KR102197717B1 - Method for multi-input multi-output communication in large-scale antenna system - Google Patents

Abstract

A multiple input multiple output communication method in a large-scale antenna system is disclosed. The MIMO transmission method includes obtaining channel information for a terminal, classifying the terminal into a class and a group dependent on the class based on the channel information, determining a group beamforming matrix for the group, and a group beamforming matrix Performing group beamforming transmission based on group beamforming to UEs belonging to the group, obtaining single SU-CQI information and interference signal information for UEs belonging to the group, UE based on SU-CQI information and interference signal information Scheduling them, and transmitting data to the terminals based on the scheduling. Therefore, it is possible to reduce the amount of feedback of the channel state information.

Description

Multi-input multi-output communication method in a large-scale antenna system {METHOD FOR MULTI-INPUT MULTI-OUTPUT COMMUNICATION IN LARGE-SCALE ANTENNA SYSTEM}

The present invention relates to a multi-input multi-output (MIMO) communication method in a large-scale antenna system, and more particularly, feedback of a small amount of channel state information in a large-scale antenna MIMO channel environment with antenna correlation. The present invention relates to a MIMO communication method for maximizing frequency efficiency of uplink and downlink while being required.

A mobile communication system after 4G (B4G) needs to increase frequency efficiency by 10 times or more compared to a 4G system such as 3GPP long term evolution (LTE) due to a rapid increase in data traffic. Current network MIMO, interference alignment, relay networks, heterogeneous networks, and large-scale antennas as the physical layer technology required to achieve the goal of increasing frequency efficiency by more than 10 times. Technology and the like are being mentioned.

The present invention relates to a massive MIMO (or large-scale antenna) system capable of obtaining a very large effect as a technique for increasing frequency efficiency. Conventional large-scale antenna systems have been limited to a time division duplex (TDD) scheme. In the frequency division duplex (FDD) scheme, there is a problem in that a large-scale antenna transmitter requires a large number of reference signals and radio resources for channel state information feedback, which is virtually impossible to obtain channel state information.

In addition, there is a problem in that scheduling and precoding calculation complexity is significantly increased compared to the conventional system due to a large increase in the number of users that can be simultaneously accommodated by a large-scale transmission antenna.

An object of the present invention for solving the above problems is to provide a MIMO transmission method suitable for a large-scale antenna system capable of minimizing a decrease in frequency efficiency while not increasing the amount of radio resources required for channel state information feedback. To provide.

Another object of the present invention for solving the above problems is a MIMO reception method suitable for a large-scale antenna system capable of minimizing a decrease in frequency efficiency at the same time without increasing the amount of radio resources required for channel state information feedback To provide.

In the MIMO transmission method of a base station according to an embodiment of the present invention for achieving the above object, obtaining channel information for at least one terminal, the at least one terminal based on the channel information at least one class Classifying into at least one group dependent on and class, determining a group beamforming matrix for each classified group, performing group beamforming transmission based on the group beamforming matrix to terminals belonging to each group The step of, obtaining SU-CQI information and interference signal information for the UEs belonging to each group, scheduling UEs based on the SU-CQI information and the interference signal information, and data based on scheduling And transmitting to the terminals.

Here, the channel information may include at least one of at least one of a transmission correlation matrix, an eigenvalue of a transmission correlation matrix, an eigenvector of a transmission correlation matrix, an AS, AOD, and at least one long period PMI selected from a fixed codebook indicating channel information. have.

Here, in the step of classifying into groups, terminals having similar transmission correlation matrices may be classified into one group.

Here, in the step of classifying into groups, terminals having similar effective eigenvectors of the transmission correlation matrix may be classified into one group, and groups having high orthogonality between effective eigenvectors of the transmission correlation matrix may be classified into one class. Can be classified.

Here, in the determining of the group beamforming matrix, the group beamforming matrix may be determined through block diagonalization so that group beamforming matrices for each group are similarly orthogonal to each other based on the channel information and the one-ring channel model. .

Here, the interference signal information may include at least one of an interference signal strength between terminals belonging to the group and an interference signal strength with another group.

Here, the interference signal information may be an offset value for the SU-CQI.

Here, the interference signal information may be a CQI expressed in an MCS level.

Here, the interference signal information may be an interference signal strength with respect to a precoding matrix indicator that will act as interference to terminals belonging to the group.

Here, the interference signal information may be a ratio to the SU-CQI.

Here, the scheduling of the terminals includes calculating MU-CQI information for each terminal based on a combination of groups belonging to each class, a combination of terminals belonging to each group, and a combination of ranks for each terminal, MU -Calculating a sum-PF metric based on CQI, selecting a multi-user combination with the largest sum-PF metric in each class as the sum-PF metric of each class, the largest sum-PF among all classes Selecting a class having a metric, and determining a multi-user combination having the least interference between each other in the selected class, and scheduling the determined multi-user combination.

Here, in the step of transmitting to the terminals, data may be transmitted to the terminals through a rank different from a rank related to the SU-CQI information and the interference signal information.

A method for receiving MIMO by a terminal according to an embodiment of the present invention for achieving the above other object includes receiving a signal to which a group beamforming matrix is applied for a group including the terminal, and a reference signal to which the group beamforming matrix is applied. Or generating SU-CQI information and interference signal information using a reference signal to which the group beamforming matrix is not applied, and feeding back the SU-CQI information and the interference signal information to a base station.

Here, the interference signal information may include at least one of an interference signal strength between terminals belonging to the group and an interference signal strength with another group.

Here, the interference signal information may be an offset value for the SU-CQI.

Here, the interference signal information may be a CQI expressed in an MCS level.

Here, the interference signal information may be an interference signal strength with respect to a precoding matrix indicator that will act as interference to terminals belonging to the group.

Here, the interference signal information may be a ratio to the SU-CQI.

Here, in the step of feeding back, the terminal may feed back the interference signal information including a PMI whose inner product is smaller than a preset value among PMIs that will act as interference to itself.

Here, in the step of feeding back, the terminal may feed back interference signal information for a subband in which the strength of the interference signal is smaller than a preset value among the subbands.

According to the present invention, it is possible to reduce the amount of feedback of channel state information in MIMO transmission. In addition, even when the amount of feedback of the channel state information is reduced, it is possible to minimize a decrease in frequency efficiency in uplink and downlink.

1 is a conceptual diagram illustrating the concept of spatial separation between user groups in a MIMO transmission/reception method according to the present invention.
2 is a conceptual diagram showing a CSI-RS pattern of 8 antenna ports.
3 is a flowchart illustrating a MIMO transmission method according to an embodiment of the present invention.
4 is a flowchart illustrating a scheduling process in a MIMO transmission method according to an embodiment of the present invention.

In the present invention, various modifications may be made and various embodiments may be provided, and specific embodiments will be illustrated in the drawings and described in detail.

However, this is not intended to limit the present invention to a specific embodiment, it is to be understood to include all changes, equivalents, and substitutes included in the spirit and scope of the present invention.

Terms such as first and second may be used to describe various components, but the components should not be limited by the terms. These terms are used only for the purpose of distinguishing one component from another component. For example, without departing from the scope of the present invention, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element. The term and/or includes a combination of a plurality of related listed items or any of a plurality of related listed items.

When a component is referred to as being "connected" or "connected" to another component, it is understood that it may be directly connected or connected to the other component, but other components may exist in the middle. Should be. On the other hand, when a component is referred to as being "directly connected" or "directly connected" to another component, it should be understood that there is no other component in the middle.

The terms used in the present application are only used to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In the present application, terms such as "comprise" or "have" are intended to designate the presence of features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, but one or more other features. It is to be understood that the presence or addition of elements or numbers, steps, actions, components, parts, or combinations thereof, does not preclude in advance.

Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs. Terms as defined in a commonly used dictionary should be interpreted as having a meaning consistent with the meaning in the context of the related technology, and should not be interpreted as an ideal or excessively formal meaning unless explicitly defined in the present application. Does not.

Hereinafter, preferred embodiments of the present invention will be described in more detail with reference to the accompanying drawings. In describing the present invention, in order to facilitate an overall understanding, the same reference numerals are used for the same elements in the drawings, and duplicate descriptions for the same elements are omitted.

Throughout the specification, a network is, for example, a wireless Internet such as WiFi (wireless fidelity), a mobile Internet such as WiBro (wireless broadband internet) or WiMax (world interoperability for microwave access), and a global system for mobile communication (GSM). ) Or a 2G mobile communication network such as code division multiple access (CDMA), a 3G mobile communication network such as wideband code division multiple access (WCDMA) or CDMA2000, a high speed downlink packet access (HSDPA) or a high speed uplink packet access (HSUPA). It may include a 3.5G mobile communication network, a 4G mobile communication network such as a long term evolution (LTE) network or an LTE-Advanced network, and a 5G mobile communication network.

Throughout the specification, a terminal is a mobile station, a mobile terminal, a subscriber station, a portable subscriber station, user equipment, and an access terminal. And the like, and may include all or part of functions such as a terminal, a mobile station, a mobile terminal, a subscriber station, a mobile subscriber station, a user equipment, and an access terminal.

Here, a desktop computer, a laptop computer, a tablet PC, a wireless phone, a mobile phone, a smart phone, and a smart watch that can communicate with a terminal. (smart watch), smart glass, e-book reader, PMP (portable multimedia player), portable game console, navigation device, digital camera, DMB (digital multimedia broadcasting) player, digital voice Recorder (digital audio recorder), digital audio player (digital audio player), digital video recorder (digital picture recorder), digital video player (digital picture player), digital video recorder (digital video recorder), digital video player (digital video player) ) Can be used.

Throughout the specification, a base station is an access point, a radio access station, a node B, an evolved node B, a base transceiver station, and an MMR. mobile multihop relay)-BS, or the like, and may include all or part of functions such as a base station, an access point, a radio access station, a nodeB, an eNodeB, a transmission/reception base station, and an MMR-BS.

Fields to which an embodiment of the present invention described below is applied are uplink and downlink of cellular communication. Here, a single cell may include a base station having M antennas and K users (ie, terminals) each having N antennas. In addition, it is assumed that the transmit antenna correlation of each terminal is high (that is, the angle spread (AS) is small). For example, in a channel environment in which the downlink urban macro and line of sight (LOS) components are strong, the correlation between the transmission antennas is high.

Also, for convenience, it is assumed that K terminals can be classified into G groups that can be spatially separated according to the similarity of transmit antenna correlations, and K'terminals exist in each group.

The channel model considered in the present invention is shown in Equation 1 below.

는 i.i.d.(independently and identically distributed) 채널 행렬이고,

는 송신 상관 행렬이고,

는 수신 상관 행렬이다. 편의상, 다중 사용자 MIMO(MU-MIMO)에서 이른바 원-링 채널 모델(one-ring channel model)을 고려하면

이다. 즉, 수신 상관도는 없다고 가정한다.here,

Is the iid (independently and identically distributed) channel matrix,

Is the transmission correlation matrix,

Is the received correlation matrix. For convenience, considering the so-called one-ring channel model in multi-user MIMO (MU-MIMO),

to be. That is, it is assumed that there is no reception correlation.

The transmission signal model proposed in the present invention is shown in Equation 2 below.

는 채널의 통계적 특성에 기반하는 빔포밍 행렬이고,

는 순시 채널 정보

에 기반하는 프리코딩(precoding) 행렬이고,

는 데이터 심볼 벡터(symbol vector)이다.here,

Is a beamforming matrix based on the statistical characteristics of the channel,

Is instantaneous channel information

Is a precoding matrix based on,

Is a data symbol vector.

The received signal model proposed in the present invention is shown in Equation 3 below.

는 잡음 신호이고,

는 아래 수학식 4와 같이 표현될 수 있다.here,

Is the noise signal,

Can be expressed as in Equation 4 below.

는 그룹 g의 전체 채널 행렬이고,

는 그룹 g의 빔포밍 행렬이다. 상기 수학식 4에서 근사 부등호는 아래 수학식 5와 같은 조건을 만족하는 경우에 해당한다.here,

Is the entire channel matrix of group g,

Is the beamforming matrix of group g. In Equation 4, the approximate inequality sign corresponds to a case where the conditions as in Equation 5 below are satisfied.

가 된다.then,

Becomes.

를 설정하고, 그러한 빔포밍 행렬을 갖는 단말들을 동시에 스케줄링하는 것이 본 발명에서 제안하는 MIMO 송수신 방법의 핵심 요소이다.
To satisfy the above conditions

And scheduling terminals having such a beamforming matrix at the same time is a key element of the MIMO transmission/reception method proposed in the present invention.

Next, the reduction in the dimension of the instantaneous channel matrix obtained by the MIMO transmission/reception method proposed in the present invention will be described. First, it is assumed that indexing is performed as shown in Equation 6 below to express the group index of K terminals.

1 is a conceptual diagram illustrating the concept of spatial separation between user groups in a MIMO transmission/reception method according to the present invention.

Referring to FIG. 1, a base station may include a large-scale antenna array 10 made of M antenna elements 10-1, 10-2, ..., 10-M. In addition, there are K active terminals 20-1, 20-2, …, 20-K, and the K active terminals may be classified into G groups. For example, the first group 30-1 may include the first terminal 20-1 and the second terminal 20-2, and the second group 30-2 is the third terminal 20-3. ) Can be included. The Gth group 20-G may include the K-1th terminal 20-(K-1) and the Kth terminal 20-K.

에 해당한다.
At this time, the channel matrix of the first group 30-1 consisting of the first terminal 20-1 and the second terminal 20-2 is

Corresponds to.

Next, the design of a code book according to the present invention will be described.

은 원-링 채널 모델로 만들어질 수 있다. 이를 위해 필요한 파라미터(parameter)는 그룹의 AS와 AoD(angle of departure)이다. 그룹의 AS는 셀 내 단말의 AS 분포에 의해 결정될 수 있고, 각 그룹 AoD는 다음과 같이 두 가지 기준을 종합적으로 판단하여 결정될 수 있다.The fixed codebook may consist of T classes and G groups belonging to each class. Here, each class may have a different number of groups. Matrix that makes up the eigenvector space of groups that make up a class

Can be built with a one-ring channel model. The necessary parameters for this are the AS of the group and the angle of departure (AoD). The AS of the group may be determined by the AS distribution of the UE in the cell, and each group AoD may be determined by comprehensively determining two criteria as follows.

-Make each group's unique vector space as orthogonal as possible. This is to reduce interference between groups.

-Make sure that the AoDs of each group are arranged as evenly as possible. This is to minimize the discrepancy between the eigen vector space of the terminal and the eigen vector space of the group to which the terminal belongs.

집합에 블록 대각화 기법을 적용하여 코드북을 이루는 빔포밍 행렬

를 생성할 수 있다. 이를 통해 그룹 간의 간섭을 보다 줄일 수 있다. 여기서, 빔포밍 행렬

는 긴 주기(혹은 와이드-밴드(wide-band)) PMI(precoding matrix indicator)로 사용되거나 짧은 주기(혹은 서브-밴드(sub-band)) PMI로 사용될 수 있다. 즉, 단말은 더 나은 성능을 얻고자하는 경우 자신에게 최적의

를 찾아서 짧은 주기로 피드백할 수 있다.
For each class acquired in the same way as above

Beamforming matrix that forms a codebook by applying block diagonalization to a set

Can be created. Through this, interference between groups can be further reduced. Here, the beamforming matrix

May be used as a long period (or wide-band) PMI (precoding matrix indicator) or a short period (or sub-band) PMI. In other words, if the terminal wants to obtain better performance,

You can find and give feedback in a short period.

Next, MU-MIMO scheduling according to the present invention will be described.

Conventional MU-MIMO scheduling was performed through a process of finding an optimal combination by obtaining a sum-PF (proportional fair) metric, for example, for a combination of terminals, but MU-MIMO scheduling according to the present invention It may be performed through a process of finding an optimal combination for a combination of groups for each class, not for individual terminals. At this time, each group can be viewed as a virtual sector.

First, the base station may determine the terminal with the largest PF metric in each group based on a multi-user-channel quality indicator (MU-CQI) as an example, and calculate all of the sum-PF metrics according to possible combinations of groups for each class. I can. For example, when the number of transmit antennas is 4, the sum-PF metric may be calculated as many as (4,1)+(4,2)+(4,3)+(4,4) cases. Here, (M,S) means the number of combinations of M to S. Based on this, the base station can perform dynamic scheduling for terminals 1 to 4 to select the most optimal group combination for each class. Finally, the base station may select a class having the largest sum-PF metric from among the T classes, and schedule the terminals belonging to the group combination selected from the corresponding class.

The MU-CQI feedback scheme enabling the MU-MIMO scheduling described above is as follows.

Each terminal may feed back a single user-channel quality indicator (SU-CQI), an interference signal strength between terminals within a group, and an interference signal strength between other groups to the base station. For example, when the number of transmit antennas is 4 and the number of groups is 4, since one beamforming vector is used per group, there is no interference between terminals within the group, and the SU-CQI is as Equation 7 below.

는 배경잡음과 다른 셀로부터의 간섭신호 세기이다. 단말이 속한 그룹을 제외한 3개의 다른 그룹의 간섭신호 세기는 아래 수학식 8과 같다.here,

Is the background noise and the strength of the interference signal from other cells. The interference signal strength of the three other groups excluding the group to which the terminal belongs is as shown in Equation 8 below.

Meanwhile, in order to reduce the number of feedback bits, the interference signal strength may be expressed as an offset from the SU-CQI.

The method of allowing the base station to calculate the MU-CQI by feeding back the strength of the interference signal as described above is not limited to the MU-MIMO method according to the present invention, but is also applied to the conventional LTE method. For example, even in the case of the Rel 8 codebook and the 8-Tx double codebook, the base station can accurately calculate the MU-CQI by receiving feedback of the SU-CQI for each terminal and the interference signal strength between all other terminals.

를 선택하였다면 나머지 15개의 PMI가 다른 단말과 조합이 될 수 있다. 따라서, 단말은 자신에게 간섭으로 작용할 15개의 PMI에 대한 간섭신호 세기(즉, 아래 수학식 9)와 상술한 SU-CQI를 피드백할 수 있고, 이를 기초로 기지국은 MU-CQI를 계산하여 최대 4명의 사용자를 동시에 스케줄링할 수 있다.In this case, the base station may reduce the number of feedback bits by reducing the number of interference combinations between terminals by placing restrictions on scheduling. For a detailed example of the Rel 8 codebook, the k-th terminal

If is selected, the remaining 15 PMIs can be combined with other terminals. Therefore, the UE can feed back the interference signal strengths (ie, Equation 9 below) and the above-described SU-CQI for 15 PMIs that will act as interference to itself, and based on this, the base station calculates the MU-CQI to a maximum of 4 You can schedule two users at the same time.

Here, the filtering process of the reception MMSE (minimum mean square error) of the terminal has been omitted. In this case, the MU-CQI is shown in Equation 10 below.

In the case of the Rel 8 codebook, the above-described interference signal may be composed of interference signals for 15 PMIs excluding the PMI selected by the terminal. In order to reduce the burden of feedback due to this, the terminal may feed back only some PMI interference signals among 15 PMIs. As an example, the terminal may select and feed back only a PMI whose absolute value of the PMI selected by itself and the value of performing the inner product among 15 PMIs is less than a specific threshold, and the base station may perform scheduling only with a combination thereof. I can. The threshold may be set to 0.3 for the Rel 8 codebook. In this case, the terminal may feed back only 5 to 9 interference signals out of 15 PMIs to the base station according to the PMI selected by the terminal. The method of reducing the feedback burden of the interference signal as described above is not limited to the Rel 8 codebook and can be applied to other codebooks.

를 피드백 받아야 한다. 혹은, 아래 수학식 11과 같이 변형된 간섭신호의 세기를 피드백 받은 경우, 기지국은 별도로

를 피드백 받을 필요 없다.In order to calculate the MU-CQI by using the above-described interference signal strength feedback, the base station uses the interference signal strength of another cell separately from the SU-CQI.

Should receive feedback. Alternatively, when the strength of the modified interference signal is fed back as in Equation 11 below, the base station is

There is no need to get feedback.

The advantage of the above-described terminal feedback is that the base station can estimate a more accurate MU-CQI by quantizing the modified interference signal strength and feeding back. A detailed description of this is as follows. The process of expressing MU-CQI as a fraction of SU-CQI may be expressed as Equation 12 below for convenience.

를 SU-CQI의 프랙션으로 표현하여 피드백하는 것이다.Here, S is a self-signal according to the PMI selected by the UE, I is an intercell interference signal, I _A is an interference signal between UEs due to PMI _A excluding the PMI selected by the UE, and α is MU-CQI of SU-CQI. It is a variable to express as a fraction. The UE feedback proposed in the present invention does not directly feed back the MU-CQI expressed as a fraction of SU-CQI, but as shown in Equation 13 below.

Is expressed as a fraction of SU-CQI and fed back.

와

를 기반으로 PMI_A, PMI_B로 인한 MU-CQI를 아래 수학식 14와 같이 추정할 수 있다.With such feedback, the base station can estimate MU-CQI for all combinations of 2 to 4 people. As an example, the base station receives from the terminal

Wow

MU-CQI due to PMI _A and PMI _B may be estimated as shown in Equation 14 below.

The difference between the above two feedback methods is as follows. Since the former α has a different dynamic range according to the SU-CQI of the terminal, it is difficult to quantize, and there is a problem in expressing an accurate MU-CQI through this. On the other hand, the latter β is easy to quantize because it has the same dynamic range regardless of the SU-CQI of the UE, and the base station can estimate more accurate MU-CQI with this feedback. Meanwhile, in the former case, since the dynamic range is affected by SU-CQI, a more accurate MU when the base station and the terminal share it by defining a mapping table of α suitable for a plurality of dynamic ranges according to SU-CQI. -CQI can be estimated. However, this method has the disadvantage of increasing the feedback burden.

는 MCS(modulation and coding scheme) 레벨로 표현되는 CQI 형태로 피드백될 수 있다. 이러한 간섭 PMI에 대한 CQI를 간섭 CQI라 한다. 이 경우, MUI(multi-user interference indicator) β와 동일한 효과를 얻기 위해 간섭 CQI는 종래 LTE가 제공하는 MCS 레벨 표의 범위 이하(즉, MCS 레벨의 범위를 벗어남)를 세분화하기 위해 낮은 MCS 레벨을 정의해야 한다. 이러한 추가 MCS 레벨의 필요성은 중간 혹은 낮은 MCS 레벨의 SU-CQI를 가지는 단말들의 MU-MIMO를 위해서이다. 즉, 중간 혹은 낮은 MCS 레벨의 단말들에 대한 간섭 CQI 레벨은 종래의 MCS 레벨의 범위 이하의 낮은 SINR(signal to interference plus noise ratio)를 가질 가능성이 매우 높다. 따라서, 정확한 간섭신호 세기를 피드백하기 위해서는 종래 MCS 레벨 범위 이하의 낮은 SINR를 표현할 수 있는 추가적인 낮은 MCS 레벨의 정의가 필요하다. 일례로, 종래 MCS 레벨 범위 밖의 -5dB SINR 이하를 통틀어 표현한다면, 추가적인 MCS 레벨은 -5 ~ -15dB의 SINR을 추가적인 몇 단계로 나누어 표현할 수 있어야 정확한 간섭신호 세기의 표현이 가능하다.Interference signal strength described above

May be fed back in the form of a CQI expressed at a modulation and coding scheme (MCS) level. The CQI for such an interfering PMI is called an interfering CQI. In this case, in order to obtain the same effect as the multi-user interference indicator (MUI) β, the interference CQI defines a low MCS level to subdivide the range of the MCS level table provided by the conventional LTE (ie, out of the range of the MCS level). Should be. The need for such an additional MCS level is for MU-MIMO of UEs having SU-CQI of a medium or low MCS level. That is, it is very likely that the interference CQI level for terminals of the medium or low MCS level has a low signal to interference plus noise ratio (SINR) below the range of the conventional MCS level. Accordingly, in order to feed back accurate interference signal strength, it is necessary to define an additional low MCS level capable of representing a low SINR below the conventional MCS level range. As an example, if a total of less than -5dB SINR outside the conventional MCS level range is expressed as a whole, the additional MCS level must be expressed by dividing the SINR of -5 ~ -15dB into several additional steps to accurately express the strength of the interference signal.

Meanwhile, the above-described interference signal strength feedback may be applied to a method of feeding back MU-CQI using an interference measurement resource (IMR). In this case, a plurality of IMRs for a plurality of MU hypotheses may be set for each subframe, and a plurality of IMRs may be set for each sub-band to reduce a channel state information (MU-CSI) feedback delay.

On the other hand, if the above-described terminal feedback is transmitted for all sub-bands, it helps to improve MU-MIMO performance, but increases the uplink feedback burden. To prevent this, if the strength of all interference signals in a specific sub-band exceeds that threshold by setting a specific threshold for the strength of the interference signal, the UE determines that the interference signal is too large to be suitable for MU-MIMO. The interference signal for the corresponding sub-band may not be fed back.

The above-described method of reducing the feedback burden is applied equally to the case of feeding back MU-CQI for each sub-band.

The SU-MIMO precoding matrix according to the present invention is composed of B _g and P _g . The precoding matrix having this structure approximates the SVD (singular value decomposition) precoding matrix by the relationship as shown in Equation 15 below.

That is, the SVD precoding matrix can be divided into a part of a very long period and a part of a short period. Since the beamforming matrix B having a very long period has almost no feedback burden, if the same feedback resources are assumed, the SU-MIMO precoding matrix according to the present invention can generate a more detailed short period precoding matrix P. Accordingly, a codebook of rank 2 or higher may be generated in the same manner as described above.

MU - MIMO Feedback for

The flexible MU-MIMO feedback scheme proposed in the present invention is characterized in that the UE feeds back the strength of the interference signal so that the base station can estimate the MU-CQI with a combination of various UEs without directly calculating and feeding back the MU-CQI. . Accordingly, each terminal can feed back SU-CQI, an interference signal strength between terminals, and an interference signal strength between different groups. Here, the strength of the interference signal is referred to as a multi-user interference indicator (MUI). For convenience, when the number of terminal antennas is 1, the SU-CQI of the transmission rank 1 of the terminal is shown in Equation 16 below.

는 배경잡음과 다른 셀로 인한 간섭신호의 세기이다. 한편, 단말이 선택한 PMI

와 MU-MIMO로 페어링(pairing)될 수 있는 PMI의 부분집합을 S라 하면, S에 속한 PMI

로 인한 MUI는 효율적인 피드백을 위해 아래 수학식 17과 같이 셀간 간섭신호의 세기와 비율로 정의될 수 있다.here,

Is the strength of the interference signal due to background noise and other cells. Meanwhile, the PMI selected by the terminal

If S is a subset of PMI that can be paired with and MU-MIMO, PMI belonging to S

The resulting MUI may be defined as the strength and ratio of the inter-cell interference signal as shown in Equation 17 below for efficient feedback.

을 양자화한 것이다. 이러한 양자화는 2비트로 할 수 있고, 서브-밴드 MUI의 피드백 부담을 줄이기 위해 와이드-밴드 MUI에 대한 1비트 오프셋으로 양자화할 수도 있다.In order to reduce the number of feedback bits, the MUI can be fed back by being expressed in a ratio with the SU-CQI. In other words, the MUI information actually fed back

Is quantized. This quantization can be performed in 2 bits, and in order to reduce the feedback burden of the sub-band MUI, quantization can be performed with a 1-bit offset for the wide-band MUI.

로 인한 MU-CQI는 아래 수학식 18과 같이 계산될 수 있다.A new PUSCH (physical uplink shared channel) feedback mode proposed in the present invention is to feed back SU-CQI and a plurality of MUIs. The base station can estimate the MU-CQI through this feedback. For example, in the case of allocating the same transmission power to two terminals, PMI

The MU-CQI caused by may be calculated as in Equation 18 below.

로 인한 MU-CQI는 아래 수학식 19와 같다.When pairing three terminals, PMI

The MU-CQI caused by is shown in Equation 19 below.

The same applies to the case of pairing four terminals.

Meanwhile, a terminal having a similar SU-CQI for two PMIs orthogonal in a PMI set of rank 1, and a similar MUI for different PMIs instead of a higher rank PMI defined in the conventional LTE codebook, only the SU-CQI and MUI of rank 1 Rank 2 transmission is possible through. The symmetry for rank 2 may be expressed as Equation 20 below.

In this case, the MMSE received vectors for PMIs 1 and 2 are as shown in Equation 21 below.

That is, one of the three PMIs orthogonal to the rank 1 PMI of the terminal may be the second PMI of rank 2. Information on the second PMI in which the UE has this syntax (for example, 2-bit information that identifies one of three PMIs orthogonal to the rank 1 PMI) and a rank indicator (RI) (here, two PMIs for RI are detailed In the case of feeding back information on (indicating that it has one symmetry relationship) to the base station in a wide-band or sub-band mode, the base station may determine this and transmit it to rank 2. Here, the MUI may be expressed as four values as in Equation 22 below by the MMSE reception vector for the remaining two orthogonal PMIs (w ₃ ,w ₄ ) excluding the second PMI.

The terminal may feed back the above-described MUI in a wide-band or sub-band mode. In this case, the two SU-CQIs for rank 2 transmission are the same, and in the case of calculating like the above-described MU-CQI, the SU-CQI is as shown in Equation 23 below.

In addition, the above-described feedback scheme enables very flexible MU-MIMO pairing in which each UE to be paired may have rank 1 or rank 2. For example, two terminals of rank 2 may be paired.

The above-described MUI feedback method has the following characteristics.

1) Transparent rank 2 transmission and pairing of two rank 2 terminals with rank 1 feedback.

Even when the terminal feedback is rank 1, the base station can calculate the MU-CQI of rank 2 transmission through the MUI, and thus can transmit the rank 2. Therefore, the base station can perform rank 2 transmission by adding the MUI feedback without the conventional rank 2 feedback. In addition, in the same manner, the base station can schedule very flexibly with only rank 1 feedback information. For example, the base station may pair two rank 2 terminals or one rank 2 terminal and a combination of two rank 1 terminals.

2) DM-RS (demodulation reference signal) resource reduction

Since the base station can know the exact interference between terminals for a specific terminal pairing through the MUI, terminals with very few MUIs can be configured to share the same DM-RS resource without a separate control signal. If necessary, the base station may use different scrambling identification (SCID).

Meanwhile, another alternative to the MUI feedback scheme for rank 2 MU-MIMO is as follows. It will be described based on the LTE Rel-8 4-Tx codebook and the number of terminal reception antennas 2, but the feedback scheme according to the present invention is not limited to a specific antenna shape or codebook.

The SU-CQI of rank 2 may be defined as in Equation 24 below for layers 1 and 2, respectively.

는 단말 K가 선택한 랭크 2에 대한 PMI를 나타내고, g_{k (1)}, g_{k (2)}는 레이어 1, 2에 대한 단말의 수신필터 계수를 나타낸다. 본 발명에 따른 피드백 방식을 설명하기 위해, 상기 SU-CQI와 달리 레이어 간의 간섭이 빠진 형태의 CQI를 SL(single layer)-CQI라 하고 아래 수학식 25과 같이 정의할 수 있다.here,

Denotes the PMI for rank 2 selected by UE K, and g _{k (1)} and g _{k (2)} denotes the reception filter coefficients of the UE for layers 1 and 2. In order to describe the feedback method according to the present invention, unlike the SU-CQI, a CQI in which inter-layer interference is omitted is referred to as SL (single layer)-CQI, and may be defined as Equation 25 below.

로 인한 레이어 1에 대한 MUI는 아래 수학식 26과 같이 재정의될 수 있다.Also, PMI

The MUI for layer 1 due to may be redefined as in Equation 26 below.

Using the above-defined feedback, it is possible to flexibly support rank 1/rank 2 MU-MIMO as follows. For convenience, it is assumed that all terminals have rank 2 and perform rank 2 CSI feedback. First, the MU-CQI for layer 1 of UE k due to pairing of two UEs having rank 2 is shown in Equation 27 below.

는 단말 k와 페어링된 다른 단말이 선택한 랭크 2 PMI를 나타낸다. 또한, 단말이 랭크 2로 피드백한 경우에도 시스템의 스케줄링 관점에서 일부 단말들에게 랭크 1로 전송하는 것이 시스템 전송 효율 측면에서 좋을 수 있다. 이 경우, 상기 피드백에 의해 정확한 MU-CQI를 통해 안정적인 링크 적응이 가능하다. 일례로, 각

의 PMI가 선택된 세 단말의 페어링으로 인한 단말 k의 MU-CQI는 아래 수학식 8과 같다.here,

Denotes a rank 2 PMI selected by another terminal paired with terminal k. In addition, even when the terminal feeds back to rank 2, transmission of rank 1 to some terminals may be good in terms of system transmission efficiency from the viewpoint of system scheduling. In this case, stable link adaptation is possible through accurate MU-CQI based on the feedback. For example, each

The MU-CQI of UE k due to the pairing of the three UEs in which the PMI is selected is shown in Equation 8 below.

On the other hand, the MU-CQI feedback, so that when g _{k (2)} and modulation when g _{k (2)} may be different may be some errors. However, since the error is underestimate, the effect on system performance will be minimal.

A problem with the above-described MUI feedback scheme is that if the MUI is fed back in units of sub-bands, the uplink resource burden may be large. One way to reduce the feedback burden is to add 1-bit information that enables/disables MUI for each sub-band and layer. This additional information is called MUI_Enb. The condition for disabling a specific sub-band MUI feedback of a specific layer for a specific terminal is when any (or all) candidate PMIs belonging to S of a corresponding layer for the corresponding terminal indicate interference exceeding a specific threshold. These thresholds can be adjusted with a system control signal (higher-layer signaling) or can be defined in the standard of the terminal. As an example, when the set S is defined as orthogonal PMIs, as shown in Table 1 below, the pairing candidate PMIs that are orthogonal may be limited to two. Table 1 shows the MU-PMI orthogonal to the PMI of rank 2.

Here, candidate PMIs that are orthogonal are phase shifted and include equivalent PMIs. For example, since w _{11 (2)} = -w _{9 (1)} , the two PMIs are equal, and w _{9 (1)} = w ₉ . The scheduler should consider all of these equivalent PMIs to find an optimal combination of terminal and PMI. The MUI_Enb may be set differently for each sub-band and for each MU-PMI (instead of a layer).

Another effect of this MUI_Enb is that when MUI is quantized into 2 bits, it indicates whether MUI_Enb is out of the range of MUI, and thus more accurate quantization can be performed from 3 to 4 steps. Alternatively, in the case of quantization with 1 bit, since two-step quantization is actually performed, 1-bit quantization can also be used without significant performance degradation. Meanwhile, when MUI_Enb is disabled, the SU-CQI for conventional rank 2 should be fed back as the sub-band CQI of the terminal instead of the aforementioned SL-CQI.

Another way to reduce the burden of feedback is the delayed MUI feedback method as follows. First, the base station may receive a conventional SU-CSI (especially, PMI) from each terminal to determine which interfering PMI can actually be used by the terminal to which it can be paired. Alternatively, the base station may instruct to select some interfering PMI and feed back to the terminal. Accordingly, the base station can inform each terminal through the downlink control information about which of the interfering PMIs to perform feedback on. Through this, each terminal can reduce uplink resources by feeding back only the MUI for an interfering PMI that can be actually used by other terminals. This technique can be applied equally to the CSI-RS-based MU-CQI.

Meanwhile, the MUI feedback scheme proposed in the present invention can be applied to both CSI-RS-based and CSI-IM-based MU-CQI schemes as a scheme to replace MU-CQI.

1) 2D array PMI

Rel. In 11, in order to adjust frequency-side granularity and feedback overhead, there are a plurality of feedback modes differently transmitted to the wide-band/sub-band PMI. It is necessary to extend this concept to a 2D antenna array and divide the PMI feedback mode into horizontal axis wide-array/sub-array PMI and vertical axis wide-array/sub-array PMI. Here, the sub-array PMI means an optimal PMI for each sub-array, and the wide-array PMI means a PMI that maximizes the sum of CQIs of all sub-arrays.

3D Beam forming

Long-period CSI-RS structure: Even when the base station has a plurality of arrays on the horizontal axis/vertical axis due to the 2D antenna array, it can be considered that the eigenvector matrix, which is the statistical channel characteristic of the terminal, exists one on the horizontal axis and one on the vertical axis. have. Therefore, the long-period CSI-RS does not need to exist for every 2D antenna element, and only one row for the horizontal axis and one column for the vertical axis are used for transmission.

Short-period CSI-RS structure: The eigenvector matrix of the terminal has the above-described structure, but a plurality of arrays may be different from each other in a horizontal axis/vertical axis in a short periodic fading channel. Therefore, the short-period CSI-RS needs to be transmitted for every 2D antenna element.

Long cycle PMI feedback: Long cycle PMI can be composed of one long cycle PMI (horizontal axis class and group ID) on the horizontal axis and one long cycle PMI (vertical axis class and group ID) on the vertical axis according to the long cycle CSI-RS. .

Short period PMI feedback: Since the short period PMI may be different for all 2D antenna elements, a plurality of short period PMIs may be fed back along the horizontal axis and a plurality of short period PMIs may be fed back along the vertical axis. In addition, the feedback period of the horizontal axis short period PMI and the vertical axis short period PMI may be different.

Next, an embodiment according to the present invention applied to the following system environment will be described.

System environment

1) Antenna configuration ( configuration )

In order to implement large-scale antenna MIMO, the number of antennas of the base station considered in the present invention is 32 co-polarization ULA, and the antenna spacing is 1/2 λ. The antenna shape is assumed to be 8(H)×4(V) in case of a two-dimensional array. Here, H means the horizontal axis in 2D and V means the vertical axis. In addition, the number of antennas of the terminal is limited to two.

2) JSDM Design variables

JSDM design variables can be defined as follows.

-T _H , T _V : Number of horizontal/vertical axis classes

-G _H , G _V : Number of horizontal/vertical axis groups

-b _H , b _V : Number of beams per horizontal/vertical axis group

The class represents a combination of terminals that can be scheduled with the same physical resource block (PRB). That is, different classes use different resources. JSDM variables used in the present invention may be defined as follows in the case of a two-dimensional array.

-T _H = 2, T _V = 2

-G _H = 4, G _V = 2

-b _H = 2, b _V = 1

In the case of a one-dimensional array, it can be defined as follows.

-T _H = 4

-G _H = 4

-b _H = 4

3) Notation ( notations )

-t _H = 0, 1,… , T _H -1; t _V = 0, 1,… , T _V -1: horizontal/vertical axis class index

-g _H = 0, 1,… , G _H -1; g _V = 0, 1,… , G _V -1: horizontal/vertical group index

-k = 0, 1,… , b _H -1: horizontal beam index

-g _k : beam k in horizontal axis group g

Reference signal ( reference signal , RS )

The RS according to the present invention must accommodate 32 transmit antennas and maintain the RS overhead at a conventional level according to the above-described system environment. In addition, the base station may limit the number of layers scheduling with the same resource to a maximum of 16 and support a maximum of 2 layers per terminal. The RS design method in which the number of RE (resource elements) of CSI-RS is set to 8 and the number of DM-RS ports is set to 4 is described below.

One) CSI Group-specific reference signal ( CSI - GRS )

2 is a conceptual diagram illustrating a CSI-RS pattern of 8 antenna ports.

2, Rel. The CSI-RS of 11 may be designed to accommodate five 8-Tx CSI-RS antenna ports in one PRB. Meanwhile, when CSI-RSs of 32 antenna ports are accommodated in one PRB, problems such as power boosting and increased CSI-RS interference between adjacent cells may occur. Therefore, in the present invention, it is possible to replace the conventional CSI-RS with the proposed CSI-GRS.

In CSI-GRS, RSs beamformed with a plurality of beam vectors orthogonal to each other between groups may be transmitted by sharing the same one RE. According to the aforementioned JSDM design variable, the total number of horizontal/vertical groups is 8, so that 8 orthogonal beamformed RSs can occupy one RE. Therefore, the eight antenna port patterns can be maintained the same as in the conventional FIG. 2.

In order that CSI-GRS between adjacent cells does not cause the worst interference that may occur due to the direction of the beam, antenna port and group ID adjustment between adjacent cells (refer to the'Port and Group Mapping' section below) is applied to CSI- Cooperative beamforming may be performed so that the GRS beams deviate from each other.

Meanwhile, in the CSI-GRS, since it may be difficult to distinguish multiple orthogonal precoded CSI-RSs, classification of multiple orthogonal precoded CSI-RSs can be easily configured using a large number of resources. At this time, the CSI-GRS may be transmitted using the same RE resource as the adjacent cell, and cooperative beamforming is performed so that the CSI-GRS beams between adjacent cells deviate from each other through inter-cell cooperation in order to mitigate CSI-GRS interference between adjacent cells. Will have to do.

Ports and groups ID Mapping

The antenna port p may be mapped to the horizontal/vertical axis class index and the beam index as shown in Equation 29 below.

의 값을 가지고, 아래 표 2와 같이 수평/수직축 그룹 인덱스로 매핑될 수 있다.Meanwhile, the group ID is

With the value of, it can be mapped to the horizontal/vertical group index as shown in Table 2 below.

sequence ( sequence ) produce

는 아래 수학식 30과 같다.The pseudo-random sequence of CSI-GRS may be redefined as follows. I.e. the reference signal sequence

Is as in Equation 30 below.

Here, n _s is a slot number in a radio frame, and l is an OFDM symbol number in the slot. The pseudo-random sequence generator may be initialized with Equation 31 below at the start of each OFDM symbol.

Here, N _CP may be defined as in Equation 32 below.

2) DM - RS

The DM-RS according to the present invention allows one DM-RS port to accommodate a maximum of 4 layers through 4 different SCIDs, and a maximum of 4 DM-RS ports to transmit a total of 16 layers. Four orthogonal DM-RS patterns follow the rank 4 DM-RS pattern and an orthogonal cover code (OCC). In this case, the layer scheduled with the same port may be determined based on the terminal feedback information as described above.

sequence produce

를 위해, 참조 신호 시퀀스 r(m)은 아래 수학식 33에 의해 정의될 수 있다.The pseudo-random sequence of the DM-RS may be defined as follows. Ie what antenna port

For, the reference signal sequence r(m) may be defined by Equation 33 below.

The pseudo-random sequence generator may be initialized with Equation 34 below at the beginning of each subframe.

양은

의해 주어질 수 있다.

German silver

Can be given by

In the case of DCI format 2C-1, n _SCID may be given by Table 3 below.

For spatial multiplexing Precoding

The precoding matrix codebook for spatial multiplexing may consist of an 8-DFT-based horizontal axis (H) codebook and a 4-DFT-based vertical axis (V) codebook, and the horizontal axis codebook follows an 8 Tx double codebook structure.

One) Rank 1 Codebook for transmission

The horizontal axis codebook of rank 1 may be defined as in Equation 35 below.

Where t = 0, 1,… , T _H -1, g = 0, 1,… , G _H -1, k = 0, 1,… , b _H -1.

The DFT beam vector v _{H ,m} may be defined as in Equation 36 below.

Where m = 0, 1,… , 15.

The vertical axis codebook of rank 1 may be defined as in Equation 37 below.

Where t = 0, 1,… , T _V -1, g = 0, 1,… , G _V -1.

The DFT beam vector v _{V ,m} may be defined as in Equation 38 below.

Where m = 0, 1,… , Is 3.

2) Rank 2 Codebook for transmission

The codebook for transmitting rank 2 (two layers per terminal) according to the above-described CQI feedback scheme reuses the rank 1 codebook, so a separate rank 2 codebook is not defined.

Downlink Control information ( downlink control information , DCI )

In order to support up to 16 layers, 16 DM-RS ports are actually required in the conventional method. In order to realize the DM-RS proposed in the present invention, the base station can know which terminal or layer interference is very small through the above-described MUI. Accordingly, the base station can demodulate the DM-RS even if the same port and different SCIDs are allocated to one or more terminals or layers without an additional control signal, thereby reducing the DM-RS resource burden.

One) DCI Format 2C-1

Table 3 shows the difference between DCI format 2C-1 supporting MU-MIMO for up to 16 layers and conventional DCI format 2C. The antenna port, n _SCID , and the number of layers may be represented by 4 bits. Table 3 shows the antenna ports and SCID.

2) PMI Justice

에 해당하고 두 개의 PMI 인덱스로 표현될 수 있다. 첫 번째 PMI 인덱스 i₁는

와 같다. 두 번째 PMI 인덱스 i₂는

와 같다. 랭크 2인 경우, 두 번째 PMI i₂는 피드백되지 않는다.PMI

It corresponds to and can be represented by two PMI indexes. The first PMI index i ₁ is

Same as The second PMI index i ₂ is

Same as In case of rank 2, the second PMI i ₂ is not fed back.

3) MUI Justice

The MUI is information reporting an interference level due to a co-scheduled PMI (Co-PMI) corresponding to the PMI selected by the terminal to the network. The targets of Co-PMI are all precoding matrices belonging to the PMI class (t _H ,t _V ).

Table 4 below shows the Co-PMI set for the PMI (t _H ,t _V ,g _H ,g _V ,k) selected by the UE. For Co-PMI, the horizontal/vertical group index has the same value as (t _H ,t _V ) of PMI, so it can be omitted, and corresponds to all groups except for the horizontal/vertical group (g _H ,g _V ) of PMI. Since the Co-PMI index 1 to 7 includes both beams belonging to the corresponding group, the index k may be omitted.

이고,

이다. G_H는 모든 수평 그룹의 집합에서 g_H를 제외한 부분 집합으로 아래 수학식 39와 같이 정의될 수 있다.here,

ego,

to be. G _H is a subset of all horizontal groups excluding g _H and may be defined as in Equation 39 below.

For MUI calculation, a virtual extended CQI as shown in Table 5 below may be defined. These CQI values are not actually used for transmission.

The MUI may be set as follows according to the CQI offset and Table 6 below.

-A CQI value is calculated for each PMI(s) belonging to the Co-PMI set, and this is defined as a virtual CQI. Here, since the Co-PMI indexes 1 to 7 include both beams (or PMIs) belonging to the corresponding group, two interference levels due to the two beams are calculated as an average to form one virtual CQI.

-The CQI offset represents the difference between the CQI according to the PMI and the virtual CQI according to the Co-PMI.

CQI offset = CQI-virtual CQI

-The CQI offset may be mapped to the MUI index as shown in Table 6 below.

-The terminal can calculate 8 MUIs corresponding to the Co-PMI set.

4) MU - CQI Estimation method

The flexible MU-CSI feedback scheme proposed in the present invention is characterized in that the UE feeds back the strength of the interference signal so that the base station can estimate the MU-CQI with a combination of various UEs without directly calculating and feeding back the MU-CQI. . Accordingly, each terminal can feed back CQI (ie, SU-CQI), an interference signal strength between terminals, and an interference signal strength between different groups. Here, the strength of the interference signal is defined as MUI. In the present invention, it is assumed that the number of terminal antennas is 2.

Rank 1 case

에 의해 결정될 수 있다.The CQI of transmission rank 1 of UE k is SINR of Equation 40 below.

Can be determined by

는 2×4 채널 행렬이고,

는 수신 빔포밍 벡터이고,

는 배경잡음과 다른 셀로 인한 간섭신호의 세기이다. 한편, 단말이 선택한 PMI

와 표 4의 MU-MIMO로 페어링될 수 있는 Co-PMI 집합을 S라고 하면, S에 속한 PMI

로 인한 간섭신호의 세기는 표 5를 이용하여 아래 수학식 41과 같은 가상 CQI로 표현될 수 있다.here,

Is a 2x4 channel matrix,

Is the received beamforming vector,

Is the strength of the interference signal due to background noise and other cells. Meanwhile, the PMI selected by the terminal

If the set of Co-PMIs that can be paired with MU-MIMO in Table 4 is S, PMI belonging to S

The intensity of the interference signal caused by may be expressed as a virtual CQI as shown in Equation 41 below using Table 5.

In order to reduce the number of feedback bits, the MUI may be expressed as an offset from the CQI as shown in Equation 42 below, and may be mapped to the MUI index through Table 6.

The MUI index defined in Table 6 can be interpreted as a power offset value compared to the effective SNR of the CQI due to the selected PMI as shown in Table 7 below.

로 인한 MU-CQI는 아래 수학식 43을 통해 계산될 수 있다.A new PUSCH feedback mode proposed in the present invention is to feed back the CQI and a plurality of MUIs. The base station can estimate the MU-CQI through this feedback. For example, when allocating the same transmission power to two terminals, Co-PMI

The MU-CQI caused by may be calculated through Equation 43 below.

로 인한 MU-CQI는 아래 수학식 44와 같다.When pairing three terminals, Co-PMI

The MU-CQI caused by is shown in Equation 44 below.

The same can be applied to the case of pairing up to 16 terminals.

The network can estimate each SINR as shown in Equation 45 below through CQI and MUI feedback, and map it to MU-CQI to perform precise MU-MIMO link adaptation and MU-MIMO scheduling.

Rank 2 case

는 제1 PMI에 속한 두 개의 제2 PMI에 해당하는 프리코딩 벡터들로 구성될 수 있다. 두 개의 프리코딩 벡터로 인한 레이어 간의 간섭을 고려하여 PMI

를 갖는 1번 레이어(혹은 코드워드)에 대한 종래 CQI는 아래 수학식 46과 같이 정의될 수 있다.Meanwhile, MU-CQI estimation for MU-MIMO operation including rank 2 follows the following scheme. PMI for rank 2 transmission according to the above-described PMI definition (i.e., a codebook for rank 2 transmission in precoding for spatial multiplexing)

May be composed of precoding vectors corresponding to two second PMIs belonging to the first PMI. PMI considering interference between layers due to two precoding vectors

The conventional CQI for layer 1 (or codeword) having a may be defined as in Equation 46 below.

은 1번 레이어에 대한 수신 빔포밍 필터 계수이다. 본 발명은 상기 CQI와 다르게 랭크 2의 경우 2번 레이어의 PMI

로 인한 간섭을 배제한 SINR을 아래 수학식 47과 같이 SL-CQI로 매핑하여 네트워크에 보고할 수 있다.here,

Is a reception beamforming filter coefficient for layer 1. Unlike the CQI, in the present invention, in the case of rank 2, the PMI of the second layer

The SINR excluding interference caused by may be mapped to SL-CQI as shown in Equation 47 below and reported to the network.

로 인한 MUI를 통해 네트워크는 아래 수학식 48과 같이 랭크 2

를 갖는 두 단말을 MU 페어링하는 경우 단말 k의 1번 레이어에 대한 MU-CQI를 추정할 수 있다.And the SL-CQI

Through the MUI caused by, the network is ranked 2 as shown in Equation 48 below.

In case of MU pairing two UEs having a, MU-CQI for layer 1 of UE k may be estimated.

에 대한 MUI는 상술한

으로 인한 그룹 간 간섭의 평균치이므로 아래 수학식 49와 같은 관계가 정의될 수 있다.Here, the Co-PMI of a terminal belonging to another group

For the MUI described above

Since it is an average value of interference between groups due to, the relationship as in Equation 49 below may be defined.

의

는 서로 유사한 방향성을 가지므로 유사한 간섭 레벨을 보이기 때문이다.The reason for expressing the MUI of interference from other groups as an average value is the Co-PMI of the interference group.

of

This is because they have similar directions and thus show similar interference levels.

As described above, through CQI (ie, SL-CQI) and MUI, the network can estimate MU-CQI of very various combinations in which ranks 1 and 2 are mixed up to 16 layers.

On the other hand, the MU-MIMO operation for the cross-polarization (X-pol) antenna array shows a more complex form due to the co-phasing element of the codebook, but the above-described cross-polarization The principle of operation can be extended to X-pol and applied. For example, if the same-parsing element of QPSK is classified into four classes, orthogonality of a precoding matrix within each class can be guaranteed.

5) Feedback mode

In the present invention, the codebook and CSI feedback are designed to be suitable for the MU-MIMO scheme. On the other hand, when the number of terminals in a cell is very small, SU-MIMO may be more advantageous in terms of frequency efficiency. Therefore, it is possible to define a codebook having a higher granularity and set a new feedback mode to be suitable for SU-MIMO. The base station may change the feedback mode through higher-layer signaling so that the UE provides feedback through SU-MIMO or MU-MIMO. In this case, if the codebook for MU-MIMO is a subset of the SU-MIMO codebook, the MU-MIMO feedback can be regarded as sub-sampling of the SU-MIMO codebook.

Scheduling

의 모든 조합에 대해 합계-PF 메트릭을 계산해야 하므로 계산 복잡도가 매우 증가한다.Conventional MU-MIMO scheduling represents a computational complexity that a network can handle when the number of terminals per cell is assumed to be 10 based on a brute-force algorithm. However, the number of terminals per cell in a massive MIMO system has a value greater than 10. As an example, if the number of terminals per cell is assumed to be 30 and the maximum number of simultaneously transmittable layers is assumed to be 16,

The computational complexity is greatly increased because the sum-PF metric must be calculated for all combinations of.

Scheduling for the massive MIMO scheme proposed in the present invention consists of two steps. Since the network first performs intra-class scheduling for each class and then inter-class scheduling between classes, computational complexity due to scheduling can be greatly reduced.

3 is a flowchart illustrating a MIMO transmission method according to an embodiment of the present invention, and FIG. 4 is a flowchart illustrating a scheduling process in a MIMO transmission method according to an embodiment of the present invention.

Referring to FIGS. 3 and 4, the base station may receive feedback from at least one terminal or measure channel information through an uplink sounding reference signal (SRS) (S300). The channel information includes at least one long period PMI selected from a fixed codebook indicating a transmit correlation matrix, an eigenvalue of a transmission correlation matrix, an eigenvector of a transmission correlation matrix, AS, AOD, and channel information. It may include at least one of.

That is, the base station may configure the CSI-RS and transmit it to the UEs, and obtain channel information through feedback of the measurement result from the UEs through the CSI-RS. Alternatively, the base station may obtain channel information by directly measuring through the uplink SRS transmitted by the terminal.

The base station may classify at least one terminal into at least one class and at least one group dependent on the class based on the channel information (S310). Specifically, the base station may generate a plurality of groups by classifying UEs having a similar effective eigenvector (eigenvector corresponding to a valid eigenvalue) matrix of the UE into one group. In addition, the base station may generate one class by grouping groups having high orthogonality between eigen vectors. Different time resources and frequency resources may be allocated to each class classified as described above, and the same time resources and frequency resources may be allocated to groups within one class.

The base station may determine a group beamforming matrix for each group based on the channel information (S320). The base station may generate a beamforming matrix of a corresponding group based on the entire set or subset of column vectors of the group eigenvector matrix selected through terminal classification. Here, the base station may determine a group beamforming matrix such that group beamforming matrices for each group are similar or orthogonal to each other based on the channel information and the one-ring channel model.

The base station may perform group beamforming transmission based on the group beamforming matrix to terminals belonging to the group for each group (S330). In addition, the base station may receive feedback from the UEs the SU-CQI information and interference signal information measured from a CSI-RS signal to which group beamforming is applied or a CSI-RS to which group beamforming is not applied (S340).

The interference signal information may include at least one of an interference signal strength between terminals belonging to a group and an interference signal strength with another group. Here, the interference signal information is expressed as an offset value for SU-CQI, a CQI defined as an MCS level, or as an interference signal strength for a PMI that will act as interference to a terminal belonging to a group, or for SU-CQI. It can be expressed as a ratio.

Meanwhile, when the terminal feeds back the interference signal information to the base station, the terminal may feed back interference signal information including a PMI whose inner product is smaller than a preset value among PMIs that will act as interference to itself. In addition, the terminal may feed back interference signal information for a subband in which the strength of the interference signal is smaller than a preset value among the subbands.

라 하면,

개의 그룹 조합이 생성됨), 각 그룹별로 가능한 모든 단말의 조합, 각 단말별 랭크의 조합을 기반으로 각 단말의 MU-CQI(s)를 계산할 수 있다(S351). The base station may schedule terminals based on the SU-CQI information and the interference signal information (S350). First, the base station can check the non-empty H/V group based on the PMI of the terminal. Thereafter, the base station can perform intra-class scheduling for each class (t _H ,t _V ). That is, the base station determines all possible combinations of non-empty H/V groups belonging to one class (that is, the number of non-empty H/V groups).

If you say,

Group combinations are generated), the MU-CQI(s) of each terminal may be calculated based on a combination of all possible terminals for each group, and a rank combination for each terminal (S351).

The base station can calculate the sum-PF metric based on the MU-CQI (S352), and select a multi-user combination (or MU pair) with the largest sum-PF metric in each class as the sum-PF metric of the class. Can be (S353).

Next, the base station may perform inter-class scheduling. That is, the base station may select a class having the largest sum-PF metric among all classes (S354).

The base station may determine a multi-user combination having the least interference with each other in the selected class, and schedule the determined multi-user combination (S355). That is, the base station can check inter-layer interference through the MUI, and can allocate the same DM-RS port to layers having very few MUIs. At this time, these livers may be classified by SCID.

On the other hand, when the number of sets of layers (or terminals) with very little interference between each other exceeds 4, the base station may drop the corresponding layers in the order of small sum-PF metric and maintain the number of DM-RS ports as four. The base station can schedule the finally determined multi-user combination.

Finally, the base station may transmit data to the terminals based on the scheduling (S360). In this case, the base station may transmit data to the terminals through a rank different from the rank related to the SU-CQI information and the interference signal information. For example, even when the base station receives feedback of SU-CQI information and interference signal information through rank 1, the base station may transmit data to the terminal through rank 2 based on this.

Although described with reference to the above embodiments, those skilled in the art will understand that various modifications and changes can be made to the present invention without departing from the spirit and scope of the present invention described in the following claims. I will be able to.

Claims (25)

As a method for transmitting multi-input multi-output (MIMO) of a base station in a wireless communication system,
Obtaining channel information for at least one terminal;
Classifying the at least one terminal into at least one class and at least one group dependent on the class based on the channel information;
Determining a group beamforming matrix for each classified group;
Performing group beamforming transmission based on the group beamforming matrix to terminals belonging to each group;
Obtaining single user-channel quality indicator (SU-CQI) information and interference signal information for terminals belonging to each group;
Scheduling terminals based on the SU-CQI information and the interference signal information; And
Including the step of transmitting data to the terminals based on the scheduling,
The interference signal information includes at least one of an interference signal strength between terminals belonging to the group and an interference signal strength with another group. The method according to claim 1,
The channel information,
Selected from a fixed codebook that means transmit correlation matrix, eigenvalue of transmission correlation matrix, eigenvector of transmission correlation matrix, AS (Angle Spread), AOD (Angle of Departure), and channel information. MIMO transmission method comprising at least one of at least one long period PMI (Precoding Matrix Indicator). delete delete The method according to claim 1,
The step of determining the group beamforming matrix,
MIMO, characterized in that, based on the channel information and a one-ring channel model, the group beamforming matrix is determined through block diagonalization so that group beamforming matrices for each group are similarly orthogonal to each other. Transmission method. The method according to claim 1,
The step of performing the group beamforming transmission to the terminals belonging to each group,
A MIMO transmission method comprising transmitting group-specific reference signals beamformed into a plurality of beam vectors orthogonal to each other between the groups through one resource element. delete delete delete The method according to claim 1,
The interference signal information,
MIMO transmission method, characterized in that the offset (offset) value for the SU-CQI. The method according to claim 1,
The interference signal information,
MIMO transmission method, characterized in that the CQI expressed in the MCS (modulation and coding scheme) level. The method according to claim 1,
The interference signal information,
MIMO transmission method, characterized in that the interference signal strength for a precoding matrix indicator (PMI) to act as interference to the terminal belonging to the group. The method according to claim 1,
The interference signal information,
MIMO transmission method, characterized in that the ratio to the SU-CQI. delete The method according to claim 1,
The step of scheduling the terminals,
A MIMO transmission method, characterized in that scheduling a plurality of terminals having less interference signal information with each other than a preset reference to share the same demodulated reference signal resource. delete The method according to claim 1,
The step of transmitting to the terminals,
And transmitting data to UEs through a rank different from a rank related to the SU-CQI information and the interference signal information. As a method for receiving MIMO of a terminal in a wireless communication system,
Receiving a signal to which a group beamforming matrix for a group including the terminal is applied;
Generating single user-channel quality indicator (SU-CQI) information and interference signal information using a reference signal to which the group beamforming matrix is applied or a reference signal to which the group beamforming matrix is not applied step; And
Including the step of feedback (feedback) the SU-CQI information and the interference signal information to a base station,
The interference signal information includes at least one of an interference signal strength between terminals belonging to the group and an interference signal strength with another group. delete delete delete delete delete delete delete

Images