KR20160066665A - Method for Feedback and Scheduling for Massive MIMO System and Apparatus Therefor - Google Patents

Abstract

Disclosed is a method for feedback and scheduling for massive MIMO system and an apparatus therefor. The present invention relates to the technology relating to a feedback and a user scheduling for a phased beamforming in a massive MIMO system, and relates to a method for feedback and scheduling for massive MIMO system and an apparatus therefor, which perform a scheduling and a two-step beamforming based on the scheduling result in which the information about a first-step beamforming of a base station gives feedback on a channel quality indicator (CQI), a channel direction indicator (CDI) and quantization error size information in a state of being provided to a terminal, and calculates a signal-to-interference and noise ratio (SINR) in a base station by using the feedback information.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a feedback and scheduling method for a large-scale MIMO system,

This embodiment relates to a method and apparatus for performing feedback and scheduling in a large-scale MIMO system.

The contents described in this section merely provide background information on the present embodiment and do not constitute the prior art.

In the following description, step-by-step beamforming of a MIMO system and feedback and user scheduling techniques are described.

A base station using a plurality of transmit antennas in a MIMO wireless communication system can improve SINR and antenna gain of a user's received signal through MIMO beamforming or MIMO beamforming. To do this, the base station must know the correct channel information and perform appropriate beamforming based on it. The channel information is estimated at each user terminal using a reference signal sent from the base station, and the estimated channel information is transmitted to the base station through feedback. At this time, since the amount of information that the user can send as feedback is limited by the time and frequency resources of the uplink, the channel information is quantized and transmitted.

However, if quantization error occurs due to insufficient amount of feedback received from the user, incorrect channel information is transmitted and accurate beamforming can not be performed, which causes a performance loss. Since the feedback load also increases as the number of transmit antennas increases, this performance loss is further exacerbated.

Staged beamforming performs two-stage beamforming using spatial correlation of channels in order to effectively reduce feedback load even in a multi-input / output system with many transmitting antennas. The spatial correlation of a channel is determined by the arrangement of transmit antennas, the location of a user and the scattering environment, etc. These factors vary considerably slowly compared to the instantaneous channel information. In addition, among the entire channel information, the instantaneous channel information excluding the spatial correlation information is only a part of the entire channel information. Therefore, the stepwise beamforming feeds back the spatial correlation information with a long period, and only the instantaneous channel information is fed back every frame. As a result, the feedback load can be significantly reduced as compared with the conventional method in which the entire channel information is fed back every frame have.

Multi-user beamforming can support multiple users at the same time to increase the aggregate data rate, but inter-user interference can degrade performance. In order to correctly obtain the performance improvement effect due to the multiuser beamforming, a proper user scheduling method for determining the users to support simultaneously is needed. To this end, the base station needs channel quality information (CQI) feedback for conveying information such as SINR and channel gain of a user in addition to channel direction information (CDI) feedback for beamforming. At this time, The information is related to the scheduling technique used by the base station. In particular, when using a step-like beamforming in a base station, since users are divided into a plurality of groups according to the spatial correlation of the channels and one-stage beamforming is already determined for each group, an effective scheduling technique Feedback method is required.

The present embodiment relates to Feedback and User Scheduling for stepwise beamforming in a large scale MIMO system. In this embodiment, channel quality information (CQI) Channel quality indicator, channel direction indicator (CDI), quantization error size information, and the like to feedback the signal-to-interference-and-noise ratio (SINR) The present invention provides a feedback and scheduling method and a device therefor in a large scale multi-input / output (I / O) system in which scheduling is performed based on a scheduling result and two-stage beamforming is performed based on a scheduling result.

According to an aspect of the present invention, there is provided a wireless communication system including: a beamforming receiver for receiving a pilot signal from a base station, the beamforming receiver feedbacking an estimated spatial correlation matrix and receiving a 1 st stage beamforming signal determined based on the feedback; A channel estimator for estimating a channel based on a reference signal received from the base station, estimating an effective channel based on the one-stage beamforming signal and the channel; And a feedback transmitter for calculating effective channel related information based on the effective channel and feeding back the effective channel related information to the base station.

According to another aspect of the present invention, there is provided an apparatus for transmitting a pilot signal to a terminal, performing grouping based on a received spatial correlation matrix based on the pilot signal, A first-stage beamforming unit for performing a beamforming operation; A feedback processor for transmitting a reference signal to the mobile station and estimating a signal-to-interference and noise ratio (SINR) by obtaining effective channel related information for an estimated effective channel based on the reference signal; A scheduling unit for calculating a sum rate (R _SUM ) using the SNR and performing a scheduling based on the sum rate; And a two-stage beamforming unit for performing two-step beamforming with the scheduled UE and transmitting data.

According to another aspect of the present invention, there is provided a receiving method for receiving a pilot signal from a base station, the method comprising: receiving a first-stage beamforming signal based on feedback based on an estimated spatial correlation matrix; A channel estimation process for estimating a channel based on a reference signal received from the base station; An effective channel estimator for estimating an effective channel based on the one-stage beamforming signal and the channel; And an effective channel related information including channel direction information (CDI), channel quality information (CQI), and quantization error information using the effective channel, and feeding back the effective channel related information to the base station And a feedback method for stepwise beamforming is provided.

According to another aspect of the present invention, there is provided an apparatus for transmitting a pilot signal to a terminal, performing grouping based on a received spatial correlation matrix based on the pilot signal, Stage beamforming process; A feedback reception unit for transmitting a reference signal to the terminal and acquiring effective channel related information including channel direction information (CDI), channel quality information (CQI), and quantization error information for an estimated effective channel based on the reference signal process; A feedback processing step of estimating a signal-to-interference noise ratio based on the effective channel-related information; A scheduling step of calculating an aggregate data rate using the SNR and performing a scheduling based on the aggregate data rate; And a two-stage beamforming process of performing two-stage beamforming with the scheduled UE and transmitting data. The present invention also provides a scheduling method for stepwise beamforming.

As described above, according to the present embodiment, when the step-by-step beamforming is used in a wireless communication system using a multi-input / output antenna, a method of feeding back the effective channel is less than a method of feeding back the spatial correlation of the channel and the instantaneous channel portion And has a positive feedback load, it is possible to support more transmission antennas with the amount of feedback limited in the existing wireless communication system.

Also, when using step-wise beamforming in a wireless communication system using a multi-input / output antenna, a method of feeding back an effective channel can reduce a quantization error rather than a method of feeding back a spatial correlation and a channel portion of a channel. Accordingly, the base station can acquire more accurate channel information and perform accurate beamforming, thereby improving the sum rate of the MIMO system.

Also, when using the step-by-step beamforming in a wireless communication system using a multi-input / output antenna, if the user feeds back the direction of the effective channel, the signal-to-noise ratio, and the quantization error size, the BS calculates the signal- It is possible to estimate the sum rate close to the actual value. The use of the estimated total rate can improve the multi-user gain of the MIMO system because the user scheduling scheme can more accurately schedule the set of users that increase the sum rate.

1 is a block diagram schematically showing a large scale MIMO system according to the present embodiment.
2 is a block diagram schematically showing a large scale MIMO system for performing user grouping and stepwise beamforming according to the present embodiment.
FIG. 3 is a block diagram schematically showing a base station and a terminal included in the large-scale MIMO system according to the present embodiment.
4A and 4B are flowcharts for explaining a step-by-step beamforming and scheduling-based wireless communication method according to the present embodiment.
FIG. 5 is a diagram comparing the accuracy of a quantization method and a general quantization method according to the present embodiment.
6 is a diagram showing a change in the sum rate based on the number of transmission antennas of the base station according to the present embodiment.
7 is a diagram showing a change in the sum rate based on the total number of users according to the present embodiment.

Hereinafter, the present embodiment will be described in detail with reference to the accompanying drawings.

1 is a block diagram schematically showing a large scale MIMO system according to the present embodiment.

로 표현할 수 있다. 여기서, α_k는 대규모 페이딩(Large-Scale Fading)이며, U_k ∈ C^{N _T×r_k}와 Λ_k ∈ C^{r _k×r_k}는 공간상관행렬(R_k = U_kΛ_kU^H _k)의 고유값 분해를 통해 얻어지는 고유벡터 행렬과 고유값 행렬이고, w_k ∈ C^r_k는 소규모 페이딩(Small-Scale Fading)으로 공분산의 단위행렬인 복소수 가우시안(Gaussian) 분포를 따른다. r_k는 공간상관행렬의 랭크(Rank)로 단말에서 N_T 개의 기지국(110)의 안테나 각각이 얼마나 공간적으로 구분되는지를 나타낸다. 따라서, 공간적 상관도는 기지국(110)의 안테나 간의 간격과 주위 반사체들의 분포 등 환경에 따라 달라질 수 있으나, 일반적으로 다중 안테나 기지국(110)의 설치 시 성능 이득을 얻기 위해 기지국(110)의 안테나 간의 거리를 공간 상관을 피하기 위한 기 설정된 거리로 설치하게 됨을 고려하면 1 보다는 기지국(110)의 안테나 개수(N_T)에 가까운 값을 지닌다. In order to explain a large scale MIMO system according to the present embodiment, the BS 110 includes N _T antennas, and a total of S terminals 120 each have a large-scale MIMO system having N _R = 1 antennas I suppose. At this time, the channel information (h _k ) between the base station 110 and the terminal 120 k is calculated in consideration of spatial correlation

. Here, α _k is a large-scale fading, and U _k ∈ C ^{N _T × r_k} and Λ _k ∈ C ^{r _k × r_k} are ^unique in the spatial correlation matrix (R _k = U _k Λ _k U ^H _k ) W _k ∈ C ^r_k is a small-scale fading, followed by a complex Gaussian distribution, which is the unit matrix of covariance. r _k is a rank of the spatial correlation matrix and indicates how spatially the antennas of the N _T base stations 110 in the terminal are separated. However, in order to obtain a performance gain in the installation of the multi-antenna base station 110, the spatial correlation between the antennas of the base station 110 may vary depending on environments such as the spacing between the antennas of the base station 110 and the distribution of the surrounding reflectors. Considering that the distance is set to a predetermined distance to avoid spatial correlation, it is closer to the number of antennas (N _T ) of the base station 110 than to 1.

The base station 110 divides all of the terminals 120 into a plurality of (G) groups according to the degree of spatial correlation in order to perform the stepwise beamforming. The base station 110 compares the spatial correlation matrix R _k of a total of S terminals 120 for group formation according to a predetermined evaluation criterion and the spatial correlation matrix R _k is used for the same terminals 120 Assign groups.

의 관계가 성립한다. g_k를 g 번째 그룹에서 스케쥴된 k 번째 단말기(120)로 정의할 때, 단말기(120) g_k의 수신 신호(y_{g _k})는 다음과 같다.Let K be the number of terminals 120 to be simultaneously scheduled by the BS 110 among the total S terminals 120 and K _g be the number of terminals 120 belonging to the gth group among the scheduled terminals 120. At this time,

. When defining the g _k a g-th group, the k-th terminal 120 in the schedule, the received signal (y _{g _k)} of the terminal 120, g _k is:

Here, P is a total transmission power, n _{g _k} is the transmitted symbol (Symbol) for the Gaussian noise with a power of σ ² _n, x _{g_k} the terminal _{(120) g k, x g} = [x g _1, ..., x _{g _k_g} ] ^T ∈ C ^{K _g} denotes the transmission symbol for the g-th group. In addition, P _{g _k} ∈ C ^{N _T} is the beamforming for a terminal _{(120) g k, P g} = [P g _1, ..., P g _k_g] T ∈ C N _T × K_g is the g-th group, User beamforming for < / RTI > In the stepped beamforming, P _g is P _g B = _g V _g = _g B is expressed can be in the form of _{[v g _1, ..., v} g _k_g].

_{^{Here, B g ∈ C N _T ×}} r ^ * is g and the second group of steps 1 beam, which is determined by the spatial correlation forming _{^{representing, v g _k ∈ C r ^}} * is a terminal for the (120) g _k 2 Step beamforming. For example, if N _T = 32 and G = 8, then r ^* = 4. Where r ^* is a system parameter determined by the number of groups, determined for effective first-order beamforming, and r ^* is pre-set or adjustable.

The base station 110 performs a step for the antenna gain improvement by group and the interference cancellation beam forming (B _g) between the groups, the terminal 120 belonging to each group by one-step beam forming (B _g) base station ( 110) appear to have r ^* antennas.

The base station 110 performs a two-step beamforming (v _{g_k} ) for eliminating interference between the terminals 120 simultaneously scheduled in the same group based on the effective channel information formed after the first-stage beamforming (B _g ) do. Here, the one-stage beam forming (B _g ) and the two-stage beam forming (v _{g_k} ) can be performed by applying one of the multi-user beam forming methods such as eigenbeam forming, zero- have.

2 is a block diagram schematically showing a large scale MIMO system for performing user grouping and stepwise beamforming according to the present embodiment.

When using a beam-forming phase, the terminal receives the signal (y _{g _k)} of (120) g _k is:

In Equation (2), the first term on the right side is the transmission signal for the terminal 120 g _k , the second term on the right side means the interference between the terminals 120 in the same group, Interferences. At this time, the terminal (120) g _k is a signal to interference noise ratio that can be obtained (SINR: Signal to Interference and Noise Ratio, hereinafter, described as SINR _{_k g)} as follows.

: 1 단계 빔포밍에 대한 신호의 세기,

: 동일 그룹 내 간섭의 세기,

: 그룹 간 간섭의 세기, σ_n ²: 노이즈)(P / K: transmission power per terminal,

: Signal intensity for one-stage beamforming,

: Intensity of interference in the same group,

: Intensity of intergroup interference, σ _n ² : noise)

The base station 110 estimates the SINR _{g _k} using the channel information received from the feedback channel information (h _{g _k)} fully understood terminal 120 can not be a. The base station 110 for feedback to using the estimated SINR _{g _k} for scheduling, calculation methods for accurate SINR estimates and SINR _{g g _k _k} estimation is required.

In general, a channel information (h _{g _k)} either there is a method for feeding back a direct quantization using a code book, this approach because the feedback load increases rapidly with an increase in the number of transmission antenna is not suitable for application to large-scale antenna system.

In order to solve this problem, it has been proposed the eigenvector matrix (U _{g _k),} eigenvalue matrix divided by (Λ _{g_k)} and the instantaneous channel information (w _{g _k)} method of each feedback cycle of the other represents the spatial correlation of the channel, eigenvector matrix (U _{g _k)} and the eigenvalue matrix (Λ _{g_k)} is determined by the arrangement and the user's location, and the scattering environment of transmission antennas, these factors vary considerably slow compared with the instantaneous channel information (w _{g _k)} . Thus, the method for feeding back to each other to give the other is, and forwards the eigenvector matrix (U _{g _k)} and the eigenvalue matrix (Λ _{g_k)} a long cycle feedback, instantaneous channel information (w _{g _k)} conveys quantizes each frame .

In this embodiment, instead of the method it was to pass information relating to the instantaneous channel information (w _{g _k),} channel information, an effective channel, which is defined by the product of (h _{g _k)} of the first stage beam-forming (B _g) Matrix (g _{g _k} = B ^H _g h _{g_k} ∈ C ^{r ^ *} ).

)을 양자화한 양자화 정보(

)를 포함하는 채널방향정보(CDI: Channel Duality Information)를 피드백으로 전달하며, 유효 채널(g_{g _k})의 신호 대 잡음비(SNR: Signal-to-Noise Ratio,

)를 포함하는 채널품질정보(CQI: Channel Quality Information)를 피드백으로 전달한다. 또한, 기지국(110)이 양자화 오류의 영향을 반영할 수 있도록 하기 위하여,

로 정의되는 양자화 오류 크기(

)를 포함하는 양자화 오류 크기정보를 기지국(110)으로 피드백한다. 양자화 오류 크기(

)는 0 내지 1 사이의 실수값(Real Number)이기 때문에 추가적인 피드백으로 인한 단말기(120)에서의 피드백 부하 증가는 미미하다. 오히려, 단말기(120)는 r^* < r_{g _ k} 의 경우, 순시적 채널정보(w_{g _k})를 전달하는 피드백 기법보다 유효 채널정보(g_{g _k})를 전달하기 위한 피드백 부하가 더 적으므로 추가적인 피드백을 고려하더라도 단말기(120)의 전체 피드백 부하를 감소시킬 수 있는 효과가 있다. Direction of each terminal (120) is effective channel g _k (g _{g _k)} in this example (

) Quantization information (

To-noise ratio (SNR) of the effective channel (g _{g - k} ), and transmits channel direction information (CDI)

(CQI: Channel Quality Information) including information on the channel quality information. In order to allow the base station 110 to reflect the influence of the quantization error,

The quantization error size (

To the base station 110. The quantization error size information includes the quantization error size information including the quantization error size information. Quantization error size (

Is a real number between 0 and 1, the increase in feedback load on the terminal 120 due to additional feedback is insignificant. Rather, the terminal 120 may determine that r ^* < In the case of r _{g _ k} , even if additional feedback is taken into consideration, the feedback load for transmitting effective channel information (g _{g _k} ) is smaller than the feedback technique for transmitting the _random channel information (w _{g -} The total feedback load can be reduced.

To this end, the base station 110 transmits to the terminal 120 belonging to a group corresponding to one step of beam forming (B _g) signal. Since the first-stage beamforming (B _g ) is determined only by the spatial correlation of the terminals 120, it is possible to transmit the long-period beamforming as well as the spatial correlation feedback. Therefore, to deliver one beam-forming step (B _g) signal from the base station 110 to the terminal 120 does not act as a large load at the base station (110) position. Further, the downlink transmission of the first beam-forming step (B _g) signal through a downlink, so has enough resources compared to uplink can be carried out relatively accurately.

Estimating method of the SINR _{g _k} from the base station 110 according to this embodiment is as follows.

)를 포함하는 채널방향정보, 유효 채널(g_{g _k})의 신호 대 잡음비(

)를 포함하는 채널품질정보 및 양자화 오류 크기(

)를 포함하는 양자화 오류 크기정보 등의 피드백 정보를 이용해서 [수학식 3]의 각 항들을 추정한다. In the base station 110, the quantized information (e.g.,

) Signal-to-noise ratio of the channel direction information, the effective channel (g _{_k g)} containing (

) And channel quality information including the quantization error size (

(3) by using feedback information such as quantization error size information including the quantization error size information.

)는 [수학식 4]와 같이 추정될 수 있다. First, the strength of the signal for the first stage beamforming (B _g ) in the base station 110

) Can be estimated as shown in Equation (4).

)는 [수학식 5]와 같이 추정될 수 있다. The strength of interference between users in the same group at the base station 110

) Can be estimated as in Equation (5).

)는 [수학식 6]과 같이 추정될 수 있다.Finally, at the base station 110,

) Can be estimated as in Equation (6).

는

에 대한 고유값 분해의 결과이며, σ_{g_k, i}는 ∑_{g_k}의 i 번째 대각선 원소를 의미한다. In Equation (6)

The

And σ _{g_k, i} means the i-th diagonal element of Σ _{g_k} .

Base station 110 may estimate the final SINR _{g _k,} as follows from the formula 4) to (Equation 6].

)를 포함하는 채널방향정보, 유효 채널(g_{g _k})의 신호 대 잡음비(

)를 포함하는 채널품질정보 및 양자화 오류 크기(

)를 포함하는 양자화 오류 크기정보 등의 피드백 정보와, 기지국(110)에서 스케쥴링을 수행하는 동안 결정되는 빔포밍, 그리고 단말기(120) 수에 대한 정보를 이용하여, [수학식 7]로 표현되는 SINR_{g _k} 추정이 가능하다. The quantization information transmitted from the terminal 120

) Signal-to-noise ratio of the channel direction information, the effective channel (g _{_k g)} containing (

) And channel quality information including the quantization error size (

(7), using feedback information such as quantization error size information including information on the number of terminals 120, beamforming determined during scheduling in the base station 110, and information on the number of terminals 120 it is possible to estimate SINR _{g _k.}

로 정의되므로, [수학식 7]을 사용하여 합 전송률을 [수학식 8]과 같이 추정할 수 있다. In this embodiment, the sum rate (R _sum ) is

, The sum rate can be estimated as in Equation (8) by using Equation (7).

In this embodiment, the base station 110 uses Equation (8) to estimate the sum rate for application to the scheduling scheme shown in FIG. 4B. Meanwhile, the base station 110 may perform the scheduling of the terminal 120 using the scheduling scheme shown in FIG. 4B in performing the scheduling of the terminal 120, but the present invention is not limited thereto, and various scheduling schemes may be applied It is possible.

FIG. 3 is a block diagram schematically showing a base station and a terminal included in the large-scale MIMO system according to the present embodiment.

The base station 110 according to the present embodiment includes a communication unit 310, a first stage beamforming module 320 and a second stage beamforming module 330. The terminal 120 includes a beamforming receiver 340, (350) and a feedback control unit (360). The first stage beamforming module 320 includes a group setting unit 322 and a first beamforming unit 324. The second stage beamforming module 330 includes a feedback processing unit 332, a scheduling unit 334, And a second beamforming unit 336.

The base station 110 and the terminal 120 shown in FIG. 3 are according to one embodiment, and not all of the blocks shown in FIG. 3 are essential elements. In other embodiments, the base station 110 and the terminal 120 Some included blocks may be added, changed or deleted.

The communication unit 310 is connected to the terminal 120 and transmits / receives various data.

The communication unit 310 transmits a pilot signal (Pilot Singal) for feedback on the spatial correlation matrix of the channel to the terminal 120 and receives information on the spatial correlation matrix corresponding to the pilot signal. Here, the information on the spatial correlation matrix means the eigenvalue matrix of the spatial correlation matrix and the eigenvector matrix feedback. The communication unit 310 transmits the information on the spatial correlation matrix received from the terminal 120 to the 1 st stage beamforming module 320 to perform 1 st stage beamforming.

In addition, the communication unit 310 transmits a reference signal for channel estimation to the terminal 120 and feedbacks the estimated effective channel related information based on the reference signal from the terminal 120. Here, the effective channel related information includes channel direction information, channel quality information, and quantization error size information calculated on the basis of the effective channel estimated by the terminal 120. The communication unit 310 transmits the effective channel related information received from the terminal 120 to the two-stage beamforming module 330 to perform two-stage beamforming.

The first-stage beamforming module 320 performs user grouping based on the received spatial correlation matrix based on the pilot signal transmitted to the terminal 120, and performs a grouping of one or a plurality of terminals 120 ) To perform one-stage beamforming.

The group setting unit 322 groups a plurality of terminals 120 located in the cell according to spatial correlation based on the spatial correlation matrix received from the terminal 110. In other words, the group setting unit 322 compares the spatial correlation matrix received from the terminal 110 with predetermined evaluation criteria, and allocates the similar terminal 120 to the same group of comparison result values. For example, the group setting unit 322 sets the terminals 120 included in the predetermined error range by comparing the spatial correlation matrix with a predetermined evaluation criterion in the same group.

The first beamforming unit 324 determines a first-stage beamforming signal for the grouped terminal 110 and performs one-step beamforming to the terminal 120. [ Here, the first beamforming unit 324 may perform one beamforming in a period of a predetermined frame (for example, several tens of frames or several hundreds of frames) according to the movement of the terminal 120 or the scattering environment.

The two-stage beamforming module 330 performs one-stage beamforming and then performs two-stage beamforming. More specifically, the two-stage beamforming module 330 estimates a signal-to-interference noise ratio based on the obtained effective channel-related information, performs scheduling, and performs a two-step beamforming to the scheduled terminal 120.

The feedback processor 332 estimates the signal-to-interference noise ratio based on the obtained effective channel-related information. To be more specific, the feedback processor 332 estimates a signal-to-interference noise ratio using effective channel-related information including channel direction information, channel quality information, and quantization error size information acquired from the terminal 120.

), 그룹 설정부(322)에서 설정된 동일 그룹 내에 포함된 단말기(120) 간 간섭의 세기(

) 및 서로 다른 그룹 간 간섭의 세기(

)를 산출한다. The feedback processing unit 332 uses the effective channel-related information to calculate the intensity of the signal for the first stage beamforming (B _g )

), The intensity of the interference between the terminals 120 included in the same group set by the group setting unit 322

) And the intensity of the different intergroup interference (

).

)의 곱을 단말기별 전송파워(P/K)와 동일 그룹 내 간섭의 세기(

)의 곱, 상기 단말기별 전송파워(P/K)와 상기 그룹 간 간섭의 세기(

)의 곱 및 노이즈(σ_n ²)의 합으로 나눈 식(예: [수학식 3])을 이용하여 신호 대 간섭 잡음비를 추정할 수 있다. The feedback processor 332 receives the transmission power P / K and the intensity of the signal

) Is calculated as the transmission power per terminal (P / K) and the intensity of interference within the same group

), The transmission power per unit (P / K) and the intensity of the intergroup interference

) And a noise (? _N ² ), for example, an equation (3)) can be used to estimate the signal-to-interference noise ratio.

)를

로 추정하고, 동일 그룹 내 단말기(120) 간 간섭의 세기(

)를

로 추정하며, 그룹 간 간섭의 세기(

)를

로 추정하고, 추정된 신호의 세기(

), 동일 그룹 내 간섭의 세기(

) 및 그룹 간 간섭의 세기(

)를 [수학식 3]에 적용하여 도출된 [수학식 7]을 이용하여 신호 대 간섭 잡음비를 추정할 수 있다. Further, the feedback processor 332 may use the acquired effective channel-related information to determine the intensity

)

And the intensity of the interference between the terminals 120 in the same group (

)

, And the intensity of intergroup interference (

)

, And the estimated signal intensity (

), Intensity of interference within the same group (

) And intensity of intergroup interference (

) Is applied to Equation (3), the signal-to-interference noise ratio can be estimated using Equation (7).

The scheduling unit 334 performs scheduling for the terminal 120 based on the estimated signal-to-interference noise ratio. In other words, the scheduling unit 334 calculates a sum rate using the estimated signal-to-interference noise ratio and performs scheduling based on the calculated sum rate.

에 적용하여 도출된 [수학식 8]을 기반으로 산출된 합 전송률을 기반으로 스케쥴링을 수행한다. 스케쥴링부(334)는 산출된 합 전송률이 최대값을 갖도록 스케쥴링을 수행할 수 있으나 반드시 이에 한정되는 것은 아니며, 다양한 방식의 스케쥴링 방식이 적용 가능하다. The scheduling unit 334 estimates the estimated signal-to-interference noise ratio

And performs a scheduling based on the sum rate calculated based on Equation (8) derived by applying Equation (8). The scheduling unit 334 may perform scheduling so that the calculated sum rate has a maximum value, but the present invention is not limited thereto, and various scheduling schemes can be applied.

The second beamforming unit 336 performs two-step beamforming to the terminal 120 scheduled by the scheduling unit 334. [ The second beamforming unit 336 transmits the actual data to the scheduled terminal 120 using the two-stage beamforming. Although it is described that the second beamforming unit 336 transmits data using only two-stage beamforming, the present invention is not limited to this. The final beamforming combining the one-stage beamforming and the two- Lt; / RTI &gt;

When receiving the pilot signal from the base station 110, the beamforming receiver 340 estimates and feeds back the spatial correlation matrix, and receives the transmitted first-stage beamforming signal based on the spatial correlation matrix at the base station 110. [ When the feedback control unit 360 feeds back effective channel related information to the base station 110, the beamforming receiver 340 receives the second-stage beamforming signal from the base station 110. [

The channel estimation unit 350 estimates a channel based on the reference signal received from the base station 110, and estimates an effective channel based on the first-step beamforming signal and the estimated channel. The channel estimator 350 identifies a reference signal in a received signal of the first-stage beamforming signal received by the beamforming receiver 340, estimates a channel using the reference signal and the received signal, It is multiplied by the channel matrix (h _{g _k)} and step beamforming matrix corresponding to the beam-forming signals phase 1 (B _g), which estimates the effective channel _(g g _{_k).}

The feedback control unit 360 calculates effective channel related information based on the effective channel estimated by the channel estimation unit 350 and feeds back the calculated effective channel related information to the base station 110. [ Here, the effective channel related information includes channel direction information, channel quality information, and quantization error size information calculated on the basis of the effective channel.

)을 산출하고, 유효 채널 방향을 양자화한 양자화 정보(

)를 포함하는 채널방향정보, 유효 채널 방향에 대한 양자화 오류정보(e_g_k)를 산출하고, 양자화 오류정보의 크기(

)를 포함하는 양자화 오류 크기정보 및 유효 채널(g_g_k)에 대한 신호 대 잡음비(

)를 포함하는 채널품질정보를 포함하는 유효채널 관련정보를 기지국(110)으로 피드백한다.The feedback controller 360 calculates the effective channel direction for the estimated effective channel (

), Quantizes the quantization information obtained by quantizing the effective channel direction (

), Quantization error information e_g _k for the effective channel direction, and outputs the quantization error information size

) And the signal-to-noise ratio (&lt; _RTI ID _{= 0.0} &gt;

To the base station 110 based on the channel quality information.

4A and 4B are flowcharts for explaining a step-by-step beamforming and scheduling-based wireless communication method according to the present embodiment.

4A illustrates the operation of feedback and scheduling between the Node B 110 and the UE 120 when performing the stepped beamforming according to the present embodiment.

The base station 110 and the terminal 120 perform a training process for estimating a spatial correlation matrix of a channel (S410). Here, the training process transmits a pilot signal from the base station 110 to the terminal 120, and confirms the difference in the size, direction, etc. of the pilot signal received from the terminal 120. For example, when the base station 110 transmits the pilot signal of '11111' to the terminal 120, the terminal 120 determines whether the pilot signal of '11111' is received or the pilot signal of which difference of '11001' And performs a training process for confirming the result.

When the training process ends, the terminal 120 feeds back the estimated spatial correlation matrix to the base station 110 based on the reception difference of the pilot signal (S412). In other words, the terminal 120 feeds back the eigenvalue matrix and the eigenvector matrix for the estimated spatial correlation matrix to the base station.

The base station 110 performs grouping on the user 120, i.e., the terminal 120 using the spatial correlation matrix received through feedback (S414). The base station 110 compares the spatial correlation matrix received from the terminal 110 with a predetermined evaluation criterion and assigns the comparison result values to the same group of the terminals 120.

The base station 110 determines the first-step beamforming using the grouped group and transmits the first-step beamforming signal to the terminal 120 (S416). Here, steps S410 to S416 may be defined as 'frame 1' for performing the first-step beamforming in the base station 110, and the period of repeating steps S410 to S416 may be defined as a period of movement of the terminal 120, (E.g., several tens of frames or several hundreds of frames) depending on the change.

After performing steps S410 to S416, the base station 110 and the terminal 120 perform a process for transmitting actual data using two-step beamforming.

The base station 110 transmits a reference signal for channel estimation to the terminal 120 (S420).

The terminal 120 estimates a channel based on the received reference signal, estimates an effective channel based on the estimated channel, and feeds back effective channel related information to the base station 110 (S422). The terminal 120 identifies the reference signal in the received signal of the first-stage beamforming signal, estimates the channel using the reference signal and the received signal, and outputs a 1-stage beamforming matrix corresponding to the 1-stage beamforming signal, Channel by the channel matrix corresponding to the channel, and feeds back the effective channel-related information to the base station 110 by estimating the effective channel.

In other words, the terminal 120 feeds effective channel related information including channel direction information, channel quality information, and quantization error size information calculated based on the effective channel to the base station 110.

The base station 110 performs scheduling for the terminal 120 using the effective channel related information fed back from the terminal 120, and determines a 2-step beamforming (S424). The base station 110 transmits data to the scheduled terminal 120 using the 2-step beamforming (S426). Although the base station 110 transmits data using two-stage beamforming, data can be transmitted through the final beamforming combining the first-stage beamforming and the second-stage beamforming.

Steps S420 to S426 may be defined as 'frames 2, 3, ..., n' for performing two-step beamforming in the base station 110, and the period of repeating steps S420 to S426 is every frame, Is shorter than the cycle of repeating steps S410 to S416.

4A, steps S410 to S426 are sequentially performed. However, this is merely illustrative of the technical idea of an embodiment of the present invention, and it is not intended to limit the scope of the present invention to conventional knowledge in the technical field to which an embodiment of the present invention belongs. Those skilled in the art will appreciate that various modifications and adaptations may be made to those skilled in the art without departing from the essential characteristics of one embodiment of the present invention by changing the order described in Figure 4a or by executing one or more of steps S410 through S426 in parallel 4A is not limited to the time series order.

FIG. 4B shows an operation procedure of the scheduling of the BS 110 when performing the stepwise beamforming according to the present embodiment.

The base station 110 estimates a data rate of each terminal 120 and performs scheduling for maximizing a sum rate. At this time, the spatial correlation of the channel is fed back with a long period, and the instantaneous channel direction and size are fed back every frame. This feedback scheme can effectively reduce the feedback load, but quantization errors due to limited feedback still exist.

In particular, when spatial correlation information of a channel already has inaccuracies due to quantization errors, the entire channel information estimated by the base station is transmitted with a quantization error, resulting in greater inaccuracy. In addition, the sum rate calculation method used in the general scheduling scheme does not reflect the inaccuracy of the channel information generated in the limited feedback situation, and the scheduling is not correctly performed due to an incorrect transmission rate calculation.

Hereinafter, the scheduling operation of the base station 110 according to the present embodiment will be described.

Using the feedback according to the present embodiment and the sum rate estimation calculated by Equation (8), the base station 110 can perform effective scheduling considering the inaccuracy of channel information even in a limited feedback state.

To this end, the base station 110 defines a standby user set S _remain and a selected (scheduled to be scheduled) user set S _select .

The base station 110 is the final sum rate (R _sum) and the temporary sum rate (R _temp) the initial set to 0, the initial set and the empty set a selected set of users (S _select) (S430), a selected set of users (S _select ) To the standby user set (S _remain ) (S432). In other words, the base station 110 sets all users not selected for S _select to S _remain .

In step S436, the base station 110 searches for a user having a maximum sum rate calculated by Equation (8) when scheduling together with users belonging to S _select for each user belonging to S _remain .

If the final sum rate (R _sum ) is larger than the sum rate (R _temp ) obtained by the users of the existing S _select, the BS 110 adds the user to the S _select (S442).

The base station 110 is the final sum rate (R _sum) the user to the sum rate (R _temp) for obtaining the existing S _select for each user until it becomes smaller than the sum rate (R _temp) to get to the user's existing S _select And the final sum rate (R _sum ) are repeated (S444, S446, S450).

In step S460, the base station 110 repeats steps S432 to S450 if the size of S _select is smaller than min {N _T , S}, which is the maximum number of users that can be scheduled. That is, the base station 110 adds one user to maximize the sum rate until the sum rate calculated by Equation (8) does not increase or reaches the maximum number of users.

The base station 110 can perform effective scheduling to maximize the sum rate even in a limited feedback state by the scheduling method of performing the steps S430 to S460 described above.

FIG. 5 is a diagram comparing the quantization method and the accuracy of a general quantization method according to the present embodiment. FIG. 6 is a diagram showing a change in the sum rate based on the number of transmit antennas of the base station according to the present embodiment, FIG. 8 is a diagram illustrating a change in the sum rate based on the total number of users according to an embodiment.

Hereinafter, a description will be given of a computer simulation process and results of confirming an increase in the sum rate when applying the feedback and scheduling technique according to the present embodiment.

의 모델에 따라 생성하였고, 공간상관행렬(R_k = U_kΛ_kU^H _k)은 송신 안테나를 균일한 선형 배열로 가정하여 다음 모델에 따라 생성하였다.For computer simulation, we set the transmission power P to 10 dB, the large fading α _k and the noise power σ ² _n to 1. The channel information (h _k )

And the spatial correlation matrix (R _k = U _k Λ _k U ^H _k ) is generated according to the following model assuming that the transmit antennas are a uniform linear array.

Here, antenna spacing D was set to 0.5, and θ _k and Δ _k were generated as random variables with an even distribution of (-60 °, 60 °) and an even distribution of (5 °, 15 °), respectively.

The quantization method used for feedback in the simulation is as follows.

로 설정한다. 고유값 행렬(Λ_k)은 실수값인 대각 성분들만 양자화하면 되므로 완벽하게 양자화가 이루어졌다고 가정하였다. 유효 채널의 크기(

)의 양자화를 위해, r^* 차원의 B_H-비트 랜덤 벡터 양자기로 생성한 코드북의 각 코드워드의 크기를 평균화하였고, 이렇게 얻어진 코드북에서 유효 채널의 크기(

)와 내적하여 크기가 가장 큰 코드워드를 양자화 정보(

)로 설정하였다. 순시적 채널의 크기(

)와 채널 정보의 크기(

)의 경우에는, 각각 r_k 차원의 B_H-비트 랜덤 벡터 양자기와 N_T 차원의 B_H-비트 랜덤 벡터 양자기를 사용하여 유효 채널의 크기(

)와 동일한 방식으로 양자화하였다.For quantization of the eigenvector matrix _{(U k), N (0} , 0.5) to about the optimal B _u - it was quantization bits using both an scalar eigenvector matrix (U _k) each element of a real part and an imaginary part of each . The orthogonal matrix generated by the QR decomposition on the quantized eigenvector matrix (U _k )

. It is assumed that the eigenvalue matrix (Λ _k ) is perfectly quantized since only the diagonal components, which are real numbers, need to be quantized. Effective channel size (

) For quantization, r ^* The size of each code word in the codebook generated by the B _H - bit random vector quantizer of the dimension is averaged, and the size of the effective channel

) And the codeword having the largest size is quantized (&quot;

). The size of the momentary channel (

) And the size of the channel information (

), The dimension of each of r _k B _H for-bit random vector both groups of N _T D B _H-bit random vector size of both using an effective channel (

). &Lt; / RTI &gt;

The user grouping process used in the simulation is as follows.

로 정의된 평가 기준에 따라 평가값이 가장 높은

번째 그룹에 속하도록 배정하였다. 단말기(120)에 대한 그룹 배정이 끝나면 각 그룹에서 대표 고유벡터 행렬(U_g)을 찾기 위해서 그룹 내 단말기(120)들의 공간상관행렬(R_k)의 평균을

로 구하고, 평균 공간상관행렬에 대한 고유벡터 행렬(U_g)을 설정한다. 여기서, S_g는 직전 과정에서 그룹 g에 배정된 단말기(120)들의 집합이다. First, each terminal 120

, The highest evaluation value

The second group. After the group assignment for the terminal 120 is completed, the average of the spatial correlation matrix R _k of the terminals 120 in the group 120 is calculated to find the representative eigenvector matrix U _g in each group

To obtain, it sets the eigenvector matrix (U _g) of the average spatial correlation matrix. Here, S _g is a set of terminals 120 assigned to the group g in the immediately preceding process.

가 이전 값의 0.001배 미만이 될 때까지 반복하여, 최종적인 단말기(120) 그룹을 결정한다.The above procedure is repeated for the total evaluation value

Until it is less than 0.001 times the previous value, to determine the final terminal 120 group.

The beamforming decision procedure used in the simulation is as follows.

First, intrinsic beamforming is used as the first step beamforming for each group. That is, 1 to r ^* th columns of the representative eigenvector matrix U _g of each group are selected and set to one-stage beamforming (B _g ). Next, block diagonalization is used for two-stage beamforming for terminals 120 in each group. At this time, only the interference to other terminals 120 scheduled in the same group is removed, and the power is uniformly allocated.

)를 사용하였다. 유효 채널 방향과 양자화된 방향 사이의 내적 크기(

)의 최대치는 1이며, 그 값이 높을수록 양자화된 방향이 실제 유효 채널 방향에 가까운 것을 의미한다. 도 5에서 나타난 것 같이, 본 발명에서 제안한 유효 채널의 크기(

)를 양자화하는 방식(510)은 채널 정보의 크기(

)를 양자화하는 방식(종래기술 1, 520)이나 순시적 채널의 크기(

)를 양자화하는 방식(종래기술 2, 530)에 비해 적은 피드백 부하를 가지므로, 같은 피드백 양으로 더 우수한 양자화 성능을 제공할 수 있다. FIG. 5 is a diagram comparing the accuracy of a quantization method and a general quantization method according to the present embodiment. In Figure 5, N _T = 32, S = 50, G = 8, and _Bi = 2 and r ^* = 4. In order to show the accuracy of the effective channel quantization, the inner size between the effective channel direction and the quantized direction

) Was used. The inner product size between the effective channel direction and the quantized direction

) Is 1, and the higher the value, the closer the quantized direction is to the actual effective channel direction. As shown in FIG. 5, the size of the effective channel proposed in the present invention

) 510 is a method of quantizing the size of the channel information (

) (Prior Art 1, 520) or the size of a random channel

(Prior art 2, 530), it is possible to provide better quantization performance with the same amount of feedback.

와 고유값 행렬을 양자화한

가 이미 부정확하므로 오류 전달로 인해

로 표현되는 전체 채널의 양자화가 부정확해진다. 따라서, 종래기술 2는 종래기술 1 보다 피드백 부하가 적지만 양자화의 정확성이 더 떨어지게 된다. In the case of Prior Art 2, the eigenvector matrix is quantized

And the eigenvalue matrix are quantized

Is already inaccurate due to error propagation

The quantization of the entire channel expressed by &lt; RTI ID = 0.0 &gt; Therefore, although the conventional art 2 has a smaller feedback load than the conventional art 1, the accuracy of the quantization is lowered.

6 is a diagram showing a change in the sum rate based on the number of transmission antennas of the base station according to the present embodiment, wherein S = 50, G = 8, B _U = 2, B _H = 10, and r ^* = 4 . 7 is a diagram showing a change in the sum rate based on the total number of users according to the present embodiment, where N _T = 32, G = 8, _BU = 2, _BH = 10, and r ^* = 4.

As shown in FIG. 6 and FIG. 7, the present invention can achieve the highest total transmission rate with respect to the number of transmission antennas 610 and the number of users 710. The performance of the prior art 1 (620, 720) and the prior art 2 (630, 730) is largely deteriorated due to the effect of the SINR estimation formula that does not reflect the quantization error and the inaccuracy that occurs in the feedback.

The foregoing description is merely illustrative of the technical idea of the present embodiment, and various modifications and changes may be made to those skilled in the art without departing from the essential characteristics of the embodiments. Therefore, the present embodiments are to be construed as illustrative rather than restrictive, and the scope of the technical idea of the present embodiment is not limited by these embodiments. The scope of protection of the present embodiment should be construed according to the following claims, and all technical ideas within the scope of equivalents thereof should be construed as being included in the scope of the present invention.

As described above, the present embodiment is applied to a wireless communication field that performs stepwise beamforming, thereby reducing a quantization error, performing accurate beamforming, estimating a sum rate close to an actual value, and performing accurate scheduling It is a useful invention that generates the effect of improving the multi-user gain of a multi-input / output system.

110: base station 120:
310: communication unit 320: first stage beamforming module
322: group setting unit 324: first beamforming unit
330: one stage beamforming module 332: feedback processing section
334: Scheduling unit 336: Second beamforming unit
340: beamforming receiver 350: channel estimator
360:

Claims (14)

A beamforming receiver configured to feedback the estimated spatial correlation matrix when receiving a pilot signal from a base station and to receive a determined first stage beamforming signal based on the feedback;
A channel estimator for estimating a channel based on a reference signal received from the base station, estimating an effective channel based on the one-stage beamforming signal and the channel; And
Related information based on the effective channel and feeds back the effective channel-related information to the base station,
And a second terminal. The method according to claim 1,
Wherein the feedback transmission unit comprises:
Related information including channel direction information (CDI), channel quality information (CQI) and quantization error size information based on the effective channel is fed back to the base station. terminal. 3. The method of claim 2,
Wherein the feedback transmission unit comprises:
Effective channel direction on the effective channel _(g g _{_k)} (

), Quantizes the quantization information obtained by quantizing the calculated effective channel direction (

(CDI) including the channel direction information (CDI) to the base station. The method of claim 3,
Wherein the feedback transmission unit comprises:
Calculating a quantization error information (e_g _k) for the effective channel direction, and the size of the calculated quantization error information (

And the quantization error size information including the quantization error size information to the base station. 3. The method of claim 2,
Wherein the feedback transmission unit comprises:
To-noise ratio (SNR) for the effective channel g_g _k , and calculates the signal-to-noise ratio (SNR)

(CQI) including the channel quality information (CQI) to the base station. The method according to claim 1,
Wherein the channel estimator comprises:
Terminal in the received signal (y _{_k g)} of step 1 beamformed signal, wherein the estimating the channel using the reference signal and the received signal to determine the reference signal, and. The method according to claim 1,
Wherein the channel estimator comprises:
Calculating a channel matrix (h _{g _k)} corresponding to the first stage beamforming matrix (B _g) and estimated the channel corresponding to the step 1, beam forming signal, and the forming matrix and the step beam by the product of the channel matrix And estimates the effective channel _{g g_k} . A one-step beamforming unit for transmitting a pilot signal to a terminal, performing grouping based on the received spatial correlation matrix based on the pilot signal, and performing one-step beamforming on the terminal included in each group;
A feedback processor for transmitting a reference signal to the mobile station and estimating a signal-to-interference and noise ratio (SINR) by obtaining effective channel related information for an estimated effective channel based on the reference signal;
A scheduling unit for calculating a sum rate (R _SUM ) using the SNR and performing a scheduling based on the sum rate; And
A two-stage beamforming unit for performing two-step beamforming on a scheduled terminal and transmitting data;
And a base station. 9. The method of claim 8,
Wherein the feedback processing unit comprises:
And estimating the signal-to-interference noise ratio by obtaining the effective channel-related information including channel direction information (CDI), channel quality information (CQI), and quantization error size information calculated based on the effective channel from the terminal . 10. The method of claim 9,
Wherein the feedback processing unit comprises:
And the strength (B _g ) of the signal (B _g ) for the first stage beamforming

), The intensity of the inter-terminal interference included in the same group among the groups

) And different intensities of the intergroup interference (

), And calculates the intensity of the signal (

), The intensity of the same intra-group interference (

) And intensity of intergroup interference (

) To estimate the signal to interference noise ratio (SNR). 11. The method of claim 10,
Wherein the feedback processing unit comprises:
(P / K) and the intensity of the signal

(P / K) and the intensity of the same intra-group interference (P / K)

), The transmission power per unit (P / K) and the intensity of the intergroup interference

) And noise (? _N ² ) to estimate the signal to interference noise ratio (SNR). 11. The method of claim 10,
Wherein the feedback processing unit comprises:
And quantization information obtained by quantizing the effective channel direction from the terminal

To-noise ratio (CDI) calculated based on the channel direction information CDI and the effective channel g_g _k ,

(CQI) including the channel quality information (CQI) and the size of the quantization error for the effective channel direction

And the quantization error size information including the quantization error size information. A receiving step of receiving a pilot signal from a base station and feedbacking an estimated spatial correlation matrix and receiving a determined one-stage beamforming signal based on the feedback;
A channel estimation process for estimating a channel based on a reference signal received from the base station;
An effective channel estimator for estimating an effective channel based on the one-stage beamforming signal and the channel; And
Related information including channel direction information (CDI), channel quality information (CQI), and quantization error information using the effective channel, and feedback-transmits the effective channel-related information to the base station
Wherein the step of generating the feedback beam comprises: A one-step beamforming process of transmitting a pilot signal to a terminal, performing grouping based on the received spatial correlation matrix based on the pilot signal, and performing one-step beamforming on the terminal included in each group;
A feedback reception unit for transmitting a reference signal to the terminal and acquiring effective channel related information including channel direction information (CDI), channel quality information (CQI), and quantization error information for an estimated effective channel based on the reference signal process;
A feedback processing step of estimating a signal-to-interference noise ratio based on the effective channel-related information;
A scheduling step of calculating an aggregate data rate using the SNR and performing a scheduling based on the aggregate data rate; And
A two-step beamforming process for performing data transmission by performing two-step beamforming on a scheduled terminal
Wherein the scheduling step comprises:

Images