KR20150112802A - Apparatus and method for channel information feedback in wireless communication system - Google Patents

Abstract

The present invention relates to channel information feedback in a wireless communication system, and a method of operating a receiving end includes the steps of transmitting an indicator indicating an indexing rule of channel values and channel information quantized in units of blocks, Received beamformed signals.

Description

[0001] APPARATUS AND METHOD FOR CHANNEL INFORMATION FEEDBACK IN WIRELESS COMMUNICATION SYSTEM [0002]

The present invention is for feedback of channel information in a wireless communication system.

The concept of adding multiple transmit antennas, called massive or large-scale multiple-input multiple-output (MIMO) systems, has attracted considerable interest from industry and academia for many years. For the beamforming gain and spatial multiplexing gain of a large-scale wireless communication system, channel state information (CSI) between a transmitter and a receiver can be provided to the transmitter. In a large-scale wireless communication system, in the case of time division duplexing (TDD), the acquisition of channel state information at the transmitting end depends on channel reciprocity without a pilot transmission and a channel estimation / feedback step. . However, in the case of frequency division duplexing (FDD), since the frequency bands of both channels are different, the channel state information is difficult to obtain depending on the channel reciprocity.

An embodiment of the present invention provides an apparatus and a method for feeding back channel information in a wireless communication system.

Another embodiment of the present invention provides an apparatus and method for indicating a characteristic of channel information fed back in a wireless communication system.

Another embodiment of the present invention provides an apparatus and method for informing a preferred domain of channel information fed back in a wireless communication system.

Another embodiment of the present invention provides an apparatus and method for indicating an indexing rule of channel entries for channel information fed back in a wireless communication system.

Yet another embodiment of the present invention provides an apparatus and method for feeding back quantized channel information in a wireless communication system.

Another embodiment of the present invention provides an apparatus and method for reducing overhead for feedback of channel information in a wireless communication system.

Another embodiment of the present invention provides an apparatus and method for feeding back block-wise configured channel information in a wireless communication system.

Another embodiment of the present invention provides an apparatus and method for selectively feeding back channel information blocked by different rules in a wireless communication system.

Another embodiment of the present invention provides an apparatus and method for feeding back trellis code quantized channel information in a wireless communication system.

Another embodiment of the present invention provides an apparatus and method for generating a trellis extended codebook to be used for quantization of channel information in a wireless communication system.

Yet another embodiment of the present invention provides an apparatus and method for feeding back quantized and phased channel information in a wireless communication system.

Another embodiment of the present invention provides an apparatus and method for reconstructing channel information from trellis code quantized information fed back in a wireless communication system.

Yet another embodiment of the present invention provides an apparatus and method for reconstructing channel information from quantized and phased information in a wireless communication system.

A method of operating a receiving end of a wireless communication system according to an embodiment of the present invention includes the steps of transmitting an indicator indicating an indexing rule of channel values and channel-wise quantized channel information, And receiving beamformed signals mapped to the antenna according to an indexing rule.

A method of operating a transmitter of a wireless communication system according to an embodiment of the present invention includes the steps of receiving an indicator indicating an indexing rule of channel values and channel information quantized on a block basis; And transmitting the shaped signals.

A receiving end apparatus of a wireless communication system according to another embodiment of the present invention includes: a transmitter for transmitting an indicator indicating an indexing rule of channel values and channel information quantized in a block unit; and a beamforming unit And a receiver for receiving signals.

According to another aspect of the present invention, there is provided a transmitting terminal apparatus of a wireless communication system, including: a receiving unit for receiving an indicator indicating an indexing rule of channel values and channel information quantized in units of blocks; and a beamforming unit And a transmitter for transmitting signals.

In a wireless communication system, in feedback of channel information quantized in block units, a beamforming gain can be improved by adaptively changing a domain for blocking a channel vector.

1 shows a transmitting end and a receiving end in a wireless communication system according to an embodiment of the present invention.
FIG. 2 illustrates a transmitting terminal configuration in a wireless communication system according to an embodiment of the present invention.
3 shows a receiving end configuration in a wireless communication system according to an embodiment of the present invention.
4 illustrates an indexing method of antenna elements in a wireless communication system according to an embodiment of the present invention.
5 illustrates mapping schemes of antenna elements in a wireless communication system according to an embodiment of the present invention.
6 illustrates an operation procedure of a receiver in a wireless communication system according to an embodiment of the present invention.
FIG. 7 illustrates an operation procedure of a transmitting end in a wireless communication system according to an embodiment of the present invention.
FIG. 8 illustrates a code word determination procedure of a receiver in a wireless communication system according to an embodiment of the present invention.
FIG. 9 illustrates a procedure of determining a mapping relationship of a transmitter in a wireless communication system according to an embodiment of the present invention.
10 illustrates a channel information feedback procedure in a wireless communication system according to an embodiment of the present invention.
FIG. 11 illustrates a channel information quantization procedure of a receiver of a wireless communication system according to an embodiment of the present invention.
12 illustrates a channel vector reconstruction procedure of a transmitter of a wireless communication system according to an embodiment of the present invention.
13 illustrates an embodiment of a quantizer for channel information quantization in a wireless communication system according to an embodiment of the present invention.
FIG. 14 shows an example of a trellis for channel information quantization in a wireless communication system according to an embodiment of the present invention.
15 shows another embodiment of a quantizer for channel information quantization in a wireless communication system according to an embodiment of the present invention.
16 shows another example of a trellis for channel information quantization in a wireless communication system according to an embodiment of the present invention.
17 shows an example of channel information in a wireless communication system according to an embodiment of the present invention.
18 shows an example of channel information quantized in a wireless communication system according to an embodiment of the present invention.
19 shows an example of a path search for reconstructing channel information in a wireless communication system according to an embodiment of the present invention.
20 shows another example of a path search for reconstructing channel information in a wireless communication system according to an embodiment of the present invention.
FIG. 21 illustrates a channel information feedback procedure in a wireless communication system according to an embodiment of the present invention.
FIG. 22 illustrates phase adjustment for channel quantization results in a wireless communication system according to an embodiment of the present invention.
23 shows phases applied to an antenna group in a wireless communication system according to an embodiment of the present invention.
24 illustrates an example of a phase adjuster implementation in a wireless communication system according to an embodiment of the present invention.
25 shows an example of a trellis for phase adjustment in a wireless communication system according to an embodiment of the present invention.
26 shows a simulation result of a wireless communication system according to an embodiment of the present invention.

Hereinafter, the operation principle of the present invention will be described in detail with reference to the accompanying drawings. In the following description, well-known functions or constructions are not described in detail since they would obscure the invention in unnecessary detail. The following terms are defined in consideration of the functions of the present invention, and these may be changed according to the intention of the user, the operator, or the like. Therefore, the definition should be based on the contents throughout this specification.

Hereinafter, a technique for feeding back channel information in a wireless communication system will be described. Terms referring to signals used in the following description, terms referring to components of a channel, terms referring to devices, terms referring to codes, groups of channel values and terms referring to groups of antenna elements, For convenience of explanation. Therefore, the present invention is not limited to the following terms, and other terms referring to objects having equivalent technical meanings can be used.

Before describing various embodiments of the present invention in detail, the related art will be described. In describing the above techniques, the following documents will be referred to.

[1] C. K. Au-Yeung and D. J. Love, "On the performance of random vector quantization limited feedback beamforming in a MISO system," IEEE Trans. Wireless Commun., Vol. 6, no. 2, pp. 458-462, Feb. 2007.

[2] J. Choi, D. J. Love, and T. Kim, "Trellis-Extended Codebooks and Successive Phase Adjustment: A Path from LTE-Advanced to FDD Massive MIMO Systems," submitted to IEEE Trans. Wireless Commun., Jan. 2014.

[3] J. Li, X. Su, J. Zeng, Y. Zhao, S. Yu, L. Xiao, and X. Xu, "Codebook Design for Uniform Rectangular Arrays of Massive Antennas," IEEE VTC Spring, .

[4] X. Su, J. Zeng, J. Li, L. Rong, L. Liu, X. Xu, and J. Wang, "Limited Feedback Precoding for Massive MIMO," International Journal of Antennas and Propagation,

[5] D. Ying, F. W. Vook, T. A. Thomas, D. J. Love, A. Ghosh, "ICC 2014, accepted as a 3D Channel Model for Feedback Correlation Model and Limited Feedback Codebook Design.

[6] D. J. Ryan, I. V. L. Clarkson, I. B. Collings, D. Guo, and M. L. Honig, "QAM and PSK codebooks for limited feedback MIMO beamforming," IEEE Transactions on Communications, vol. 57, no. 4, pp. 1184-1196, Apr. 2009.

[7] J. Choi, Z. Chance, D. J. Love, and U. Madow, "Noncoherent trellis-coded quantization for massive MIMO limited feedback beamforming," UCSD Information Theory and Applications Workshop, Feb. 2013.

[8] J. Choi, D. J. Love, and U. Madow, "Limited feedback in massive MIMO systems-exploiting channel correlations via noncoherent trellis-coded quantization," Proceedings of the Conference on Information Sciences and Systems, Mar. 2013.

[9] J. Choi, Z. Chance, D. J. Love, and U. Madhaw, "Noncoherent Trellis Coded Quantization: A Practical Limited Feedback Technique for Massive MIMO Systems," submitted to IEEE Transactions on Communications.

[10] W. Sweldens, "Fast block noncoherent decoding," IEEE Communications Letters, vol. 5, no. 4, pp. 132-134, Apr. 2001.

For the sake of simplicity of the description, the present invention exemplifies a case such as that shown in Fig. 1, in which the transmitting end has a plurality of (for example, M) transmitting antennas and the receiving end has a single receiving antenna. 1 shows a transmitting end and a receiving end in a wireless communication system according to an embodiment of the present invention.

를 추정하고, 채널 벡터 h[k]를 상기 송신단 110으로 송신할 수 있다. 상기 채널 벡터 h[k]를 전달하는 채널은 '피드백 채널(feedback channel)'이라 지칭될 수 있다. 상기 채널 벡터 h[k]는 채널 정보로서, 상기 채널 정보는 '채널 상태 정보(CSI: channel state information)'라 지칭될 수 있다. 1, the transmitter 110 generates a transmission signal x [k] by processing the data s [k] and transmits the transmission signal x [k]. The receiving end 120 receives the signal y [k]

And transmit the channel vector h [k] to the transmitter 110. [ The channel carrying the channel vector h [k] may be referred to as a 'feedback channel'. The channel vector h [k] is channel information, and the channel information may be referred to as 'channel state information (CSI)'.

Here, the channel vector h [k] may be quantized. The quantization means a process of determining a codeword corresponding to the channel vector h [k] estimated in the codebook. Accordingly, the channel information may include at least one of a channel matrix / vector, a codeword index, a precoding matrix index (PMI), and a rank index (RI). According to an embodiment of the present invention, the quantization may be performed based on trellis-coded. In this case, the receiver 120 generates a binary vector b [k] of the B _tot dimension by quantizing the channel information, and transmits the binary vector b [k] to the transmitter 110 through the feedback channel . Accordingly, the transmitter 110 may receive the channel information from the receiver 120 and construct a beamforming channel vector f [k] from the binary vector b [k]. The transmitting terminal 110 and the receiving terminal 120 shown in FIG. 1 may be configured as shown in FIG. 2 and FIG.

FIG. 2 illustrates a transmitting terminal configuration in a wireless communication system according to an embodiment of the present invention. Referring to FIG. 2, the transmitter 110 includes a communication unit 210, a storage unit 220, and a controller 230. Used below '... Wealth, '... Quot; and the like denote a unit for processing at least one function or operation, and may be implemented by hardware, software, or a combination of hardware and software.

The communication unit 210 performs functions for transmitting and receiving signals through a wireless channel. For example, the communication unit 210 performs a function of converting a baseband signal and a bit string according to a physical layer specification of the system. For example, at the time of data transmission, the communication unit 210 generates complex symbols by encoding and modulating transmission bit streams. Also, upon receiving the data, the communication unit 210 demodulates and decodes the baseband signal to recover the received bit stream. In addition, the communication unit 210 up-converts the baseband signal to a radio frequency (RF) band signal, transmits the signal through an antenna, and downconverts the RF band signal received through the antenna to a baseband signal. For example, the communication unit 210 may include a transmission filter, a reception filter, an amplifier, a mixer, an oscillator, a digital to analog converter (DAC), an analog to digital converter .

In addition, the communication unit 210 may include a plurality of RF chains. Further, the communication unit 210 may perform beamforming. For the beamforming, the communication unit 210 may adjust the phase and size of signals transmitted / received through a plurality of antennas or antenna elements.

The communication unit 210 transmits and receives signals as described above. Accordingly, the communication unit 210 may be referred to as a transmitting unit, a receiving unit, or a transmitting / receiving unit.

The storage unit 220 stores data such as a basic program, an application program, and setting information for the operation of the transmitter 110. In particular, the storage unit 220 may store a codebook for beamforming of a data signal. The storage unit 220 provides the stored data at the request of the controller 230.

The controller 230 controls overall operations of the transmitter 110. For example, the control unit 230 transmits and receives signals through the communication unit 210. Also, the control unit 230 records and reads data in the storage unit 220. For this, the controller 230 may include at least one processor. According to an embodiment of the present invention, the controller 230 includes a feedback analyzer 232 that analyzes feedback information received from the receiver 120. For example, the controller 230 may control the transmitting terminal 110 to perform the procedure shown in FIG. 7, FIG. 9, FIG. 10, FIG. 12, and FIG.

3, the receiving terminal 120 includes a communication unit 310, a storage unit 320, and a control unit 330. Referring to FIG. Used below '... Wealth, '... Quot; and the like denote a unit for processing at least one function or operation, and may be implemented by hardware, software, or a combination of hardware and software.

The communication unit 310 performs functions for transmitting and receiving signals through a wireless channel. For example, the communication unit 310 performs a function of converting a baseband signal and a bit string according to a physical layer specification of the system. For example, at the time of data transmission, the communication unit 310 generates complex symbols by encoding and modulating transmission bit streams. In addition, upon receiving the data, the communication unit 310 demodulates and decodes the baseband signal to recover the received bit stream. Also, the communication unit 310 up-converts the baseband signal to an RF band signal, transmits the RF band signal through the antenna, and down converts the RF band signal received through the antenna to a baseband signal. For example, the communication unit 310 may include a transmission filter, a reception filter, an amplifier, a mixer, an oscillator, a DAC, an ADC, and the like. The communication unit 310 transmits and receives signals as described above. Accordingly, the communication unit 310 may be referred to as a transmitting unit, a receiving unit, or a transmitting / receiving unit. 3, the receiving terminal 120 is illustrated as having one antenna. However, according to another embodiment of the present invention, the receiver 120 may include a plurality of antennas.

The storage unit 320 stores data such as a basic program, an application program, and setting information for the operation of the receiver 120. In particular, the storage unit 320 may store a codebook for feedback of channel information. The storage unit 320 provides the stored data according to the request of the controller 330. [

The controller 330 controls the overall operations of the receiver 120. For example, the control unit 330 transmits / receives a signal through the communication unit 310. In addition, the control unit 330 records and reads data in the storage unit 320. FIG. For this, the controller 330 may include at least one processor. For example, the controller 330 may include a communication processor (CP) for performing communication control and an application processor (AP) for controlling an upper layer such as an application program. According to an embodiment of the present invention, the controller 330 includes a feedback generator 332 for generating feedback information for providing channel information to the transmitter 120. For example, the controller 330 may control the receiving terminal 120 to perform the procedure shown in FIG. 6, FIG. 8, FIG. 10, FIG. 11, and FIG.

에 의존하며, 상기 VQ 코드북은 상기 송신단 110 및 상기 수신단 120 간 공유될 수 있다. 여기서, c_i는, 모든 i에 대하여,

인 Mx1 크기의 복소 벡터(complex vector)를 의미한다. 상기 VQ 코드북 방식(approach)을 이용하면, 상기 수신단 120은 하기 <수학식 1>과 같이 최적의 코드워드(optimal codeword)를 선택할 수 있다.In the case of a frequency division duplexing (FDD) system, a feedback channel for transmitting channel information from the receiving end 120 to the transmitting end 110 may have a limited capacity. Systems using most limited feedback channels, including the 3rd Generation Partnership Project (3GPP) Long-Term-Evolution (LTE), use a common vector quantized (VQ) codebook,

And the VQ codebook may be shared between the transmitting end 110 and the receiving end 120. Here, c _i , for all i,

Quot; means a complex vector of Mx1 size. Using the VQ codebook approach, the receiver 120 can select an optimal codeword as shown in Equation (1) below.

In Equation (1), C _opt denotes an optimal codeword, C denotes a codebook, c denotes a codeword in a codebook, and h denotes a channel vector.

로서, 코드워드들의 개수에 대해 지수적으로 증가한다. 즉, 상기 코드워드들의 개수 증가에 따른 복잡도의 급격한 증가로 인해, 큰 B_tot를 갖는 코드북에 대하여 전수 검색을 실시간으로 수행하는 것은 용이하지 아니하다.However, the whole number search can be performed when the total number of codewords is small (e.g., B _tot = 4), as in an LTE system, It is feasible. The computational complexity of the integer search for the optimal codeword is

, Which exponentially increases with respect to the number of codewords. That is, due to a rapid increase in the complexity due to the increase in the number of codewords, it is not easy to perform full-scale search for a codebook having a large B _tot in real time.

이고,

일 때, 고정 비율(fixed ratio)

을 갖는 최적의 VQ 코드북인 랜덤 VQ(RVQ: random vector quantization) 코드북을 이용하면, 정규화된(normalized) 빔포밍 이득(beamforming gain)에서의 손실은, [1]에 나타난 바와 같이, 하기 <수학식 2>와 같을 수 있다. In order for the quantization error of the channel information to have a certain level, the number of bits for the codebook must increase in proportion to the number of transmission antennas.

ego,

The fixed ratio,

The loss in the normalized beamforming gain can be expressed by the following equation (1), as shown in [1], by using the random VQ codebook (RQQ) 2>.

는 정규화된 채널 벡터, c_opt는 최적의 코드워드를 의미한다.In Equation (2), M is the number of antennas, B _tot is the codeword size, i.e., the number of bits of the codeword index,

Denotes a normalized channel vector, and c _opt denotes an optimal codeword.

Referring to Equation (2), it is apparent that the feedback overhead must be increased in proportion to M in order to maintain the beamforming loss normalized to a certain level. Therefore, the way of using the VQ codebook in combination with the whole number search is not useful for a large-scale MIMO system due to the complexity issue.

One way to solve the above problem is to quantize the channel vector in a block manner. For example, the receiver 120 may divide an M × 1 size channel vector into N blocks. Here, M / N is an integer. Then, the receiver 120 may quantize N × 1 size blocks using B _tot / N quantization bits per block. This can effectively reduce the complexity of channel information quantization. Recently, trellis-extended codebook (TEC) and trellis-extended successive phase adjustment (TE-SPA) have been proposed as block-based quantization techniques such as [2].

At this time, it is desirable to consider the structure of the channel vector and the antenna array. In the case of a planar antenna array, different channel quantization strategies can be performed by using the structure of the plane antenna array. The channel vector of the plane antenna array can be expressed as Equation (3).

이고, h_h∈

이다. 여기서,

는 복소수로 구성되는 M_h×1 크기의 벡터 집합,

는 복소수로 구성되는 M_v×1 크기의 벡터 집합을 의미한다.In Equation (3), h denotes a channel vector, h _v denotes a channel vector representing a vertical domain, and h _h denotes a channel vector representing a horizontal domain. Further, the h _v?

, H _h ∈

to be. here,

Is a vector set of M _h × 1 size composed of complex numbers,

Denotes a vector set of M _v × 1 size which is composed of complex numbers.

As in [3] - [5], the approximation may be based on the approximation of the spatial correlation matrix of the plane antenna array. The product of the dimension of h _v and the dimension of h _h must be equal to the number M of elements of the channel vector, that is, M = M _v × M _h . In this case, the easiest and most intuitive way of quantizing channel information is to rely on a kronecker product codebook. In the case of the Kronecker codebook, the receiver 120 may quantize the horizontal channel domain and the vertical channel domain using common or possibly different codebooks, respectively. At this time, codewords for the horizontal domain and the vertical domain can be determined as shown in Equation (4) below.

이고, h_h∈

이다. 여기서,

는 복소수로 구성되는 M_h×1 크기의 벡터 집합,

는 복소수로 구성되는 M_v×1 크기의 벡터 집합을 의미한다.In Equation (4), c _v is a codeword for a vertical domain, c _h is a codeword for a horizontal domain, c is a codeword candidate, h _v is a channel vector representing a vertical domain, h _h denotes a channel vector representing a horizontal domain. Further, the h _v?

, H _h ∈

to be. here,

Is a vector set of M _h × 1 size composed of complex numbers,

Denotes a vector set of M _v × 1 size which is composed of complex numbers.

When the receiver 120 feeds back indexes of two (2) codewords to the transmitter 110, the transmitter 110 may generate an M × 1 quantized channel. The quantized channel is expressed by Equation (5).

In Equation (5), c _v denotes a codeword for a vertical domain, c _h denotes a codeword for a horizontal domain, and c denotes a codeword for the entire channel.

As described above, the channel vector may be represented by a vertical domain or a horizontal domain. Thus, according to various embodiments of the present invention, the receiving terminal 120 may transmit additional feedback information indicating a preferred domain. For example, the additional information may be composed of one bit. The additional feedback information may be transmitted in a long-term or short-term depending on the state of the receiving terminal 120 or the transmitting terminal 110. For example, if the receiver 120 moves at the ground level, it may be preferable to select the horizontal domain. On the other hand, if the receiving terminal 120 is located in a tall building and a user descends the stairs, it may be preferable to select the vertical domain. Based on the selected domain, the receiver 120 re-indexes the channel entries to be quantized so as to have a better channel directionality of the channel. Thus, by adapting the channel to the preferred domain, a better quantization result for the channel information can be obtained.

For ease of explanation, a planar antenna array including 32 antenna elements of 4x8 size is illustrated. Each effective antenna element may be composed of four physical antenna elements. However, various embodiments of the present invention may be implemented for antenna arrays comprised of a different number of antenna elements. In the antenna array, the antenna elements may be indexed as shown in FIG.

4 illustrates an indexing method of antenna elements in a wireless communication system according to an embodiment of the present invention. In FIG. 4, one antenna element means an effective antenna element. 4, there are eight antenna elements in the first column 401, eight antenna elements in the second column 402, eight antenna elements in the third column 403, eight antenna elements in the fourth column 404 . Distance between the vertical direction between the antenna element is d1, the horizontal spacing d _2. At this time, the antenna elements may be indexed in the horizontal direction priority, as indicated by the arrows in FIG. Thus, the first column 401 includes the antenna elements of indices 1 through 8, the second column 402 represents the antenna elements of indices 9 through 16, the third column 403 represents the antenna elements of indices 17 through 24, Fourth column 404 includes the antenna elements of indices 25-32. In this case, the channel vector indexed in the same manner as the antenna elements is [h ₁ , ... , h ₃₂ ] ^T.

Hereinafter, for convenience of explanation, it is exemplified that a channel vector is quantized on a block-by-block basis using a 4TX codebook. However, various embodiments of the present invention can be applied to quantization of block-based channel information in any other manner.

Standardized codebooks such as LTE or discrete fourier transform (DFT) have good directionality. Therefore, in consideration of the preferred domain, considering the channel vector using the codebooks from the codebook and the plane antenna array as shown in FIG. 4, the selection of the preferred domain may be performed as follows.

From the codebook, two candidate codewords may be selected, i.e., a first candidate and a second candidate.

는 임의의(arbitrary) 교란된(perturbed) 채널 벡터로서, h_{p ,k}∈

(k=1, …, 8) 이고, 상기 C_{opt ,1}은 상기 h에 대응하는 코드워드의 제1후보, 상기 C_opt,2는 상기 g에 대응하는 코드워드의 제2후보를 의미한다.In Equation (6), C denotes a codebook, c denotes a codeword in a codebook, h denotes a channel vector,

Is an arbitrary perturbed channel vector, h _{p , k} ∈

(k = 1, ..., 8), where C _{opt , 1} is a first candidate of a codeword corresponding to h, and C _{opt, 2} is a second candidate of codeword corresponding to g.

Each block h _{p , k} can be quantized using 4TX LTE or DFT codebook. If different codebooks such as 2TX or 8TX codebooks are adopted for channel information quantization, the block size can be changed as appropriate. Using the first candidate and the second candidate, the receiving terminal 120 can compare two metrics. The above indices can be compared as shown in Equation (7) below.

In Equation (7), h denotes a channel vector indexed in the horizontal direction first, g denotes a channel vector indexed vertically, c _{opt , 1} denotes a first candidate of codeword corresponding to h, c _{opt , 2} denotes the second candidate of the codeword corresponding to g.

According to the comparison result, additional information indicating what the feedback codeword is among the first candidate and the second candidate may be transmitted. That is, the receiving terminal 120 may select one of the first candidate and the second candidate according to the comparison result, and transmit the selected channel information to the transmitting terminal 110. Here, the channel information is information indicating selected codewords or codewords, and may include an index of at least one codeword or values generated from the index. The values generated from the index may be composed of a smaller number of bits than the index.

The physical meaning of the above processes is as follows.

5 illustrates mapping schemes of antenna elements in a wireless communication system according to an embodiment of the present invention. FIG. 5 illustrates mappings between antenna elements and codewords according to a selected domain. In FIG. 5, (a) illustrates a case where a horizontal domain is selected, and (b) illustrates a case where a vertical domain is selected.

Referring to FIG. 5A, when the horizontal domain is selected, each codeword from the codebook is mapped to a group of four horizontally adjacent antenna elements. In other words, the antenna group includes adjacent antenna elements in the horizontal direction. That is, when the horizontal domain is selected, the channel values of the antenna elements included in the channel vector are indexed in the horizontal direction. On the other hand, referring to FIG. 5B, when the vertical domain is selected, each codeword from the codebook is mapped to a group of four vertically adjacent antenna elements. In other words, the antenna group includes vertically adjacent antenna elements. That is, when the vertical domain is selected, the channel values of the antenna elements included in the channel vector are indexed in the vertical direction. With the mapping as illustrated in FIG. 5, the finally selected code word has greater directionality in the selected domain.

6 illustrates an operation procedure of a receiver in a wireless communication system according to an embodiment of the present invention. 6 illustrates an operation method of the receiver 120. FIG.

Referring to FIG. 6, in step 601, the receiver transmits an indication indicating an indexing rule and channel information quantized (block-wise quantized) on a block-by-block basis. The indexing rule indicates the order of arrangement of channel values constituting the channel vector in the channel vector. In other words, the indexing rule indicates the indexing order of the channel values, i.e., channel entries, in the code word determination corresponding to the channel information at the receiving end. In other words, the indexing rule indicates correspondence between channel values and antenna elements included in the channel vector used in the codeword determination. That is, the channel information includes indices of codewords for each antenna group, and the indicator indicates information indicating a preferred domain used in determining the codewords. Here, the preferred domain includes a vertical domain or a horizontal domain. In other words, the indicator indicates whether the antenna elements were grouped horizontally prioritized or vertically prioritized in code word determination.

In step 603, the receiver receives the beamformed signal mapped to the antennas according to the indexing rule. The beamforming is performed by the transmitting end, and the beamformed signals with a plurality of codewords are transmitted through the antennas determined according to the indexing rule. That is, in mapping the beamformed signals, the transmitting terminal maps the beamformed signals to antenna elements in a vertical direction priority or a horizontal direction priority according to the indexing rule.

7 illustrates an operation procedure of a receiving end in a wireless communication system according to an embodiment of the present invention. 7 illustrates an operation method of the transmitting terminal 110. FIG.

Referring to FIG. 7, in step 701, the transmitter receives quantized channel information on a block-by-block basis and an indicator indicating an indexing rule. The indexing rule indicates the indexing order of the channel values, that is, the channel entries, in the code word determination corresponding to the channel information at the receiving end. In other words, the indexing rule indicates correspondence between channel values and antenna elements included in the channel vector used in the codeword determination. That is, the channel information includes indices of codewords for each antenna group, and the indicator indicates information indicating a preferred domain used in determining the codewords. Here, the preferred domain includes a vertical domain or a horizontal domain. In other words, the indicator indicates whether the antenna elements are blocked in the horizontal direction or the vertical direction in the codeword determination in the receiving end.

In step 703, the transmitter transmits a beamformed signal mapped to the antennas according to the indexing rule. The beamforming is performed by the transmitting end, and the beamformed signals with a plurality of codewords are transmitted through the antennas determined according to the indexing rule. That is, in mapping the beamformed signals, the transmitting terminal maps the beamformed signals to antenna elements in a vertical direction priority or a horizontal direction priority according to the indexing rule.

FIG. 8 illustrates a code word determination procedure of a receiver in a wireless communication system according to an embodiment of the present invention. 8 illustrates an operation method of the receiver 120. FIG.

Referring to FIG. 8, in step 801, the receiver estimates a channel between the transmitter and the receiver. To this end, the receiving end may receive a signal transmitted from the transmitting end. The signal may include at least one of a pilot signal, a reference signal, a training signal, a synchronization signal, and a preamble. At this time, the channel may be a matrix or vector having a size corresponding to the number of antennas of the transmitter and the number of antennas of the receiver. Hereinafter, for convenience of explanation, it is assumed that the channel has a form of a vector.

In step 803, the receiver quantizes the channel values on a block-by-block basis. Specifically, the receiving end divides a channel vector composed of channel entries, that is, channel values, by the number of antenna elements included in the antenna array of the transmitting end, and determines an optimal codeword for each block. Here, the size of the block may vary according to the size of the codebook used. At this time, the receiver quantizes each of a plurality of channel vectors in which the channel values are arranged according to a plurality of domains, that is, in a horizontal direction and a vertical direction. Accordingly, the channel values may be blocked to correspond to the grouped antenna elements as shown in FIG. 5 (a) or (b). As a result, quantization results corresponding to the number of selectable domains are obtained.

In step 805, the receiving end determines a preferred domain. That is, the receiver selects channel information having a larger beamforming gain among the quantized channel information corresponding to each domain, and determines a domain corresponding to the selected channel information as the preferred domain. For example, the receiver may select the channel information, i.e., codewords, that maximize the size of the effective channel. Specifically, if the condition of Equation (7) is satisfied, the receiving end can select a horizontal domain as a preferred domain.

FIG. 9 illustrates a procedure of determining a mapping relationship of a transmitter in a wireless communication system according to an embodiment of the present invention. 9 illustrates an operation method of the transmitting terminal 110. FIG.

Referring to FIG. 9, in step 901, the transmitting terminal checks an indexing rule for channel information fed back from a receiving end. The indexing rule indicates the indexing order of the channel values, i.e., channel entries, in the channel vector used in determining the codeword corresponding to the channel information at the receiving end. The transmitting end can check the indexing rule through an indicator fed back from the receiving end. Accordingly, the transmitting terminal can determine the preferred domain of the receiving terminal.

In step 903, the transmitter determines a mapping relationship between the codewords and the antenna elements according to the indexing rule. The indexing rule indicates correspondence between channel values and antenna elements included in the channel vector used in the codeword determination. Therefore, the transmitting terminal can determine whether to apply the beamformed signals with one codeword to the antenna group in the vertical direction or the antenna group in the horizontal direction according to the domain of the channel vector indicated by the indexing rule . In other words, the transmitting terminal determines, based on the indexing rule, which of the codewords indicated by the fed back channel information corresponds to channels of which antenna elements. For example, when the indexing rule indicates a horizontal domain, one codeword corresponds to the antenna group as shown in FIG. 5 (a). On the other hand, when the indexing rule indicates the vertical domain, one code word corresponds to the antenna group as shown in FIG. 5 (b).

10 illustrates a channel information feedback procedure of a wireless communication system according to an embodiment of the present invention. For example, the procedure shown in FIG. 6 may be performed by the transmitting terminal 110 and the receiving terminal 120.

Referring to FIG. 10, in step 1001, the transmitting terminal 110 transmits a pilot signal to the receiving terminal 120. The pilot signal has a predetermined value as a signal for channel estimation, and can be transmitted through a predetermined resource in advance. The pilot signal may be referred to as a reference signal, a training signal, or the like, and may be replaced with a synchronization signal, a preamble, or the like.

In step 1003, the receiving terminal 120 estimates a channel between the transmitting terminal 110 and the receiving terminal 120. The channel may be a matrix or a vector having a size corresponding to the number of antennas of the transmitter 110 and the number of antennas of the receiver 120. Hereinafter, for convenience of explanation, it is assumed that the channel is composed of a vector.

In step 1005, the receiving end 120 calculates a first candidate of a codeword corresponding to the channel. The codeword may be referred to as a precoding matrix, a beamforming matrix, or the like. In this case, the first candidate means a codeword determined based on a channel vector when channel values in the channel vector are arranged in the horizontal direction. For example, in the case of the planar antenna array shown in FIG. 4, the channel vector may include channel values at antenna element indices 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, , 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, At this time, the receiver 120 may determine a codeword that maximizes the size of the effective channel as the first candidate. Specifically, the receiver 120 may determine a codeword that maximizes a square of a product of the Hermitian of the horizontal domain channel vector as the first candidate.

In step 1007, the receiving terminal 120 calculates a second candidate of a codeword corresponding to the channel. The codeword may be referred to as a precoding matrix, a beamforming matrix, or the like. In this case, the second candidate means a codeword determined based on a channel vector in a case where channel values in the channel vector are arranged in the vertical direction. For example, in the case of the plane antenna array as shown in FIG. 4, the channel vector may include channel values at antenna element indices 1, 9, 17, 25, 2, 10, 18, 26, 3, 11, 19, , 12, 20, 28, 5, 13, 21, 29, 6, 14, 22, 30, 7, 15, 23, 31, 8, 16, 24, At this time, the receiver 120 may determine a codeword for maximizing the size of the effective channel as the second candidate. Specifically, the receiver 120 may determine a codeword that maximizes a square of a product of the hermetian of the vertical domain channel vector as the second candidate.

In step 1009, the receiving terminal 120 selects one of the first candidate and the second candidate. For this, the receiving terminal 120 may compare the indicators corresponding to the first candidate and the second candidate. Here, the indicator may be defined as a size of an effective channel, which is a square of a product of a hermetian of a channel vector. If the index of the first candidate is more than or equal to the index of the second candidate, the receiving terminal 120 selects the first candidate. On the other hand, if the index of the first candidate is less than the index of the second candidate, the receiving terminal 120 selects the second candidate.

If the first candidate is selected, the receiver 120 determines the first candidate as quantized channel information in step 1011. [ On the other hand, if the second candidate is selected, the receiver 120 determines the second candidate as the quantized channel information in step 1013. At this time, the quantized channel information represents a plurality of codewords corresponding one to one with the channel blocks. The channel information may directly indicate indices of the plurality of codewords, or may include values generated from the indexes. For example, the channel information may include information obtained by processing the indices based on a trellis code.

Thereafter, in step 1015, the receiver 120 transmits channel information quantized to the transmitter. In other words, the receiver 120 feeds back channel information indicating codeword indices. That is, the receiver 120 transmits indexes of a plurality of codewords for each of the antenna groups. At this time, according to the embodiment of the present invention, the receiver 120 may transmit additional feedback information indicating whether the channel information to be transmitted is generated based on a horizontal domain channel vector or based on a vertical domain channel vector . Alternatively, according to another embodiment of the present invention, the receiving terminal 120 may transmit the additional feedback information prior to transmission of the channel information.

In step 1017, the transmitter 110 transmits a data signal using the feedback channel information. In other words, the transmitter 110 may beamform the data signal using the codeword indicated by the codeword indices and transmit the beamformed data signal. At this time, the transmitter 110 determines the mapping relationship between the codewords and the antenna elements based on the channel information or an indicator fed back separately from the channel information. That is, the indicator indicates which domain the channel values are indexed based on which domain the codeword is determined by the receiver 120, and accordingly, the transmitter 110 determines a mapping relationship between the codewords and the antenna elements based on the indicator And transmit the beamformed signals mapped to the antenna elements according to the mapping relationship.

In the procedure described with reference to FIG. 10, in steps 1011 and 1013 of quantizing the channel information, the codewords included in the channel information and the codebook are divided into blocks of a plurality of channel vectors and groups of a plurality of codewords And performing trellis code quantization on the blocks of the channel vectors using the groups of codewords. Here, the step of Trellis code quantizing the channel information may include dividing the channel information and the codewords into M / L blocks of the channel vectors and groups of codewords . &Lt; / RTI &gt; Where M is the total number of antenna elements of the transmitter 110 and L is the number of antenna elements per group. According to an embodiment of the present invention, the step of performing trellis code quantization on the channel information may include: dividing groups of the codewords into outputs in a trellis structure corresponding to a predefined convolutional encoder; Searching for a path with respect to the trellis structure; and outputting, as a quantization result corresponding to the blocks of the channel vectors, information indicating an optimal codeword corresponding to the optimal path as a result of the path search, . &Lt; / RTI &gt;

According to an embodiment of the present invention, the step of searching for the path may include searching for a path to the trellis structure in a predefined search range of the entire search range. According to an exemplary embodiment of the present invention, the step of assigning groups of the code words to outputs in a trellis structure corresponding to a predefined convolutional code may include the steps of: And allocating the groups of the code words to the outputs in the trellis structure such that the minimum Euclidean distance between the assigned codewords is maximized. According to one embodiment, the convolutional code includes one of a 3/4 rate convolutional code, a 2/3 rate convolutional code or a convolutional code having an arbitrary rate.

As described above, the receiving end according to the various embodiments of the present invention can feed back channel information composed of block units, that is, indexes of codewords for each antenna group. Further, the receiving end may further transmit an indexing rule for channel values included in the channel vector used in determining the codewords. At this time, the channel information may be quantized. The quantization may be performed based on a trellis code. Hereinafter, the present invention will specifically describe the quantization.

As described in [6] to [9], methods for quantizing a channel vector having a low complexity in a MISO (multiple input single output) system have been proposed. The schemes proposed in [6] to [9] depend on the duality of the beamforming VQ problem and the non-coherent sequence detection problem in AWGN (additive white Gaussian noise) channel. The duality indicates that the two problems are equivalent to Equation (8).

In Equation (8), C denotes a codebook, c denotes a codeword in a codebook, and h denotes a channel vector.

Referring to Equation (8), the left side is a beamforming VQ problem and the right side is a non-coherent sequence detection problem. In the case of the incoherent sequence detection problem, h is the received signal, e j [ ^theta] is the channel coefficient (assuming block fading with unit amplitude), c is the candidate transmitted codeword, .

구조와 유사하다. 여기서 W1은 광대역/긴-주기(wideband/long-term) 채널 정보이고, W2는 부대역/짧은-주기(subband/ short-term) 채널 정보이다. 상기 TEC 및 상기 TE-SPA는 각각 W1, W2로 간주될 수 있다.Various embodiments of the present invention include a scheme for generating a TEC using a codebook, using the generated TEC for trellis code quantization, and a TE-SPA scheme for adjusting the phase with respect to a Trellis code quantized result do. These schemes are described in 3GPP LTE-A 8TX antenna codebook

Structure. Where W1 is the wideband / long-term channel information and W2 is the subband / short-term channel information. The TEC and the TE-SPA can be regarded as W1 and W2, respectively.

In describing the TEC, an LTE 4TX codebook is exemplified. However, any other VQ codebook such as Discrete Fourier Transform (DFT), Residual Vector Quantization (RVQ), and Grassmannian-line-packing (GLP) codebook may be used.

The characteristics of the receiver for quantizing the channel information will be described below.

The receiving end according to the embodiment of the present invention can store the codebook. For example, the codebook may be a 3GPP LTE 4TX codebook as shown in Table 1 below

으로부터의 세트 {s}에 의해 주어지는 열(column)들에 의해 정의되는 행렬을 나타낸다. 여기서, I는 4×4 단위 행렬(identity matrix)이고, 벡터 u_n은 상기 <표 1>에 나타난 바와 같이 주어진다. 4TX 랭크(rank) 또는 레이어(layer) 1 코드북으로서의

은

와 같다. In Table 1, W _n ^{s}

&Lt; RTI ID = 0.0 > {s}. &Lt; / RTI > Here, I is a 4 × 4 identity matrix, and the vector u _n is given as shown in Table 1 above. 4TX rank or layer 1 codebook

silver

.

) 중 미리 정의된 검색 범위(예:

,

)에서 상기 트렐리스 구조에 대하여 경로를 검색한다. 일 실시 예에서, 상기 수신단은 상기 트렐리스 구조의 홀수 번째 출력과 짝수 번째 출력에 할당되는 코드워드들 사이의 최소 유클리디언 거리(minimum Euclidean distance)가 최대가 되도록, 상기 각 코드워드들의 그룹들을 상기 트렐리스 구조에서의 출력들에 할당한다. 여기서, 상기 컨볼루션 부호는 3/4 레이트(rate) 컨볼루션 부호, 2/3 레이트 컨볼루션 부호 또는 임의의 레이트를 가지는 컨볼루션 부호 중의 하나를 포함할 수 있다.
The receiving end performs trellis code quantization on channel information using a codebook selected from a plurality of codebooks stored in the codebook. The receiving end truncates the channel information and codewords included in the selected codebook into a plurality of blocks of channel vectors and a plurality of groups of codewords, Quantized using Trellis code. In one embodiment, the receiving end divides the channel information and the codewords into blocks of a predefined number of channel vectors and groups of codewords. In one embodiment, the receiving end allocates groups of the codewords to outputs in a trellis structure corresponding to a predefined convolutional code, searches a path for the trellis structure, And outputs information indicating an optimum code word corresponding to the search result optimal path as a quantization result corresponding to the blocks of the channel vectors. In one embodiment, the receiving end has a full search range (e.g.,

) Of predefined search ranges (e.g.,

,

) Searches the path for the trellis structure. In one embodiment, the receiving end is configured such that the minimum Euclidean distance between the codewords allocated to the odd-numbered output and the even-numbered output of the trellis structure is maximized, To the outputs in the trellis structure. Here, the convolutional code may include one of a 3/4 rate convolutional code, a 2/3 rate convolutional code, or a convolutional code having an arbitrary rate.

The characteristics of the transmitter for processing the quantized channel information will be described below.

The transmitting end receives the quantized channel information. That is, the transmitter receives the feedback information including the trellis code quantized channel information. The feedback information is generated by the receiving end by performing trellis code quantization on channel information using a codebook selected from a plurality of codebooks. In one embodiment, the trellis code quantization operation allocates groups of the respective code words to outputs in a trellis structure corresponding to the convolution code, searches for a path for the trellis structure And outputting, as the quantization result corresponding to the blocks of the channel vectors, information indicating an optimal codeword corresponding to the optimal path as a result of the path search. In one embodiment, searching for the path includes searching for a path for the trellis structure in a predefined search range of the entire search range. In one embodiment, the groups of the respective codewords are arranged such that the minimum Euclidean distance between the codewords allocated to the odd-numbered output and the even-numbered output of the trellis structure is maximized. Lt; / RTI &gt; are assigned to outputs in the learl structure. The transmitting end convolutionally encodes the feedback information. The convolutional code may include one of a 3/4 rate convolutional code, a 2/3 rate convolutional code, or a convolutional code having an arbitrary rate. The transmitting end maps the convolutional coding result to codewords according to a predefined mapping rule. Also, the transmitter reconstructs the channel vector according to the mapped codewords.

FIG. 11 illustrates a channel information quantization procedure of a receiver of a wireless communication system according to an embodiment of the present invention. FIG. 10 illustrates a channel information quantization method of the receiver 120. FIG.

를 갖는 채널 벡터 h에 대한 하기 <수학식 9>와 같이 경로 메트릭을 최소화할 수 있다. Referring to FIG. 11, in step 1101, the receiving terminal minimizes a path metric using a Viterbi algorithm. For example,

The path metric can be minimized as shown in Equation (9) below with respect to the channel vector h having < EMI ID = 13.0 >

는 경로 메트릭의 최소 값, L은 블록 크기, M은 안테나 개수, c[m:n]는 코드워드 c의 m번째 엔트리로부터 n번째 엔트리 사이의 분할 벡터(truncated vector),

는 코드북, h[m:n]는 채널 벡터 h의 m번째 엔트리로부터 n번째 엔트리 사이의 분할 벡터를 의미한다.In Equation (9) above,

Is the minimum value of the path metric, L is the block size, M is the number of antennas, c [m: n] is the truncated vector between the mth entry and the nth entry of the codeword c,

Denotes a codebook, and h [m: n] denotes a division vector between the mth entry and the nth entry of the channel vector h.

를 제공하는 최적의 경로(best path)를 저장한다. 그리고, 상기 수신단은 1105단계로 진행하여

에 대한 최소값들

중에서 최소값을 제공하는 θ_opt을 선택한다. 예를 들어, 상기 수신단은 하기 <수학식 10>과 같이 상기 θ_opt을 선택할 수 있다. Then, the receiving end proceeds to step 1103, where the minimum value of the path metric

(Best path) to provide the best path. Then, the receiver proceeds to step 1105

Minimum values for

Among selects θ _opt which provides the minimum value. For example, the receiving end can select θ _opt as shown in Equation (10).

는 경로 메트릭의 최소 값을 의미한다.In Equation (10),? _Opt is an optimal path,? _K is a kth path,? Is a set of paths,

Means the minimum value of the path metric.

의 최적 경로를 2진 값 b_opt로 표현한다. 상기 2진 값은 트렐리스에서 최적 경로의 출력이 아닌, 입력을 구성한다.
Then, the receiver proceeds to step 1107

Is represented by a binary value b _opt . The binary value constitutes the input, not the output of the optimal path in the trellis.

12 illustrates a channel vector reconstruction procedure of a transmitter of a wireless communication system according to an embodiment of the present invention. FIG. 12 illustrates a channel vector reconstruction method of the transmitter 110. FIG.

Referring to FIG. 12, in step 1201, the transmitting end receives a binary value b _opt included in the feedback information from the receiving end. The transmitter proceeds to step 1203 and inputs the received binary value b _opt to the convolutional encoder. The transmitting terminal proceeds to step 1205 and acquires the output sequence of the convolutional encoder. In other words, the convolutional encoder outputs the sequence corresponding to the binary value b _opt . In step 1207, the transmitting terminal maps the output sequence of the convolutional encoder to a codeword (e.g., LTE codeword) according to a predefined mapping rule. In step 1209, the transmitter reconstructs the channel vector based on the mapping result.

Hereinafter, examples of trellis code quantization operations using a trellis extension codebook according to an embodiment of the present invention will be described with reference to FIG. 13 through FIG. 20 as follows. In describing the quantization operation, a trellis extension codebook using the LTE 4TX codebook as shown in Table 1 will be used.

13 illustrates an embodiment of a quantizer for channel information quantization in a wireless communication system according to an embodiment of the present invention. FIG. 13 shows an example of implementing a trellis code quantizer using a 3/4 rate convolutional encoder, through which 3/4 bits per channel entry quantization can be output. 13, the smaller the index, the less significant the bits are. For example, b _{in , 1} is the input bit of LSB (Least Significant Bit), b _{in , and 3} is the input bit of MSB (Most Significant Bit).

FIG. 14 shows an example of a trellis for channel information quantization in a wireless communication system according to an embodiment of the present invention. FIG. 14 shows a representation of the trellis structure corresponding to the 3/4 rate convolutional encoder shown in FIG. Each state transition shown on the right side of FIG. 14 is constituted by parallel transitions shown in the input / output relationship using the decimal numbers of the left boxes. For example, 1/4 (decimal) corresponds to input = 001 / output = 0100 (binary number) in transition 1402 indicated by the dotted line from state 0 to state 1, Quot; corresponds to input = 101 / output = 1100 (binary number).

Table 2 below shows a table for mapping the 3GPP LTE 4TX rank 1 codewords to the outputs of the trellis structure shown in FIG.

Equation (8), which shows the duality of the beamforming VQ problem and incoherent sequence detection problem in the AWGN channel, will be described again. The incoherent sequence detection problem in the AWGN channel is similar to the source coding problem as shown in Equation (11) below.

In Equation (11), c _opt denotes an optimal code word, C denotes a codebook, c denotes a codeword in the codebook, and h denotes a channel vector.

Equation (11) is a source coding problem for finding an optimal codeword c _opt that minimizes a mean squared error with h for a given [theta]. Therefore, the embodiment of the present invention relies on the concept of trellis-coded quantization (TCQ) [10] to extend the LTE codebook for a large-scale wireless communication system.

The TCQ uses a trellis decoder and a convolutional encoder in channel coding as a source encoder and a source decoder in source coding, respectively.

According to an embodiment of the present invention, a TEC including a 3/4 rate convolutional encoder as shown in Fig. 13 and a TEC having a trellis representation as shown in Fig. 14 corresponding thereto can be used. However, various types of convolutional encoders and their trellis representations can likewise be used for TEC according to embodiments of the present invention. For example, the 2/3 rate convolutional encoder shown in FIG. 15 and the corresponding TEC with the trellis representation shown in FIG. 16 can be used.

The object function in Equation (11) with given θ is decomposed as Equation (12) below.

Where h is the channel vector, c is the codeword, M is the number of antennas, L is the design parameter, block size, h [m: n] is the nth entry from the mth entry of the channel vector h , C [m: n] denotes a partition vector between the mth entry and the nth entry of the code word c.

Where h _{[m: n]} and c _{[m: n]} are the truncated vectors between the mth entry and the nth entry in the channel vector h and the code vector c, respectively, vector). For example, when the channel vector h has a size of M (the number of transmit antennas of the transmitter) (for example, 16) as shown in FIG. 12A, the channel vector h is M / L (e.g., 16/4 = 4) And each of the divided channel vectors 1401 to 1404 may have four codewords.

In this case, when one of the domains of the channel vector, i.e., one of the vertical domain and the horizontal domain, is selectively used, two candidates for the optimal codeword can be calculated as Equation (13).

이고,

이다.In Equation (13), C denotes a codebook, c denotes a codeword in the codebook, h denotes a channel vector indexed in the horizontal direction, g denotes a channel vector indexed in the vertical direction, c _{opt ,} C _{opt , 2} denotes the second candidate of the codeword corresponding to the g. here,

ego,

to be.

The operations described above provide the same codeword to antenna group mapping. The codeword candidates, i.e., the first candidate C _{opt , 1} and the second candidate C _{opt , 2} , correspond when the vertical domain is selected and the horizontal domain is selected. The two optimization problems can be solved by the Viterbi algorithm. The final codeword may then be fed back with an indicator indicating the selected domain.

Equation (12) or Equation (13) can be effectively calculated using a Viterbi algorithm. In other words, in each of the state transition t, one of the split-channel vector h having a Lx1 _{size:: [Lt L (t-} 1) +1] [L (t-1) +1 Lt] is the corresponding code vector c And quantized. After optimizing the Viterbi algorithm after the M / L state transition, an optimal code word that minimizes Equation (12) can be found. The path search using the Viterbi algorithm starts from state 0 in the trellis. If this is not done, the receiver needs to explicitly feed back information about the starting state of the optimal path to the transmitting end, which may increase the overall feedback overhead.

라고 가정한다. 여기서,

은 상기 <표 1>과 같은 3GPP LTE 4TX 랭크 1 코드북이다. 이 경우, L=4이다. 이와 달리, 예를 들어, 3GPP LTE 2TX 랭크 1 코드북과 같이 상이한 L 값을 갖는 임의의 코드북이 사용될 수도 있다. 또한, 다중 수신 안테나를 갖는 공간 다중화 (spatial multiplexing)의 경우, 또한 프로베니우스 놈(Frobenius norm) 연산을 이용하여 c_{[L(t-1)+1:Lt]}에 대한 상위의 랭크 코드북을 선택할 수도 있을 것이다. t = 1, ... , C _{[L (t-1) + 1: Lt ] <}

. here,

Is a 3GPP LTE 4TX rank 1 codebook as shown in Table 1 above. In this case, L = 4. Alternatively, any codebook having a different L value, such as, for example, a 3GPP LTE 2TX rank 1 codebook, may be used. Further, in the case of spatial multiplexing with multiple receive antennas, it is also possible to select an upper rank codebook for c _{[L (t-1) +1: Lt]} using a Frobenius norm operation It might be.

14, the transition from states 0, 2, 4, 6 (decimal) has only outputs of even 0, 2, 4, 6, 8, 10, 12, 14 (decimal) Transitions from 1, 3, 5, and 7 (decimal numbers) have only outputs of odd numbers 1, 3, 5, 7, 9, 11, 13, and 15 (decimal). Therefore, although a total of 16 codewords are available, embodiments of the present invention use only 3 bits to quantize the L = 4 entry of the channel vector h in each state transition, and only 3/4 bits per entry quantization &Lt; / RTI &gt;

Now, as shown in Table 1 The code word W _k ^{1} (k = 0, ..., 15) of the code word W _k is assigned to the output of the convolutional encoder shown in FIG. 13 or the output of the trellis representation shown in FIG. To minimize quantization errors, the minimum Euclidean distance between code words assigned to each set of odd and even outputs in the trellis is maximized.

및

은 하기 <수학식 14>에 나타낸 바와 같이 동일한 집합 원소 개수(cardinality)를 가지는

의 모든 가능한 분할들을 나타낸다고 가정한다.

And

Has the same cardinality as shown in Equation (14): < EMI ID = 14.0 >

&Lt; / RTI &gt; and all possible divisions of &lt; RTI ID =

는 코드북,

및

은

의 부분 집합들, card(·)는 관련 집합의 집합 원소 개수를 나타내고,

는 공집합을 의미한다.

이고,

라고 가정한다. 만약, C_odd 및 C_even을 각각 홀수 및 짝수의 출력에 할당된 코드워드들의 집합으로 정의한다면, C_odd 및 C_even은 이하 <수학식 15>와 나타낼 수 있다. In Equation (14) above,

A codebook,

And

silver

, Card (·) denotes the number of set elements of the related set,

Means an empty set.

ego,

. If C _odd and C _even are defined as a set of codewords allocated to odd and even outputs, respectively, C _odd and C _even may be expressed by Equation (15) below.

는 코드북,

및

은

의 부분 집합들,

는

의 m번째 코드워드를 의미한다. In Equation (15), C _odd denotes a codeword allocated to an odd output, C _even denotes a codeword allocated to an even output,

A codebook,

And

silver

&Lt; / RTI >

The

&Lt; / RTI &gt;

Through exhaustive search, the LTE codeword can be assigned to both odd and even trellis outputs as shown in Table 2 above. The even-numbered trellis outputs 0, 2, 4, 6, 8, 10, 12 and 14 are assigned to the LTE codewords index 0, 4, 2, 6, 1, 5, 3 and 7, respectively. The odd trellis outputs 1, 3, 5, 7, 9, 11, 13 and 15 are assigned to the LTE codewords index 8, 12, 10, 14, 9, 13, 11 and 15, respectively. Table 2 is merely an example of the mapping between the LTE codeword and the trellis output. Thus, mapping between different types of codeword and trellis output will be possible.

에 대하여 검색하는 대신, θ를

(여기서,

)와 같이 파라미터화할 수 있고, 지정된 범위 Θ에 대한 검색은 하기 <수학식 16>와 같이 표현될 수 있다. On the other hand, in the above explanation, it is assumed that θ is given in advance. However,? Is a parameter to be optimized in Equation (12). Whole space

Instead of searching for

(here,

), And a search for the specified range [theta] can be expressed as Equation (16).

In Equation (16), C denotes a codebook, c denotes a codeword in a codebook,? Denotes a set of paths,? Denotes a path, and h denotes a channel vector.

As a result, the solution of Equation (16) is obtained by performing the Viterbi algorithm K times when each Viterbi algorithm is executed at a given θ. This parallel search only increases the complexity and does not increase the feedback overhead. This is because θ is not required in the channel reconstruction process at the transmitting end.

15 shows another embodiment of a quantizer for channel information quantization in a wireless communication system according to an embodiment of the present invention. 15 shows an example of implementing a trellis code quantizer using a 2/3 rate convolutional encoder, through which 1/2 bit per channel entry quantization can be output. In FIG. 15, the smaller the index, the less significant the bits are. For example, b _{in , 1} is the input bit of LSB (Least Significant Bit), b _{in , 2} is the input bit of MSB (Most Significant Bit).

16 shows another example of a trellis for channel information quantization in a wireless communication system according to an embodiment of the present invention. 16 shows a representation of the trellis structure corresponding to the 2/3 rate convolutional encoder shown in FIG. The state transitions shown on the right side of FIG. 16 consist of parallel transitions shown in the input / output relationship using the decimal numbers of the left boxes. For example, 1/4 (decimal) in transition 1602, indicated by the dotted line from state 0 to state 1, corresponds to input = 01 / output = 100 (binary).

Table 3 below shows a table for mapping the 3GPP LTE 4TX rank 1 codewords to the outputs of the trellis structure shown in FIG.

By changing the 3/4 rate convolutional encoder shown in FIG. 13 to the 2/3 rate convolutional encoder as shown in FIG. 15, a TEC having 1/2 bit per channel entry quantization can be easily implemented .

For 1/2 bit per channel entry quantization, LTE codewords are assigned to the odd and even trellis outputs as shown in Table 3. < tb >< TABLE > The even trellis outputs 0, 2, 4, and 6 are assigned to the LTE codewords index 0, 1, 2, and 3, respectively. The odd trellis outputs 1, 3, 5, and 7 are assigned to LTE codewords indexes 4, 5, 6, and 7, respectively. Table 3 is merely an example of the mapping between the LTE codeword and the trellis output. Thus, mapping between different types of codeword and trellis output will be possible.

Referring to FIG. 17 to FIG. 20, the present invention describes a process of reconstructing channel information. 17 through 20 illustrate channel information reconstruction corresponding to a quantization operation with a trellis extension codebook using an LTE 4TX codebook and outputting 1/2 bit per channel entry quantization.

라고 가정한다. 그러므로 채널 벡터 h를 양자화하기 위하여 비터비 알고리즘은 2회 실행된다. M=16의 크기를 가지는 채널 벡터 h는 M/L(=16/4=4)개의 그룹들(또는 채널 엔트리) 1701 내지 1704로 분할된다. 각 그룹의 채널 벡터는 코드북으로부터 선택된 코드북 중에서 대응하는 코드 벡터들의 분할된 그룹들을 이용하여 트렐리스 부호 양자화된다. The number M of transmission antennas is 16, the channel vector h is as shown in FIG. 17,

. Therefore, the Viterbi algorithm is executed twice to quantize the channel vector h. A channel vector h having a size of M = 16 is divided into M / L (= 16/4 = 4) groups (or channel entries) 1701 to 1704. The channel vector of each group is Trellis-code quantized using the divided groups of corresponding code vectors from among the codebooks selected from the codebook.

In the case of quantizing the channel information using the existing LTE code back as it is, the number of bits of feedback information is determined in proportion to the number of transmission antennas. For example, if the number of transmit antennas is 16, the feedback information is determined to be 16 bits. However, according to the embodiment of the present invention, when the 2/3 rate convolutional encoder is used for Trellis code quantization, the quantization result of each channel vector group can be determined to be 2 bits. In this case, since the channel vector is divided into four channels, The feedback information can be implemented with 8 bits.

19 and 20 show path search by the Viterbi algorithm when? = 0 and? =?, Respectively. The path indicated by the dashed line in Fig. 19 represents the optimal path having the minimum path metric m (0) = 4.8357. The input sequence of the optimal path is binary [10,10,01,01]. In FIG. 20, a path indicated by a dotted line represents an optimal path having a minimum path metric. The input sequence of the optimal path is binary [01, 00, 11, 11]. Since m (0) <m (?) and? _opt = 0, the optimal path becomes the path shown in FIG. Therefore, the feedback sequence is b _opt = [10,10,01,01].

In the transmitting end, b _opt = [10,10,01,01] is the input of the convolutional encoder shown in Fig. 11A, and the output sequence of the corresponding convolutional encoder is [010, 110, 000, 101 ].

는 도 18에 도시된 바와 같다. 결과적으로, 최적의 코드워드는 이하 <수학식 17>과 같다. According to the mapping table shown in Table 3, the output sequence corresponds to the LTE codewords W ₁ ^{1} , W ₃ ^{1} , W ₀ ^{1} and W ₆ ^{1} Channel vector

Is as shown in Fig. As a result, the optimal code word is expressed by Equation (17) below.

는 양자화된 채널 벡터를 의미한다.
In Equation (17), c _opt is an optimal codeword,

Denotes a quantized channel vector.

Although the embodiments of the present invention as described above are applied to the 3GPP LTE 4TX rank 1 codebook as an example, the embodiments of the present invention can be similarly applied to higher rank cases.

It is possible to extend the trellis extended codebook (TEC) in the upper rank case by mapping the upper rank code word to the trellis output as shown in Table 2 or Table 3. [ The TEC can maintain orthogonal properties of the codebook reused for TEC. The nested property of the LTE codebook can be imitated by using the same mapping rule as shown in Table 2 or Table 3 for the case of the upper rank. These mapping rules are merely examples, and other mapping rules can be used.

구조와 유사하다. 상기 TEC 및 상기 TE-SPA는 각각 긴-주기/광대역 채널 정보 및 짧은-주기/부대역 채널 정보로 간주될 수 있다.
Channel information quantization and reconstruction according to the TE-SPA scheme can be performed as follows. According to the TE-SPA scheme, the receiver adjusts the phase of the Trellis code quantized result and feeds back the result. The TE-SPA scheme can further improve performance by using a trellis structure for phase adjustment subsequently to the block of quantized channels in addition to the TEC scheme. The TE-SPA is an LTE-

Structure. The TEC and the TE-SPA may be regarded as long-period / wide-band channel information and short-period / sub-band channel information, respectively.

For the TE-SPA scheme, the receiver can store a codebook as shown in Table 1 below. The receiving end performs trellis code quantization on channel information using a codebook selected from a plurality of codebooks stored in the codebook. Then, the receiver adjusts the quantization result by a predefined phase, and generates a phase-adjusted quantization result. The receiving end generates feedback information including the quantized channel information. In addition, the receiver generates secondary feedback information including quantized and phase adjusted channel information. The receiving end transmits the generated feedback information and / or secondary feedback information to the transmitting end. The receiving end transmits feedback information at intervals of a first period, and transmits secondary feedback information at intervals of a second period. In one embodiment, the second periodic interval is set to be shorter than the first periodic interval. According to an embodiment, the feedback information generating operation of the receiving end may be performed in a manner as shown in FIG. That is, the feedback information including the result quantized at the first time point is generated and transmitted, and at the second and third time points, the secondary feedback information including the phase adjustment result for the quantized result by the quantizer 230 is Can be generated and transmitted.

For the TE-SPA scheme, the transmitting end receives feedback information from the receiving end. The feedback information includes a trellis code quantized result. The feedback information is generated by the receiver using the codebook selected from the plurality of codebooks, and by performing the trellis code quantization on the channel information. Wherein the trellis code quantization operation truncates the channel information and codewords included in the selected codebook into a plurality of blocks of a plurality of channel vectors and a plurality of groups of codewords, For each of the groups of codewords. The transmitting end convolutionally encodes the feedback information. For example, the convolutional code may include one of a 3/4 rate convolutional code, a 2/3 rate convolutional code, or a convolutional code having an arbitrary rate. Then, the transmitting terminal maps the convolutional coding result, i.e., the trellis output, to codewords according to a predefined mapping rule (e.g., Table 2, Table 3). Also, the transmitting terminal maps the convolutional coding result, i.e., the trellis output, to the phases according to a predefined mapping rule (e.g., Table 4 below, Table 5). The transmitting end reconstructs the channel vector according to the mapped codewords or phases.

FIG. 21 illustrates a channel information feedback procedure in a wireless communication system according to an embodiment of the present invention.

를 피드백 정보로서 상기 송신단 110으로 피드백한다. 2109단계에서, 상기 송신단 110은 상기 수신단 120으로부터 피드백된 정보에 기초하여 채널 벡터를 재구성하고, 이 재구성된 채널 벡터를 이용하여 데이터를 전송한다. Referring to FIG. 21, in step 2101, the transmitting terminal 110 transmits a pilot signal, and the receiving terminal 120 receives a pilot signal transmitted from the transmitting terminal 110. In step 2103, the receiver 120 estimates a channel using the received pilot signal. In step 2105, the receiver 120 quantizes the received channel using the trellis extension codebook. In step 2107, the receiving end 120 extracts the optimal codeword index

To the transmitting terminal 110 as feedback information. In step 2109, the transmitter 110 reconstructs the channel vector based on the information fed back from the receiver 120, and transmits the reconstructed channel vector using the reconstructed channel vector.

을 위상조정 행렬 R1을 이용하여 회전(이하 도 22의 가운데 도면 참고)시킴으로써 채널 양자화 결과를 출력한다. 2117단계에서, 상기 수신단 120은 양자화 결과에 따른 코드워드 인덱스

을 추가 피드백 정보로서 상기 송신단 110으로 피드백한다. 2119단계에서, 상기 송신단 110은 상기 수신단 120으로부터 피드백된 정보에 기초하여 채널 벡터를 재구성하고, 이 재구성된 채널 벡터를 이용하여 데이터를 전송한다.In step 2111, the transmitting terminal 110 transmits a pilot signal, and the receiving terminal 120 receives a pilot signal transmitted from the transmitting terminal 110. In step 2113, the receiver 120 estimates a channel using the pilot signal. In step 2115, the receiving terminal 120 transmits the previously transmitted optimum codeword index &lt; RTI ID = 0.0 &gt;

(See the middle drawing in FIG. 22) using the phase adjustment matrix R1 to output the channel quantization result. In step 2117, the receiver 120 calculates a codeword index

To the transmitting terminal 110 as additional feedback information. In step 2119, the transmitter 110 reconstructs the channel vector based on the information fed back from the receiver 120, and transmits the reconstructed channel vector using the reconstructed channel vector.

을 위상조정 행렬 R2를 이용하여 회전(도 22의 오른쪽 도면 참고)시킴으로써 채널 양자화 결과를 출력한다. 2127단계에서, 상기 수신단 120은 양자화 결과에 따른 코드워드 인덱스

을 추가 피드백 정보로서 상기 송신단 110으로 피드백한다. 2129단계에서, 상기 송신단 110은 수신단 120으로부터 피드백된 정보에 기초하여 채널 벡터를 재구성하고, 재구성된 채널 벡터를 이용하여 데이터를 전송한다.Thereafter, in step 2121, the transmitting terminal 110 transmits a pilot signal, and the receiving terminal 120 receives a pilot signal transmitted from the transmitting terminal 110. In step 2123, the receiver 120 estimates a channel using the received pilot signal. In step 2125, the receiver 120 transmits the previously transmitted optimal codeword index &lt; RTI ID = 0.0 &gt;

(See the right drawing of FIG. 22) using the phase adjustment matrix R2 to output the channel quantization result. In step 2127, the receiver 120 calculates a codeword index

To the transmitting terminal 110 as additional feedback information. In step 2129, the transmitting terminal 110 reconstructs the channel vector based on the information fed back from the receiving end 120, and transmits the data using the reconstructed channel vector.

Here, it is described that the receiving end 120 feeds back additional feedback information (subband / short-period feedback information) twice after feedback information (broadband / long-term feedback information) is fed back once. However, the transmission of the additional feedback information at the receiving end 120 may be performed an appropriate number of times.

In the procedure shown in FIG. 21, the receiver 120 generates channel information quantized in steps 2105, 2115, and 2125. In this case, according to an embodiment of the present invention, the receiver can generate a plurality of candidates using channel vectors having different indexing rules, and quantize one of the candidates. Here, the channel vectors composed of the different indexing rules include a horizontal domain channel vector and a vertical domain channel vector.

According to the embodiment of the present invention, when the channel information is as shown in FIG. 17, that is, when the number of transmit antennas of the transmitter is M = 16, the 2/3 rate convolutional code is Trellis code quantized , The feedback information transmitted in the long-period can be implemented with 8 bits. On the other hand, in the case of the TE-SPA, since only phase adjustment information is provided as compared with the case of the TEC, the transmission additional feedback information can be implemented with a smaller number of bits (e.g., 4 bits) in a short period.

For the sake of simplicity of explanation, it is assumed that the size of the block for phase adjustment is the same as the VQ codebook used for TEC, but different sizes are also possible.

은 하기 <수학식 18>과 같은 블록 단위 위상 조정 행렬(block-wise phase adjustment matrix) R_k에 의해 회전된다. Previously quantized channel information

Is rotated by a block-wise phase adjustment matrix R _k as shown in Equation (18).

는 크로네커 곱, 1_L은 길이(L)를 가지는 모두 1인 열 벡터(all 1 column vector)를 의미한다..In Equation (18), R _k is a phase adjustment matrix, M is the number of antennas, L is a block size,

Is a Kronecker product, and 1 _L is an all 1 column vector with a length (L).

는 하기 <수학식 19>와 같다. Next, the current quantized channel information

Is expressed by Equation (19) below.

는 R_k에 의해 위상 조정된 양자화된 채널 정보, R_k는 위상 조정 행렬을 의미한다.In Equation (19) above,

Denotes quantized channel information phase-adjusted by R _k , and R _k denotes a phase adjustment matrix.

값들을 계산하도록 사용되어진다. 이 구조는 TEC를 위한 트렐리스와 상이할 수 있다. This trellis structure uses the Viterbi algorithm to minimize the < _{RTI ID = 0.0} > _Rk &lt; / _RTI &gt;

Are used to calculate values. This structure can be different from Trellis for TEC.

는 R_k에 의해 위상 조정된 양자화된 채널 정보를 의미한다.In Equation (20), R _k is a phase adjustment matrix,? Is a set of paths,? Is a path, h _k is a channel vector,

Denotes quantized channel information that is phase-adjusted by R _k .

및

를 각각 k의 엘리먼트들을 가지는 벡터 a 및 행렬 A의 대각 엔트리들(diagonal entries)의 순환 쉬프팅(circular shifting)으로 가정한다. 예를 들어,

이면,

이다. The first state transition experiences a limited number of branches as shown in Figures 19 and 20. [ 'Block shifting' is applied before trellis optimization to mitigate these losses and further reduce the time quantization error.

And

Is assumed to be a circular shifting of a vector a having elements of k and diagonal entries of matrix A, respectively. E.g,

If so,

to be.

Next, the optimization problem in Equation (20) can be rewritten as Equation (21), and the quantized channel vector is Equation (22).

는 R_k에 의해 위상 조정된 양자화된 채널 정보를 의미한다.In Equation (21) and Equation (22), R _k is a phase adjustment matrix,? Is a path set,? Is a path, h _k is a channel vector, L is a block size,

Denotes quantized channel information that is phase-adjusted by R _k .

이 k에 따라 서로 다른 블록들에 곱해진다. 예를 들어, 시간 k=0에서 위상 조정

은 채널 벡터의 4개 코드워드들의 블록에 곱해지며, 시간 k=1에서 위상 조정

은 채널 벡터의 5개 코드워드들의 블록에 곱해지며, 시간 k=2에서 위상 조정

은 채널 벡터의 4개 코드워드들의 블록에 곱해진다. A conceptual description of the TE-SPA with shifting is shown in FIG. 22 below. FIG. 22 illustrates phase adjustment for channel quantization results in a wireless communication system according to an embodiment of the present invention. Referring to FIG. 22, from the first state transition at time k,

Different blocks are multiplied according to k. For example, at time k = 0,

Is multiplied by the block of four codewords of the channel vector and the phase adjustment &lt; RTI ID = 0.0 &gt;

Is multiplied by the block of five codewords of the channel vector and the phase adjustment at time k =

Is multiplied by a block of four code words of the channel vector.

의 교란된(perturbed) 버전(version)인

에 대하여도 수행된다. 여기서,

이다. 이에 따라, 하기 도 23과 같이, 2개의 도메인들에 따라 서로 다르게 블록화된 안테나들에 위상 조정이 적용될 수 있다. According to an embodiment of the present invention, since two domains for a channel vector are considered,

Which is a perturbed version of

. here,

to be. Accordingly, as shown in FIG. 23, phase adjustment can be applied to antennas that are differentially blocked according to two domains.

23 shows phases applied to an antenna group in a wireless communication system according to an embodiment of the present invention. In FIG. 23, (a) shows a case of blocking in a horizontal domain, and (b) shows a case of blocking in a vertical domain. A common phase is applied to one antenna group. When the vertical domain is selected as the preferred domain, the receiving end can adjust the relation of the phases of the antenna groups of the antenna groups in the vertical direction as shown in FIG. 5 (a). Similarly, if the horizontal domain is selected as the preferred domain, the receiving end can adjust the relation of the phases of the antenna groups of the antenna groups in the horizontal direction as shown in FIG. 5B.

For each domain, the quantization of the channel information may be performed as follows. For the TE-SPA according to the embodiment of the present invention, two different trellis structures may be applied: a case where the channel entry quantization result is ½ bit and a case where the channel entry quantization result is ¼ bit. If the channel entry quantization result is 1/2 bit for the TE-SPA, the correspondence relationship between the trellis output and the phase can be defined as shown in Table 4 below.

The convolutional encoder corresponding to the phase values illustrated in Table 4 is a 2/3 rate convolutional encoder as shown in FIG. 13, and the corresponding trellis structure is as shown in FIG. 14 .

For 1/2 bit per channel entry quantization, phases may be assigned to the odd and even trellis outputs as shown in Table 5. [ The even trellis outputs 0, 2, 4, and 6 are assigned to the phase indices 0, 2 / 8π, 4 / 8π, 6 / 8π, respectively. The odd trellis outputs 1, 3, 5 and 7 are assigned to the phase indices 1 / 8π, 3 / 8π, 5 / 8π and 7 / 8π, respectively. Table 5 is merely an example of a mapping between phases and a trellis output. Thus, mapping between different types of phases and trellis output will be possible.

FIG. 24 shows another embodiment of a quantizer for quantizing channel information in a wireless communication system according to an embodiment of the present invention. FIG. 25 shows a quantizer for quantizing channel information in a wireless communication system according to an embodiment of the present invention. &Lt; / RTI &gt; FIG. 24 shows a case where the channel entry quantization result for the TE-SPA is 1/4 bit. 24 illustrates a 1/2 rate convolutional encoder, and the trellis structure corresponding to the quantizer illustrated in FIG. 24 is as shown in FIG.

For 142 bits per channel entry quantization, the correspondence between the trellis output and the phase can be defined as shown in Table 5 below.

As shown in Table 5, phases may be assigned to the odd and even Trellis outputs. The even trellis outputs 0, 2 are assigned to the phase indices 0, 2/4 [pi], respectively. The odd trellis outputs 1 and 3 are assigned to the phase indices 1 / 4π and 3 / 4π, respectively. Table 5 is merely an example of a mapping between phases and a trellis output. Thus, mapping between different types of phases and trellis output will be possible.

The simulation results for evaluating the performance of the wireless communication system according to the embodiment of the present invention are as follows. In the simulation, the number of antennas M of the transmitting end was set to 32, and the number of antennas of the receiving end was set to one. The channel is defined as having the following characteristics (Equation 23).

은 시간 n에서의 이노베이션 프로세스(innovation process)를 의미한다. 상기 R은 [5]의 모델에 따르며, η=0.9881이다.In the <Equation 23>, h [n] is the channel vector, η at time n has a M × 1 size is temporarily correlation coefficient (temporal correlation coefficient), R = E [h [n] H h [n] ] Is a spatial correlation matrix,

Means the innovation process at time n. The R depends on the model of [5], and η = 0.9881.

The beamforming gain can be defined as a performance metric such as Equation (24).

In Equation 24, h [n] denotes a channel vector at time n having an M × 1 size, and c _opt denotes an optimal codeword at time n.

The results of the simulation performed on the assumption of the above conditions are shown in FIG. 26 below. 26 shows a simulation result of a wireless communication system according to an embodiment of the present invention. Figure 26 illustrates four different techniques in comparison. The above four techniques are a technique using a 32 TX DFT codebook, a technique using a Kronecker product-based approach using a 4TX / 8 TX DFT codebook, a technique based on a TEC / TE-SPA technique, and the present invention.

The technique using the 32 TX DFT codebook and the TEC / TE-SPA technique have B _tot = 16, i.e., 16 bits of feedback overhead. On the other hand, the technique of the Kronecker-based scheme using the 4TX / 8TX DFT codebook and the technique of the present invention has a feedback overhead of B _tot = 17, that is, 17 bits. The 17 bits can be divided into 7 bits and 10 bits for a 4 TX DFT codebook and an 8 TX DFT codebook. For the technique of the present invention, the indicator may indicate the preferred domain of the horizontal domain and the vertical domain as additional feedback.

Through the above FIGS. 26A and 26B, four techniques having different standard deviations of horizontal and vertical angular perturbations are compared. The technique according to the present invention has better performance than the method based on the Kronecker product with the same overhead. In addition, the technique according to the present invention is comparable to a 32 TX DFT codebook consisting of 65536 vector-quantized codewords. The complexity of the technique using the 32 TX DFT codebook is not feasible in practice.

Methods according to the claims or the embodiments described in the specification may be implemented in hardware, software, or a combination of hardware and software.

When implemented in software, a computer-readable storage medium storing one or more programs (software modules) may be provided. One or more programs stored on a computer-readable storage medium are configured for execution by one or more processors in an electronic device. The one or more programs include instructions that cause the electronic device to perform the methods in accordance with the embodiments of the invention or the claims of the present invention.

Such programs (software modules, software) may be stored in a computer readable medium such as a random access memory, a non-volatile memory including a flash memory, a ROM (Read Only Memory), an electrically erasable programmable ROM (EEPROM), a magnetic disc storage device, a compact disc-ROM (CD-ROM), a digital versatile disc (DVDs) An optical storage device, or a magnetic cassette. Or a combination of some or all of these. In addition, a plurality of constituent memories may be included.

In addition, the program may be transmitted through a communication network composed of a communication network such as the Internet, an Intranet, a LAN (Local Area Network), a WLAN (Wide LAN), or a SAN (Storage Area Network) And can be stored in an attachable storage device that can be accessed. Such a storage device may be connected to an apparatus performing an embodiment of the present invention via an external port. In addition, a separate storage device on the communication network may be connected to an apparatus that performs an embodiment of the present invention.

In the concrete embodiments of the present invention described above, the elements included in the invention are expressed singular or plural in accordance with the specific embodiment shown. It should be understood, however, that the singular or plural representations are selected appropriately according to the situations presented for the convenience of description, and the present invention is not limited to the singular or plural constituent elements, Even the expressed components may be composed of a plurality.

While the present invention has been described in connection with what is presently considered to be the most practical and preferred embodiment, it is to be understood that the invention is not limited to the disclosed embodiments, but is capable of various modifications within the scope of the invention. Therefore, the scope of the present invention should not be limited by the illustrated embodiments, but should be determined by the scope of the appended claims and equivalents thereof.

Claims (20)

A method of operating a receiving end of a wireless communication system,
Transmitting an indicator indicating an indexing rule of channel values and channel information quantized in units of blocks;
And receiving beamformed signals mapped to the antenna according to the indexing rule.
The method according to claim 1,
Wherein the indexing rule represents a correspondence between channel values and antenna elements included in a channel vector used in code word determination.
The method according to claim 1,
Comprising the steps of: receiving a signal for estimating a channel;
And quantizing the channel information determined based on the signal on a block-by-block basis.
The method of claim 3,
The quantization step may include:
And Trellis code quantization of channel information using a codebook selected from a plurality of codebooks.
The method of claim 4,
The quantization step may include:
Truncating the channel information and codewords included in the selected codebook into blocks of a plurality of channel vectors and groups of a plurality of codewords,
And Trellis code quantizing the blocks of the channel vectors using the groups of the codewords.
The method of claim 3,
And transmitting the phase adjusted channel information generated by adjusting the channel information quantized in units of blocks by a predefined phase to the transmitting terminal.
A method of operating a transmitter of a wireless communication system,
Receiving an indicator indicating an indexing rule of channel values and channel information quantized in a block unit;
And transmitting beamformed signals mapped to the antenna according to the indexing rule.
The method of claim 7,
Wherein the indexing rule represents a correspondence between channel values and antenna elements included in a channel vector used in code word determination.
The method of claim 7,
Wherein the channel information is channel information quantized in block units.
The method of claim 9,
And receiving phase adjusted channel information generated by adjusting the channel information quantized in units of blocks by a predefined phase.
A receiving terminal apparatus of a wireless communication system,
An indicator indicating an indexing rule of channel values, and a transmitter for transmitting quantized channel information on a block-
And a receiver for receiving the beamformed signals mapped to the antenna according to the indexing rule.
The method of claim 11,
Wherein the indexing rule indicates correspondence between channel values and antenna elements contained in a channel vector used in code word determination.
The method of claim 11,
The receiver receives a signal for estimating a channel,
And a controller for quantizing the channel information determined based on the signal on a block-by-block basis.
14. The method of claim 13,
Wherein the control unit trellis code quantizes channel information using a codebook selected from a plurality of codebooks.
15. The method of claim 14,
Wherein the control unit truncates the channel information and codewords included in the selected codebook into a plurality of groups of channel vectors and a plurality of groups of codewords, An apparatus for Trellis code quantization using groups of words.
14. The method of claim 13,
Wherein the transmitter transmits the phase-adjusted channel information generated by adjusting the quantized channel information by a predefined phase to the transmitting terminal.
A transmitting terminal apparatus of a wireless communication system,
An indicator for indicating an indexing rule of channel values, and a receiver for receiving channel information quantized in block units;
And a transmitter for transmitting the beamformed signals mapped to the antenna according to the indexing rule.
18. The method of claim 17,
Wherein the indexing rule indicates correspondence between channel values and antenna elements contained in a channel vector used in code word determination.
18. The method of claim 17,
Wherein the channel information is channel information quantized in units of blocks.
The method of claim 19,
Wherein the receiving unit receives the phase adjusted channel information generated by adjusting the channel information quantized in units of blocks by a predefined phase.

Images