WO2018149503A1

WO2018149503A1 - Device and method for wireless communication network transmission

Info

Publication number: WO2018149503A1
Application number: PCT/EP2017/053676
Authority: WO
Inventors: Chaitanya TUMULA
Original assignee: Huawei Technologies Co., Ltd.
Priority date: 2017-02-17
Filing date: 2017-02-17
Publication date: 2018-08-23

Abstract

A system and method for wireless communication network using multiple-input and multiple-output transmission is disclosed. In the system and method, at least one data stream is transmitted using multiple transmit antennas from a transmitting device to a receiving device. The transmission device uses one or more precoding matrices for transmission using multiple subcarriers. The precoding matrices are generated based on information of estimates of the channel of the reserve link so that phase variation in the effective channel seen at the receiving device between adjacent subcarriers is minimized.

Description

DEVICE AND METHOD FOR WIRELESS COMMUNICATION NETWORK

TRANSMISSION

FIELD [0001 ] The present disclosure relates to the field of wireless communications, and more particularly to transmission of signals using multiple-input and multiple- output systems.

BACKGROUND

[0002] Multiple-input and multiple output (MIMO) systems are used for increasing system capacity and/or improved reliability of data transmission. Hence MIMO transmission schemes are included in many telecommunication system standards specifications, such as, Long Term Evolution (LTE), LTE- Advanced, Wi- Fi, Asymmetric Digital Subscriber Line (ADSL) and similar.

[0003] MIMO systems are used either to transmit a single or multiple data streams using multiple transmit antennas. Typically, when the number of data streams to be transmitted is smaller than or equal to the number of transmit antennas, precoding is used to transmit a linear combination of data streams from each transmit antenna. The linear combining weights used for each transmit antenna are chosen such that the data reception at the receiver is improved.

[0004] Figure 1 illustrates a MIMO-OFDM (Orthogonal Frequency-Division

Multiplexing) system with N_sub subcarriers and T OFDM symbols per frame. Let BWsub denote the OFDM subcarrier spacing and Ts denote the symbol duration of one OFDM symbol. A subcarrier in one OFDM symbol is called a resource element (RE). Typically, some of the REs in the MIMO-OFDM frame are used for reference signal transmission and remaining for data transmission. Assume that the transmitter is equipped with Nr transmit antennas and the receiver is equipped with NR receive antennas. Let us also assume that N_S denote the number of data streams to be transmitted on each subcarrier, N_S<NT. Let W_{k i} denote the precoding matrix of size NT^XN_S used by the transmitter for sending data on the ^th subcarrier of the z^'th OFDM symbol, and 1<ζ^'<Γ. We can write the received signal model on the fc^th subcarrier of the 2^th MIMO-OFDM symbol as given by Equation 1. y_k,i = H _iW _ix _i + n _i = H^e _k ^f/ x _i + n_kX Equation 1

[0005] Where, x_{k i} is the vector of transmitted symbols with size N_s X 1, in which each element in x_{k i} belongs to the fmite-alphabet set Ω , e.g., M-QAM constellation; y_{k i} is the vector of received symbol of size N_R x 1; H_{k i} is the MIMO channel matrix with size N_R X N_T on the fc^th subcarrier of the i^th MIMO-OFDM symbol; H^e _k [ is the effective channel matrix of size N_R X N_s seen by the receiver on the fc^th subcarrier of the ith MIMO-OFDM symbol; n_{k i} is the vector of noise samples.

[0006] Channel state information (CSI) plays an important role in the performance of precoded MIMO transmissions. With the knowledge of H_{k i} , the optimal precoding scheme is to transmit data streams in the directions corresponding to the channel "eigenmodes". For this, the transmitter performs the SVD decomposition of the MIMO channel matrix as H_{k i} = U_{k i}D_ktiV_{k i} with U_{k i} being an N_R X N_R unitary matrix and V_{k i} being an N_T X N_T unitary matrix and D_{k i} being an N_R X N_T diagonal matrix of singular values. The transmitter chooses the optimal precoding matrix as the N_s columns of V_{k i} corresponding to the largest singular values. In the following, we assume that N_s = N_T and hence W_{k i} = V_{k i}, i.e., the number of columns in the precoding matrix are the same as the number of transmit antennas. However, the embodiments of the invention and its effects hold true in other cases as well.

[0007] Typically in wireless systems using the time-division duplex (TDD) operation, reference signals received in the reverse link are used to estimate the channel of the reverse link. Then, using the reciprocity and calibration, the knowledge of the forward channel can be obtained for computing the precoding matrices used for forward link transmission.

[0008] One of the important aspects of precoding methods is that the effective channel seen by the receiver on a given transmit-receive antenna pair can have a large phase variation across subcarriers as shown in figure 2b. In the figure 2a, the plot shows the phase variation of the wireless channel across the subcarriers of an OFDM symbol corresponding to a transmit-receive antenna pair. The plot of figure 2b shows the phase variation of the effective channel (precoded) for the same transmit-receive antenna pair. As we can see, in case of precoded transmission, the phase variation of the effective channel is much more random and hence, less efficient narrowband frequency domain channel estimation methods have to be used to estimate the effective channel at the receiver.

SUMMARY

[0009] It is an object of the present invention to provide an improved concept for precoding. This object is solved by the invention as defined in the independent claims. Further embodiments can be found in the dependent claims.

[001 0] Embodiments of the present invention are based on the finding that if precoding at the transmitter can minimize the phase variations of the effective channel, the receiver may use wideband channel estimation methods (eg: time- domain algorithms) to obtain processing gain as well as reducing the complexity of channel estimation algorithm at the forward link receiver. As will be described in the following, embodiments of the present invention reduce said phase variation of the effective channel which thereby enables the use of the wideband channel estimation methods at receiver side.

[001 1 ] In the following description, a system and method for wireless communication network using multiple-input and multiple-output transmission is disclosed. In the system and method, at least one data stream is transmitted using multiple transmit antennas from a transmitting device to a receiving device. The transmission device uses one or more precoding matrices for transmission using multiple subcarriers. The precoding matrices are generated based on information of estimates of the channel of the reserve link so that phase variation in the effective channel seen at the receiving device between adjacent subcarriers is minimized.

[001 2] In the first aspect, a network node comprising a transceiver and a processor is disclosed. The transceiver is configured to receive reference signals in a reverse link from a receiving device, and the processor is configured to obtain estimates of the channel of the reverse link for a plurality of subcarriers, and generate a first precoding matrix for a first subcarrier of the plurality of subcarriers based on at least one of the obtained estimates of the reverse link. The processor is further configured to rotate columns of the first precoding matrix to obtain a first rotated precoding matrix for a second subcarrier of the plurality of subcarriers so that phase variation in effective channels seen at the receiving device across the first and second subcarriers is minimized; or generate a second precoding matrix based on at least one of the obtained estimates of the reverse link, and rotate columns of the second precoding matrix to obtain a second rotated precoding matrix so that phase variation in effective channels seen at the receiving device across the first and second subcarriers is minimized. Then the transceiver is configured to transmit a data stream to the receiving device on the first subcarrier using the first precoding matrix and on the second subcarrier using either the first rotated precoding matrix or the second rotated precoding matrix.

[001 3] According to the first aspect, a precoding matrix for the second subcarrier is generated either by rotating the columns of the first precoding matrix or by generating a second precoding matrix and rotating the columns of the generated second precoding matrix, which are used to precode at least one data stream. Then the transmitter of the network node transmits at least one of the precoded data stream using multiple antennas. The network node is used to designate any network device that is connected to a network and is capable of transmitting data using multiple antennas. Thus, the transmission can be initiated by a base station, wireless router, mobile device or any other similar device.

[001 4] It is advantageous to use a precoding approach described above as it allows minimizing the phase variations in the effective channel seen at the receiver across the subcarriers of an OFDM symbol. This facilitates the use of a wideband channel estimation algorithm at the receiver, resulting in a better performance with reduced complexity.

[001 5] In the first implementation of the first aspect, wherein in the case in which the processor is configured to obtain the second rotated precoding matrix, the processor is configured to obtain an estimate of the channel of the reverse link for the first subcarrier and then to generate the first precoding matrix based on the obtained estimate of the first subcarrier. It is advantageous to generate the precoding matrix for each of the subcarriers based on the respective estimate of the reverse link as this approach provides a better precoding matrix for each of the subcarriers in terms of maximizing the precoding gain.

[001 6] In the second implementation of the first aspect, wherein in the case in which the processor is configured to obtain the second rotated precoding matrix, the processor is configured to obtain an estimate of the channel of the reverse link for the second subcarrier and to generate the second precoding matrix based on the obtained estimate of the second subcarrier. It is advantageous to generate the second precoding matrix for the second subcarrier based on the respective estimate as it provides a more accurate starting point for obtaining a second rotated precoding matrix than using estimates of other subcarriers.

[001 7] In the third implementation of the first aspect the processor is further configured to generate a precoding matrix for each subcarrier in an OFDM symbol, wherein the processor is further configured to rotate the second and subsequent precoding matrices. The third implementation form is advantageous, as it is applied for each subcarrier in an OFDM symbol so that the whole OFDM symbol can be received with phase continuity of the effective channel seen at the receiving device. [001 8] In the fourth implementation of the first aspect, the processor is configured to obtain the first rotated precoding matrix, the processor is further configured to obtain a combined estimate of the channel of the reverse link for a sub- band formed by a plurality of subcarriers and to generate the first precoding matrix based on the obtained estimate of the sub-band.

[001 9] A sub-band here refers to a grid of resource elements in time and frequency dimension. The size of a sub-band can vary from one RE to all the REs in an OFDM frame. Typically the sub-band size is configurable, and it depends on the properties of the wireless channel between the transmitting and receiving devices. The size of a sub-band (number of REs) along the time dimension depends on the value of channel Doppler and the size of the sub-band along the frequency dimension depends on the value of channel delay spread. The fourth implementation form is advantageous because it provides a simplified option of using a combined estimate to obtain precoding matrices. The combined estimate provides a possibility to use only one precoding matrix that is then rotated for each of the subcarriers.

[0020] In the fifth implementation of the first aspect, the processor is configured to perform the rotation of the first precoding matrix for the second and each subsequent subcarrier in an OFDM symbol. The rotation of the precoding matrix for each subcarrier provides minimized phase variations in the effective channel seen at the receiver across the subcarriers of an OFDM symbol.

[0021 ] In the sixth implementation of the first aspect, the processor is configured to generate the precoding matrices based on singular value decomposition of the channel estimates obtained in the reverse link. It is advantageous to perform singular value decomposition, and choosing the right- singular matrix as the precoding matrix for each subcarrier is theoretically optimal. Even though, in this disclosure, we use the SVD based precoding to present the embodiments of the invention, one can use any other precoding method.

[0022] In the seventh implementation of the first aspect, the processor is configured to determine a rotation angle φι for rotating the Zth column of each precoding matrix the processor is configured to rotate. By choosing a rotation angle for each column, we could minimize the phase variation in the effective channel seen across all receiving antennas for each of the transmit antennas (streams).

[0023] In the eighth implementation of the first aspect, the processor is configured to determine the rotation angle φι such that the sum of the absolute phase difference in the effective channel vectors of the two adjacent subcarriers seen at all receiving antennas of the receive device corresponding to the Zth transmit antenna is minimized. It is advantageous to determine a rotation angle based on the minimization of the sum of the absolute phase difference in the effective channel vectors as it provides an angle that reduces phase variation of the effective channel between adjacent subcarriers. Furthermore, one can also minimize the average phase differences in the effective channel vectors and obtain the same rotation angle φι.

[0024] In the ninth implementation of the first aspect, the processor is configured to determine the rotation angle φι such that the Euclidean distance between effective channel vectors of the two adjacent subcarriers seen at all receiving antennas of the receiving device corresponding to the Zth transmit antenna is minimized. It is advantageous to determine a rotation angle based on minimizing the Euclidean distance between effective channel vectors of the two adjacent subcarriers as the channel magnitude changes in a smooth manner across all the subcarriers of an OFDM symbol.

[0025] In the second aspect, a method for transmitting data using at least one precoding matrix is disclosed. In the method, first, reference signals on a reverse link are received from a receiver. Then estimates of the channel of the reverse link for a plurality of subcarriers are obtained. Based on the obtained estimates of the reverse link, a first precoding matrix for the first subcarrier of the plurality of subcarriers is generated. Then at least one data stream is transmitted on the first subcarrier using multiple transmit antennas and the first precoding matrix. The method according to the second aspect further comprises either rotating columns of the first precoding matrix to obtain a first rotated precoding matrix for a second subcarrier of the plurality of subcarriers so that phase variation in effective channels seen at the receiving device across the first and second subcarriers is minimized, or generating a second precoding matrix based on at least one of the obtained estimates of the reverse link and rotate columns of the second precoding matrix to obtain a second rotated precoding matrix so that phase variation in effective channels seen at the receiving device across the first and second subcarriers is minimized. Lastly at least one data stream is transmitted on the second subcarrier using multiple transmit antennas using either the first rotated precoding matrix or second rotated precoding matrix.

[0026] According to the second aspect, a precoding matrix for the second subcarrier is generated either by rotating the columns of the first precoding matrix or by generating a second precoding matrix and rotating the columns of the generated second precoding matrix, which are used to precode at least one data stream. Then a transmitter of a network node transmits at least one precoded data stream using multip le antennas .

[0027] It is advantageous to use a precoding approach described above as it allows minimizing the phase variations in the effective channel seen at the receiver across the subcarriers of an OFDM symbol. This facilitates the use of a wideband channel estimation algorithm at the receiver possibly resulting in a better performance with reduced complexity.

[0028] In the first implementation of the second aspect, the method further comprises obtaining an estimate of the channel of the reverse link for the first subcarrier and generating the first precoding matrix based on the obtained estimate of the first subcarrier; and obtaining an estimate of the channel of the reverse link for the second subcarrier and to generate the second precoding matrix based on the obtained estimate of the second subcarrier. This is advantageous to generate the precoding matrix for each of the two subcarriers based on the respective estimates as this approach provides a better precoding matrix for each of the subcarriers in terms of maximizing the precoding gain. Furthermore, it is advantageous to generate the second matrix for the second subcarrier based on the respective estimate as it provides a more accurate starting point for rotating the second precoding matrix to minimize the phase variations in the effective channel seen at the receiving device across the two subcarriers.

[0029] In the second implementation of the second aspect the method further comprises generating a precoding matrix for each subcarrier in an OFDM symbol and rotating the second and subsequent precoding matrices. This is advantageous as the described method is applied for each subcarrier in an OFDM symbol so that whole OFDM symbol can be received with phase continuity of the effective channel seen at the receiving device.

[0030] In the third implementation of the second aspect, the rotating columns of the first precoding matrix further comprises obtaining a combined estimate of the channel of the reverse link for a sub-band formed by a plurality of subcarriers and generating the first precoding matrix based on the obtained estimate of the sub-band. This is advantageous because it provides a simplified option of using a combined estimate to obtain precoding matrices. The combined estimate provides a possibility to use only one precoding matrix that is then rotated for each of the subcarriers.

[0031 ] In the fourth implementation of the second aspect the method further comprises rotating the first precoding matrix for the second and each subsequent subcarrier in an OFDM symbol. The rotation of the precoding matrix for each subcarrier provides minimized phase variations in the effective channel seen at the receiver across the subcarriers of an OFDM symbol.

[0032] In the third aspect a computer program comprising a program code configured to perform the method as described above when executed in a transmitting device is disclosed. It is advantageous to provide a computer program for implementing the method as it is possible to enable the transmission as described above in devices that already have all necessary hardware components required for the transmission. [0033] In the following new precoding methods are introduced for reducing the phase variations in the effective channel observed across the subcarriers of an OFDM symbol at the receiver. With the described methods applied at the transmitter, the receiver can use low-complexity wideband channel estimation algorithms to estimate the effective channel across all the data REs with improved performance and demodulate the data.

[0034] The device and method for transmitting a data stream in a multiple- input multiple-output communication network provides an efficient mechanism for transmitting the data stream from a transmitting device to the receiving device. These and other aspects of the system and method for wireless communication network using multiple-input and multiple-output transmission will be apparent from and the embodiment(s) described below.

DESCRIPTION OF THE DRAWINGS

[0035] The present description will be better understood from the following detailed description read in light of the accompanying drawings, wherein:

FIG. 1 illustrates a MIMO-OFDM frame structure;

FIG. 2a illustrates a plot showing the phase variation of the wireless channel between a transmit-receive antenna pair of a MIMO-OFDM wireless

communication system without precoding;

FIG. 2b illustrates a plot showing the phase variation of the wireless channel between a transmit-receive antenna pair of a MIMO-OFDM wireless

communication system with a conventional precoding;

FIG 3 illustrates an example of a transmitter;

FIG 4 illustrates an example of a transmission method; and

FIG 5 illustrates results of the first embodiment. DETAILED DESCRIPTION

[0036] The detailed description provided below in connection with the appended drawings is intended as a description of the embodiments and is not intended to represent the only forms in which the embodiment may be constructed or utilized. However, the same or equivalent functions and structures may be accomplished by different embodiments.

[0037] In figure 3 a, an example of a network node 10 comprising a transceiver

11 and a processor 12 is illustrated. In figure 3b the example of the network node 10 is illustrated in use and communicating with a receiving device 13. The transceiver 11 is configured to receive reference signals in a reverse link from a receiving device 13 and the processor 12 is configured to obtain estimates of the channel of the reverse link for a plurality of subcarriers and generate a first precoding matrix for a first subcarrier of the plurality of subcarriers based on at least one of the obtained estimates of the reverse link. The processor 12 is further configured to rotate columns of the first precoding matrix to obtain a first rotated precoding matrix for a second subcarrier of the plurality of subcarriers so that phase variation in effective channels seen at the receiving device 13 across the first and second subcarriers is minimized. Alternatively the processor 12 can be configure to generate a second precoding matrix based on at least one of the obtained estimates of the reverse link and rotate columns of the second precoding matrix to obtain a second rotated precoding matrix so that phase variation in effective channels seen at the receiving device 13 across the first and second subcarriers is minimized. Then the transceiver 11 is configured to transmit a data stream to the receiving device 13 on the first subcarrier using the first precoding matrix and on the second subcarrier using either the first rotated precoding matrix or the second rotated precoding matrix. Examples on how the processor 12 minimizes the phase variation in effective channels seen at the receiving device 13 across the first and second subcarriers will be provided later in this document. [0038] Hence in embodiments of the present invention a precoding matrix for the second subcarrier is generated either by rotating the columns of the first precoding matrix or by generating a second precoding matrix and rotating the columns of the generated second precoding matrix, which are used to precode at least one data stream. Then the transceiver 11 of the network node 10 transmits at least one precoded data stream using multiple antennas. The network node 10 is used to designate any network device that is connected to a network and is capable of transmitting data using multiple antennas. Thus, the transmission can be initiated by a base station, wireless router, mobile device or any other similar device.

[0039] It is advantageous to use a precoding approach described above as it allows minimizing the phase variations in the effective channel seen at the receiver across the subcarriers of an OFDM symbol. This facilitates the use of a wideband channel estimation algorithm at the receiver possibly resulting in a better performance with reduced complexity.

[0040] In TDD systems, the MIMO-OFDM forward-link transmitter (such as network node 10) receives the reference signals transmitted on the reverse link from the forward-link receiver (such as receiving device 13) so it can obtain estimates of the reverse link. Using the wireless channel reciprocity and correcting for the hardware mismatch between the forward and reverse links, the forward-link transmitter can obtain the knowledge of the forward channel to the forward-link receiver. Once the knowledge of the forward-link channel corresponding to all subcarriers of all OFDM symbols is obtained, the transmitter can transmit data and reference signals on the forward-link using precoding.

[0041 ] The forward-link transmitter can either perform precoding on each subcarrier of each OFDM symbol or use a single precoding matrix for all the subcarriers in a grid of REs (sub-band) with different precoders for different sub- bands. The option of sub-band based precoding is typically used in practical systems to reduce the complexity and signaling overhead. [0042] The forward-link receiver receives MIMO-OFDM signal transmitted from the forward-link transmitter using the described precoding methods. It then estimates the effective channel on all the data REs by employing a wideband channel estimation algorithm on the received reference signal REs. For example, the receiver can perform Fast Fourier Transform (FFT) or Discrete Fourier Transform (DFT) and use time-domain channel estimation algorithms to estimate the effective channel corresponding to the data REs and demodulate the data.

[0043] In a first embodiment of the example of figures 3a and 3b, wherein in the case in which the processor 12 is configured to obtain the second rotated precoding matrix, the processor 12 is configured to obtain an estimate of the channel of the reverse link for the first subcarrier and then to generate the first precoding matrix based on the obtained estimate of the first subcarrier.

[0044] Furthermore, in the first embodiment, wherein in the case in which the processor is configured to obtain the second rotated precoding matrix, the processor 12 may be further configured to obtain an estimate of the channel of the reverse link for the second subcarrier and to generate the second precoding matrix based on the obtained estimate of the second subcarrier. It is advantageous to generate the precoding matrix for each of the subcarriers based on the respective estimate of the reverse link as this approach provides a better precoding matrix for each of the subcarriers in terms of maximizing the precoding gain.

[0045] Furthermore, in the first embodiment, the processor 12 may be further configured to generate a precoding matrix for each subcarrier in an OFDM symbol, wherein the processor is further configured to rotate the second and subsequent precoding matrices. This is advantageous, as it is applied for each subcarrier in an OFDM symbol so that whole OFDM symbol can be received with phase continuity of the effective channel seen at the receiving device.

[0046] In a second embodiment according to the example of figures 3a and 3b, in which the processor 12 is configured to obtain the first rotated precoding matrix, the processor 12 is further configured to obtain a combined estimate of the channel of the reverse link for a sub-band formed by a plurality of subcarriers and to generate the first precoding matrix based on the obtained estimate of the sub-band. Hence, in the second embodiment one and the same precoding matrix is rotated for the several subcarriers of the sub-band.

[0047] A sub-band here refers to a grid of resource elements in time and frequency dimension. The size of a sub-band can vary from one RE to all the REs in an OFDM frame. Typically the sub-band size is configurable, and it depends on the properties of wireless channel between the transmitting and receiving devices. The size of a sub-band (number of REs) along the time dimension depends on the value of channel Doppler and the size of the sub-band along the frequency dimension depends on the value of channel delay spread. The fourth implementation form is advantageous because it provides a simplified option of using a combined estimate to obtain precoding matrices. The combined estimate provides a possibility to use only one precoding matrix that is then rotated for each of the subcarriers.

[0048] In the second embodiment, the processor 12 of the network node 10, can compute the average of the MIMO channel matrix of all the REs in each sub- band and use the average MIMO channel matrix to generate a single precoding matrix for all the REs in the sub-band. In other words, the processor 12 obtains a combined estimate of the channel of the reverse link for the sub-band and generates the first precoding matrix based on this estimate. As an example, let us assume that all the REs in the MIMO-OFDM frame are divided into M = N_Sub/N non- overlapping sub-bands, with the m ¹¹¹ sub-band containing REs with subcarrier indices k = m— 1)N + 1,2, ... , mN and OFDM symbol indices i = 1,2, ... , T. For the mth sub-band, the processor 12 computes the average MIMO channel matrix as H_{avg m} = ^∑_fc₌₁∑[₌₁ H_{k i}, and uses it to generate a single precoding matrix (e.g used as the first precoding matrix which will be rotated later) for all the REs in the sub-band. [0049] Hence, in the second embodiment described above, the processor 12 is configured to perform the rotation of the first precoding matrix for the second and each subsequent subcarrier in an OFDM symbol.

[0050] The rotation of the precoding matrices (respective precoding matrices for each subcarrier (first embodiment) or the first precoding matrix for each subcarrier (second embodiment)) provides minimized phase variations in the effective channel seen at the receiver across the subcarriers of an OFDM symbol.

[0051 ] In embodiments described above the processor 12 is configured to generate the precoding matrices based on singular value decomposition (SVD) of the channel estimates obtained in the reverse link.

[0052] The processor 12 of the network node 10 can use SVD based precoding matrix on each subcarrier of each OFDM symbol by performing the SVD decomposition of the estimated MIMO channel matrix and choosing the right- singular matrix as the precoding matrix.

[0053] For the first embodiment, if the estimated MIMO channel matrix of the fc^th subcarrier of the i^th MIMO-OFDM symbol is denoted by H_k>i, the processor 12 performs the SVD decomposition of the MIMO channel matrix as H_{k i} = U_ktiD_{k i}V_{k i} and selects the precoding matrix W_{k i} = V_{k i}.

[0054] In case of the sub-band based precoding method (second embodiment), the processor 12 performs the SVD decomposition of the average MIMO channel matrix as _{avg m} =

(corresponding to a combined estimate of the sub-band), and selects the precoding matrix for the mth sub-band as V_{avg m} , i.e., W_k = V_avg>m, for (m - 1)N + 1 < k≤ mN, 1 < i≤ T. [0055] It is advantageous to perform singular value decomposition and choosing the right-singular matrix as the precoding matrix for each subcarrier, as this approach maximizes the precoding gain for each subcarrier. Even though, in this disclosure, we use the SVD based precoding to present the embodiments of the invention, the skilled in the art can use any other precoding method.

[0056] The forward-link receiver, which is the receiving device 13, can estimate the effective channel across all the REs carrying the data by using the precoded reference signals transmitted in the forward link. For the effective channel estimation, the receiving device 13 can use either narrowband channel estimation or wideband channel estimation algorithms. In case of narrowband channel estimation, the receiving device 13 for example, may use only reference signals in each sub- band to estimate the effective MIMO channel of the REs carrying data and perform demodulation. In case of wideband channel estimation, the receiving device 13 may for example use all the reference signals of all the sub-bands and use an FFT based time-domain estimation algorithm to estimate the MIMO channel of the REs carrying data and perform demodulation.

[0057] In the embodiments described above, the processor 12 is configured to determine a rotation angle φι for rotating the Zth column of each precoding matrix the processor is configured to rotate (as for example the first and second precoding matrices). By choosing a rotation angle for each column, we could minimize the phase variation seen across all receiving antennas for each of the transmit antennas (streams).

[0058] In the following, the presented methods and the discussion assume that different precoding matrices are used for data transmission on each data RE (first embodiment). However, the described methods and the corresponding arguments can be used in conjunction with the sub-band based precoding methods as well (second embodiment) where the same (first) precoding matrix is rotated for the different subcarriers of the sub-band. Furthermore, in the following two different methods for determining the rotation angles are discussed. However, the disclosed embodiments maybe used in conjunction with any other method for minimizing phase variations in the effective channel seen at the receiver across the adjacent subcarriers of an OFDM symbol.

[0059] Assuming that the precoder for the subcarrier with index k on the i^th

MIMO-OFDM symbol is designed, we now describe two methods in which we rotate the columns of the SVD based precoding matrix of subcarrier with index (k + 1) on the ith MIMO-OFDM symbol to minimize the phase variations in the effective channel seen at the receiver across the two subcarriers.

[0060] On the fcth subcarrier of the i^th MIMO-OFDM symbol, let W_{k i} = ... denote the precoding matrix used for data transmission, then the effective channel seen at the receiver on the fcth subcarrier of the ith MIMO-OFDM symbol is given by H ( = H_k>iW_k>i = [h^e/' ¹ ... h^e _k ^Ντ] =

] . In the subcarrier based precoding case (first embodiment), W_{k i} = V_{k i}, whereas, in the sub-band based precoding case (second embodiment), W_{k i} = V_{avg m}, where m denotes the sub-band index in which the fc^th subcarrier of the i^th MIMO-OFDM symbol lies.

[0061 ] Designing the precoder for the (k + l)^th subcarrier of the i^th MIMO-

OFDM symbol as W_k+ = V_k+ R_k+ , where

[0062] That is, the Zth column of precoder on k + 1)^Λ subcarrier of the ith

MIMO-OFDM symbol w_k ^l _{+1 1} is obtained by rotating the Z^th column of the SVD based precoder v_k ^l _{+l i} by φι. [0063] In case of subcarrier based precoding, V_{k+l i} is obtained by performing the SVD decomposition of the channel matrix H_{k+l i} = U _{k+l i}D _{k+l i}V_{k+l i} , whereas, in case of sub-band based precoding, V_{k+l i} = V_{avg m}, where m denotes the sub-band index in which the (k + l) th subcarrier of the i th MIMO-OFDM symbol lies.

[0064] The effective channel seen by the receiver on the (k + l)th subcarrier of the ith MIMO-OFDM symbol is given by

Ηι ίι_,ί ⁼ H_k+i_,iW_k+1)i = H _{k+l i}V _{k+l i}R_{k+l i} = [ j i t ...

=

[0065] In embodiments described above the processor 12, using the first alternative method for determining rotation angles, is configured to determine the rotation angle φι such that the sum of the absolute phase difference in the effective channel vectors of the two adjacent subcarriers seen at all receiving antennas of the receiving device corresponding to the I ^th transmit antenna is minimized. It is advantageous to determine a rotation angle based on the minimization of the sum of the absolute phase difference in the effective channel vectors as it provides an angle that reduces phase variation of the effective channel between adjacent subcarriers. Furthermore, one can also minimize the average phase differences in the effective channel vectors and obtain the same rotation angle φι.

[0066] In the first method, the rotation angle φι is chosen such that it minimizes the sum or the average of absolute phase difference between the Zth columns of the H^e _k [ and U^e _kl_{x t} , given by h^e/f'¹ = H_{k i}w_k ^l _i and h^e _k^_t = H_{k i}v_k ^l _ie^¹, respectively. This can be formulated using an optimization problem as ) = arq min d^Tl Equation 1

^{1 a} θε[0,2π] ^

[0067] Where d = p - and p = (angle(H_{k i}w_k ^l _i)—

angle{H_{k+l i}v_{k+1 <έ})— Θ1) with 1 = [1 ... 1]^T denoting a column vector of size N_R . In the above, the p vector accounts for the phase difference on the Z^th columns of the effective channel on two (adjacent) subcarrier indices k and k + 1 after rotating the Z^th column of the SVD based precoding matrix of (fc + 1) subcarrier by Θ radians. The vector d denotes the wrapping of the phase differences between — π and π radians and taking the absolute values of the resulting phase differences.

[0068] The principle of this method is to rotate the Z^th column of the SVD based precoding matrix corresponding to the (k + l)th subcarrier such that the sum or the average of the absolute phase variation in the effective channel seen on subcarrier indices k and k + 1 across all of the receiving antennas corresponding to the Z^th transmit antenna is minimized.

[0069] The solution to optimization problem in Equation 2 can be obtained by sampling the angular space or using other efficient search methods.

[0070] In embodiments described above the processor 12, using the second alternative method for determining rotation angles, is configured to determine the rotation angle φι such that the Euclidean distance between effective channel vectors of the two adjacent subcarriers seen at all receiving antennas of the receiving device corresponding to the Z th transmit antenna is minimized. It is advantageous to determine a rotation angle based on minimizing the Euclidean distance between effective channel vectors of the two adjacent subcarriers as the channel magnitude changes in a smooth manner across all the subcarriers of an OFDM symbol.

[0071 ] In the second method, the rotation angle φι is chosen such that it minimizes the Euclidean distance between the effective channel vectors seen on subcarrier indices k and k + 1 of the i th MIMO-OFDM symbol across all the receiving antennas corresponding to the Zth transmit antenna is minimized. This can be formulated as

φ_ι = arg min

- H_k+liiv_k ^l _+1>ie>^e \\ Equation 2

[0072] If we denote H _iw_k ^l _i = a and H_k+ v_k ^l _+U = b , then the optimization problem in Equation 2 can equivalently be written as

φ_ι = arg ^max ^ Re{a^Hbe^ie) Equation 3

[0073] Where Re(. ) denotes the real part of a complex number. Now, letting a^Hb = x + j^'y, we can write Re(a^Hbe^{je ^}) = (x * cos Θ— y sin Θ) hence we can write the optimization problem in Equation 3 as

φι = arg ^ max ^(x cos 0— y sin #) Equation 4

[0074] The solution for the optimization problem in Equation 4 can be expressed in closed-form. The motivation for the formulation in Equation 2 is that since the channel changes slowly across the bandwidth, we try to minimize the channel variation across two adjacent subcarriers.

[0075] In figure 4, an example of a method for transmitting data using at least one precoding matrix is disclosed. In the method, first, reference signals on a reverse link are received from a receiver, step 20. Then estimates of the channel of the reverse link for a plurality of subcarriers are obtained, step 21. Based on the obtained estimates of the reverse link, a first precoding matrix for the first subcarrier of the plurality of subcarriers is generated, step 22. Then at least one data stream is transmitted on the first subcarrier using multiple transmit antennas and the first precoding matrix, step 23. The method according to the second aspect further comprises either rotating columns of the first precoding matrix to obtain a first rotated precoding matrix for a second subcarrier of the plurality of subcarriers so that phase variation in effective channels seen at the receiving device across the first and second subcarriers is minimized, step 24 or generating a second precoding matrix based on at least one of the obtained estimates of the reverse link and rotate columns of the second precoding matrix to obtain a second rotated precoding matrix so that phase variation in effective channels seen at the receiving across the first and second subcarriers is minimized, step 25. Lastly at least one data stream is transmitted on the second subcarrier using multiple transmit antennas using either the first rotated precoding matrix or second rotated precoding matrix, step 26.

[0076] It should be noted that even if the method illustrated in figure 4 is a sequential method, typically it is executed in parallel. Thus, subcarriers are sent simultaneously. Furthermore, if a precoding matrix is generated for each subcarrier these precoding matrices may be generated and rotated simultaneously. If only one precoding matrix is generated for the whole sub-band, the generation and rotation is done in sequential order so that each generation and rotation being a perquisite for subsequent rotations is done on time.

[0077] According to the example of figure 4 a precoding matrix for the second subcarrier is generated either by rotating the columns of the first precoding matrix or by generating a second precoding matrix and rotating the columns of the generated second precoding matrix, which are used to precode at least one data stream. Then a transmitter of a network node transmits the precoded at least one data stream using multiple antennas.

[0078] It is advantageous to use a precoding approach described above as it allows minimizing the phase variations in the effective channel seen at the receiver across the subcarriers of an OFDM symbol. This facilitates the use of a wideband channel estimation algorithm at the receiver possibly resulting in a better performance with reduced complexity. [0079] As it can be seen in the figure 4, steps 24 and 25 are optional to each other and applied similarly as in the example of figure 3a and 3b described above. These two optional steps may be implemented in a similar device, however, when the option has been chosen it will be used for communication.

[0080] According to the example of figure 4, when step 25 is executed, the method further comprises obtaining an estimate of the channel of the reverse link for the first subcarrier and generating the first precoding matrix based on the obtained estimate of the first subcarrier; and obtaining an estimate of the channel of the reverse link for the second subcarrier and to generate the second precoding matrix based on the obtained estimate of the second subcarrier. This is advantageous to generate the precoding matrix for each of the two subcarriers based on the respective estimates as this approach provides a better precoding matrix for each of the subcarriers in terms of maximizing the precoding gain. Furthermore, it is advantageous to generate the second matrix for the second subcarrier based on the respective estimate as it provides more accurate starting point for rotating the second precoding matrix to minimize the phase variations in the effective channel seen at the receiving device across the two subcarriers.

[0081 ] According to the example of figure 4, when step 25 is executed, the method further comprises generating a precoding matrix for each subcarrier in an OFDM symbol and rotating the second and subsequent precoding matrices. This is advantageous as the described method is applied for each subcarrier in an OFDM symbol so that whole OFDM symbol can be received with phase continuity of the effective channel seen at the receiving device.

[0082] According to the example of figure 4, when step 24 is executed, rotating columns of the first precoding matrix further comprises obtaining a combined estimate of the channel of the reverse link for a sub-band formed by a plurality of subcarriers and generating the first precoding matrix based on the obtained estimate of the sub-band. This is advantageous because it provides a simplified option of using a combined estimate to obtain precoding matrices. The combined estimate provides a possibility to use only one precoding matrix that is then rotated for each of the subcarriers.

[0083] According to the example of figure 4, when step 24 is executed, the method further comprises rotating the first precoding matrix for the second and each subsequent subcarrier in an OFDM symbol. This is advantageous because it provides a simplified option of using a combined estimate to obtain precoding matrices. The combined estimate provides a possibility to use only one precoding matrix that is then rotated for each of the subcarriers.

[0084] The method described above may be implemented as a computer program comprising a program code configured to perform the method as described above when executed in a transmitting device is disclosed. It is advantageous to provide a computer program for implementing the method as it is possible to enable the transmission as described above in devices that already have all necessary hardware components required for the transmission.

[0085] Figure 5 illustrates results achieved using the first embodiment. The figure illustrates the probability that the phase difference in the effective channels or effective channel vectors of two adjacent subcarriers corresponding to a transmit- receive antenna pair exceeds a given value X (in degrees). For comparison, the figure also illustrates a curve corresponding to the SVD based precoding method in which a right-singular matrix of the estimated channel matrix is used as a precoder for each subcarrier (without any rotation). The label 'Methodl ' in the figure corresponds to the case in which the SVD based precoding matrix is rotated in each subcarrier such that the sum of absolute phase differences between effective channels of two adjacent subcarriers is minimized. Similarly, the label 'Method2' corresponds to the case in which the SVD based precoding matrix is rotated in each subcarrier such that the Euclidean distance between effective channels of two adjacent subcarriers is minimized. As can be noted from the figure, the described methods minimize the phase variation in the effective channel compared to the SVD based precoding method without rotation. [0086] The system and method for wireless communication network synchronization has been described in conjunction with various embodiments herein. However, other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. A single processor or other unit may fulfill the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measured cannot be used to advantage. A computer program may be stored/distributed on a suitable medium, such as an optical storage medium or a solid-state medium supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems.

Claims

1. A network node comprising:

a transceiver (11) configured to:

receive reference signals in a reverse link from a receiving device;

a processor (12) configured to:

obtain estimates of the channel of the reverse link for a plurality of subcarriers;

generate a first precoding matrix for a first subcarrier of the plurality of subcarriers based on at least one of the obtained estimates of the reverse link; and rotate columns of the first precoding matrix to obtain a first rotated precoding matrix for a second subcarrier of the plurality of subcarriers so that phase variation in effective channels seen at the receiving device across the first and second subcarriers is minimized, or

generate a second precoding matrix based on at least one of the obtained estimates of the reverse link and rotate columns of the second precoding matrix to obtain a second rotated precoding matrix so that phase variation in effective channels seen at the receiving device across the first and second subcarriers is minimized; and

wherein the transceiver (11) is configured to transmit a data stream to the receiving device on the first subcarrier using the first precoding matrix and on the second subcarrier using either the first rotated precoding matrix or the second rotated precoding matrix.

2. The network node according to claim 1, wherein in the case in which the processor (12) is configured to obtain the second rotated precoding matrix, the processor (12) is further configured to: obtain an estimate of the channel of the reverse link for the first subcarrier and generate the first precoding matrix based on the obtained estimate of the first subcarrier.

3. The network node according to claim 1 or 2, wherein in the case in which the processor (12) is configured to obtain the second rotated precoding matrix: the processor is further configured to obtain an estimate of the channel of the reverse link for the second subcarrier and to generate the second precoding matrix based on the obtained estimate of the second subcarrier.

4. The network node according to claim 1, 2 or 3, wherein the processor (12) is configured to generate a precoding matrix for each subcarrier in an OFDM symbol, wherein the processor is further configured to rotate the second and subsequent precoding matrices.

5. The network node according to claim 1, wherein in the case in which the processor (12) is configured to obtain the first rotated precoding matrix the processor is further configured to obtain a combined estimate of the channel of the reverse link for a sub-band formed by a plurality of subcarriers and to generate the first precoding matrix based on the obtained estimate of the sub-band.

6. The network node according to any of claims 1 or 5, wherein the processor (12) is configured to perform the rotation of the first precoding matrix for the second and each subsequent subcarrier in an OFDM symbol.

7. The network node according to any preceding claims 1 - 6, wherein the processor (12) is configured to generate the precoding matrices based on singular value decomposition of the channel estimates obtained in the reverse link.

8. The network node according to any preceding claims 1 - 7 wherein the processor (12) is configured to determine a rotation angle φι for rotating the Zth column of each precoding matrix the processor is configured to rotate.

9. The network node according to claim 8, wherein the processor (12) is configured to determine the rotation angle φι such that the sum of the absolute phase difference in the effective channel vectors of the two adjacent subcarriers seen at all receiving antennas of the receive device corresponding to the Zth transmit antenna is minimized.

10. The network node according to claim 8, wherein the processor (12) is configured to determine the rotation angle φι such that the Euclidean distance between effective channel vectors of the two adjacent subcarriers seen at all receiving antennas of the receiving device corresponding to the Zth transmit antenna is minimized.

11. A method for transmitting data using at least one precoding matrix, the method comprising:

receiving (20) reference signals on a reverse link from a receiver;

obtaining (21) estimates of the channel of the reverse link for a plurality of subcarriers;

generating (22) a first precoding matrix for the first subcarrier of the plurality of subcarriers based on the obtained estimates of the reverse link;

transmitting (23) a data stream to the receiver on the first subcarrier using the first precoding matrix, and wherein the method further comprises:

rotating (24) columns of the first precoding matrix to obtain a first rotated precoding matrix for a second subcarrier of the plurality of subcarriers so that phase variation in effective channels seen at the receiving device across the first and second subcarriers is minimized, or generating (25) a second precoding matrix based on at least one of the obtained estimates of the reverse link and rotate columns of the second precoding matrix to obtain a second rotated precoding matrix so that phase variation in effective channels seen at the receiving device across the first and second subcarriers is minimized; and

transmitting (26) a data stream to the receiver on the second subcarrier using either the first rotated precoding matrix or second rotated precoding matrix.

12. The method according to claim 11, wherein the method further comprises:

obtaining an estimate of the channel of the reverse link for the first subcarrier and generating the first precoding matrix based on the obtained estimate of the first subcarrier; and

obtaining an estimate of the channel of the reverse link for the second subcarrier and generating the second precoding matrix based on the obtained estimate of the second subcarrier.

13. The method according to claim 11 or 12, wherein the method further comprises generating a precoding matrix for each subcarrier in an OFDM symbol and rotating the second and subsequent precoding matrices.

14. The method according to claim 11, wherein the method further comprises obtaining a combined estimate of the channel of the reverse link for a sub-band formed by a plurality of subcarriers and generating the first precoding matrix based on the obtained estimate of the sub-band.

15. The method according to claim 11 or 14, wherein the method further comprises rotating the first precoding matrix for the second and each subsequent subcarrier in an OFDM symbol.

16. The method according to any preceding claim 11 - 15, wherein the method further comprises determining a rotation angle φι for rotating the Zth column of each precoding matrix.

17. A computer program comprising a program code for performing a method according to any one of claims 11 - 16, when said computer program runs on a computer.