detection unit for detecting an event;
A transmission unit that transmits complete precoding information at a first time in response to detection of occurrence of the event, and transmits differential precoding information at a second time different from the first time. is there,
The storage medium according to any one of claims 24 to 27.   The detection unit detects, as the event, that a predetermined time has elapsed since the previous time at which the complete precoding information was transmitted, or for the communication link, the complete precoding information than the differential precoding information. 28. A storage medium according to claim 27, wherein the event detects that coding information is advantageous. A storage medium provided in a transmission device that transmits a signal using precoding,
The precoding is updated based on precoding information;
The transmitting device is
A detection unit for detecting the occurrence of an event;
Updating is performed using the complete precoding information at a first time in response to detection of occurrence of an event, and updating is performed using the differential precoding information at a second time different from the first time. Means and
The storage medium according to any one of claims 24 to 29.   The detection unit detects, as an event, that a predetermined time has passed since the previous time at which the complete precoding information was received, or the complete precoding information is more than the differential precoding information for a communication link. The storage medium according to claim 30, wherein the storage medium is detected as an event.

Images