WO2008114158A1

WO2008114158A1 - Method of precoding in a multi-antenna system

Info

Publication number: WO2008114158A1
Application number: PCT/IB2008/050809
Authority: WO
Inventors: Gang Wu; Xun Fan; Ni Ma; Xiaobo Zhang; Qi Zhou
Original assignee: Koninklijke Philips Electronics N.V.
Priority date: 2007-03-19
Filing date: 2008-03-06
Publication date: 2008-09-25
Also published as: TW200840256A

Abstract

The present invention provides a method of precoding in a multi-antenna system. The method generates precoding codebooks with extension Fourier series, based on the number of transmit antennas in the system and the number of groups of user data streams, so as to obtain a larger range for selection of precoding codebooks, which is helpful to find out which is the most suitable codebook.

Description

METHOD OF PRECODING IN A MULTI-ANTENNA SYSTEM

TECHNICAL FIELD

The present invention relates to a radio communication technology and, more particularly, to a method of precoding in a multi-antenna system.

BACKGROUND ART

Compared with a SISO (Single Input Single Output) wireless communication system, a MIMO (Multiple Input Multiple Output) wireless communication system can provide higher channel capacity. For a multi-user MIMO system, precoding technology is technology for transmitter optimization, which could yield higher system performance and lower complexity in the receiving process. With precoding, a transmitter could pre-correct transmitted signals in order to mitigate some effects in a wireless broadcast channel, including inter-symbol interference, fading, channel correlation etc.

For precoding, the ideal case would be to transmit full MIMO channel status information back to a transmitter. However, this is almost impossible to achieve due to heavy loads involved in feedback transmissions. Therefore, codebooks are considered to express feedback information in limited transmissions for feedback information, i.e., a certain number of codebooks are generated when precoding and a receiver selects a suitable codebook to transmit back to the transmitter. Accordingly, how to optimize the design of codebooks in the phase of precoding becomes a key problem. In the prior art precoding schemes, the scheme of PU²RC (Per-User Unitary Rate Control) is a dominant scheme. Assuming that Node B of a MIMO system has M transmit antennas, the corresponding PU²RC system may be as shown in Figure 1. In Figure 1, a group of precoding matrixes is E = {E⁽⁰⁾ ... E^(G~υ }, where E^(g) = [e⁽ _o ^g) ... ^f₁] is the g*

precoding matrix, and ej is the m* precoding vector in the group. Each user device will calculate a CQI (Channel Quality Indicator) for each vector in each matrix of the group of precoding matrixes E. Precoding vectors in a precoding matrix may be designed as a plurality of data streams of multiple users or one user, so as to support SDM (Spatial Division Multiplexing) or SDMA (Spatial Division Multiple Access) technology.

Node B may select the amount of feedback information by selecting an appropriate G value and deciding the amount of information required to be fed back to Node B by a user, in order to achieve different adaptabilities.

For example, to achieve the highest adaptability, a user equipment may feed back each CQI value of each of the matrixes in E. Therefore, each user needs to feed back up to GM CQI values.

Nevertheless, to achieve the lowest management cost, a user equipment only needs to feed back the best CQI value and the corresponding vector index. In this case, each user equipment needs to transmit log₂(GM) bits of data in addition to the actual CQI value. This type of feedback can only support limited SDMA since each user may only be allocated to a certain beam. When a user is arranged as a time-frequency unit, this type of feedback could only support the transmission of beam formings. Each user equipment may also feed back M CQI values corresponding to a preferred or the best group, in order to obtain a better mixture of SDM and SDMA.

Node B may select a group g and a corresponding Έ!^8> , and allocate up to M independent data streams to different user equipments or a single user equipment, based on a predetermined design. The requirement to be met by a codebook is that it should be unique so as to assure that the precoding cannot change channel capacity.

The design of a codebook in a PU²RC system is based on the Fourier transform in the following Equation (1): , wherein, _*&^> =

where G is the number of user groups, M is the number of transmit antennas, 0 ≤ n ≤ M — 1 and O ≤ m ≤ M -1 .

There are two examples in the codebook of PU²RC, i.e., l) when M=2, G=2,

2) When M=4, G=I,

Referring to Figure 1, the corresponding PU RC process includes:

1. a controller 100 collecting feedback information containing the index of a preferred precoding matrix and corresponding CQI values of all the precoding vectors in the matrix;

2. a user grouping unit 200 grouping users that require the same preferred precoding matrix into one group; 3. a multiplexing unit 300 selecting a group with the highest priority among different groups and a data stream with the highest priority (which may be from different users or a same user) in the selected group;

4. an AMC (Adaptive Modulation and Coding) unit 400 performing an appropriate AMC process on the selected data stream according to an associated CQI; and 5. a precoding unit 500 performing a precoding process on the selected group.

The precoding matrix mentioned above is the so-called precoding codebook. The disadvantage of the PU²RC process and the codebook is that: the number of available codebooks is limited, and in a practical system, optimized outputs of the system could only be obtained by calculating feasibility of all the possible codebooks. Since the codebook is unique after the number of transmit antennas, the number of users and the number of codebook groups are predetermined, a change of the wireless transmission channel results in the changeable range for selection of codebooks being limited. For example, when

M=2 and G=2, the codebook is fixed to the one shown in Equation (2).

SUMMARY OF THE INVENTION

An object of the present invention is to provide a method of precoding in a multi-antenna system in order to optimize codebooks with extension Fourier series. A method of precoding in a multi-antenna system according to the present invention comprises the steps of: a. selecting an appropriate number of groups of user data streams; b. generating precoding codebooks with extension Fourier series based on the number of transmit antennas in the system and the number of groups of user data streams; c. obtaining feedback information from the users, which contains a preferred precoding codebook and an associated CQI value; d. grouping the users, based on the feedback information; and e. selecting a user group according to a predetermined criterion and performing a precoding process on the selected user group. Compared with the prior art, the method of precoding in a multi-antenna system according to the present invention can obtain a larger range for selection of codebooks with extension Fourier series, which is helpful to find out the most suitable codebook.

Other objects and effects of the present invention will become apparent from the following detailed description and the claims when taken in connection with the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS Now a more detailed description will be given in conjunction with the drawings, where: Figure 1 illustrates a schematic view of the construction of a PU RC transmitter; and Figure 2 illustrates a flow chart of the method of precoding according to the present invention.

BEST MODES FOR CARRYING OUT THE INVENTION

Compared with PU²RC schemes in the prior art, the method of precoding according to the present invention optimizes codebooks with extension Fourier series. The principle of the extension Fourier series is expressed in the following Equation (5):

_Λ 1 where Δ is defined as a start point and may vary from 0 to — according to requirements

G for grouping precoding codebooks. The definitions of other variants in the Equation (5) is the same as that of Equation (1). Thus, an extension Fourier series is obtained.

It will be illustrated in conjunction with the following detailed embodiments how codebooks can be optimized with the extension Fourier series according to the present invention.

Assuming that the number of transmit antennas in the system is M, the symbol of the k^th user transmitted via the M transmit antennas is S^ = [££ , ..., ££ ] where k=l,...,K, K denotes the number of users, the superscript T denotes a transposition operation on the matrix, and the channel response matrix of different users is H^, so received signals of the k^th user could be modeled as: τζ

% = H^ e^ +n_έ (6)

Jk=I

The definition of G^ G.?) may be as shown in Equation (1). For a better understanding it is assumed here that M=K, which is a general case, however, in fact, processes in the case of M>K are similar to this. As for ZF (Zero Forcing) detection, this is shown in the following Equation (7):

where H and -1 denote conjugate transposition operations and inverse operations on the matrix.

The most suitable codebook is the one that can select a symbol from the detected symbols such that the largest possible amount of SINR (Signal to Inference and Noise Ratio) is obtained, i.e.,

K g = argmax [ ∑ SINR(s_k)] (8) gφ,...G) k=l

To find out which is the best codebook, the values in Equation (8) will be calculated several times based on different codebooks for different sizes of groups. Taking the case of M=2 as an example, when M=2 and G=A, according to Equation (1) in the prior art, codebooks are:

1 1 1 1 1 1

,E₁ = 1 -1 42^~ exp(jπ/4) exp(jπ5/4) (9)

It can be noted that two of the codebooks in the case of G=4 are identical to those in the case of G=2, although the size of the group changes from 2 to 4. This repetition is disadvantageous to effectively select the most suitable codebook. However, the extension Fourier series in Equation (5) may enlarge the number of codebooks in each group, i.e., the start point Δ is added to the Fourier series in Equation (1). The advantages of this situation include: more Fourier series can be generated for codebooks so as to facilitate finding the most suitable codebook, because in a practical system, ideal data throughput of the system can only be obtained by calculating feasibility of all the possible codebooks; and when the wireless transmission channel changes, the changeable range for selection of codebooks becomes so large as to be more adaptive to actual conditions of a MIMO precoding transmission. The advantages can be described in detail by means of the following two examples, where M=2:

(1) Assuming that G=A and o ' codebooks are: o 1 1 1

¹^o exp(jπ/8) exp(jπ9/8)

(10)

1 1 1

E₁ exp(jπ3/8) exp(jπll/8)

1 1 1

E₂ exp(jπ5/8) exp(jπl3/8)

1 1 1

^ exp(jπ7/8) exp(jπl5/8)

1

Assuming that G=S and ^- codebooks are:

Io

1 1 1

^E°^"V2 exp(jπ/16) exp(jπ9/16)

(H)

1 1 1

E, = l^~4i exp(jπ3/16) exp(jπll/16)

1 1 1

E, =

^ exp(jπ7/16) exp(jπl5/16)

1 1 1

E₅ =

^ exp(jπll/16) exp(jπ27/16) 1 1 1

E, = exp(jπl3/16) exp(jπ29 /16)

It can be seen from Equations (2), (10) and (11) that when G=2, G=4 and G=8, the codebooks are absolutely different. This proves that extension Fourier series could be used to obtain a larger range for selection of codebooks, so that it will be helpful to find out the most suitable codebook.

(2) Assuming that G=A and Δ = — — _? codebooks are:

1 1 1

E₀ exp(jπ/32) exp(jπ33/32)

(12)

1 1 1

E₁ exp(jπ9/32) exp(jπ41/32)

1 1 1

E₂

V2^~ exp(jπl7/32) exp(jπ49/32)

n 1 1 1

Λ/2^" exp(jπ25/32) exp(jπ57/32)

It can be seen from Equations (10) and (12) that even when G=4, absolutely different groups of codebooks could be obtained.

Therefore, with reference to Figure 2, the method of precoding in a multi-antenna system according to the present invention includes: first, selecting an appropriate number of groups of user data streams to determine the amount of feedback information (step 10); then, generating precoding codebooks with the above extension Fourier series, based on the number of transmit antennas in the system and the number of groups of user data streams (step 11); receiving feedback information from the user terminals by means of Node B, which contains the index of a preferred precoding codebook and an associated CQI value (step 12); grouping the users, based on the selection of the preferred precoding codebook in the feedback information (step 13); and selecting a user group with the highest priority and performing a precoding process on the selected user group according to a predetermined precoding scheme.

The method of precoding in a multi-antenna system according to the present invention may be implemented in a hardware construction such as the PU²RC transmitter in Figure 1.

Since the method of precoding in a multi-antenna system according to the present invention has a larger range for selection of codebooks, it is better in terms of performance than PU²RC schemes in the prior art.

It should be noted that the above embodiments are used to illustrate but not limit the invention. It should be appreciated by a person skilled in the art that various modifications are possible in the method of precoding in a multi-antenna system disclosed in the present invention without departing from the scope of the present invention. Therefore, the scope of protection of the invention should be defined by the appended claims. Furthermore, any of the reference numbers in the claims should not be interpreted as limiting the scope of the claims.

Claims

What is claimed is:

1. A method of precoding in a multi-antenna system, including the steps of: a. selecting an appropriate number of groups of user data streams; b. generating precoding codebooks with extension Fourier series, based on the number of transmit antennas in the system and the number of groups of user data streams; c. obtaining feedback information from the users, which contains a preferred precoding codebook and an associated CQI value; d. grouping the users, based on the feedback information; and e. selecting a user group according to a predetermined criterion and performing a precoding process on the selected user group.

2. The method of claim 1, wherein, in step b, the precoding codebook is generated with the following extension Fourier series:

0 < Δ < — G

where, G is the number of user groups, M is the number of the transmit antennas,

O ≤ n ≤ M -1 and 0 < m ≤ M -l .

3. The method of claim 2, wherein, Δ is defined as a start point and may

vary from 0 to — according to requirements for grouping precoding

G

codebooks.

4. The method of claim 1, wherein, in step d, users that require a same

preferred precoding matrix are grouped into one group.

5. The method of claim 1, wherein, step e further includes: applying AMC (Adaptive Modulation and Coding) according to the associated CQI, to process data streams in the selected user group.