WO2011085695A1 - 获取预编码矩阵指示以及预编码矩阵的方法和装置 - Google Patents

获取预编码矩阵指示以及预编码矩阵的方法和装置 Download PDF

Info

Publication number
WO2011085695A1
WO2011085695A1 PCT/CN2011/070323 CN2011070323W WO2011085695A1 WO 2011085695 A1 WO2011085695 A1 WO 2011085695A1 CN 2011070323 W CN2011070323 W CN 2011070323W WO 2011085695 A1 WO2011085695 A1 WO 2011085695A1
Authority
WO
WIPO (PCT)
Prior art keywords
differential
precoding matrix
codebook
differential codebook
matrix indication
Prior art date
Application number
PCT/CN2011/070323
Other languages
English (en)
French (fr)
Inventor
王建国
周永行
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Priority to EP11732685.0A priority Critical patent/EP2525505B1/en
Priority to EP15182630.2A priority patent/EP2985924B1/en
Publication of WO2011085695A1 publication Critical patent/WO2011085695A1/zh
Priority to US13/549,810 priority patent/US8848768B2/en
Priority to US13/730,044 priority patent/US8625711B2/en
Priority to US14/340,971 priority patent/US9054754B2/en
Priority to US14/713,502 priority patent/US9350432B2/en

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04BTRANSMISSION
    • H04B7/00Radio transmission systems, i.e. using radiation field
    • H04B7/02Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/0413MIMO systems
    • H04B7/0456Selection of precoding matrices or codebooks, e.g. using matrices antenna weighting
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04BTRANSMISSION
    • H04B7/00Radio transmission systems, i.e. using radiation field
    • H04B7/02Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/0413MIMO systems
    • H04B7/0456Selection of precoding matrices or codebooks, e.g. using matrices antenna weighting
    • H04B7/0478Special codebook structures directed to feedback optimisation
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04BTRANSMISSION
    • H04B7/00Radio transmission systems, i.e. using radiation field
    • H04B7/02Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/06Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station
    • H04B7/0613Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission
    • H04B7/0615Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission of weighted versions of same signal
    • H04B7/0619Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission of weighted versions of same signal using feedback from receiving side
    • H04B7/0636Feedback format
    • H04B7/0641Differential feedback
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00Baseband systems
    • H04L25/02Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/03Shaping networks in transmitter or receiver, e.g. adaptive shaping networks
    • H04L25/03891Spatial equalizers
    • H04L25/03898Spatial equalizers codebook-based design
    • H04L25/0391Spatial equalizers codebook-based design construction details of matrices
    • H04L25/03923Spatial equalizers codebook-based design construction details of matrices according to the rank
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00Baseband systems
    • H04L25/02Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/03Shaping networks in transmitter or receiver, e.g. adaptive shaping networks
    • H04L25/03891Spatial equalizers
    • H04L25/03898Spatial equalizers codebook-based design
    • H04L25/03936Spatial equalizers codebook-based design multi-resolution codebooks
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04BTRANSMISSION
    • H04B7/00Radio transmission systems, i.e. using radiation field
    • H04B7/02Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/0413MIMO systems
    • H04B7/0456Selection of precoding matrices or codebooks, e.g. using matrices antenna weighting
    • H04B7/046Selection of precoding matrices or codebooks, e.g. using matrices antenna weighting taking physical layer constraints into account
    • H04B7/0465Selection of precoding matrices or codebooks, e.g. using matrices antenna weighting taking physical layer constraints into account taking power constraints at power amplifier or emission constraints, e.g. constant modulus, into account
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04BTRANSMISSION
    • H04B7/00Radio transmission systems, i.e. using radiation field
    • H04B7/02Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/0413MIMO systems
    • H04B7/0456Selection of precoding matrices or codebooks, e.g. using matrices antenna weighting
    • H04B7/046Selection of precoding matrices or codebooks, e.g. using matrices antenna weighting taking physical layer constraints into account
    • H04B7/0469Selection of precoding matrices or codebooks, e.g. using matrices antenna weighting taking physical layer constraints into account taking special antenna structures, e.g. cross polarized antennas into account
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04BTRANSMISSION
    • H04B7/00Radio transmission systems, i.e. using radiation field
    • H04B7/02Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/06Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station
    • H04B7/0613Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission
    • H04B7/0615Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission of weighted versions of same signal
    • H04B7/0619Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission of weighted versions of same signal using feedback from receiving side
    • H04B7/0636Feedback format
    • H04B7/0639Using selective indices, e.g. of a codebook, e.g. pre-distortion matrix index [PMI] or for beam selection

Definitions

  • the present invention relates to the field of communications technologies, and in particular, to a method and apparatus for obtaining a precoding matrix indication and a precoding matrix. Background technique
  • the data sender (such as
  • Node B Node B
  • BS Base station, etc.
  • PMI Precoding Matrix Indicator, Pre The coding matrix indicates
  • PMI Precoding Matrix Indicator, Pre The coding matrix indicates
  • PMI Precoding Matrix Indicator, Pre The coding matrix indicates
  • PMI Precoding Matrix Indicator, Pre The coding matrix indicates
  • PMI Precoding Matrix Indicator, Pre The coding matrix indicates
  • PMI Precoding Matrix Indicator, Pre The coding matrix indicates
  • the precoding matrix which is preprocessed by the precoding matrix and then sent to the data receiving end, so that the data transmission process can be adaptively changed by the channel state, thereby improving the data.
  • the performance of the transmission Therefore, how to get the PMI and what kind of codebook to use is very important.
  • the existing LTE R8 Long Term Evolution Release 8
  • the existing IEEE The Institute of Electrical and Electronics Engineers 802.16m system
  • two non-differential and differential codebooks are used, mainly using differential mode for feedback.
  • the feedback process is as follows: During the feedback period, the UE first feeds back a non-differential PMI based on the non-differential codebook, and then sequentially feeds back the differential PMI based on the differential codebook according to the non-differential PMI.
  • the inventors have found that at least the following problems exist in the prior art:
  • a fixed single codebook is used, the subband feedback overhead is large, and the broadband feedback accuracy is low.
  • the existing IEEE 802.16m system uses time domain differential feedback. The acquisition of the precoding matrix depends on the historical information of the feedback, which leads to error propagation problems.
  • each code included in the differential codebook used by the existing IEEE 802.16m system The characteristics of the elements in the word will make the resulting precoding matrix not necessarily have constant model, especially it is difficult to satisfy the finite character set (such as 8PSK (Phase Shift Keying)) constraint characteristics of each element;
  • the real risk discovery when the above differential feedback form directly uses the codebook in the LTE R8 system as the non-differential codebook, and uses the differential codebook of the IEEE 802.16m system, the overhead is large, compared to the feedback using only the LTE R8 codebook. At the time, its improvement in feedback performance is not large, resulting in inefficiency of each feedback bit. Therefore, it is necessary to enter one Step by step study the structure of the feedback and the codebook used to improve the feedback performance of the system. Summary of the invention
  • an embodiment of the present invention provides a method and apparatus for obtaining a precoding matrix indication and a precoding matrix.
  • the technical solution is as follows:
  • an embodiment of the present invention provides a method for obtaining a precoding matrix indication, where the method includes:
  • an embodiment of the present invention provides a method for obtaining a precoding matrix, where the method includes:
  • the reference precoding matrix indication and the differential precoding matrix indication are determined by the data receiving end according to the first non-differential codebook and the first diagonal differential codebook Obtaining and transmitting; the codeword included in the first diagonalized differential codebook is a diagonalization matrix;
  • Decoding a precoding matrix by using a second pre-stored second non-differential codebook and a second diagonally differentiated codebook according to the reference precoding matrix indication and the differential precoding matrix indication, where the second non- The differential codebook and the second diagonalized differential codebook are respectively consistent with the first non-differential codebook and the first diagonalized differential codebook.
  • an embodiment of the present invention provides an apparatus for obtaining a precoding matrix indication, where the apparatus includes:
  • a precoding matrix indication acquiring module configured to obtain a reference precoding matrix indication and a differential precoding matrix indication according to the first non-differential codebook and the first diagonalized differential codebook, where the first diagonalized differential code
  • the codewords included in this document are diagonalized matrices.
  • an embodiment of the present invention provides an apparatus for acquiring a precoding matrix, where the apparatus includes:
  • a precoding matrix indication receiving module configured to receive a reference precoding matrix indication and a differential precoding matrix indication; wherein, the reference precoding matrix indication and the differential precoding matrix indication are determined by the data receiving end according to the first non-differential codebook And the first diagonalized differential codebook is obtained and sent; the first The codewords included in the diagonalized differential codebook are diagonalized matrices;
  • a precoding matrix obtaining module configured to: after the precoding matrix indication receiving module receives the reference precoding matrix indication and the differential precoding matrix indication, according to the reference precoding matrix indication and the differential precoding The matrix indicates that the pre-coding matrix is obtained by using the second pre-stored second non-differential codebook and the second diagonally differentiated codebook, wherein the second non-differential codebook and the second diagonally differentiated codebook Consistent with the first non-differential codebook and the first diagonalized differential codebook, respectively.
  • an embodiment of the present invention provides a method for obtaining a precoding matrix indication, where the method includes:
  • the non-differential precoding matrix indication is calculated based on the preset criterion; wherein the codeword included in the first non-differential codebook is obtained from the rotated Hadamard matrix.
  • an embodiment of the present invention provides a method for obtaining a precoding matrix, where the method includes:
  • Non-differential precoding matrix indication Receiving a non-differential precoding matrix indication; wherein the non-differential precoding matrix indication is obtained and transmitted by the data receiving end according to the first non-differential codebook; the codeword included in the first non-differential codebook is rotated from the Hadamard Matrix is obtained;
  • an embodiment of the present invention provides an apparatus for obtaining a precoding matrix indication, where the apparatus includes:
  • a non-differential precoding matrix indication acquiring module configured to calculate, according to the first non-differential codebook, a non-differential precoding matrix indication according to a preset quasi-side, where the codeword included in the first non-differential codebook Obtained from the rotating Hadamard matrix.
  • an embodiment of the present invention provides an apparatus for acquiring a precoding matrix, where the apparatus includes:
  • a non-differential precoding matrix indication receiving module configured to receive a non-differential precoding matrix indication, where the non-differential precoding matrix indication is obtained and transmitted by the data receiving end according to the first non-differential codebook; the first non-differential The codewords included in the codebook are obtained from a rotating Hadamard matrix;
  • a non-differential precoding matrix processing module configured to: after the precoding matrix indication receiving module receives the non-differential precoding matrix indication, use the non-differential precoding matrix indication Pre-stored second non-differential codebook, to obtain a precoding matrix, wherein the second non-differential codebook is consistent with the first non-differential codebook.
  • the reference PMI and the differential PMI are obtained, which can save feedback overhead and improve feedback precision, thereby further improving feedback performance;
  • the codewords included in the diagonal differential codebook are diagonal
  • the matrix can maintain the amplitude characteristics of the elements of the non-differential codebook itself (such as constant model property, finite character set constraint characteristics) or facilitate the power distribution between the antennas.
  • FIG. 1 is a flowchart of a method for obtaining a precoding matrix indication according to Embodiment 1 of the present invention
  • FIG. 2 is a flowchart of a method for obtaining a precoding matrix indication according to Embodiment 2 of the present invention
  • FIG. 4 is a flowchart of a method for obtaining a precoding matrix indication according to Embodiment 4 of the present invention
  • FIG. 5 is a flowchart of a method for obtaining a precoding matrix indication according to Embodiment 4 of the present invention
  • FIG. 6 is a schematic structural diagram of an apparatus for acquiring a precoding matrix indication according to Embodiment 8 of the present invention
  • FIG. 7 is a schematic structural diagram of an apparatus for acquiring a precoding matrix according to Embodiment 9 of the present invention
  • FIG. 8 is a flowchart of a method for acquiring a precoding matrix indication according to Embodiment 10 of the present invention
  • FIG. 11 is a schematic diagram of a device structure for obtaining a precoding matrix indication
  • FIG. 10 is a flowchart of a method for acquiring a precoding matrix according to Embodiment 12 of the present invention
  • FIG. 11 is a schematic structural diagram of an apparatus for acquiring a precoding matrix according to Embodiment 13 of the present invention.
  • an embodiment of the present invention provides a method for obtaining a precoding matrix indication, including: 101: Acquire a reference precoding matrix indication and a differential preamble according to a first non-differential codebook and a first diagonalized differential codebook.
  • An encoding matrix indication, where the codeword included in the first diagonalized differential codebook is a pair Keratinization matrix.
  • the obtaining the reference precoding matrix indication and the differential precoding matrix indication according to the first non-differential codebook and the first diagonalized differential codebook may specifically include:
  • Obtaining the reference precoding matrix indication and the differential precoding matrix indication according to the first non-differential codebook and the first diagonalized differential codebook which may specifically include:
  • a reference precoding matrix indication and a differential precoding matrix indication are calculated based on a preset criterion.
  • yt 0, l...
  • Z indicates that the first diagonalized differential codebook D is included
  • k
  • the first N/2 transmit antennas of the dual-polarized transmit antenna array including N transmit antennas are a set of co-polarized transmit antennas
  • the latter N/2 transmit antennas are another set of co-polarized transmit antennas
  • e j represents the differential precoding matrix indication included in the first diagonalized differential codebook D
  • fc 0, l...
  • represents the number of codewords included in the first diagonalized differential codebook D
  • -1 represents the number of transmit antennas
  • N is an even number
  • mi represents phase shift
  • m and "is a natural number, l, 2,..., N/2.
  • the structure of the first non-differential code can be as follows:
  • r represents the rank of the first codeword W w non-differential codebook included
  • 4 denotes an index corresponding to a diagonal matrix
  • R represents i k corresponds to the n-th order diagonal matrix
  • denotes the index of the corresponding normalized Hadamard matrix
  • ⁇ ⁇ denotes the corresponding n-th order normalized Hadamard matrix
  • the value of ⁇ is equal to the number of transmitting antennas.
  • n is equal to the number of transmitting antennas.
  • the r column is selected from the rotating Hadamard matrix R IT H Rail A to form a matrix (R 4 H Rail ; T ) W , and the specific method can be used to determine which r column to select according to the actual application situation. For example: For non-correlated channels, the choice of columns should be such that each set of r-column sub-matrices is extracted from a different matrix R ; H File , ⁇ (R 4 H Rail, ;t ⁇ each matrix satisfies the maximum chord distance; For the relevant channel, the column should be selected such that a set of r column sub-matrices is extracted from different matrices 1 ; ⁇ vide
  • the structure of the codeword included in the first non-differential codebook can be as follows:
  • W w represents the number of code words included in the first non-differential codebook W w
  • r represents the rank of the codeword included in the first non-differential codebook W w
  • 4 represents the index of the corresponding diagonal matrix R ; , R ; means with 4 pairs
  • the n-order diagonal matrix, A denotes the index of the corresponding normalized Hadamard matrix
  • ⁇ adjuvant ⁇ denotes the corresponding n-order normalized Hadamard matrix
  • the value of ⁇ is equal to the number of transmitting antennas.
  • the r-column is selected from the rotating Hadamard matrix R ; H Rail, R to form a matrix.
  • the method may further include:
  • the data transmitting end receives the reference precoding matrix indication and the differential precoding matrix indication, and utilizes the second non-differential codebook and the second diagonalized differential codebook pre-stored in itself, according to the reference precoding matrix indication and the difference
  • the precoding matrix indicates that a precoding matrix is obtained.
  • the second non-differential codebook and the second diagonalized differential codebook are respectively consistent with the first non-differential codebook and the first diagonalized differential codebook.
  • the first non-differential codebook is a first base codebook
  • the first diagonalized differential codebook is a first transform codebook
  • the second non-differential codebook is a second base codebook
  • the second diagonalized differential codebook is a second transform codebook
  • the method for obtaining the precoding matrix indication obtains the reference PMI and the differential PMI, which can save feedback overhead and improve feedback precision, thereby further improving feedback performance;
  • the codeword included in the differential codebook is a diagonalized matrix, which can maintain the amplitude characteristics of the elements already possessed by the non-differential codebook (such as constant model property, limited character set constraint characteristics) or facilitate power distribution between the antennas;
  • the non-differential codebook obtained from the rotating Hadamard matrix can be used, wherein the columns of the Hadamard matrix can be matched with the eigenvectors of the strong correlation and the low correlation channel respectively in the configuration of the uniform hook line array antenna and the dual polarization antenna configuration.
  • the rotation matrix uses a diagonalization matrix to keep the space quantified on the basis of the Hadamard matrix quantization.
  • the elements of the Hadamard matrix are +1 or -1, which can satisfy Constant model, and the multiplication of the Hadamard matrix with other matrices or vectors can be simplified to addition or subtraction, which can greatly reduce the CIRI (Channel Quality Indicator) based on SINR (Signal to Interference Noise Ratio). Quality indication) Computational complexity of calculations, PMI selection, and rank adaptation.
  • the columns of the Hadamard matrix are orthogonal to each other.
  • the ⁇ matrix characteristics of the codewords in the non-differential codebook can be maintained, and the power distribution of each spatial stream in the transmission can be ensured to be the same. In the case of full rank, the power distribution of each transmitting antenna is the same;
  • the method of obtaining a non-differential codebook from a rotating Hadamard matrix can be applied to a number of transmitting antennas of 2, 4, 8, 16, 32, 64, etc., and the obtained non-differential codebook can satisfy the constant model property and the matrix property. And the computational complexity of SINR-based CQI calculation, PMI selection, and rank adaptation can be reduced.
  • an embodiment of the present invention provides a method for obtaining a precoding matrix indication, including: 201: A UE selects a differential PMI based on a reference PMI for feedback, and obtains a reference PMI.
  • the reference PMI may be: The non-differential PMI that is recently fed back by the UE is used as a reference PMI, and may be a non-differential broadband PMI or a non-difference molecular band PMI, and may be selected according to actual application conditions.
  • the reference PMI is in the form of a non-differential PMI, and the reference PMI can be obtained by using at least one of the following two methods:
  • the reference PMI is calculated as n (can be recorded as reference PMI n ), as shown in equation (1):
  • W W represents the magnitude of non-differential codebook W w, i.e., number of non-codewords differential codebook W w included;
  • R & lt represents the rank non-differential codebook W w included in the code word,
  • represents a non- W w differential codebook corresponding to the PMI and the reference codeword; represents the objective function corresponding to preset criteria.
  • preset criterion may be a throughput maximization criterion
  • the target function corresponding to the criterion may be a throughput maximization function, which may be implemented based on information capacity calculation, or may be based on mutual The transformation of information or mutual information (such as the weighting of mutual information) is implemented.
  • the preset criterion may also be a chord maximization criterion.
  • the objective function corresponding to the preset criterion may be flexibly set according to the actual application condition, which is not specifically limited. The presupposition criteria in other places are similar here, no longer - repeat.
  • the non-differential codebook w w may be a non-differential codebook set by the embodiment of the present invention, and the codeword in the non-differential codebook w w is obtained from a rotated Hadamard matrix, specifically a non-differential codebook w w
  • the structure of the code word w ⁇ ) included in it can be as follows: Wherein, indicating a non-differential precoding matrix indication included in the non-differential codebook w w ,
  • r represents the rank codewords non-differential codebook comprises, expressed with the corresponding diagonal matrix R; t index, R; t represents 4 corresponding The n-th order diagonal matrix, A represents the index of the corresponding normalized Hadamard matrix, ⁇ condiment ⁇ denotes the corresponding n-th order normalized Hadamard matrix, and the value of ⁇ is equal to the number of transmitting antennas.
  • n is equal to the number of transmit antennas.
  • the r-column is selected from the rotating Hadamard matrix RH ⁇ ⁇ to form a matrix (R 4 H Rail ⁇ f, depending on the actual application, the relevant method is used to determine which r-column to choose, for example:
  • the column should be selected such that a set of r-column sub-matrices is extracted from different matrices R ; H Directory], ⁇ ( Rit H Computer, ;t ⁇ each matrix satisfies the maximum chord distance; for the relevant channel, The column should be selected such that a set of r-column sub-matrices is extracted from different matrices 1 ; ⁇ vide
  • Each matrix in ⁇ (R 4 H Hook, ;t ⁇ satisfies the maximum gain in the zero direction of the space projection of the antenna array, and can also be selected by any other feasible method, which is not specifically limited.
  • the diagonal element of R can be selected to be consistent with the 8-point DFT (Discrete Fourier Transform) vector, as follows
  • R 0 ⁇ ⁇ 1,1, 1,1, 1,1, 1,1 ⁇
  • R v diag ⁇ - ⁇ -jXj - ⁇ -j ⁇
  • R 3 diagU,ee 4 ,e 4 ,-1,-e 4 ,-e 4 ,-e 4
  • rank precoding codebook can be given as shown in Table 1 or Table 2 below.
  • low rank precoding codebooks may be derived from the full rank precoding codebook described above, subject to nesting characteristics.
  • the structure of the codeword w included in the non-differential codebook can be as follows:
  • n is equal to the number of transmitting antennas
  • R ; H ", Rf f) represents the Hadamard matrix rotation r ,. H", r r columns selected matrix configuration.
  • the r-column is selected from the rotating Hadamard matrix R ; H Rail, R to form a matrix.
  • Each matrix in I V 3 ⁇ 4 ⁇ I satisfies the maximum gain in the zero direction of the space projection of the antenna array, and can be selected by any other feasible method, which is not specifically limited.
  • any other feasible method which is not specifically limited.
  • a 4-bit 8-antenna non-differential codebook there is
  • H sn , H sl , HH are as follows:
  • the diagonal elements of R can be selected to be consistent with the 8-point DFT vector, as follows:
  • a full rank precoding codebook can be given as shown in Table 3 or Table 4 below.
  • low rank precoding codebooks may be derived from the full rank precoding codebook described above, subject to nesting characteristics.
  • structure of the codewords included in the non-differential codebook can be as follows:
  • n is equal to the number of transmitting antennas.
  • non-differential codebook can also use any non-differential codebook in the prior art, for example
  • the codebook of the LTE R8 can be flexibly selected according to the actual application, and is not specifically limited.
  • the Node B may indicate that the UE uses the reference PMI-based differential PMI to perform feedback through the high-layer signaling or the downlink physical control channel, and after receiving the indication of the Node B, the UE selects to use the reference PMI-based differential PMI for feedback.
  • the Node B and the UE agree in advance that the UE uses the differential PMI based on the reference PMI for feedback, and when the UE performs feedback, automatically selects the differential PMI based on the reference PMI for feedback. It is not limited to the above two methods, and any other feasible method can be set according to the actual application condition, so that the UE can choose to use the reference PMI based differential PMI for feedback.
  • the UE feeds back the reference PMI to the NodeB, and calculates a differential PMI according to the preset criterion according to the reference PMI, the non-differential codebook W w, and the diagonalized differential codebook 0.
  • the non-differential codebook W w can use a non-differential codebook similar to that in step 201, that is,
  • the non-differential codebook obtained from the rotated Hadamard matrix of the codeword set in the embodiment of the present invention may also be any non-differential codebook in the prior art.
  • the diagonalized differential codebook is a differential codebook provided in the embodiment of the present invention, and the codeword included in the diagonalized differential codebook D set in the embodiment of the present invention is C 3 ⁇ 4
  • C represents D diagonalization differential codebook with PMI ⁇ ⁇ corresponding to the differential codeword
  • W represents a non-differential
  • the codeword corresponding to the reference PMI n in the codebook W w ; /(C; tenu) represents the objective function corresponding to the preset criterion.
  • the UE feeds back the differential PMI to the NodeB.
  • the differential PMI is , so specifically k is fed back to the NodeB.
  • the NodeB receives the reference PMI and the differential PMI, and calculates the precoding matrix V by using the non-differential codebook ⁇ ) and the diagonalized differential codebook D according to the reference PMI and the differential PMI.
  • the UE since the UE has fed back the reference PMI to the NodeB in step 202, the UE feeds the differential PMI to the NodeB in step 203, so the NodeB first receives the reference PMI and then receives the differential PMI, so the NodeB
  • the reference PMI received first may be stored, and when the differential PMI is received, the precoding matrix is calculated according to the reference PMI and the differential PMI.
  • the NodeB pre-stores the non-differential codebook W w from the NodeB according to the reference PMI (without pre-stored non-differential in the UE)
  • the codebook is consistently obtained.
  • the query obtains the codeword corresponding to the reference PMI, and according to the differential PMI, the diagonally differentiated differential codebook D pre-stored from the NodeB (consistent with the diagonally-distributed differential codebook pre-stored in the UE)
  • the query obtains the codeword corresponding to the differential PMI; multiplies the codeword corresponding to the differential PMI by the codeword corresponding to the reference PMI, and uses the product result as a precoding matrix.
  • reference PMI is n, the code word and the reference PMI n-corresponding to W "; differential PMI is k, a codeword difference PMIk corresponding to C 3 ⁇ 4, thus pre-coding matrix of the present invention implemented as Equation (4):
  • the NodeB uses the non-differential codebook W w and the diagonalized differential codebook according to the reference PMI and the differential PMI fed back by the respective UEs that are simultaneously paired.
  • D based on the ZF-BF (Zero-Forcing Beam Forming) algorithm or based on the maximum SLNR (Signal-Leakage Plus Noise Ratio) criterion to calculate the precoding matrix.
  • the corresponding differential PMI is 13 ⁇ 4 (and the codeword corresponding to the reference PMI 13 ⁇ 4 is C 3 ⁇ 41)
  • the reference PMI corresponding to another UE is n 2 (and the codeword corresponding to the reference PMI n 2 is W detergent 2 ), and the corresponding differential PMI is k 2 (and the codeword corresponding to the reference PMI k 2 is C k2 )
  • the method of the above SU-MIMO system is used to calculate a precoding matrix corresponding to one UE as shown in equation (5), and the precoding matrix 2 corresponding to another UE is as shown in equation (6):
  • the Node uses the precoding matrix to preprocess the data to be sent, and sends the preprocessed data to be sent to the UE through the transmitting antenna.
  • the UE receives the received signal and performs data detection on the received signal y.
  • the received signal y received by the UE is as shown in the formula (8):
  • the reference PMI and the differential PMI are respectively fed back to the Node B through the step 202 and the step 203.
  • the reference PMI and the differential PMI may be simultaneously fed back to the Node B, specifically, the UE.
  • the reference PMI may be stored first, and then in step 203, the UE simultaneously feeds back the reference PMI and the differential PMI to the NodeB.
  • step 201 according to the non-differential codebook W w and the diagonalized differential codebook D, based on the preset criterion, the reference PMI is calculated as n and the differential PMI is k, as shown in the formula (9). :
  • step 204 is directly executed.
  • the PMI may be fed back to the entire system bandwidth, or the system may be divided into multiple BPs (Bandwidth Part). , bandwidth part), each BP contains multiple sub-bands, one PMI for each sub-band or one PMI for multiple sub-bands (eg Best-M mode: feedback one PMI for the selected M sub-bands) .
  • the above reference PMI and differential PMI can be either a wideband PMI or a subband PMI, or a PMI can be fed back to multiple subbands according to the Best-M method.
  • the method for obtaining the precoding matrix indication obtains the reference PMI and the differential PMI according to the non-differential codebook and the diagonalized differential codebook, which can save feedback overhead, improve feedback precision, and improve feedback performance;
  • the codewords included in the diagonalized difference codebook are diagonalized matrices, and can maintain the amplitude characteristics of the elements that the non-differential codebook already has (such as constant model property, finite character set constraint characteristics).
  • a non-differential codebook obtained from a rotating Hadamard matrix can be used, wherein each column of the Hadamard matrix can be matched with a eigenvector of a strong correlation and a low correlation channel in a uniform linear array transmit antenna configuration and a dual polarization transmit antenna configuration, respectively. Therefore, the accuracy of the quantization is improved, and the rotation matrix is a diagonalization matrix, and the space can be finely quantized based on the Hadamard matrix quantization; and the elements of the Hadamard matrix are +1 or -1, which can satisfy the constant model.
  • the multiplication of the Hadamard matrix with other matrices or vectors can be simplified to addition or subtraction, which can greatly reduce the computational complexity of SINR-based CQI calculation, PMI selection and rank adaptation.
  • the columns of the Hadamard matrix are orthogonal to each other. After normalization, the ⁇ matrix characteristics of the codewords in the non-differential codebook can be maintained, and the power distribution of each spatial stream can be ensured to be the same. In the case of full rank, the power distribution of each transmit antenna is the same.
  • the method of obtaining a non-differential codebook from the rotating Hadamard matrix can be applied to 2, 4, 8, 16, 32, 64 and other number of transmit antennas and the obtained non-differential codebook can satisfy the above-mentioned constant model property, ⁇ matrix characteristics and reduce the computational complexity of SINR-based CQI calculation, PMI selection and rank adaptation;
  • Using the non-differential PMI recently feedbacked by the UE as the reference PMI can not only reduce the overhead, but also make full use of the correlation of the channel in the frequency domain and the time domain or in the airspace, thereby improving the feedback precision.
  • an embodiment of the present invention provides a method for obtaining a precoding matrix indication, including: 301: A UE selects a differential PMI based on a reference PMI for feedback, and obtains a reference PMI.
  • Reference PMI can be: Use the non-differential PMI and differential PMI that the UE recently fed back as a reference PML
  • the reference PMI may specifically include a non-differential PMI and m reference differential PMIs (m is a natural number, and the value of m may be selected according to actual application conditions, for example, the value of m may be selected according to the number of subbands) .
  • the non-differential PMI can be obtained by using the method in step 201 of Embodiment 2.
  • the m reference differential PMIs are respectively calculated as n Q , n!...n m (m reference differential PMIs are respectively recorded as reference differences PMI n.
  • the reference differential PMIn...reference differential PMIn m ) is specifically as shown in equation (10):
  • the UE feeds back the reference PMI to the NodeB, and calculates a differential PMI according to the preset criterion according to the reference PMI, the non-differential codebook W w, and the diagonalized differential codebook 0.
  • the non-differential codebook W w and the diagonalized differential codebook D are the same as the non-differential codebook W w and the diagonalized differential codebook D in the step 202 of the embodiment 2, and are not described herein again.
  • the differential PMI is calculated as (differential PMI k ) , as shown in equation (11):
  • C diagonalization represents the difference D in a differential codebook PMI ⁇ ⁇ corresponding codeword; Table C “C” ... C "
  • the UE feeds back the differential PMI to the NodeB.
  • the differential PMI is, so specifically k is fed back to the NodeB.
  • the NodeB receives the reference PMI and the differential PMI, and calculates the precoding matrix by using the non-differential codebook ⁇ ) and the diagonalized differential codebook D according to the reference PMI and the differential PMI.
  • the NodeB queries, according to the reference PMI, the codeword corresponding to the reference PMI from the non-differential codebook W w (which is consistent with the pre-stored non-differential codebook stored in the UE) pre-stored in the NodeB, And according to the differential PMI, the diagonally differentiated codebook D pre-stored in the NodeB (consistent with the diagonally differentiated codebook pre-stored in the UE) is used to query the codeword corresponding to the differential PMI; the code corresponding to the differential PMI The word is multiplied by a codeword corresponding to the reference PMI, and the product result is used as a precoding matrix.
  • the reference PMI includes a non-differential PMI and m reference differential PMIs, the non-differential PMI is n, and the m reference differential PMIs are respectively n. , n!...n m , the codeword corresponding to the non-differential PMI n is W raw, and the codewords corresponding to the reference difference PMI n., ! ⁇ ...! ⁇ are respectively ⁇ , C Economics-C tripod; differential PMI For k, the codeword corresponding to the differential PMI k is C 3 ⁇ 4 , so the precoding matrix obtained by the embodiment of the present invention is as shown in the formula (12):
  • the NodeB uses the non-differential codebook W w and the diagonalized differential codebook D according to the reference PMI and the differential PMI fed back by the respective UEs that are simultaneously paired, based on the ZF-BF algorithm or based on the maximum SLNR criterion. Encoding matrix.
  • the specific process is similar to the step 204 of Embodiment 2, and details are not described herein again.
  • the Node B uses the precoding matrix to preprocess the data to be sent, and sends the preprocessed data to be sent to the UE through the transmitting antenna.
  • the UE receives the received signal and performs data detection on the received signal y.
  • the received signal y received by the UE is as shown in the formula (13):
  • y HVs + n ( 13 )
  • y the received signal received by the UE
  • H the channel matrix
  • s the data to be transmitted
  • n additive white Gaussian noise
  • the reference PMI and the differential PMI are respectively fed back to the Node B through the step 302 and the step 303.
  • the reference PMI and the differential PMI may be simultaneously fed back to the Node B, specifically, the UE may be After the reference PMI is obtained, the reference PMI may be stored first, and then in step 303, the UE simultaneously feeds back the reference PMI and the differential PMI to the NodeB.
  • the reference PMI is calculated to be n, ⁇ at the same time.
  • ⁇ . ⁇ and differential PMI are as follows, as shown in equation (14a):
  • step 304 is directly executed.
  • the PMI when feeding back the reference PMI and the differential PMI, whether the feedback is a differential PMI or a reference PMI, the PMI may be fed back to the entire system bandwidth, or the system may be divided into multiple BPs, each BP contains multiple subbands, one PMI is fed back for each subband, or multiple subbands feed back one PMI (for example, Best-M mode: feedback one MMI for selecting M subbands). That is, the above reference PMI and differential PMI may be a broadband PMI, a sub-band PMI, or a PMI may be fed back to multiple sub-bands according to the Best-M method.
  • the method for obtaining the precoding matrix indication obtains the reference PMI and the differential PMI according to the non-differential codebook and the diagonalized differential codebook, which can save feedback overhead, improve feedback precision, and improve feedback performance;
  • the codewords included in the diagonalized differential codebook are diagonalized matrices, which can maintain the amplitude characteristics of the elements that the non-differential codebook already has, such as constant model property and finite character set constraint characteristics.
  • a non-differential codebook obtained from a rotating Hadamard matrix can be used, wherein each column of the Hadamard matrix can be matched with a eigenvector of a strong correlation and a low correlation channel in a uniform linear array transmit antenna configuration and a dual polarization transmit antenna configuration, respectively.
  • the accuracy of the quantization is improved, and the rotation matrix is a diagonalized matrix, and the space can be finely quantized on the basis of the Hadamard matrix quantization; and the elements of the Hadamard matrix are +1 or -1, which can be kept non-
  • the constant model of the differential codebook; and the multiplication of the Hadamard matrix with other matrices or vectors can be simplified to addition or subtraction, which can greatly reduce the computational complexity of SINR-based CQI calculation, PMI selection and rank adaptation.
  • the columns of the Hadamard matrix are orthogonal to each other, and the ⁇ matrix characteristic of the codewords in the non-differential codebook can be maintained, and the power distribution of each spatial stream of the transmission can be ensured to be the same, and the power distribution of each transmitting antenna is the same under full rank; further,
  • the method of rotating the Hadamard matrix to obtain a non-differential codebook can be applied to the number of transmitting antennas of 2, 4, 8, 32, 64, etc. and the obtained non-differential codebook can satisfy the constant model property, the ⁇ matrix characteristic and the SINR-based reduction.
  • the methods described in the second embodiment and the third embodiment are also applicable to a CoMP (Coordinated Multiple Point Transmission) system. The following is an example in which the CoMP system is applied to the CoMP system.
  • an embodiment of the present invention provides a method for obtaining a precoding matrix indication, including: 401: A UE selects a differential PMI based on a reference PMI for feedback, and acquires a reference PMI of each cell.
  • the base station of all cells of the CoMP system where the UE is located may indicate that the UE uses the reference PMI-based differential PMI to perform feedback through the high layer signaling or the downlink physical control channel, and after receiving the indication, the UE selects to use the reference PMI based differential PMI for feedback.
  • the base stations of all cells of the CoMP system are uniformly managed by the eNode B (evolved Node B).
  • Embodiment of the present invention assumes that there are M CoMP cell system, M is a positive integer embodiment, each reference cell PMI were ⁇ , n 2,... N M. It should be noted that the reference PMI of each cell in the embodiment of the present invention may be in the form of the reference PMI described in step 201 of Embodiment 2, or the reference described in step 301 of step embodiment 3 may be used. The form of PMI can be flexibly selected according to the actual application.
  • the UE feeds back the reference PMI of each cell to the eNode B, and calculates each cell according to a preset criterion according to a reference PMI, a non-differential codebook, and a diagonalized differential codebook D of each cell. Differential PMI.
  • non-differential codebook in step 2 of Example W w w 202 in the same non-differential codebook W is not repeated here.
  • the diagonalized differential codebook D may be the same as the diagonalized differential codebook D in step 202 of Embodiment 2, or may be multiplied by a phase shift matrix in each of the diagonalized differential codebooks in step 202 of Embodiment 2. For example, you can multiply the diagonalized matrix obtained by diag ⁇ e k , e j " e k e k ⁇ .
  • the differential PMI of each cell is calculated as . M , as shown in equation (15) Show:
  • M represents the number of cells, M is a positive integer; C 4 C i2 ... C M represents a codeword corresponding to the difference PMI z; , .. J M in the diagonalized difference codebook D; WW 3 ⁇ 4 . ..W nM denotes a codeword corresponding to the reference PMI /3 ⁇ 4 n 2 3 ⁇ 4 in the non-differential codebook W w ; /C) shows an objective function corresponding to the preset criterion; Pi denotes a large base station of the UE to the cell i The scale fading corresponds to a power control parameter well known to the eNode B and the UE; ⁇ rm ⁇ . ⁇ indicates normalization of each column of the matrix;
  • the UE feeds back the differential ⁇ of each cell to the eNode B.
  • the differential PMI of all cells is, in order, K, so specifically, k 2 ---k M is fed back to the eNode B.
  • the eNode B receives the reference PMI of each cell and the differential PMI of each cell, and calculates a pre-preparation of each cell by using a non-differential codebook and a diagonally differentiated codebook D according to the reference PMI and the differential PMI. Coding matrix V
  • the precoding matrix of each cell is sequentially a precoding matrix, a precoding matrix 2 ... a precoding matrix.
  • the eNode B calculates the precoding matrix of each cell according to the reference PMI and the differential PMI using the non-differential codebook ⁇ ) and the diagonalized differential codebook D as shown in the equation (16):
  • M represents the number of cells, and M is a positive integer;
  • C ; i C i2 ... C iM represents a codeword corresponding to the differential PMI ⁇ ;, 2 ... M in the diagonalized differential codebook D;
  • WW ...W nM denotes a codeword corresponding to the reference PMI /3 ⁇ 4 in the non-differential codebook W w ;
  • A. A large-scale fading corresponding to the base station of the UE to the cell ⁇ , which is a well-known power of the eNode B and the UE Control parameters; "orm ⁇ . ⁇ means normalizing the columns of the matrix; (widely indicating the conjugate transpose operation of the matrix or vector;
  • the eNode B uses the precoding matrix of each cell to preprocess the data to be sent s of each cell, and sends the preprocessed data to be sent s to the UE through the transmitting antenna.
  • the UE receives the received signal and performs data detection on the received signal y.
  • the method for obtaining the precoding matrix indication obtains the reference PMI and the differential PMI according to the non-differential codebook and the diagonalized differential codebook, which can save feedback overhead, improve feedback precision, and improve feedback performance;
  • the codewords included in the diagonalized difference codebook are diagonalized matrices, and can maintain the amplitude characteristics of the elements that the non-differential codebook already has (such as constant model property, finite character set constraint characteristics).
  • a non-differential codebook obtained from a rotating Hadamard matrix can be used, wherein each column of the Hadamard matrix can be matched with a eigenvector of a strong correlation and a low correlation channel in a uniform linear array transmit antenna configuration and a dual polarization transmit antenna configuration, respectively. Therefore, the accuracy of the quantization is improved, and the rotation matrix is a diagonalization matrix, and the space can be finely quantized on the basis of the Hadamard matrix quantization; the elements of the Hadamard matrix are +1 or -1, and the non-differential codebook can be maintained.
  • the constant model property; and, the multiplication of the Hadamard matrix with other matrices or vectors can be simplified to addition or subtraction, which can greatly reduce the computational complexity of SINR-based CQI calculation, PMI selection and rank adaptation.
  • the columns of the Hadamard matrix are orthogonal to each other, and the ⁇ matrix characteristics of the codewords in the non-differential codebook can be maintained, and the power distribution of each spatial stream of the transmission can be ensured to be the same, and the power distribution of each transmitting antenna is the same under full rank; further, The method of rotating the Hadamard matrix to obtain a non-differential codebook can be applied to the number of transmitting antennas of 2, 4, 8, 16, 32, 64, etc.
  • the structure of the codeword included in the diagonalized differential codebook in the foregoing embodiment 2-4 may also be as shown in the formula (18):
  • the structure of the codeword in the diagonalized differential codebook is as shown in equation (18), the power allocation of the transmitting antenna can be achieved.
  • diagonal elements may also be exchanged according to a specific antenna configuration, such as cyclic shifting along a diagonal position, or all diagonal elements multiplied by a phase shift factor.
  • the value of the phase shift difference m0 should be symmetrically distributed with respect to 0. Under the condition of the feedback overhead, more differential matrices can be allocated near the 0 phase shift.
  • the 2-bit 4-antenna diagonal differential codebook provided in the embodiment of the present invention is shown in Tables 5 and 8.
  • the 3-bit 4-antenna diagonalized differential codebook is shown in Tables 6 and 9, and the 4-bit 4-antenna is shown.
  • the diagonalized differential codebook is shown in Tables 7 and 10.
  • the 2-bit 8-antenna diagonal differential codebook provided in the embodiment of the present invention is shown in Tables 11 and 14.
  • the 3-bit 8-antenna diagonalized differential codebook is shown in Tables 12 and 15, and the 4-bit 8-antenna diagonal is shown.
  • the differential codebook is shown in Tables 13 and 16. table 5.
  • the method for obtaining the precoding matrix indication obtains the reference PMI and the differential PMI according to the non-differential codebook and the diagonalized differential codebook, which can save feedback overhead, improve feedback precision, and improve feedback performance;
  • the codewords included in the diagonalized difference codebook are diagonalized matrices, and can maintain the amplitude characteristics of the elements that the non-differential codebook already has (such as constant model property, finite character set constraint characteristics).
  • a non-differential codebook obtained from a rotating Hadamard matrix can be used, wherein each column of the Hadamard matrix can be matched with a eigenvector of a strong correlation and a low correlation channel in a uniform linear array transmit antenna configuration and a dual polarization transmit antenna configuration, respectively. Therefore, the accuracy of the quantization is improved, and the rotation matrix is a diagonalized matrix, and the space can be finely quantized on the basis of the Hadamard matrix quantization; the elements of the Hadamard matrix are +1 or -1, which can satisfy the constant model.
  • the multiplication of the Hadamard matrix with other matrices or vectors can be simplified to addition or subtraction, which can greatly reduce the computational complexity of SINR-based CQI calculation, PMI selection and rank adaptation; normalize the columns of Hadamard matrix Orthogonal to each other, the ⁇ matrix characteristic of the codeword in the non-differential codebook can be maintained, and the power distribution of each spatial stream transmitted can be ensured to be the same, and the power distribution of each transmitting antenna is the same under full rank; further, the non-rotating Hadamard matrix is obtained.
  • the method of differential codebook can be applied to 2, 4, 8, 16, 32, 64, etc.
  • the embodiment of the present invention provides a method for obtaining a precoding matrix indication, and the method described in Embodiment 2-5.
  • the embodiment of the present invention utilizes the diagonalized differential codebook of the single-polarized transmit antenna in Embodiment 2-5 (which may be referred to as a single-polarization diagonalized differential codebook) to obtain a dual-polarized transmit antenna array.
  • Diagonal differential codebook (which can be called a dual-polarization diagonal differential codebook).
  • the differential PMI included in codebook D, k 0, l ... ⁇ D ⁇ - l ,
  • the first N/2 diagonal elements of the matrix in the above formula (20) correspond to a set of homopolar transmit antennas, and the last N/2 diagonal elements correspond to another set of co-polarized transmit antennas. And if the arrangement position of the dual-polarized transmitting antenna array changes, the position of the above element is exchanged accordingly, and the specific structure is similar, and details are not described herein again.
  • the method for obtaining the precoding matrix indication obtains the reference PMI and the differential PMI according to the non-differential codebook and the diagonalized differential codebook, which can save feedback overhead, improve feedback precision, and improve feedback performance;
  • the codewords included in the diagonalized difference codebook are diagonalized matrices, and can maintain the amplitude characteristics of the elements that the non-differential codebook already has (such as constant model property, finite character set constraint characteristics).
  • a non-differential codebook obtained from a rotating Hadamard matrix can be used, wherein each column of the Hadamard matrix can be matched with a eigenvector of a strong correlation and a low correlation channel in a uniform linear array transmit antenna configuration and a dual polarization transmit antenna configuration, respectively. Therefore, the accuracy of the quantization is improved, and the rotation matrix is a diagonalized matrix, and the space can be finely quantized based on the Hadamard matrix quantization; and the elements of the Hadamard matrix are +1 or -1, which can remain non-differential.
  • the constant model of the codebook, and the multiplication of the Hadamard matrix with other matrices or vectors can be simplified to addition or subtraction, which can greatly reduce the computational complexity of SINR-based CQI calculation, PMI selection and rank adaptation.
  • the columns of the Hadamard matrix are orthogonal to each other, and the ⁇ matrix characteristic of the codewords in the non-differential codebook can be maintained, and the power distribution of each spatial stream of the transmission can be ensured to be the same, and the power distribution of each transmitting antenna is the same under full rank; further,
  • the method of rotating the Hadamard matrix to obtain a non-differential codebook can be applied to the number of transmitting antennas of 2, 4, 8, 16, 32, 64, etc.
  • the diagonalized differential codebook of the dual-polarization transmit antenna array is constructed by the diagonalized differential codebook of the single-polarized transmit antenna, which can be fully utilized.
  • the characteristics of the dual-polarized transmit antenna array improve the performance of the diagonalized differential codebook in the configuration of the dual-polarized transmit antenna array.
  • Embodiments 2-6 are also applicable to the uplink system ( For example, the UE sends Node B data to).
  • the difference from the application to the downlink system is that the Node B acquires the reference PMI and the difference according to steps similar to Embodiment 2-6.
  • the PMI is notified to the UE, and the UE receives the notification of the NodeB to perform precoding and transmits data to the Node B, and the Node B receives the data and performs data detection.
  • an embodiment of the present invention provides a method for obtaining a precoding matrix, where the method includes:
  • the reference precoding matrix indication and the differential precoding matrix indication are obtained and transmitted by the data receiving end according to the first non-differential codebook and the first diagonalized differential codebook, and the codeword included in the first diagonalized differential codebook For the diagonalization matrix.
  • the first non-differential codebook (4 may be a non-differential codebook set in the embodiment of the present invention, and the codeword in the first non-differential codebook W w is obtained from a rotating Hadamard matrix, specifically the first non-
  • the structure of the codewords included in the differential codebook w w can be as follows:
  • denotes the number of codewords of a first non-differential codebook included in W w
  • r represents the rank of the first non-differential codewords included in the codebook
  • the index corresponding 4 represents a diagonal matrix
  • n represents the corresponding 4
  • denotes the index of the corresponding normalized Hadamard matrix
  • ⁇ resume ⁇ denotes the corresponding n-order normalized Hadamard matrix
  • the value of ⁇ is equal to the number of transmitting antennas
  • (R it H Directory, ;t f) denotes a matrix consisting of r columns selected from the rotating Hadamard matrix R it H Meeting, A.
  • the r columns are selected to form a matrix (R 4 H Chapter, A f).
  • the relevant method can be used to determine which r column to select, for example: For non-correlated channels, The choice of columns should be such that a set of r-column sub-matrices is extracted from different matrices R ; ⁇ vide
  • Each matrix in ⁇ (R it H Computer, ;t f) ⁇ satisfies the maximum gain in the zero direction of the space projection of the antenna array, and can also be selected by any other feasible method, which is not specifically limited.
  • the structure of the codeword included in the first non-differential codebook can be as follows:
  • W w represents the number of code words included in the first non-differential codebook W w
  • r represents the rank of the codeword included in the first non-differential codebook W w
  • 4 represents the index of the corresponding diagonal matrix.
  • R ; indicates with 4 pairs
  • the n-th order diagonal matrix, ⁇ denotes the index of the corresponding normalized Hadamard matrix, ⁇ ⁇ denotes the n-th order normalized Hadamard matrix corresponding to ⁇ , and the value of ⁇ is equal to the number of transmitting antennas.
  • R ; H Attention , Rf f denotes a matrix composed of r columns selected from the rotating Hadamard matrix ⁇ H Chapter , R .
  • the structure of the diagonal matrix 1 can be expressed as:
  • the method is selected, and no specific limitation is made thereto.
  • the second non-differential codebook and the second diagonalized differential codebook are respectively consistent with the first non-differential codebook and the first diagonalized differential codebook.
  • the pre-coding matrix is obtained by using the second pre-stored second non-differential codebook and the second diagonally differentiated codebook according to the reference precoding matrix indication and the differential precoding matrix indication.
  • the method may include:
  • the differential precoding matrix indicates that the corresponding codeword is multiplied by the codeword corresponding to the reference precoding matrix, and the product result is used as a precoding matrix.
  • the pre-coding matrix is obtained by using the second pre-stored second non-differential codebook and the second diagonally differentiated codebook according to the reference precoding matrix indication and the differential precoding matrix indication.
  • the method may include:
  • the precoding matrix is calculated by using a zero-forcing beamforming algorithm or based on a maximum signal-to-noise ratio criterion according to the reference precoding matrix indication and the differential precoding matrix indication of each data receiving end.
  • the second non-differential codebook can use a non-differential codebook obtained from the rotated Hadamard matrix, wherein the columns of the Hadamard matrix can be strongly correlated and low respectively with the uniform linear array transmit antenna configuration and the dual polarized transmit antenna configuration.
  • the eigenvectors of the relevant channels are matched to improve the accuracy of the quantization.
  • the rotation matrix is a diagonal matrix, which can be used to quantize the space on the basis of the Hadamard matrix quantization.
  • each element of the Hadamard matrix is +1 or- 1, can maintain the constant model of the non-differential codebook; and the multiplication of the Hadamard matrix with other matrices or vectors can be simplified to addition or subtraction, which can greatly reduce the basis
  • the computational complexity of CIRI calculation, PMI selection and rank adaptation of SINR The columns of the normalized Hadamard matrix are orthogonal to each other, and the ⁇ matrix characteristic of the codewords in the non-differential codebook can be maintained, and the power distribution of each spatial stream in the transmission can be ensured to be the same.
  • the power distribution of each transmitting antenna is the same, Ensure that the spatial power distribution of each spatial stream is the same, and the power distribution of each transmitting antenna is the same under full rank; in addition, the method of obtaining a non-differential codebook from the rotating Hadamard matrix can be applied to 2, 4, 8, 16, 32, 64, etc.
  • the number of transmit antennas and the resulting non-differential codebook can satisfy the constant model property, the ⁇ matrix characteristics, and the computational complexity of reducing SINR-based CQI calculation, PMI selection, and rank adaptation.
  • an embodiment of the present invention provides an apparatus for obtaining an indication of a precoding matrix, where the apparatus includes:
  • the precoding matrix indication obtaining module 601 is configured to obtain a reference precoding matrix indication and a differential precoding matrix indication according to the first non-differential codebook and the first diagonalized differential codebook, where the first diagonalized differential codebook The codewords included are diagonalized matrices.
  • the first non-differential codebook (which may be a non-differential codebook set in the embodiment of the present invention, the codeword in the first non-differential codebook W w is obtained from the rotated Hadamard matrix, specifically the first non-differential
  • the structure of the codeword included in the codebook W w can be as follows:
  • the structure of the diagonal matrix can be expressed as:
  • the structure of the codeword w ⁇ ) included in the non-differential codebook can be as follows:
  • W w represents the number of code words included in the first non-differential codebook W w
  • r represents the first non-differential codebook
  • the rank of the codeword included in W w , 4 denotes an index with the corresponding diagonal matrix
  • 3 ⁇ 4 denotes an n-th order diagonal matrix corresponding to 4
  • A denotes an index with a corresponding normalized Hadamard matrix
  • disregard ⁇ Representing an n-th order normalized Hadamard matrix corresponding to ⁇ , the value of ⁇ is equal to the number of transmitting antennas
  • R ; H Attention , Rf f represents a matrix consisting of r columns selected from a rotating Hadamard matrix! ⁇ H Fundamental , R .
  • the structure of the diagonal matrix can be expressed as:
  • the column should be selected such that a set i of r sub-matrices is extracted from different matrices R ; H Guide, Rf, respectively (VR ; H Chapter , Rf " [I matrix in each case satisfies the maximum chord distance;
  • the channel and column shall be selected such that each matrix of the set i i (VR ; H Rail, Rff ⁇ I is extracted from different matrices R ; H Directory , R respectively to satisfy the zero direction of the space projection in the antenna array space.
  • the gain is the largest, and any other feasible method can be used for selection, which is not specifically limited.
  • the precoding matrix indication acquiring module 601 may specifically include:
  • the reference precoding matrix indication calculation unit is configured to calculate a reference precoding matrix indication according to the preset criterion according to the first non-differential codebook.
  • a differential precoding matrix indication calculation unit configured to: after the reference precoding matrix indication calculation unit obtains the reference precoding matrix indication, according to the first non-differential codebook, the first diagonalized differential codebook, and the reference precoding matrix indication, based on The preset criterion is used to calculate the differential precoding matrix indication; or the precoding matrix indication obtaining module 601 may specifically include:
  • a reference precoding matrix indication and a differential precoding matrix indication calculating unit configured to calculate a reference precoding matrix indication and a differential precoding according to a preset criterion according to the first non-differential codebook and the first diagonalized differential codebook Matrix indication.
  • the device further includes:
  • the precoding matrix indication feedback module 602 is configured to: after the precoding matrix indication obtaining module 601 obtains the reference precoding matrix indication and the differential precoding matrix indication, feed back the reference precoding matrix indication and the differential precoding matrix indication to the data sending end, And causing the data transmitting end to obtain a precoding matrix according to the reference precoding matrix indication and the differential precoding matrix indication, by using the second non-differential codebook and the second diagonalized differential codebook stored in the foregoing, wherein the second non- The differential codebook and the second diagonalized differential codebook are respectively consistent with the first non-differential codebook and the first diagonalized differential codebook.
  • the device for obtaining the precoding matrix indication obtains the reference PMI and the differential PMI according to the non-differential codebook and the diagonalized differential codebook, thereby saving feedback overhead and improving feedback precision, thereby improving feedback performance;
  • the codewords included in the diagonalized difference codebook are diagonalized matrices, and can maintain the amplitude characteristics of the elements that the non-differential codebook already has (such as constant model property, finite character set constraint characteristics).
  • a non-differential codebook obtained from a rotating Hadamard matrix can be used, wherein the columns of the Hadamard matrix can be respectively associated with the eigenvectors of the strong correlation and low correlation channels of the uniform hook line array antenna configuration and the dual polarization antenna configuration.
  • the rotation matrix is a diagonalization matrix, and the space can be further quantified based on the Hadamard matrix quantization; and, the elements of the Hadamard matrix are +1 or -1, which can be kept non- The constant model of the differential codebook; in addition, the multiplication of the Hadamard matrix with other matrices or vectors can be simplified to addition or subtraction, which can greatly reduce the computational complexity of SINR-based CQI calculation, PMI selection and rank adaptation.
  • the columns of the normalized Hadamard matrix are orthogonal to each other, and the ⁇ matrix characteristic of the codewords in the non-differential codebook can be maintained, and the power distribution of each spatial stream of the transmission can be ensured to be the same.
  • the power distribution of each transmitting antenna is the same, Ensure that the power distribution of each spatial stream transmitted is the same, in the case of full rank
  • the transmit power of the transmit antennas is the same; in addition, the method of obtaining the non-differential codebook from the rotated Hadamard matrix can be applied to the transmit antennas of 2, 4, 8, 16, 32, 64, etc. and the obtained non-differential codebook can satisfy Constant model properties, ⁇ matrix properties, and computational complexity for reducing SINR-based CQI calculations, PMI selection, and rank adaptation.
  • an embodiment of the present invention provides an apparatus for acquiring a precoding matrix, where the apparatus includes: a precoding matrix indication receiving module 701, configured to receive a reference precoding matrix indication and a differential precoding matrix indication; The coding matrix indication and the differential precoding matrix indication are obtained and transmitted by the data receiving end according to the first non-differential codebook and the first diagonalized differential codebook, and the codeword included in the first diagonalized differential codebook is diagonalized matrix.
  • a precoding matrix indication receiving module 701 configured to receive a reference precoding matrix indication and a differential precoding matrix indication
  • the coding matrix indication and the differential precoding matrix indication are obtained and transmitted by the data receiving end according to the first non-differential codebook and the first diagonalized differential codebook, and the codeword included in the first diagonalized differential codebook is diagonalized matrix.
  • the first non-differential codebook ⁇ ) may be a non-differential codebook set by the embodiment of the present invention, and the codeword in the first non-differential codebook W w is obtained from a rotating Hadamard matrix, specifically the first non-
  • the structure of the codewords included in the differential codebook W w can be as follows:
  • the structure of the diagonal matrix can be expressed as:
  • n is equal to the number of transmit antennas.
  • R IT H Computer A From the rotating Hadamard matrix R IT H Computer A , select r columns to form a matrix (R IT H Chapter, ;t f), depending on the actual application, use the relevant method to determine which r column to choose, for example: For the relevant channel, the column should be selected such that a set of r-column sub-matrices is extracted from different matrices R ; ⁇ vide
  • each matrix in ⁇ (R ; H Hook , f) ⁇ satisfies the maximum gain in the zero direction of the space projection of the antenna array, and can also ⁇ Choose any other feasible method, which is not specifically limited.
  • the structure of the codeword w ⁇ ) included in the first non-differential codebook may be as follows:
  • W w represents the number of code words included in the first non-differential codebook W w
  • r represents the first non-differential codebook
  • the index, ⁇ alt ⁇ denotes the n-th order normalized Hadamard matrix corresponding to ⁇ , and the value of ⁇ is equal to the number of transmitting antennas.
  • the structure of the diagonal matrix can be expressed as:
  • Each matrix in I V I satisfies the maximum gain in the zero direction of the space projected by the antenna array, and can be selected by any other feasible method, which is not specifically limited.
  • the precoding matrix obtaining module 702 is configured to: after the precoding matrix indication receiving module 701 receives the reference precoding matrix indication and the differential precoding matrix indication, according to the reference precoding matrix indication and the differential precoding matrix indication, using local pre-stored a second non-differential codebook and a second diagonally differentiated codebook, to obtain a precoding matrix, wherein the second non-differential codebook and the second diagonalized differential codebook are respectively associated with the first non-differential codebook and the first pair The keratinized differential codebook is consistent.
  • the precoding matrix acquisition module 702 can include:
  • a codeword obtaining unit configured to: after the precoding matrix indication receiving module 701 receives the reference precoding matrix indication and the differential precoding matrix indication, obtain a reference precoding from the second non-differential codebook according to the reference precoding matrix indication The matrix indicates the corresponding codeword, and according to the differential precoding matrix indication, the differential precoding matrix indicates the corresponding codeword from the second diagonalized differential codebook; the first precoding matrix acquiring unit is used in the codeword After the obtaining unit obtains the differential precoding matrix indicating that the corresponding codeword and the reference precoding matrix indication correspond to the codeword, the differential precoding matrix indicates The corresponding codeword is multiplied by the corresponding codeword indicated by the reference precoding matrix, and the product result is used as a precoding matrix.
  • the pre-coding matrix acquisition module 702 specifically includes:
  • a second precoding matrix acquiring unit configured to: after the precoding matrix indication receiving module 701 receives the reference precoding matrix indication and the differential precoding matrix indication, according to the reference precoding matrix indication and difference of each data receiving end that are simultaneously paired
  • the precoding matrix indicates that the precoding matrix is calculated using a zero-forcing beamforming algorithm or based on a maximum signal-to-noise ratio criterion.
  • the apparatus for acquiring a precoding matrix obtains a reference PMI and a differential PMI according to the non-differential codebook and the diagonalized differential codebook, thereby saving feedback overhead and improving feedback precision, thereby further improving feedback performance;
  • the codewords included in the diagonalized difference codebook are diagonalized matrices, and can maintain the amplitude characteristics of the elements that the non-differential codebook already has (such as constant model property, finite character set constraint characteristics).
  • a non-differential codebook obtained from a rotating Hadamard matrix can be used, wherein each column of the Hadamard matrix can be matched with a eigenvector of a strong correlation and a low correlation channel in a uniform linear array transmit antenna configuration and a dual polarization transmit antenna configuration, respectively. Therefore, the accuracy of the quantization is improved, and the rotation matrix is a diagonalized matrix, and the space can be finely quantized based on the Hadamard matrix quantization; and the elements of the Hadamard matrix are +1 or -1, which can remain non-differential.
  • the constant model of the codebook in addition, the multiplication of the Hadamard matrix with other matrices or vectors can be simplified to addition or subtraction, which can greatly reduce the SINR (signal to noise ratio) based CQI (channel quality indicator) calculation, PMI
  • the computational complexity of selection and rank adaptation The columns of the normalized Hadamard matrix are orthogonal to each other, and the ⁇ matrix characteristic of the codewords in the non-differential codebook can be maintained, and the power distribution of each spatial stream in the transmission can be ensured to be the same.
  • the power distribution of each transmitting antenna is the same, Ensure that the spatial power distribution of each spatial stream is the same, and the power distribution of each transmitting antenna is the same under full rank; further, the method of obtaining a non-differential codebook from the rotating Hadamard matrix can be applied to 2, 4, 8, 16, 32, 64
  • the equal number of transmit antennas and the resulting non-differential codebook can satisfy the constant model property, the ⁇ matrix characteristics, and the computational complexity of reducing SINR-based CQI calculation, PMI selection, and rank adaptation.
  • an embodiment of the present invention provides a method for obtaining a precoding matrix indication, which includes: 801: Calculate a non-differential precoding matrix indication according to a preset criterion according to a first non-differential codebook.
  • the first non-differential codebook can be a non-differential codebook set by the embodiment of the present invention.
  • the codewords in the first non-differential codebook w w are obtained from the rotated Hadamard matrix.
  • the structure of the codewords included in the first non-differential codebook w w can be as follows:
  • r represents a first code word includes a non-differential codebook of rank 4 are the corresponding diagonal matrix R; t index, R it represents An n-th order diagonal matrix corresponding to 4, A denotes an index with a corresponding normalized Hadamard matrix, ⁇ ⁇ denotes a corresponding normalized Hadamard matrix of n-order, and the value of ⁇ is equal to the number of transmitting antennas , ( R ; H Directory , ) (r) denotes a matrix consisting of r columns selected from the rotating Hadamard matrix R ; ⁇ adj.
  • the structure of the diagonal matrix can be expressed as:
  • R. diag ⁇ b ik S '" ,b i 2 e iS ' k ' 2 ,...,b it e iS ' k ,...,b in e iS ' k ' n ),
  • IV [I can satisfy the maximum gain in the zero direction of the space projection of the antenna array in I, and can also be selected by any other feasible method, which is not specifically limited.
  • the structure of the codeword included in the first non-differential codebook can be as follows:
  • r represents the rank of the first codeword W w non-differential codebook included
  • 4 denotes an index corresponding to a diagonal matrix
  • R represents i k corresponds to the n-th order diagonal matrix
  • A denotes the index of the corresponding normalized Hadamard matrix
  • ⁇ ⁇ denotes the corresponding n-order normalized Hadamard matrix
  • the value of ⁇ is equal to the number of transmitting antennas
  • the structure of the diagonal matrix can be expressed as:
  • the structure of the codeword included in the first non-differential codebook may also be as follows:
  • W w represents the number of code words included in the non-differential codebook W w
  • r represents the rank of the codeword included in the non-differential codebook W w
  • 4 represents the index with the corresponding diagonal matrix R it
  • R 4 represents An n-th order diagonal matrix corresponding to 4, where B represents a codeword corresponding to an index of a r-transmitted non-differential codebook of rank r.
  • the method may further include: 802: feeding back the non-differential precoding matrix indication to the data transmitting end.
  • the data sending end receives the non-differential precoding matrix indication, and uses the second non-differential codebook pre-stored in itself to obtain the precoding matrix according to the non-differential precoding matrix indication.
  • the second non-differential codebook is consistent with the first non-differential codebook.
  • the method for obtaining the indication of the precoding matrix uses the feature vector matching of the strong correlation and the low correlation channel from the rotating Hadamard moment configuration and the dual polarization transmitting antenna configuration, thereby improving the accuracy of the quantization, and the rotation matrix
  • the space can be carefully quantized based on the Hadamard matrix quantization; moreover, each element of the Hadamard matrix is +1 or -1, which can maintain the constant model of the non-differential codebook;
  • the multiplication of the horse matrix with other matrices or vectors can be simplified to addition or subtraction, which can greatly reduce the computational complexity of SINR-based CQI calculation, PMI selection and rank adaptation.
  • the columns of the normalized Hadamard matrix are orthogonal to each other, and the ⁇ matrix characteristic of the codewords in the non-differential codebook can be maintained, and the power distribution of each spatial stream in the transmission can be ensured to be the same.
  • the power distribution of each transmitting antenna is the same, Ensure that the spatial power distribution of each spatial stream is the same, and the power distribution of each transmitting antenna is the same under full rank; further, the method of obtaining a non-differential codebook from the rotating Hadamard matrix can be applied to 2, 4, 8, 16, 32, 64 Wait
  • the number of transmit antennas and the resulting non-differential codebook can satisfy the constant model property, the ⁇ matrix characteristics, and the computational complexity of reducing SINR-based CQI calculation, PMI selection, and rank adaptation.
  • an embodiment of the present invention provides an apparatus for acquiring an indication of a precoding matrix, where the apparatus includes:
  • the non-differential precoding matrix indication obtaining module 901 is configured to calculate, according to the preset criterion, a non-differential precoding matrix indication according to the first non-differential codebook, where the codeword included in the first non-differential codebook is rotated The Hadamard matrix is obtained.
  • the preset criterion is similar to the preset side preset in Embodiment 2, and details are not described herein again.
  • the first non-differential codebook W w may be a non-differential codebook set by the embodiment of the present invention, and the codeword in the first non-differential codebook w (r) is obtained from the rotated Hadamard matrix, specifically the first non-
  • the structure of the codewords included in the differential codebook w w can be as follows:
  • the values of 2, ..., «, n are equal to the number of transmitting antennas.
  • W w represents the number of code words included in the first non-differential codebook W w
  • r represents the first non-differential codebook
  • the rank of the codeword included in W w , 4 represents the index with the corresponding diagonal matrix 3 ⁇ 4
  • R 4 represents the n-th order diagonal matrix corresponding to 4, and represents the index of the corresponding Hadamard matrix, ⁇ critique ⁇
  • is equal to the number of transmitting antennas, (R ; H Register , Rf f
  • V represents a matrix consisting of r columns selected from the rotated Hadamard matrix R, . H Chapter , .
  • the structure of the diagonal matrix can be expressed as:
  • Each matrix in I V I satisfies the maximum gain in the zero direction of the space projected by the antenna array, and can be selected by any other feasible method, which is not specifically limited.
  • the device may further include:
  • the non-differential precoding matrix indication feedback module 902 is configured to: after the non-differential precoding matrix indication obtaining module 901 obtains the non-differential precoding matrix indication, feed back the non-differential precoding matrix indication to the data sending end, so that the data sending end uses the data transmitting end thereof
  • the apparatus for obtaining the precoding matrix indication uses the eigenvector matching of the strong correlation and the low correlation channel from the rotating Hadamard moment configuration and the dual polarization transmitting antenna configuration, thereby improving the accuracy of the quantization, and the rotation matrix
  • the space can be carefully quantized based on the Hadamard matrix quantization; and, the elements of the Hadamard matrix are +1 or -1, which can maintain the constant model of the non-differential codebook;
  • the multiplication of the Hadamard matrix with other matrices or vectors can be simplified to addition or subtraction, which can greatly reduce the computational complexity of SINR-based CQI calculation, PMI selection and rank adaptation.
  • the columns of the normalized Hadamard matrix are orthogonal to each other,
  • the ⁇ matrix characteristic of the codeword in the non-differential codebook can be maintained, and the power distribution of each spatial stream in the transmission can be ensured to be the same.
  • the power distribution of each transmitting antenna is the same, and the spatial power distribution of each spatial stream can be guaranteed to be the same.
  • the power distribution of each of the transmitting antennas is the same; further, the method of obtaining the non-differential codebook from the rotating Hadamard matrix can be applied to the number of transmitting antennas of 2, 4, 8, 16, 32, 64, etc. and the obtained non-differential codebook
  • the computational complexity of constant model, ⁇ matrix characteristics, and reduction of SINR-based CQI calculation, PMI selection, and rank adaptation can be satisfied.
  • an embodiment of the present invention provides a method for obtaining a precoding matrix, where the method includes:
  • 1001 Receive a non-differential precoding matrix indication; where the non-differential precoding matrix indication is obtained and sent by the data receiving end according to the first non-differential codebook; the codeword included in the first non-differential codebook is obtained from the rotating Hadamard matrix .
  • the first non-differential codebook (which may be a non-differential codebook set in the embodiment of the present invention, the codeword in the first non-differential codebook w w is obtained from a rotating Hadamard matrix, specifically the first non-differential
  • the structure of the codeword included in the codebook w w can be as follows:
  • Wi r ( R , W
  • the r columns are selected from the rotating Hadamard matrix R it H Chapter A to form a matrix (R it H Chapter A f), which can be used according to the actual application.
  • the method determines which r-column to choose, for example: For non-correlated channels, the column should be selected such that a set of r-column sub-matrices is extracted from different matrices R ; ⁇ supervise, respectively ⁇ (R it H Computer ;t ) W j Each matrix in the matrix satisfies the maximum chord distance; for the relevant channel, the column should be selected such that a set of r-column sub-matrices is extracted from different matrices R ; H Rail, respectively ( ⁇ ; ⁇ spirit, " The array satisfies the maximum gain in the zero direction of the space projection of the antenna array, and can also be selected by any other feasible method, which is not specifically limited.
  • the first non-differential code constitutive can be as follows:
  • the structure of the diagonal matrix can be expressed as:
  • the value is equal to the number of transmitting antennas.
  • the r-column is selected from the rotating Hadamard matrix R ; H Fundamental, Rf to form a matrix (R it H Chapter ; t R ) W , which can be determined according to the actual application conditions.
  • r column For non-correlated channels, the column should be selected such that a set of r-column sub-matrices is extracted from different matrices R ; H Directory , Rf, respectively, "f(R ; H Chapter , Rf f)" [ Each matrix satisfies the maximum chord distance; for the relevant channel, the column should be selected such that a set i of r sub-matrices is extracted from different matrices R ; H Directory, Rf, respectively (R ; H Directory , Rf "
  • I V I [I can satisfy the maximum gain in the zero direction of the antenna array space projection in I. I can also use any other feasible method to select, which is not specifically limited.
  • 1002 Obtain a precoding matrix by using a locally pre-stored second non-differential codebook according to the non-differential precoding matrix indication, where the second non-differential codebook is consistent with the first non-differential codebook.
  • obtaining the precoding matrix specifically includes: according to the non-differential precoding matrix indication, from the second The non-differential codebook obtains the non-differential precoding matrix to indicate the corresponding codeword; and the non-differential precoding matrix indicates the corresponding codeword as the precoding matrix.
  • the precoding matrix is specifically included: according to the non-differential pre-preparation of each data receiving end of the simultaneous pairing
  • the coding matrix indicates that the precoding matrix is calculated using a zero-forcing beamforming algorithm or based on a maximum signal-to-noise ratio criterion.
  • the method for acquiring a precoding matrix obtains a precoding matrix according to a non-differential precoding matrix indication, and the non-differential codebook is obtained from a rotated Hadamard matrix, wherein each column of the Hadamard matrix can be respectively hooked
  • the eigenvector matching of the strong correlation and low correlation channels in the array transmit antenna configuration and the dual polarization transmit antenna configuration improves the accuracy of the quantization.
  • the rotation matrix is a diagonalized matrix, which can be maintained on the basis of the Hadamard matrix quantization.
  • each element of the Hadamard matrix is +1 or -1, which can maintain the constant model of the non-differential codebook; in addition, the multiplication of the Hadamard matrix with other matrices or vectors can be simplified to addition or subtraction.
  • the computational complexity of SINR-based CQI calculation, PMI selection, and rank adaptation can be greatly reduced.
  • the columns of the normalized Hadamard matrix are orthogonal to each other, and the ⁇ matrix characteristic of the codewords in the non-differential codebook can be maintained, and the power distribution of each spatial stream of the transmission can be ensured to be the same, and the power distribution of each transmitting antenna is the same under full rank;
  • the method of obtaining a non-differential codebook from a rotating Hadamard matrix can be applied to a number of transmitting antennas of 2, 4, 8, 16, 32, 64, etc. and the obtained non-differential codebook can satisfy the constant model property, the matrix property and Reduce the computational complexity of SINR-based CQI calculation, PMI selection, and rank adaptation.
  • an embodiment of the present invention provides an apparatus for acquiring a precoding matrix, where the apparatus includes:
  • the non-differential precoding matrix indication receiving module 1101 is configured to receive the non-differential precoding matrix indication, where the non-differential precoding matrix indication is obtained by the data receiving end according to the first non-differential codebook, and is sent in the first non-differential codebook.
  • the included codewords are obtained from a rotating Hadamard matrix.
  • the first non-differential codebook (which may be a non-differential codebook set in the embodiment of the present invention, the codeword in the first non-differential codebook W w is obtained from the rotated Hadamard matrix, specifically the first non-differential
  • the structure of the codeword included in the codebook W w can be as follows:
  • Wi r ( R , W
  • r represents a first code word includes a non-differential codebook of rank 4 are the corresponding diagonal matrix R; t index, R it represents An n-th order diagonal matrix corresponding to 4, representing an index with a corresponding Hadamard matrix, H Rule A represents a corresponding n-th order Hadamard matrix, and the value of n is equal to the number of transmitting antennas, (R ; H Directory , W represents a matrix consisting of r columns selected from the rotating Hadamard matrix R ; ⁇ resort , .
  • the structure of the diagonal matrix R 4 can be specifically expressed as:
  • the first non-differential structure can be as follows:
  • W w represents the number of code words included in the first non-differential codebook W w
  • r represents the first non-differential codebook
  • the rank of the codeword included in W w , 4 denotes an index with the corresponding diagonal matrix 3 ⁇ 4, R it denotes an n-th order diagonal matrix corresponding to 4, and denotes an index with the corresponding Hadamard matrix, ⁇ critique ⁇ denotes For the corresponding n-th order Hadamard matrix, the value of ⁇ is equal to the number of transmitting antennas, and (R ; H Chapter , Rf f ) represents a matrix composed of r columns selected from the rotating Hadamard matrix R, H, and R.
  • the structure of the diagonal matrix can be expressed as:
  • R H Computer, Rf, select r columns to form a matrix (R 4 H Chapter ; t R ), which can be used according to practical applications.
  • the column should be selected to extract a set of r sub-matrices from different matrices R ; H Rail, Rf, respectively, Hf, Rf "[Each matrix satisfies the largest chord; for the relevant channel, the column should be chosen such that a set of r-column sub-matrices i (V ; H; , Rf ⁇ I) is extracted from different matrices R ; H, Rf
  • Each matrix in the matrix satisfies the maximum gain in the zero direction of the space projection of the antenna array, and can also be selected by any other feasible method, which is not specifically limited.
  • the non-differential precoding matrix processing module 1102 is configured to: after the non-differential precoding matrix indication receiving module 1101 receives the non-differential precoding matrix indication, use the locally pre-stored second non-differential codebook according to the non-differential precoding matrix indication Calculating a precoding matrix, wherein the second non-differential codebook is consistent with the first non-differential codebook. Further, for a single-user multiple-input multiple-output system, a non-differential precoding matrix acquisition module
  • the non-differential precoding matrix indication receiving module 1101 receives the non-differential precoding matrix indication
  • the non-differential precoding matrix indication is obtained from the second non-differential codebook according to the non-differential precoding matrix indication.
  • the non-differential precoding matrix indicates the corresponding codeword as a precoding matrix.
  • the pre-coding matrix acquisition module 1102 is specifically configured to: after the non-differential pre-coding matrix indication receiving module 1101 receives the non-differential pre-coding matrix indication, according to each data receiving end of the simultaneous pairing
  • the non-differential precoding matrix indicates that the precoding matrix is calculated using a zero-forcing beamforming algorithm or based on a maximum signal-to-noise ratio criterion.
  • the apparatus for acquiring a precoding matrix obtains a precoding matrix according to a non-differential precoding matrix indication, and the non-differential codebook is obtained from a rotating Hadamard matrix, wherein each column of the Hadamard matrix can be respectively hooked
  • the eigenvector matching of the strong correlation and low correlation channels in the array transmit antenna configuration and the dual polarization transmit antenna configuration improves the accuracy of the quantization.
  • the rotation matrix is a diagonalized matrix, which can be maintained on the basis of the Hadamard matrix quantization.
  • each element of the Hadamard matrix is +1 or -1, which can maintain the constant model of the non-differential codebook; in addition, the multiplication of the Hadamard matrix with other matrices or vectors can be simplified to addition or subtraction.
  • the computational complexity of SINR-based CQI calculation, PMI selection, and rank adaptation can be greatly reduced.
  • the columns of the normalized Hadamard matrix are orthogonal to each other, and the ⁇ matrix characteristic of the codewords in the non-differential codebook can be maintained, and the power distribution of each spatial stream of the transmission can be ensured to be the same, and the power distribution of each transmitting antenna is the same under full rank;
  • the method of obtaining a non-differential codebook from a rotating Hadamard matrix can be applied to a number of transmitting antennas of 2, 4, 8, 16, 32, 64, etc. and the obtained non-differential codebook can satisfy the constant model property, the matrix property and Reduce the computational complexity of SINR-based CQI calculation, PMI selection, and rank adaptation.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Power Engineering (AREA)
  • Radio Transmission System (AREA)
  • Compression, Expansion, Code Conversion, And Decoders (AREA)
  • Error Detection And Correction (AREA)

Description

获取预编码矩阵指示以及预编码矩阵的方法和装置 技术领域
本发明涉及通信技术领域, 特别涉及一种获取预编码矩阵指示以及预编 码矩阵的方法和装置。 背景技术
随着通信技术的不断发展, 为了改善数据传输的性能, 数据发送端 (如
NodeB (节点 B )、 BS ( Base station, 基站 )等 )可以根据数据接收端(如 UE ( User Equipment, 用户设备)、 MS ( Mobile Station, 移动台)等)反馈的 PMI ( Precoding Matrix Indicator, 预编码矩阵指示)以及本地预存的码本, 得到预 编码矩阵, 通过预编码矩阵对待发送数据进行预处理后再发送到数据接收端, 这样可以使数据发送过程自适应信道状态的变化, 从而提高数据传输的性能。 因此, 如何获取 PMI以及釆用何种码本至关重要。
现有 LTE R8 ( Long Term Evolution Release 8 , 长期演进发布版本 8 ) 系 统支持 4天线传输,釆用固定的单一码本,其码本基于 Householder反射得到。 现有 IEEE ( The Institute of Electrical and Electronics Engineers, 国际电子电气 工程师协会) 802.16m系统中, 釆用非差分和差分两种码本, 主要利用差分模 式进行反馈, 反馈过程如下: 在每个预设的反馈周期内, UE先反馈一个基于 非差分码本的非差分 PMI, 然后再根据非差分 PMI依次反馈几个基于差分码 本的差分 PMI。
在实现本发明的过程中, 发明人发现现有技术至少存在以下问题: 现有 LTE R8系统, 釆用固定的单一码本, 子带反馈开销大, 宽带反馈精 度低。 现有 IEEE 802.16m系统釆用时间域差分反馈, 预编码矩阵的获取依赖 于反馈的历史信息, 会导致误差传播问题; 而且由于现有 IEEE 802.16m系统 使用的差分码本中包括的每个码字中的元素的特性, 会使得最终得到的预编 码矩阵不一定具有恒模特性,特别是难以满足各个元素具有的有限字符集(如 8PSK(Phase Shift Keying, 相移键控))约束特性; 并且实险发现, 当上述差分 反馈形式直接使用 LTE R8 系统中的码本作为非差分码本, 并釆用 IEEE 802.16m系统的差分码本时, 开销大, 相对于仅使用 LTE R8码本反馈时, 其 对反馈性能的改进不大, 致使每个反馈比特的效率低下。 因此, 有必要进一 步研究反馈的结构及其使用的码本, 以提高系统的反馈性能。 发明内容
为了进一步提高反馈的性能, 本发明实施例提供了一种获取预编码矩阵 指示以及预编码矩阵的方法和装置。 所述技术方案如下:
一方面, 本发明实施例提供了一种获取预编码矩阵指示的方法, 所述方 法包括:
根据第一非差分码本和第一对角化差分码本, 获取参考预编码矩阵指示 和差分预编码矩阵指示, 其中, 所述第一对角化差分码本中包括的码字为对 角化矩阵。
一方面, 本发明实施例提供了一种获取预编码矩阵的方法, 所述方法包 括:
接收参考预编码矩阵指示和差分预编码矩阵指示; 其中, 所述参考预编 码矩阵指示和所述差分预编码矩阵指示由数据接收端根据第一非差分码本和 第一对角化差分码本得到并发送; 所述第一对角化差分码本中包括的码字为 对角化矩阵;
根据所述参考预编码矩阵指示和所述差分预编码矩阵指示, 利用本地预 先存储的第二非差分码本和第二对角化差分码本, 得到预编码矩阵, 其中, 所述第二非差分码本和所述第二对角化差分码本分别与所述第一非差分码本 和所述第一对角化差分码本相一致。
一方面, 本发明实施例提供了一种获取预编码矩阵指示的装置, 所述装 置包括:
预编码矩阵指示获取模块, 用于根据第一非差分码本和第一对角化差分 码本, 获取参考预编码矩阵指示和差分预编码矩阵指示, 其中, 所述第一对 角化差分码本中包括的码字为对角化矩阵。
一方面, 本发明实施例提供了一种获取预编码矩阵的装置, 所述装置包 括:
预编码矩阵指示接收模块, 用于接收参考预编码矩阵指示和差分预编码 矩阵指示; 其中, 所述参考预编码矩阵指示和所述差分预编码矩阵指示由数 据接收端根据第一非差分码本和第一对角化差分码本得到并发送; 所述第一 对角化差分码本中包括的码字为对角化矩阵;
预编码矩阵获取模块, 用于在所述预编码矩阵指示接收模块接收到所述 参考预编码矩阵指示和所述差分预编码矩阵指示后, 根据所述参考预编码矩 阵指示和所述差分预编码矩阵指示, 利用本地预先存储的第二非差分码本和 第二对角化差分码本, 得到预编码矩阵, 其中, 所述第二非差分码本和所述 第二对角化差分码本分别与所述第一非差分码本和所述第一对角化差分码本 相一致。
一方面, 本发明实施例提供了一种获取预编码矩阵指示的方法, 所述方 法包括:
根据第一非差分码本, 基于预设的准则, 计算得到非差分预编码矩阵指 示; 其中, 所述第一非差分码本中包括的码字从旋转哈达马矩阵中得到。
一方面, 本发明实施例提供了一种获取预编码矩阵的方法, 所述方法包 括:
接收非差分预编码矩阵指示; 其中, 所述非差分预编码矩阵指示由数据 接收端根据第一非差分码本得到并发送; 所述第一非差分码本中包括的码字 从旋转哈达马矩阵得到;
根据所述非差分预编码矩阵指示, 利用本地预先存储的第二非差分码本, 得到预编码矩阵, 其中, 所述第二非差分码本与所述第一非差分码本相一致。
一方面, 本发明实施例提供了一种获取预编码矩阵指示的装置, 所述装 置包括:
非差分预编码矩阵指示获取模块, 用于根据第一非差分码本, 基于预设 的准侧, 计算得到非差分预编码矩阵指示, 其中, 所述第一非差分码本中包 括的码字从旋转哈达马矩阵得到。
一方面, 本发明实施例提供了一种获取预编码矩阵的装置, 所述装置包 括:
非差分预编码矩阵指示接收模块, 用于接收非差分预编码矩阵指示; 其 中, 所述非差分预编码矩阵指示由数据接收端根据第一非差分码本得到并发 送; 所述第一非差分码本中包括的码字从旋转哈达马矩阵得到;
非差分预编码矩阵处理模块, 用于在所述预编码矩阵指示接收模块接收 到所述非差分预编码矩阵指示后, 根据所述非差分预编码矩阵指示, 利用本 地预先存储的第二非差分码本, 得到预编码矩阵, 其中, 所述第二非差分码 本与所述第一非差分码本相一致。
本发明实施例提供的技术方案的有益效果是:
根据非差分码本和对角化差分码本, 获取参考 PMI和差分 PMI, 可以节 省反馈开销, 提高反馈精度, 从而进一步改进了反馈性能; 对角化差分码本 中包括的码字为对角化矩阵, 可以保持非差分码本本身具有的元素的幅度特 性(如恒模特性, 有限字符集约束特性)或者便于实现天线之间的功率分配。 附图说明
图 1是本发明实施例 1提供的一种获取预编码矩阵指示的方法流程图; 图 2是本发明实施例 2提供的一种获取预编码矩阵指示的方法流程图; 图 3是本发明实施例 3提供的一种获取预编码矩阵指示的方法流程图; 图 4是本发明实施例 4提供的一种获取预编码矩阵指示的方法流程图; 图 5是本发明实施例 7提供的一种获取预编码矩阵的方法流程图; 图 6是本发明实施例 8提供的一种获取预编码矩阵指示的装置结构示意 图;
图 7是本发明实施例 9提供的一种获取预编码矩阵的装置结构示意图; 图 8是本发明实施例 10提供的一种获取预编码矩阵指示的方法流程图; 图 9是本发明实施例 11提供的一种获取预编码矩阵指示的装置结构示意 图;
图 10是本发明实施例 12提供的一种获取预编码矩阵的方法流程图; 图 11是本发明实施例 13提供的一种获取预编码矩阵的装置结构示意图。 具体实施方式
为使本发明的目的、 技术方案和优点更加清楚, 下面将结合附图对本发 明实施方式作进一步地详细描述。
实施例 1
参见图 1, 本发明实施例提供了一种获取预编码矩阵指示的方法, 包括: 101 : 根据第一非差分码本和第一对角化差分码本, 获取参考预编码矩阵 指示和差分预编码矩阵指示, 其中, 第一对角化差分码本中包括的码字为对 角化矩阵。
进一步地, 根据第一非差分码本和第一对角化差分码本, 获取参考预编 码矩阵指示和差分预编码矩阵指示, 具体可以包括:
根据第一非差分码本, 基于预设的准则, 计算得到参考预编码矩阵指示; 根据第一非差分码本、 第一对角化差分码本和参考预编码矩阵指示, 基 于预设的准则, 计算得到差分预编码矩阵指示; 或
根据第一非差分码本和第一对角化差分码本, 获取参考预编码矩阵指示 和差分预编码矩阵指示, 具体可以包括:
根据第一非差分码本和第一对角化差分码本, 基于预设的准则, 计算得 到参考预编码矩阵指示和差分预编码矩阵指示。
进一步地, 第一 如下:
Figure imgf000007_0001
其中, 表示第一对角化差分码本 D 中包括的差分预编码矩阵指示, yt = 0,l...| |-l, |Z)|表示第一对角化差分码本 D中包括的码字的数目, =- 1, N表示发射天线的数目, N为正整数, ;表示相移, i' = l,2...N。
进一步地, 一对角化差分码本中包括的码字 的结构如下:
Figure imgf000007_0002
其中, 表示第一对角化差分码本 D 中包括的差分预编码矩阵指示, k =
Figure imgf000007_0003
|Ζ)|表示第一对角化差分码本 D中包括的码字的数目, =-1, Ν表示发射天线的数目, Ν为正整数, *(«^)表示相移, i = \,2』, m0表示相 邻发射天线间的相移差。
进一步地, 当包含 N个发射天线的双极化发射天线阵的前 N/2个发射天 线为一组同极化发射天线, 后 N/2个发射天线为另外一组同极化发射天线时, 第一对角化差分码本中包括的码字 的结构如下:
Ck =diag{Sm,e Sm}
其中, Sm ,
Figure imgf000007_0004
e j ,..., ",..., 、, 表示第一对角化差分码 本 D中包括的差分预编码矩阵指示, fc = 0,l...|Z)|-l, |D|表示第一对角化差分码 本 D中包括的码字的数目, =-1, N表示发射天线的数目, N为偶数, 和 6>mi表示相移, m和"为自然数, = l,2,...,N/2。
进一步地, 第一对角化差分码本中包括的码字 C¾的结构如下: = diag、ak eAi ,ak, ,… ,akieA ,...,akNeA'N }
其中, 表示第一对角化差分码本 D中的差分预编码矩阵指示, k = (),l...\D\-l, |Ζ)|表示第一对角化差分码本 D中包括的码字的数目, =-1 Ν表示发射天线的数目, Ν为正整数, 为实数, 表示相移, i = \,l』 进一步地, 第一非差分码 的结构可以如下:
Figure imgf000008_0001
其中, 表示第一非差分码本 Ww中包括的非差分预编码矩阵指示 t = 0,l
Figure imgf000008_0002
表示第一非差分码本 Ww中包括的码字的数目, r表示第 一非差分码本 Ww中包括的码字的秩, 4表示与 对应的对角矩阵 的索引, Rit表示与 ik对应的 n阶对角矩阵, Λ表示与 对应的归一化哈达马矩阵的索引, Ηηλ表示与 对应的 η阶归一化哈达马矩阵, η的取值与发射天线的数目相等, ( R; H„ , )(r)表示从旋转哈达马矩阵 R; Η„ ,中选择 r列构成的矩阵。
需要说明的是, 对角化矩阵 将11„ , 中的列矢量旋转得到 R; H„ ,, 故可 将1 ; Η„ ,称为旋转哈达马矩阵。 }
Figure imgf000008_0003
其中, 为实数, ,表示相移, =- 1, ί = 1,2,...,/!, n的取值与发射天 线的数目相等。
需要说明的是, 从旋转哈达马矩阵 RITH„ A中选择 r列构成矩阵 (R4H„ ;T)W, 具体可以根据实际应用状况, 釆用相关的方法确定具体选择哪 r列, 例如: 对 于非相关信道,列的选择应使得从不同的矩阵 R; H„ ,中分别抽取一个 r列子矩 阵构成的集合 {(R4H„,;t }中各个矩阵满足弦距最大; 对于相关信道, 列的选 择应使得从不同的矩阵1 ; Η„ ,中分别抽取一个 r列子矩阵构成的集合
{(RITH„,;t }中各个矩阵满足在天线阵列空间投影的零方向增益最大, 还可以 釆用其他任何可行的方法进行选择, 对此不做具体限定。
或进一步地, 第一非差分码本中包括的码字 的结构可以如下:
其中, 表示第一非差分码本 Ww中包括的预编码矩阵指示, fc = 0,l — 1
Figure imgf000008_0004
Ww|表示第一非差分码本 Ww中包括的码字的数目, r表示第一非差分码本 Ww中包括的码字的秩, 4表示与 对应的对角矩阵 R;的索引, R;表示与 4对 应的 n阶对角矩阵, A表示与 对应的归一化哈达马矩阵的索引, Η„Λ表示与 对应的 η阶归一化哈达马矩阵, η的取值与发射天线的数目相等,
( V R; ¾ H„ , Rf )(r)表示从旋转哈达马矩阵 R; H„ , R 中选择 r列构成的矩阵。
需要说明的是,对角化矩阵 R;将 Η„ , 中的列矢量旋转得到 R; H„ , Rf, 因此 R; H„ , R 也是一种旋转哈达马矩阵。
需要说明的是, 对角矩阵1 ;的结构具体可以表示为:
Figure imgf000009_0001
其中, 为实数, ,表示相移, = -1, = 1,2, ... , ;ί, n的取值与发射天线 的数目相等。
需要说明的是, 从旋转哈达马矩阵 R; H„ , R 中选择 r列构成矩阵
(R. H , R )W , 具体可以根据实际应用状况, 釆用相关的方法确定具体选择哪 r列, 例如: 对于非相关信道, 列的选择应使得从不同的矩阵 R; H„, Rf 中分别 抽取一个 r列子矩阵构成的集合 i I( VR; H„ , Rf " [ I中各个矩阵满足弦距最大; 对 于相关信道, 列的选择应使得从不同的矩阵 R; H„ , R 中分别抽取一个 r列子 矩阵构成的集合 i I( VR; H„ , Rf } I中各个矩阵满足在天线阵列空间投影的零方向 增益最大, 还可以釆用其他任何可行的方法进行选择, 对此不做具体限定。
进一步地, 参见图 1, 获取参考预编码矩阵指示和差分预编码矩阵指示之 后, 还可以包括:
102: 将参考预编码矩阵指示和差分预编码矩阵指示反馈给数据发送端。
103: 数据发送端接收参考预编码矩阵指示和差分预编码矩阵指示, 并利 用其自身中预先存储的第二非差分码本和第二对角化差分码本, 根据参考预 编码矩阵指示和差分预编码矩阵指示, 得到预编码矩阵。
其中, 第二非差分码本和第二对角化差分码本分别与第一非差分码本和 第一对角化差分码本相一致。
进一步地, 第一非差分码本为第一基码本, 第一对角化差分码本为第一 变换码本。
进一步地, 第二非差分码本为第二基码本, 第二对角化差分码本为第二 变换码本。
本发明实施例所述的获取预编码矩阵指示的方法, 获取参考 PMI和差分 PMI, 可以节省反馈开销, 提高反馈精度, 从而进一步改进了反馈性能; 对角 化差分码本中包括的码字为对角化矩阵, 可以保持非差分码本已具有的元素 的幅度特性(如恒模特性, 有限字符集约束特性)或者便于实现天线间的功 率分配; 并且, 可以使用从旋转哈达马矩阵得到的非差分码本, 其中哈达马 矩阵的各列可以分别与均勾线阵发射天线配置和双极化发射天线配置下强相 关和低相关信道的特征向量匹配, 从而提高量化的准确性, 旋转矩阵釆用对 角化矩阵, 可以保持在哈达马矩阵量化的基础上对空间进行细致量化; 而且, 哈达马矩阵的各个元素为 +1 或者 -1, 可以满足恒模特性, 且哈达马矩阵与其 它矩阵或者矢量的乘法运算可以简化为加法或者减法运算, 从而可以大大降 低基于 SINR ( Signal to Interference Noise Ratio , 信干噪比) 的 CQI ( Channel Quality Indicator,信道质量指示)计算、 PMI选择以及秩自适应的计算复杂性。 哈达马矩阵的各个列相互正交, 归一化之后可以保持非差分码本中码字的酉 矩阵特性, 可以保证发射的各个空间流功率分配相同, 满秩情况下各个发射 天线功率分配相同; 另外, 从旋转哈达马矩阵得到非差分码本的方法可以应 用于 2、 4、 8、 16, 32、 64等个数的发射天线且得到的非差分码本可以满足 恒模特性、 酉矩阵特性以及可以降低基于 SINR的 CQI计算、 PMI选择和秩 自适应的计算复杂度。 实施例 2
参见图 2, 本发明实施例提供了一种获取预编码矩阵指示的方法, 包括: 201 : UE选择使用基于参考 PMI的差分 PMI进行反馈,并获取参考 PMI。 参考 PMI可以是: 将 UE最近反馈的非差分 PMI作为参考 PMI, 具体可 以是非差分宽带 PMI或非差分子带 PMI等,可以根据实际应用状况进行选择。
本发明实施例中参考 PMI釆用非差分 PMI的形式,参考 PMI至少可以釆 用下面 2种方法中的一种得到:
1 )根据非差分码本, 基于预设的准则, 计算得到参考 PMI为 n (可记为 参考 PMI n ), 具体如式( 1 ) 所示:
n = arg max ( W. ) (1) i'=。,l...|w(r)卜 l'w^ww
其中, |WW |表示非差分码本 Ww的大小, 即非差分码本 Ww中包括的码字 的数目; r表示非差分码本 Ww中包括的码字的秩, ^表示非差分码本 Ww中 与参考 PMI 对应的码字; 表示与预设的准则对应的目标函数。 需要说明的是, 上述预设的准则可以是吞吐量最大化准则, 该准则对应 的目标函数可以是吞吐量最大化函数, 该吞吐量最大化函数可以是基于信息 容量计算实现, 也可以基于互信息或互信息的变形 (如互信息的加权)等实 现。 此外, 上述预设的准则也可以是弦距最大化准则, 当然还可以根据实际 应用状况, 对与预设的准则对应的目标函数进行灵活设置, 对此不做具体限 定。 其他地方的预设的准则与此处类似, 不再——赘述。
具体地, 非差分码本 ww可以是本发明实施例设置的一种非差分码本, 非 差分码本 ww中的码字从旋转哈达马矩阵中得到, 具体地非差分码本 ww中包 括的码字 w^)的结构可以如 :
Figure imgf000011_0001
其中, 表示非差分码本 ww中包括的非差分预编码矩阵指示,
k = 0,l
Figure imgf000011_0002
表示非差分码本 Ww中包括的码字的数目, r表示非差分 码本中包括的码字的秩, 表示与 对应的对角矩阵 R;t的索引, R;t表示与 4对 应的 n阶对角矩阵, A表示与 对应的归一化哈达马矩阵的索引, Η„Λ表示 与 对应的 η阶归一化哈达马矩阵, η的取值与发射天线的数目相等,
(R; H„ , f)表示从旋转哈达马矩阵 . Η„ ,中选择 r列构成的矩阵。
需要说明
b.
Figure imgf000011_0003
lk'n J
其中, bi f为实数, ,表示相移, =-1, = 1,2,...,«, n的取值与发射天 线的数目相等。
需要说明的是, 从旋转哈达马矩阵 R H„ Λ中选择 r列构成矩阵 (R4H„ Λ f, 具体可以根据实际应用状况, 釆用相关的方法确定具体选择哪 r列, 例如: 对 于非相关信道,列的选择应使得从不同的矩阵 R;H„】,中分别抽取一个 r列子矩 阵构成的集合 {(RitH„,;t }中各个矩阵满足弦距最大; 对于相关信道, 列的选 择应使得从不同的矩阵1 ; Η„ ,中分别抽取一个 r列子矩阵构成的集合
{(R4H„,;t }中各个矩阵满足在天线阵列空间投影的零方向增益最大, 还可以 釆用其他任何可行的方法进行选择, 对此不做具体限定。 例如, 对于 4比特 8 天线非差分码本, 有, ^)= (Ι ;Η8; f)。其中, 4=0,1,2,3、 =0,1,2,3, 并设 k = jk*4 + ik (需要说明的是, 并不限于此种形式, 可以根据实际应用状况设计 其他任何可行的形式, 如还可以设为 = 4*4+ ), r = l,2...8。如 r=l时, W,(1) = { R, H 、( R, HR , Γ表示取 R,.!^,.的某一列例如第一列。 r = 2,3...8与 r=l 不再 -, Η 8,0、 Η8 1、 Η8 2、 H83具体如下 -
Γι v 、 ' 1 1 1 1 1 1 1 一 ― 1 1 — 1 一 1 1 1 — 1 - 1—
1 — 1 1 一 1 1 - 1 1 一 1 1 一 1 — 1 1 1 -1 — 1 1
1 1 1 1 一 1 一 1 -1 -1 -1 -1 1 1 1 1 — 1 -1
1 1 一 1 -1 1 -1 1 1 -1 — 1
H 1
ϊ 1 -1 1 1 -1 -1 •1 = 8 1 1 1 1 1 1 1 1
1 — 1 — 1 1 1 - 1 -1 1 1 一 1 1 一 1 1 -1 1 一 1
1 1 — 1 -1 一 1 一 1 1 1 -1 -1 — 1 一 1 1 1 1 1
1 — 1 — 1 1 一 1 1 1 — 1 -1 1 — 1 1 1 -1 1 — 1 - 一 1 1 1 1 一 1 一 1 -1 - 1— 一 1 1 — 1 一 1 - - 1 — 1 1 1一
1 -1 1 -1 一 1 1 -1 1 一 1 — 1 1 - - 1 1 1 -1
1 1 1 1 1 1 1 1 -1 -1 1 1 - - 1 — 1 1 1
1 1 -1 1 -1 1 - 1 1 一 1 1 -1 1 1 一 1 - - 1 1 1 -1
H
8 一 1 一 1 1 1 1 1 -1 -1 •3 = , V Λ8 -1 -1 — 1 -1 1 1 1 1 一 1 1 1 -1 1 - 1 -1 1 -1 1 — 1 1 1 一 1 1 -1 一 1 一 1 1 1 一 1 一 1 1 1 1 1 1 1 1 1 1 1 一 1 1 1 -1 一 1 1 1 一 1 1 一 1 1 一 1 1 一 1 1 -1
R, 的对角线元素可以选择与 8点 DFT ( Discrete Fourier Transformation, 离散 傅立叶变换) 矢量一致, 具体如下
R0= ^ {1,1, 1,1, 1,1, 1,1}, Rv =diag{\ -\ -jXj -\ -j}
Figure imgf000012_0001
R3 =diagU,e e 4 ,e 4 ,-1,-e 4 ,-e 4 ,-e 4
此外, 一个 秩预编码码本可以给出如下表 1或表 2所示,
表 1
对角化旋转矩阵 归一化哈达马矩阵
H8,0 H8,l H8,2 H8,3 ύ¾¾{ΐ, 1,1, 1,1, 1,1,1} w。(8) w ) w8(8) i
w w6 (8) diag ,e 2 ,e 2 ,e 2 ,e 2 ,e 2 ,e 2 w/8) w3 (8) w9 (8)
Figure imgf000012_0002
「 3π .6π .9π Χΐπ \ίπ Χίπ .2\π Ί
diag , e 4 , e 4 , e 4 , e 4 , e 4 , e 4 , e 4 w5 (8) w7 (8) i i'flg {1,1, 1,-1, 1,1, 1,-1} w "1(28) w "1(48) w "1(38)
或者
表 2
Figure imgf000013_0003
其它低秩预编码码本可以在满足嵌套特性的条件下从上述满秩预编码码本导 出。
或具体地, 非差分码本中包括的码字 w 的结构可以如下:
)
Figure imgf000013_0001
其中, 表示非差分码本 Ww中包括的预编码矩阵指示, fc = 0,l...
表示非差分码本 Ww中包括的码字的数目, r表示非差分码本 Ww中包括的码 字的秩, 4表示与 对应的对角矩阵 Rit的索引, ¾表示与 4对应的 n阶对角 矩阵, A表示与 对应的归一化哈达马矩阵的索引, Η„Λ表示与 A对应的 n 阶归一化哈达马矩阵, n的取值与发射天线的数目相等, (R; H„ , Rf f)表示从 旋转哈达马矩阵 R,. H„ ,. R 中选择 r列构成的矩阵。
需要说明
Figure imgf000013_0002
其中, 为实数, ,表示相移, =-1, = 1,2,...,;ί, n的取值与发射天线 的数目相等。
需要说明的是, 从旋转哈达马矩阵 R; H„ , R 中选择 r列构成矩阵
(R.H , R )W, 具体可以根据实际应用状况, 釆用相关的方法确定具体选择哪 r列, 例如: 对于非相关信道, 列的选择应使得从不同的矩阵 R; H„ , Rf 中分别 抽取一个 r列子矩阵构成的集合 i(R; H„ , Rf "}中各个矩阵满足弦距最大; 对 于相关信道, 列的选择应使得从不同的矩阵 R; H„ , R 中分别抽取一个 r列子 矩阵构成的集合 i(R; H„ , R
I V ¾ } I中各个矩阵满足在天线阵列空间投影的零方向 增益最大, 还可以釆用其他任何可行的方法进行选择, 对此不做具体限定。 例如, 对于 4比特 8天线非差分码本, 有
Figure imgf000014_0001
4=0,1,2,3、 =0,1,2,3, 并设 fc = j 4 (需要说明的是, 并不限于此种形式,
),
Figure imgf000014_0002
一 列, 例如第一列 与 r=l时类似, 不再——赘述; Hsn、 Hsl、 H H 具体如下:
H。„ 丄
Figure imgf000014_0003
-
1 1 1 1 -1 -1 -1 - 1— 一 1 1 -1 一 1 - - 1 — 1 1 1一
1 一 1 1 -1 -1 1 -1 1 1 一 1 -1 1 - -1 1 1 -1
1 1 1 1 1 1 1 1 -1 -1 1 1 - - 1 — 1 1 1
1 1 一 1 1 -1 1 一 1 一 1 , 1 -1 1 1 一 1 - - 1 1 1 -1
H。, = 8 — 1 一 1 1 1 1 1 -1 -1 , = V8 -1 -1 -1 -1 1 1 1 1
— 1 1 1 -1 1 一 1 -1 1 -1 1 -1 1 1 一 1 1 -1
— 1 一 1 1 1 -1 -1 1 1 1 1 1 1 1 1 1 1
— 1 1 1 -1 -1 1 1 一 1 1 一 1 1 一 1 1 一 1 1 -1
R;的对角线元素可以选择与 8点 DFT矢量一致, 具体如下:
-1,-j, 1, j,—1, -j]
Figure imgf000014_0004
此外, 一个满秩预编码码本可以给出如下表 3或表 4所示,
表 3
或者
Figure imgf000015_0001
其它低秩预编码码本可以在满足嵌套特性的条件下从上述满秩预编码码本导 出。 或具体地, 非差分码本中包括的码字 的结构可以如下:
W,(r) = R Β('
其中, 表示非差分码本 Ww中包括的预编码矩阵指示, fc = 0,l
Figure imgf000016_0001
表示非差分码本 Ww中包括的码字的数目, r表示非差分码本 Ww中包括的码 字的秩, 4表示与 对应的对角矩阵 的索引, R4表示与 4对应的 n阶对角 矩阵, 表示与索引 对应的一个秩为 r的 n发射天线非差分码本中的一个 码字, 与 对应。
需要说明的
Figure imgf000016_0002
其中, 为实数, ,表示相移, j2 =—l , = 1, 2,...,«, n的取值与发射天线 的数目相等。
并且, 非差分码本还可以釆用现有技术中的任意一种非差分码本, 例如
LTE R8的码本, 可以根据实际应用状况进行灵活选择, 对此不做具体限定。
2 )选择 UE最近反馈的一个非差分宽带 PMI (设该非差分宽带 PMI为 n ) 作为参考 PMI。
具体地, Node B可以通过高层信令或者下行物理控制信道指示 UE使用 基于参考 PMI的差分 PMI进行反馈, UE收到 Node B的指示后选择使用基于 参考 PMI的差分 PMI进行反馈。或者 Node B和 UE事先约定 UE使用基于参 考 PMI的差分 PMI进行反馈,当 UE进行反馈时,自动选择使用基于参考 PMI 的差分 PMI进行反馈。 并不限于上述二种方法, 可以根据实际应用状况设置 其他任何可行的方法, 使 UE可以选择使用基于参考 PMI的差分 PMI进行反 馈。
202: UE将参考 PMI反馈给 NodeB,并根据参考 PMI、非差分码本 Ww和 对角化差分码本0, 基于预设的准则, 计算得到差分 PMI。
其中, 非差分码本 Ww可以釆用与步骤 201中类似的非差分码本, 即可以 疋
本发明实施例设置的码字从旋转哈达马矩阵中得到的非差分码本, 也可以是 现有技术中的任意一种非差分码本。
其中, 对角化差分码本是本发明实施例设置的包含的码字为对角化矩阵 的差分码本, 本发明实施例中设置的对角化差分码本 D中包括的码字 C¾的结 构如式(2 )所示: Ck = diag } ( 2 ) 其中, 表示对角化差分码本 D中包括的差分 ΡΜΙ, yt = 0,l...| |-l, |Z)|表 示差分码本 D中包括的码字的数目; =-1, N表示发射天线的数目, N为 正整数, , · = 1,2,3...Λ 表示相移, 可以根据具体的发射天线数目和具体发 射天线的配置得到。
根据参考 ΡΜΙ、 非差分码本 (^和对角化差分码本0, 基于预设的准则, 计算得到差分 ΡΜΙ为 (记作差分 PMI k ), 具体如式(3) 所示:
k= argmax (C;W„) ( 3 )
=0,1...|D|- 1'C'eD
其中, C;表示对角化差分码本 D中与差分 PMI ί·对应的码字; W„表示非 差
分码本 Ww中与参考 PMI n对应的码字; /(C; „)表示与预设的准则对应的目 标函数。
203: UE将差分 PMI反馈给 NodeB。
本发明实施例中差分 PMI为 , 所以具体是将 k反馈给 NodeB。
204: NodeB接收参考 PMI和差分 PMI, 并根据参考 PMI和差分 PMI, 利用非差分码本 ^)和对角化差分码本 D, 计算得到预编码矩阵 V。
需要说明的是, 由于在步骤 202 中 UE已将参考 PMI反馈给 NodeB, 在 步骤 203中 UE才将差分 PMI反馈给 NodeB,因此 NodeB会先接收到参考 PMI, 然后再接收到差分 PMI, 因此 NodeB可以将先接收到的参考 PMI进行存储, 等接收到差分 PMI时, 再根据参考 PMI和差分 PMI计算得到预编码矩阵 。
具体地, 对于 SU-MIMO ( Single User Multiple Input Multiple Output, 单 用户多输入多输出) 系统, NodeB根据参考 PMI, 从 NodeB中预先存储的非 差分码本 Ww (与 UE中预先存储的非差分码本相一致)中查询得到参考 PMI 对应的码字,并根据差分 PMI,从 NodeB中预先存储的对角化差分码本 D (与 UE中预先存储的对角化差分码本相一致) 中查询得到差分 PMI对应的码字; 将差分 PMI对应的码字与参考 PMI对应的码字进行相乘, 将乘积结果作为预 编码矩阵 。
本发明实施例中, 参考 PMI为 n, 与参考 PMI n对应的码字为 W„; 差分 PMI为 k, 与差分 PMIk对应的码字为 C¾ , 因此本发明实施例得到的预编码矩 阵 如式(4)所示:
V = C¾W„ (4) 对于 MU-MIMO ( Multiple User Multiple Input Multiple Output, 多用户多 输多输出)系统, NodeB根据同时配对的各个 UE反馈的参考 PMI和差分 PMI, 利用非差分码本 Ww和对角化差分码本 D, 基于 ZF-BF ( Zero-Forcing Beam Forming, 迫零波束赋形) 算法或基于最大 SLNR ( Signal-Leakage Plus Noise Ratio, 信泄噪比) 准则计算得到预编码矩阵 。
其中,各个 UE反馈 PMI的具体过程与步骤 201-203类似,此处不再赘述。 并且, 为了便于理解 MU-MIMO系统时的具体算法, 下面以利用 ZF-BF算法 计算得到预编码矩阵 为例进行说明:
对于同时配对的两个 UE,假定其中一个 UE对应的参考 PMI为 (并设 参考 PMI 对应的码字为 Wnl ),对应的差分 PMI为 1¾ (并设参考 PMI 1¾对应 的码字为 C¾1 ), 另一个 UE对应的参考 PMI为 n2 (并设参考 PMI n2对应的码 字为 W„2 )、对应的差分 PMI为 k2 (并设参考 PMI k2对应的码字为 Ck2 ), 釆用 上述 SU-MIMO系统的方法,计算得到一个 UE对应的预编码矩阵 如式(5) 所示, 另一个 UE对应的预编码矩阵 2如式(6)所示:
^C^W, (5)
,=C,,,W„, (6)
设置 H V, V, G = Hfl HH
根据 ή和 G得到预编 矩阵 如式(7)所示:
^ 。r,|gl|- ^, Γ. — i (7) 其中, P表示 NodeB的发射总功率, S表示两个 UE的空间复用的总层 数, §ί · = 0,1.^- 1)表示 G的第 i列矢量, |^|( = 0,1.^- 1)表示 G的殴氏范数。
釆用基于最大 SLNR准则计算得到预编码矩阵 的过程与现有技术类似, 此处不再赘述。
205: Node Β利用预编码矩阵 ,对待发送数据 s进行预处理, 并将预处 理后的待发送数据 s通过发射天线发送给 UE。
206: UE接收到接收信号 , 并对接收信号 y进行数据检测。
具体地, UE接收到的接收信号 y如式 (8)所示:
y = HVs + n (8)
其中, y表示 UE接收到的接收信号; H表示信道矩阵; 表示预编码矩 阵; s表示待发送数据; n表示加性高斯白噪声。 需要说明的是, 在上述过程中, 参考 PMI和差分 PMI通过步骤 202和步 骤 203分别反馈给 Node B,实际应用中也可以将参考 PMI和差分 PMI同时反 馈给 Node B, 具体地可以是, UE在得到参考 PMI后, 可以将参考 PMI先存 储, 然后在步骤 203 中 UE将参考 PMI和差分 PMI同时反馈给 NodeB。 或者 可以是在步骤 201中根据非差分码本 Ww和对角化差分码本 D,基于预设的准 则, 同时计算得到参考 PMI为 n和差分 PMI为 k, 具体如式 (9)所示:
(k, n) = arg max (C;W. ) ( 9 )
i=0,l,...,|D|-l,C,eD
=0,l,...,|w(')|-l,WjeW(')
然后在步骤 202中将参考 PMI和差分 PMI同时反馈给 NodeB后,直接执 行步骤 204。
并且需要说明的是, 上述在反馈参考 PMI和差分 PMI时, 无论反馈的是 差分 PMI还是参考 PMI, 均可以是对整个系统带宽反馈一个 PMI, 也可以是 将系统分为多个 BP(Bandwidth Part, 带宽部分), 每个 BP 中含有多个子带 (sub-band),为每个子带反馈一个 PMI或者多个子带反馈一个 PMI (例如 Best-M 方式: 对选择的 M个子带反馈一个 PMI)。 即上述参考 PMI和差分 PMI即可 以是宽带 PMI, 也可以是子带 PMI, 也可以按照 Best-M方式对多个子带反馈 一个 PMI。
本发明实施例所述的获取预编码矩阵指示的方法,根据非差分码本和对角 化差分码本, 获取参考 PMI和差分 PMI, 可以节省反馈开销, 提高反馈精度, 从而改进了反馈性能; 对角化差分码本中包括的码字为对角化矩阵, 可以保 持非差分码本已具有的元素的幅度特性(如恒模特性,有限字符集约束特性)。 并且, 可以使用从旋转哈达马矩阵得到的非差分码本, 其中哈达马矩阵的各 列可以分别与均匀线阵发射天线配置和双极化发射天线配置下强相关和低相 关信道的特征向量匹配, 从而提高量化的准确性, 旋转矩阵为对角化矩阵, 可以保持在哈达马矩阵量化的基础上对空间进行细致量化; 而且, 哈达马矩 阵各个元素为 +1或者 -1, 可以满足恒模特性; 且哈达马矩阵与其它矩阵或者 矢量的乘法运算可以简化为加法或者减法运算, 从而可以大大降低基于 SINR 的 CQI计算、 PMI选择以及秩自适应的计算复杂性。 哈达马矩阵的各个列相 互正交, 归一化之后可以保持非差分码本中码字的酉矩阵特性, 可以保证发 射的各个空间流功率分配相同, 满秩情况下各个发射天线功率分配相同, 可 以保证发射的各个空间流功率分配相同, 满秩情况下各个发射天线功率分配 相同; 进一步地, 从旋转哈达马矩阵得到非差分码本的方法可以应用于 2、 4、 8、 16、 32、 64等个数的发射天线且得到的非差分码本可以满足上述恒模特性、 酉矩阵特性以及降低基于 SINR的 CQI计算、 PMI选择和秩自适应的计算复 杂度; 另外, 使用 UE最近反馈的非差分 PMI作为参考 PMI, 不仅可以减少 开销, 而且可以充分利用信道在频域和时域上或者空域上的相关性, 从而提 高反馈精度。 实施例 3
参见图 3, 本发明实施例提供了一种获取预编码矩阵指示的方法, 包括: 301: UE选择使用基于参考 PMI的差分 PMI进行反馈,并获取参考 PMI。 参考 PMI可以是: 将 UE最近反馈的非差分 PMI和差分 PMI作为参考 PML
本发明实施例中参考 PMI具体可以包括非差分 PMI和 m个参考差分 PMI (m是自然数, 可以根据实际应用状况, 选择 m的取值, 如可以根据子带的 个数选择 m的取值)。 其中, 可以釆用实施例 2步骤 201中的方法获取非差分 PMI为 n。 并根据对角化差分码本0, 基于预设的准则, 计算得到 m个参考 差分 PMI分别为 nQ、 n!...nm ( m个参考差分 PMI分别记为参考差分 PMI n。、 参考差分 PMIn…参考差分 PMInm), 具体如式(10)所示:
^-¾)= argmax (C....C.C,W„) (10)
(¾,¾,...,ί„),;-ο,ι...|ο|-ι,ο,.£θ, =ο,ι
302: UE将参考 PMI反馈给 NodeB,并根据参考 PMI、非差分码本 Ww和 对角化差分码本0, 基于预设的准则, 计算得到差分 PMI。
其中, 非差分码本 Ww和对角化差分码本 D与实施例 2步骤 202中的非 差分码本 Ww和对角化差分码本 D相同, 此处不再赘述。
根据参考 PMI ( n, n0、 ni...nm), 非差分码本 Ww和对角化差分码本 D, 基于预设的准则, 计算得到差分 PMI为 (记作差分 PMI k ), 具体如式(11 ) 所示:
k = argmax / (C, (C 〜 \¥„ )) (11) i=0,l...|D|-l,C,eD V V1 。 〃
其中, C;表示对角化差分码本 D中与差分 PMI ζ·对应的码字; C„C„ ...C„ 表
示对角化差分码本 D中分别与参考差分 PMI n0、参考差分 PMI ? ..参考差分 PMI 对应的码字; \¥„表示非差分码本 Ww中与非差分 PMIn对应的码字。
303: UE将差分 PMI反馈给 NodeB。 本发明实施例中差分 PMI为 , 所以具体是将 k反馈给 NodeB。
304: NodeB接收参考 PMI和差分 PMI, 并根据参考 PMI和差分 PMI, 利用非差分码本 ^)和对角化差分码本 D, 计算得到预编码矩阵 。
具体地, 对于 SU-MIMO系统, NodeB根据参考 PMI, 从 NodeB中预先 存储的非差分码本 Ww (与 UE中预先存储的非差分码本相一致)中查询得到 参考 PMI对应的码字, 并根据差分 PMI, 从 NodeB中预先存储的对角化差分 码本 D (与 UE中预先存储的对角化差分码本相一致) 中查询得到差分 PMI 对应的码字; 将差分 PMI对应的码字与参考 PMI对应的码字进行相乘, 将乘 积结果作为预编码矩阵 。
本发明实施例中,参考 PMI包括非差分 PMI和 m个参考差分 PMI,非差 分 PMI为 n, m个参考差分 PMI分别为 n。、 n!...nm, 与非差分 PMI n对应的 码字为 W„、与参考差分 PMI n。、 !^…!^对应的码字分别为^、 C„—C„ ; 差分 PMI为 k, 与差分 PMI k对应的码字为 C¾ , 因此本发明实施例得到的预编码矩 阵 如式(12 ) 所示:
= C¾ (C„m ( 12 )
对于 MU-MIMO系统, NodeB根据同时配对的各个 UE反馈的参考 PMI 和差分 PMI, 利用非差分码本 Ww和对角化差分码本 D, 基于 ZF-BF算法或 基于最大 SLNR准则计算得到预编码矩阵 。具体过程与实施例 2步骤 204类 似, 此处不再赘述。
305: Node B利用预编码矩阵 ,对待发送数据 s进行预处理, 并将预处 理后的待发送数据 s通过发射天线发送给 UE。
306: UE接收到接收信号 , 并对接收信号 y进行数据检测。
具体地, UE接收到的接收信号 y如式(13 )所示:
y = HVs + n ( 13 ) 其中, y表示 UE接收到的接收信号; H表示信道矩阵; 表示预编码矩 阵; s表示待发送数据; n表示加性高斯白噪声。
需要说明的是, 在上述过程中, 参考 PMI和差分 PMI通过步骤 302和步 骤 303分别反馈给 Node B,实际应用中也可以将参考 PMI和差分 PMI同时反 馈给 Node B, 具体地可以是, UE在得到参考 PMI后, 可以将参考 PMI先存 储, 然后在步骤 303 中 UE将参考 PMI和差分 PMI同时反馈给 NodeB。 或者 可以是在步骤 301中根据非差分码本 Ww和对角化差分码本 D,基于预设的准 则, 同时计算得到参考 PMI为 n、 η。、 η^ .ι 和差分 PMI为 , 具体如式 (14a) 所示:
Figure imgf000022_0001
( 14a ) 并且需要说明的是, 步骤 301中参考差分 PMI也可以递归实现, 具体 (14b) 所示:
arg max f l C^ (C^ -^ W ,j=0,l, ..,m ( 14b ) 然后在步骤 302中将参考 PMI和差分 PMI同时反馈给 NodeB后, 直接执 行步骤 304。
并且需要说明的是, 上述在反馈参考 PMI和差分 PMI时, 无论反馈的是 差分 PMI还是参考 PMI, 均可以是对整个系统带宽反馈一个 PMI, 也可以是 将系统分为多个 BP, 每个 BP中含有多个子带, 为每个子带反馈一个 PMI, 或者多个子带反馈一个 PMI (例如 Best-M方式: 对选择 M个子带反馈一个 PMI)。 即上述参考 PMI和差分 PMI即可以是宽带 PMI, 也可以是子带 PMI, 也可以按照 Best-M方式对多个子带反馈一个 PMI。
本发明实施例所述的获取预编码矩阵指示的方法,根据非差分码本和对角 化差分码本, 获取参考 PMI和差分 PMI, 可以节省反馈开销, 提高反馈精度, 从而改进了反馈性能; 对角化差分码本中包括的码字为对角化矩阵, 可以保 持非差分码本已具有的元素的幅度特性, 如恒模特性和有限字符集约束特性。 并且, 可以使用从旋转哈达马矩阵得到的非差分码本, 其中哈达马矩阵的各 列可以分别与均匀线阵发射天线配置和双极化发射天线配置下强相关和低相 关信道的特征向量匹配, 从而提高量化的准确性, 旋转矩阵为对角化矩阵, 可以保持在哈达马矩阵量化的基础上对空间进行细致量化; 而且, 哈达马矩 阵的各个元素为 +1或者 -1, 可以保持非差分码本的恒模特性; 且, 哈达马矩 阵与其它矩阵或者矢量的乘法运算可以简化为加法或者减法运算, 从而可以 大大降低基于 SINR的 CQI计算、 PMI选择和秩自适应的计算复杂性。 哈达 马矩阵的各个列相互正交, 可以保持非差分码本中码字的酉矩阵特性, 可以 保证发射的各个空间流功率分配相同, 满秩情况下各个发射天线功率分配相 同; 进一步地,从旋转哈达马矩阵得到非差分码本的方法可以应用于 2、 4、 8、 32、 64等个数的发射天线且得到的非差分码本可以满足恒模特性、 酉矩阵特 性以及降低基于 SINR的 CQI计算、 PMI选择和秩自适应的计算复杂度; 另 夕卜,使用 UE最近反馈的非差分 PMI和 M个参考差分 PMI作为参考 PMI, 不 仅可以进一步地减少开销, 而且可以更充分地利用信道的频域和时域或者空 域相关性, 从而进一步地提高反馈精度。 实施例 2和实施例 3所述的方法也适用于 CoMP ( Coordinated Multiple Point transmission, 协作多点传输 ) 系统, 下面以其应用于 CoMP系统为例进 行进一步地说明。
实施例 4
参见图 4, 本发明实施例提供了一种获取预编码矩阵指示的方法, 包括: 401 : UE选择使用基于参考 PMI的差分 PMI进行反馈, 并获取每个小区 的参考 PMI。
具体地, UE所在 CoMP系统的所有小区的基站可以通过高层信令或者下 行物理控制信道指示 UE使用基于参考 PMI的差分 PMI进行反馈, UE收到 指示后选择使用基于参考 PMI的差分 PMI进行反馈。 并且需要说明的是, 本 发明实施例中 CoMP系统的所有小区的基站通过 eNode B ( evolved Node B, 演进型基站)进行统一管理。
本发明实施例中假定 CoMP系统共有 M个小区, M是正整数, 每个小区 的参考 PMI依次为 ηι、 n2、 . . . nM。 需要说明的是, 本发明实施例中的每个小 区的参考 PMI即可以釆用实施例 2步骤 201中所述的参考 PMI的形式, 也可 以釆用步骤实施例 3步骤 301中所述的参考 PMI的形式, 可以根据实际应用 状况进行灵活选择。
402: UE将每个小区的参考 PMI反馈给 eNode B, 并根据每个小区的参 考 PMI、 非差分码本 ^)和对角化差分码本 D, 基于预设的准则, 计算得到每 个小区的差分 PMI。
其中, 非差分码本 Ww与实施例 2步骤 202中的非差分码本 Ww相同, 此 处不再赘述。 对角化差分码本 D可以与实施例 2步骤 202中的对角化差分码 本 D相同, 也可以是实施例 2步骤 202中的对角化差分码本中矩阵各乘以一 个相移矩阵如 可以乘以 diag {e k , ej" e k e k }得到的对角化矩阵。
根据每个小区的参考 PMI、 非差分码本 Ww和对角化差分码本 D,基于预 设的准则, 计算得到每个小区的差分 PMI 分别为 、 . M, 具体如式(15 ) 所示:
Figure imgf000024_0001
其中, M表示小区的个数, M是正整数; C4 Ci2...CM表示对角化差分码 本 D中分别与差分 PMI z; ,.. JM对应的码字; W W¾ ...WnM表示非差分码 本 Ww中分别与参考 PMI /¾ n2 ¾对应的码字; /C) 示与预设的准 则对应的目标函数; Pi表示 UE到小区 i的基站的大尺度衰落对应,为 eNode B 和 UE共知的功率控制参数; ^rm{.}表示对矩阵的各列进行归一化;
¾ = 0,1...|Z)|-1,C. e i2 =
Figure imgf000024_0002
1,C½ e Z) ; (·)Η表示矩 阵或者向量的共轭转置操作。
403: UE将每个小区的差分 ΡΜΙ反馈给 eNode B
本发明实施例中所有小区的差分 PMI依次为 、 K ,所以具体是将 k2---kM反馈给 eNode B
404: eNode B接收每个小区的参考 PMI和每个小区的差分 PMI, 并根据 参考 PMI和差分 PMI, 利用非差分码本 ^)和对角化差分码本 D, 计算得到 每个小区的预编码矩阵 V
本发明实施例中共有 M个小区, 设每个小区的预编码矩阵 依次为预编 码矩阵 、预编码矩阵 2...预编码矩阵 。 eNode B根据参考 PMI和差分 PMI 利用非差分码本^)和对角化差分码本 D,计算得到每个小区的预编码矩阵 如式 ( 16)所示:
(J chwJ… (V^c,Mw¾)ff
Figure imgf000024_0003
其中, M表示小区的个数, M是正整数; C;i Ci2...CiM表示对角化差分码 本 D中分别与差分 PMI ί;、 2... M对应的码字; W W ...WnM表示非差分码 本 Ww中分别与参考 PMI /¾、 对应的码字; A.表示 UE到小区 ί·的基站 的大尺度衰落对应, 为 eNode B和 UE共知的功率控制参数; 《orm{.}表示对矩 阵的各列进行归一化; ( 广表示矩阵或者向量的共轭转置操作;
¾ = 0,1...|Z)|-1,C. e ί·2 = 0,1...|Ζ)|— l,Ci2e Z .. M =0,1...|Z)|— l,CiM e Z)
405: eNode B利用每个小区的预编码矩阵 ,对每个小区的待发送数据 s 进行预处理, 并将预处理后的待发送数据 s通过发射天线发送给 UE
406: UE接收到接收信号 , 并对接收信号 y进行数据检测。
具体地, UE接收到的接收信号 y如式(17)所示: y = i i7g{H1,H2,...,H,...,HM}/io/77i
Figure imgf000025_0001
(17) 其中, H;( = 1,2...M)表示 UE到小区 的基站的信道矩阵, 其他符合的含义 与式(15)和式(16)相同, 此处不再赘述。
本发明实施例所述的获取预编码矩阵指示的方法,根据非差分码本和对角 化差分码本, 获取参考 PMI和差分 PMI, 可以节省反馈开销, 提高反馈精度, 从而改进了反馈性能; 对角化差分码本中包括的码字为对角化矩阵, 可以保 持非差分码本已具有的元素的幅度特性(如恒模特性,有限字符集约束特性)。 并且, 可以使用从旋转哈达马矩阵得到的非差分码本, 其中哈达马矩阵的各 列可以分别与均匀线阵发射天线配置和双极化发射天线配置下强相关和低相 关信道的特征向量匹配, 从而提高量化的准确性, 旋转矩阵为对角化矩阵, 可以保持在哈达马矩阵量化的基础上对空间进行细致量化; 哈达马矩阵各个 元素为 +1或者 -1, 可以保持非差分码本的恒模特性; 且, 哈达马矩阵与其它 矩阵或者矢量的乘法运算可以简化为加法或者减法运算, 从而可以大大降低 基于 SINR的 CQI计算、 PMI选择和秩自适应的计算复杂性。 哈达马矩阵的 各个列相互正交, 可以保持非差分码本中码字的酉矩阵特性, 可以保证发射 的各个空间流功率分配相同, 满秩情况下各个发射天线功率分配相同; 进一 步地, 从旋转哈达马矩阵得到非差分码本的方法可以应用于 2、 4、 8、 16、 32、 64等个数的发射天线且得到的非差分码本可以满足恒模特性、 酉矩阵特性以 及可以降低基于 SINR的 CQI计算、 PMI选择和秩自适应的计算复杂度; 另 外, 可以获取每个 d、区的参考 PMI和差分 PMI, 可以应用于 CoMP系统, 提 高了应用范围。 另夕卜, 需要说明的是, 当考虑到发射天线的功率分配时,上述实施例 2-4 中的对角化差分码本中包括的码字 的结构还可以如式(18)所示:
C { 1, 2^ '2,…, '、.." '"} ( 18 ) 其中, 表示对角化差分码本 D中的差分 PMI, yt = 0,l...| |-l, |Z)|表示对 角化差分码本 D 中包括的码字的数目; f=-l , N表示发射天线的数目, N 为正整数, 为实数, , · = 1,2,3...Λ 表示第 个发射天线的相移, 可以根 据具体的发射天线数目和具体发射天线的配置得到。 其中, ,'( = l,2...N)的 值可以根据实际应用状况进行设置, 例如 (i = 1,2...N)可以从 16QAM(Quadrature Amplitude Modulation, 正交幅度调制)或者 64QAM等星 座图的星座点中选择得到。 当对角化差分码本中的码字的结构如式(18 ) 所 示时, 可以实现发射天线的功率分配。 实施例 5
需要说明的是,本发明实施例与实施例 1-4的区别在于, 当针对强相关发 射天线配置时, 上述实施例 1-4中的对角化差分码本中包括的码字 C¾的结构 可以进一步优化为如式(19)所示:
Ck =diag{eMme e] me ...,e]iir{me ..^ ( 19 ) 其中, 表示对角化差分码本 D中包括的差分 PMI, k =
Figure imgf000026_0001
|Ζ)|表 示对角化差分码本 D中包括的码字的数目, =-1, N表示发射天线的数目, N为正整数, *(«^)表示第 ί·个发射天线的相移, i = \,2』, m0表示相邻发射天 线间的相移差。
此外, 需要说明的是, 上述对角线元素也可以根据具体的天线配置交换 位置, 如沿着对角线位置循环移位等, 或者所有对角线元素乘以一个相移因 子。
需要说明的是, 相移差 m0的取值应尽量关于 0对称分布;在反馈开销容 许的条件下,可以分配更多的差分矩阵在 0相移附近,例如: m的取值可以为: m = 0,±1,±2,±4,±8,±16,±32,±64....。例如: 本发明实施例中提供的 2比特的 4天线 对角化差分码本如表 5、 8所示, 3比特 4天线对角化差分码本如表 6、 9所示, 4比特 4天线对角化差分码本如表 7、 10所示。 本发明实施例中提供的 2比特 的 8天线对角化差分码本如表 11、 14所示, 3比特 8天线对角化差分码本如 表 12、 15所示, 4比特 8天线对角化差分码本如表 13、 16所示。 表 5.
Figure imgf000026_0002
表 6.
Figure imgf000027_0001
表 7.
Figure imgf000027_0002
表 8.
Figure imgf000028_0001
表 10
Figure imgf000028_0002
Figure imgf000029_0001
表 11
Figure imgf000029_0002
6 diag {1, -j,—1, j, 1,—j,—1, j}
7 ώ'α^ {1,-1, 1,-1, 1,-1, 1,-1} 表 13
Figure imgf000030_0001
Figure imgf000031_0001
表 14
Figure imgf000031_0002
表 15
Figure imgf000031_0003
表 16
Figure imgf000032_0001
本发明实施例所述的获取预编码矩阵指示的方法,根据非差分码本和对角 化差分码本, 获取参考 PMI和差分 PMI, 可以节省反馈开销, 提高反馈精度, 从而改进了反馈性能; 对角化差分码本中包括的码字为对角化矩阵, 可以保 持非差分码本已具有的元素的幅度特性(如恒模特性,有限字符集约束特性)。 并且, 可以使用从旋转哈达马矩阵得到的非差分码本, 其中哈达马矩阵的各 列可以分别与均匀线阵发射天线配置和双极化发射天线配置下强相关和低相 关信道的特征向量匹配, 从而提高量化的准确性, 旋转矩阵为对角化矩阵, 可以保持在哈达马矩阵量化的基础上对空间进行细致量化; 哈达马矩阵各个 元素为 +1或者 -1, 可以满足恒模特性, 且哈达马矩阵与其它矩阵或者矢量的 乘法运算可以简化为加法或者减法运算,从而可以大大降低基于 SINR的 CQI 计算、 PMI选择和秩自适应的计算复杂度; 归一化哈达马矩阵的各个列相互 正交, 可以保持非差分码本中码字的酉矩阵特性, 可以保证发射的各个空间 流功率分配相同, 满秩情况下各个发射天线功率分配相同; 进一步地, 从旋 转哈达马矩阵得到非差分码本的方法可以应用于 2、 4、 8、 16、 32、 64等个 数的发射天线且得到的非差分码本可以满足恒模特性、 酉矩阵特性以及降低 基于 SINR的 CQI计算、 PMI选择和秩自适应的计算复杂度。 另外, 通过对 对角化差分码本中的码字进行进一步地优化, 可以进一步地提高反馈的性能。 实施例 6
随着发射天线的增加, 双极化发射天线阵将成为一种普遍的配置, 为此 本发明实施例提供了一种获取预编码矩阵指示的方法,与实施例 2-5中所述的 方法的区别在于,本发明实施例利用实施例 2-5中的单极化发射天线的对角化 差分码本(可称为单极化对角化差分码本), 获取双极化发射天线阵的对角化 差分码本(可称为双极化对角化差分码本)。
具体地,当包括 N个发射天线的双极化发射天线阵的前 N/2( 1、2、3〜Ν/2 ) 个发射天线为一组同极化发射天线, 后 Ν/2 ( Ν/2+1、 N/2+2〜N )个发射天线 为另外一组同极化发射天线时, 对角化差分码本中包括的码字 C¾的结构如式 ( 20 )所示: 其中, Sm =
Figure imgf000033_0001
码本 D中包括 的差分 PMI, k = 0, l ...\D\ - l , |Ζ)|表示对角化差分码本 D中包括的码字的数目, j'2 = -1, N表示发射天线的数目, N为偶数, 和 6>m i表示相移参数, m和"为 自然数, = l, 2, ..., N/2。
需要说明的是, 上述式(20 ) 中矩阵的前 N/2个对角元素对应于一组同极 化发射天线, 后 N/2个对角元素对应于另外一组同极化发射天线。 并且如果 双极化发射天线阵的排列位置发生了变化, 则将上述元素的位置作相应的交 换即可, 具体结构类似, 此处不再赘述。
本发明实施例所述的获取预编码矩阵指示的方法, 根据非差分码本和对角 化差分码本, 获取参考 PMI和差分 PMI, 可以节省反馈开销, 提高反馈精度, 从而改进了反馈性能; 对角化差分码本中包括的码字为对角化矩阵, 可以保 持非差分码本已具有的元素的幅度特性(如恒模特性,有限字符集约束特性)。 并且, 可以使用从旋转哈达马矩阵得到的非差分码本, 其中哈达马矩阵的各 列可以分别与均匀线阵发射天线配置和双极化发射天线配置下强相关和低相 关信道的特征向量匹配, 从而提高量化的准确性, 旋转矩阵为对角化矩阵, 可以保持在哈达马矩阵量化的基础上对空间进行细致量化; 而且, 哈达马矩 阵各个元素为 +1或者 -1, 可以保持非差分码本的恒模特性, 且哈达马矩阵与 其它矩阵或者矢量的乘法运算可以简化为加法或者减法运算, 从而可以大大 降低基于 SINR的 CQI计算、 PMI选择和秩自适应的计算复杂度。 哈达马矩 阵的各个列相互正交, 可以保持非差分码本中码字的酉矩阵特性, 可以保证 发射的各个空间流功率分配相同, 满秩情况下各个发射天线功率分配相同; 进一步地,从旋转哈达马矩阵得到非差分码本的方法可以应用于 2、 4、 8、 16、 32、 64等个数的发射天线且得到的非差分码本可以满足恒模特性、 酉矩阵特 性以及降低基于 SINR的 CQI计算、 PMI选择和秩自适应的计算复杂度; 另 外, 通过单极化发射天线的对角化差分码本构造出双极化发射天线阵的对角 化差分码本, 可以充分利用双极化发射天线阵的特点, 提高双极化发射天线 阵配置下对角化差分码本的性能。 上述实施例 2-6所述的方法, 以下行系统( Node B向 UE发送数据) 为 例进行了说明, 需要说明的是, 上述实施例 2-6所述的方法, 同样适用于上行 系统(如 UE向发送 Node B数据)。 当将实施例 2-6所述的方法应用于上行系 统如 UE向发送 Node B数据时, 与应用于下行系统的区别在于, Node B按照 与实施例 2-6类似的步骤获取参考 PMI和差分 PMI并通知给 UE即可, UE接 收 NodeB的通知进行预编码并发送数据给 Node B, Node B接收数据并进行 数据检测。 实施例 7
参见图 5,本发明实施例提供了一种获取预编码矩阵的方法,该方法包括:
501 : 接收参考预编码矩阵指示和差分预编码矩阵指示。
其中, 参考预编码矩阵指示和差分预编码矩阵指示由数据接收端根据第 一非差分码本和第一对角化差分码本得到并发送, 第一对角化差分码本中包 括的码字为对角化矩阵。
其中, 第一非差分码本 (4可以是本发明实施例设置的一种非差分码本, 第一非差分码本 Ww中的码字从旋转哈达马矩阵中得到,具体地第一非差分码 本 ww中包括的码字 的结构可以如下:
Wir) =^(R ¾ H»,J(r)
其中, 表示第一非差分码本 Ww中包括的非差分预编码矩阵指示, k = 0, l
Figure imgf000035_0001
表示第一非差分码本 Ww中包括的码字的数目, r表示第 一非差分码本中包括的码字的秩, 4表示与 对应的对角矩阵 的索引, 表 示与 4对应的 n阶对角矩阵, Λ表示与 对应的归一化哈达马矩阵的索引, Η„Λ 表示与 对应的 η阶归一化哈达马矩阵, η的取值与发射天线的数目相等, (RitH„,;t f)表示从旋转哈达马矩阵 RitH„,A中选择 r列构成的矩阵。 对角矩阵 Rit 的结构 4具体可以表示为: Rik二^^^ ^ ^, ^^'2.. "…^ ' '" ) ¾ 为 实数, = -1, 表示相移, = 1, 2,...,;ί, η的取值与发射天线的数目相等。从 旋转哈达马矩阵 R4H„,;t中选择 r列构成矩阵 (R4H„,A f), 具体可以根据实际应 用状况, 釆用相关的方法确定具体选择哪 r列, 例如: 对于非相关信道, 列的 选择应使得从不同的矩阵 R; Η„ ,中分别抽取一个 r列子矩阵构成的集合
{(R4H„,;t }中各个矩阵满足弦距最大(对于非相关信道); 对于相关信道, 列 的选择应使得从不同的矩阵 R; Η„ ,中分别抽取一个 r列子矩阵构成的集合
{(RitH„,;t f)}中各个矩阵满足在天线阵列空间投影的零方向增益最大, 还可以 釆用其他任何可行的方法进行选择, 对此不做具体限定。
或具体地, 第一非差分码本中包括的码字 的结构可以如下:
其中, 表示第一非差分码本 Ww中包括的预编码矩阵指示, = 0,1 — 1
Figure imgf000035_0002
Ww|表示第一非差分码本 Ww中包括的码字的数目, r表示第一非差分码本 Ww中包括的码字的秩, 4表示与 对应的对角矩阵 .的索引, R;表示与 4对 应的 n阶对角矩阵, Λ表示与 对应的归一化哈达马矩阵的索引, Η„Λ表示与 Λ对应的 η阶归一化哈达马矩阵, η的取值与发射天线的数目相等。
(R; H„ , Rf f)表示从旋转哈达马矩阵^ H„ , R 中选择 r列构成的矩阵。
对角矩阵1 ;的结构具体可以表示为:
R. = diag
Figure imgf000036_0001
" }, 为实数, = _1, 表示相移, t = l, 2, ..., n , n的取值与发射天线的数目相等。从旋转哈达马矩阵 R; H„, Rf 中 选择 r列构成矩阵 (R4H„,;tR )W, 具体可以根据实际应用状况, 釆用相关的方 法确定具体选择哪 r列, 例如: 对于非相关信道, 列的选择应使得从不同的矩 阵^ H„ , Rf 中分别抽取一个 r列子矩阵构成的集合 "[(R,. H„ , Rf f )"[中各个矩阵 满足弦距最大; 对于相关信道, 列的选择应使得从不同的矩阵 R; H„, Rf 中分 别抽取一个 r列子矩阵构成的集合 i I( VR; H„ , Rf l I中各个矩阵满足在天线阵列 空间投影的零方向增益最大, 还可以釆用其他任何可行的方法进行选择, 对 此不做具体限定。
502: 根据参考预编码矩阵指示和差分预编码矩阵指示, 利用本地预先存 储的第二非差分码本和第二对角化差分码本, 得到预编码矩阵。
其中, 第二非差分码本和第二对角化差分码本分别与第一非差分码本和 第一对角化差分码本相一致。
进一步地, 对于单用户多输入多输出系统, 根据参考预编码矩阵指示和 差分预编码矩阵指示, 利用本地预先存储的第二非差分码本和第二对角化差 分码本, 得到预编码矩阵, 具体可以包括:
根据参考预编码矩阵指示, 从第二非差分码本中查询得到参考预编码矩 阵指示对应的码字, 并根据差分预编码矩阵指示, 从第二对角化差分码本中 查询得到差分预编码矩阵指示对应的码字;
将差分预编码矩阵指示对应的码字与参考预编码矩阵指示对应的码字进 行相乘, 将乘积结果作为预编码矩阵。
进一步地,对于多用户多输多输出系统,根据参考预编码矩阵指示和差分 预编码矩阵指示, 利用本地预先存储的第二非差分码本和第二对角化差分码 本, 得到预编码矩阵, 具体可以包括:
根据同时配对的各个数据接收端的参考预编码矩阵指示和差分预编码矩 阵指示, 利用迫零波束赋形算法或基于最大信泄噪比准则, 计算得到预编码 矩阵。 本发明实施例所述的获取预编码矩阵的方法, 根据参考预编码矩阵指示 和差分预编码矩阵指示, 利用本地预先存储的第二非差分码本和第二对角化 差分码本, 获取预编码矩阵, 其中第二对角化差分码本中包括的码字为对角 化矩阵, 可以保持非差分码本已具有的元素的幅度特性(如恒模特性, 有限 字符集约束特性)或者便于实现天线间功率分配。 并且, 第二非差分码本可 以使用从旋转哈达马矩阵得到的非差分码本, 其中哈达马矩阵的各列可以分 别与均匀线阵发射天线配置和双极化发射天线配置下强相关和低相关信道的 特征向量匹配, 从而提高量化的准确性, 旋转矩阵为对角化矩阵, 可以保持 在哈达马矩阵量化的基础上对空间进行细致量化; 而且, 哈达马矩阵各个元 素为 +1或者 -1, 可以保持非差分码本的恒模特性; 且哈达马矩阵与其它矩阵 或者矢量的乘法运算可以简化为加法或者减法运算, 从而可以大大降低基于
SINR的 CQI计算、 PMI选择和秩自适应的计算复杂度。 归一化哈达马矩阵的 各个列相互正交, 可以保持非差分码本中码字的酉矩阵特性, 可以保证发射 的各个空间流功率分配相同, 满秩情况下各个发射天线功率分配相同, 可以 保证发射的各个空间流功率分配相同, 满秩情况下各个发射天线功率分配相 同; 另外, 从旋转哈达马矩阵得到非差分码本的方法可以应用于 2、 4、 8、 16、 32、 64等个数的发射天线且得到的非差分码本可以满足恒模特性、 酉矩阵特 性以及降低基于 SINR的 CQI计算、 PMI选择和秩自适应的计算复杂度。 实施例 8
参见图 6, 本发明实施例提供了一种获取预编码矩阵指示的装置, 所述装 置包括:
预编码矩阵指示获取模块 601,用于根据第一非差分码本和第一对角化差 分码本, 获取参考预编码矩阵指示和差分预编码矩阵指示, 其中, 第一对角 化差分码本中包括的码字为对角化矩阵。
其中, 第一非差分码本 ( 可以是本发明实施例设置的一种非差分码本, 第一非差分码本 Ww中的码字从旋转哈达马矩阵中得到,具体地第一非差分码 本 Ww中包括的码字 的结构可以如下:
wir) = (R HJ(r)
其中, 表示第一非差分码本 Ww中包括的非差分预编码矩阵指示, k = 0,l
Figure imgf000037_0001
表示第一非差分码本 Ww中包括的码字的数目, r表示第 一非差分码本中包括的码字的秩, 4表示与 对应的对角矩阵 ¾的索引, Rik表 示与 4对应的 n阶对角矩阵, A表示与 对应的归一化哈达马矩阵的索引, Η„Λ 表示与 对应的 η阶归一化哈达马矩阵, η的取值与发射天线的数目相等, ( R; H„ , )(r)表示从旋转哈达马矩阵 R; Η„ ,中选择 r列构成的矩阵。
对角矩阵 .的结构具体可以表示为:
Figure imgf000038_0001
b,. ,为实数, =- 1, ,表示相移, = 1,2,... ,η , η的取值与发射天线的数目 相等。从旋转哈达马矩阵 RitH„,;t中选择 r列构成矩阵 (RitH„Af),具体可以根据 实际应用状况,釆用相关的方法确定具体选择哪 r列,例如:对于非相关信道, 列的选择应使得从不同的矩阵 R; Η„ ,中分别抽取一个 r列子矩阵构成的集合
{(RitH„,;t }中各个矩阵满足弦距最大(对于非相关信道); 对于相关信道, 列 的选择应使得从不同的矩阵 R; Η„ ,中分别抽取一个 r列子矩阵构成的集合
{(RitH„,;t }中各个矩阵满足在天线阵列空间投影的零方向增益最大, 还可以 釆用其他任何可行的方法进行选择, 对此不做具体限定。 或具体地, 第一非 差分码本中包括的码字 w^)的结构可以如下:
W = (R' "'W
其中, 表示第一非差分码本 Ww中包括的预编码矩阵指示, fc = 0,l — 1
Figure imgf000038_0002
Ww|表示第一非差分码本 Ww中包括的码字的数目, r表示第一非差分码本
Ww中包括的码字的秩, 4表示与 对应的对角矩阵 的索引, ¾表示与 4对 应的 n阶对角矩阵, A表示与 对应的归一化哈达马矩阵的索引, Η„Λ表示与 Α对应的 η阶归一化哈达马矩阵, η的取值与发射天线的数目相等,
(R; H„ , Rf f)表示从旋转哈达马矩阵!^ H„ , R 中选择 r列构成的矩阵。
对角矩阵 .的结构具体可以表示为:
R. = diag [bik lei5'k bi2ei5'kl ...bite5'kl ...bine5'k" }, 为实数, ,表示相移, ί = 1,2, ...,", η的取值与发射天线的数目相等。从旋转哈达马矩阵 R; H„ , R 中选择 r列构成 矩阵 (RitH„AR )W, 具体可以根据实际应用状况, 釆用相关的方法确定具体选 择哪 r列, 例如: 对于非相关信道, 列的选择应使得从不同的矩阵 R;H„,Rf 中 分别抽取一个 r列子矩阵构成的集合 i I( VR; H„ , Rf " [ I中各个矩阵满足弦距最大; 对于相关信道, 列的选择应使得从不同的矩阵 R; H„ , R 中分别抽取一个 r列 子矩阵构成的集合 i I( VR; H„ , Rff} I中各个矩阵满足在天线阵列空间投影的零方 向增益最大, 还可以釆用其他任何可行的方法进行选择, 对此不做具体限定。 进一步地, 预编码矩阵指示获取模块 601具体可以包括:
参考预编码矩阵指示计算单元, 用于根据第一非差分码本, 基于预设的 准则, 计算得到参考预编码矩阵指示。
差分预编码矩阵指示计算单元, 用于在参考预编码矩阵指示计算单元得 到参考预编码矩阵指示后, 根据第一非差分码本、 第一对角化差分码本和参 考预编码矩阵指示, 基于预设的准则, 计算得到差分预编码矩阵指示; 或 预编码矩阵指示获取模块 601具体可以包括:
参考预编码矩阵指示和差分预编码矩阵指示计算单元, 用于根据第一非 差分码本和第一对角化差分码本, 基于预设的准则, 计算得到参考预编码矩 阵指示和差分预编码矩阵指示。
进一步地, 参见图 6, 该装置还包括:
预编码矩阵指示反馈模块 602,用于在预编码矩阵指示获取模块 601得到 参考预编码矩阵指示和差分预编码矩阵指示后, 将参考预编码矩阵指示和差 分预编码矩阵指示反馈给数据发送端, 使数据发送端利用其自身中预先存储 的第二非差分码本和第二对角化差分码本, 根据参考预编码矩阵指示和差分 预编码矩阵指示, 得到预编码矩阵, 其中, 第二非差分码本和第二对角化差 分码本分别与第一非差分码本和第一对角化差分码本相一致。
本发明实施例所述的获取预编码矩阵指示的装置, 根据非差分码本和对 角化差分码本, 获取参考 PMI和差分 PMI, 可以节省反馈开销, 提高反馈精 度, 从而改进了反馈性能; 对角化差分码本中包括的码字为对角化矩阵, 可 以保持非差分码本已具有的元素的幅度特性(如恒模特性, 有限字符集约束 特性)。 并且, 可以使用从旋转哈达马矩阵得到的非差分码本, 其中哈达马矩 阵的各列可以分别与均勾线阵发射天线配置和双极化发射天线配置下强相关 和低相关信道的特征向量匹配, 从而提高量化的准确性, 旋转矩阵为对角化 矩阵, 可以在哈达马矩阵量化的基础上对空间进一步进行细致量化; 而且, 哈达马矩阵各个元素为 +1或者 -1, 可以保持非差分码本的恒模特性; 此外, 哈达马矩阵与其它矩阵或者矢量的乘法运算可以简化为加法或者减法运算, 从而可以大大降低基于 SINR的 CQI计算、 PMI选择和秩自适应的计算复杂 度。 归一化哈达马矩阵的各列相互正交, 可以保持非差分码本中码字的酉矩 阵特性, 可以保证发射的各个空间流功率分配相同, 满秩情况下各个发射天 线功率分配相同, 可以保证发射的各个空间流功率分配相同, 满秩情况下各 个发射天线功率分配相同; 另外, 从旋转哈达马矩阵得到非差分码本的方法 可以应用于 2、 4、 8、 16、 32、 64等个数的发射天线且得到的非差分码本可 以满足恒模特性、 酉矩阵特性以及降低基于 SINR的 CQI计算、 PMI选择和 秩自适应的计算复杂度。 实施例 9
参见图 7,本发明实施例提供了一种获取预编码矩阵的装置,该装置包括: 预编码矩阵指示接收模块 701,用于接收参考预编码矩阵指示和差分预编 码矩阵指示; 其中, 参考预编码矩阵指示和差分预编码矩阵指示由数据接收 端根据第一非差分码本和第一对角化差分码本得到并发送, 第一对角化差分 码本中包括的码字为对角化矩阵。
其中, 第一非差分码本 ^)可以是本发明实施例设置的一种非差分码本, 第一非差分码本 Ww中的码字从旋转哈达马矩阵中得到,具体地第一非差分码 本 Ww中包括的码字 的结构可以如下:
其中, 表示第一非差分码本 Ww中包括的非差分预编码矩阵指示, k = 0, l
Figure imgf000040_0001
表示第一非差分码本 Ww中包括的码字的数目, r表示第 一非差分码本中包括的码字的秩, 4表示与 对应的对角矩阵 ¾的索引, Rit表 示与 4对应的 n阶对角矩阵, A表示与 对应的哈达马矩阵的索引, H„ A表示 与 对应的 n阶哈达马矩阵, n的取值与发射天线的数目相等, (R; H„, )W表 示从旋转哈达马矩阵 Η„ ,.中选择 r列构成的矩阵。
对角矩阵 .的结构具体可以表示为:
Figure imgf000040_0002
为实数, = -1, ,表示相移, = 1,2, ..., /ί, n的取值与发射天线的数目相 等。从旋转哈达马矩阵 RITH„A中选择 r列构成矩阵 (RITH„,;t f),具体可以根据实 际应用状况, 釆用相关的方法确定具体选择哪 r列, 例如: 对于非相关信道, 列的选择应使得从不同的矩阵 R; Η„ ,中分别抽取一个 r列子矩阵构成的集合
{(RITH„,;t }中各个矩阵满足弦距最大(对于非相关信道); 对于相关信道, 列 的选择应使得从不同的矩阵 R; Η„ ,中分别抽取一个 r列子矩阵构成的集合
{(R; H„ , f)}中各个矩阵满足在天线阵列空间投影的零方向增益最大, 还可以 釆用其他任何可行的方法进行选择, 对此不做具体限定。 或具体地, 第一非 差分码本中包括的码字 w^)的结构可以如下:
W = (R' "'W
其中, 表示第一非差分码本 Ww中包括的预编码矩阵指示, fc = 0, l — 1
Figure imgf000041_0001
Ww |表示第一非差分码本 Ww中包括的码字的数目, r表示第一非差分码本
Ww中包括的码字的秩, 4表示与 对应的对角矩阵 R;t的索引, R;t表示与 4对 应的 n阶对角矩阵, A表示与 对应的归一化哈达马矩阵的索引, Η„Λ表示与 Α对应的 η阶归一化哈达马矩阵, η的取值与发射天线的数目相等。
(R; H„ , Rf 表示从旋转哈达马矩阵^ H„ , R 中选择 r列构成的矩阵。
对角矩阵 .的结构具体可以表示为:
Rik = diag
Figure imgf000041_0002
}, \,t为实数, = -1, S t表示相移, t = l, 2, ..., n , n的取值与发射天线的数目相等。从旋转哈达马矩阵 R; H„ , Rf 中选 择 r列构成矩阵 (R4H„;tR ), 具体可以根据实际应用状况, 釆用相关的方法 确定具体选择哪 r列, 例如: 对于非相关信道, 列的选择应使得从不同的矩阵 R; H„ , Rf 中分别抽取一个 r列子矩阵构成的集合 i(R; H„ , Rf "[中各个矩阵满 足弦距最大; 对于相关信道, 列的选择应使得从不同的矩阵 R; H„ , Rf 中分别 抽取一个 r列子矩阵构成的集合 i(R; H„ , Rf }
I V I中各个矩阵满足在天线阵列空 间投影的零方向增益最大, 还可以釆用其他任何可行的方法进行选择, 对此 不做具体限定。
预编码矩阵获取模块 702, 用于在预编码矩阵指示接收模块 701接收到 参考预编码矩阵指示和差分预编码矩阵指示后, 根据参考预编码矩阵指示和 差分预编码矩阵指示, 利用本地预先存储的第二非差分码本和第二对角化差 分码本, 得到预编码矩阵, 其中, 第二非差分码本和第二对角化差分码本分 别与第一非差分码本和第一对角化差分码本相一致。
进一步地, 对于单用户多输入多输出系统, 预编码矩阵获取模块 702具 体可以包括:
码字获取单元, 用于在预编码矩阵指示接收模块 701接收到参考预编码 矩阵指示和差分预编码矩阵指示后, 根据参考预编码矩阵指示, 从第二非差 分码本中查询得到参考预编码矩阵指示对应的码字, 并根据差分预编码矩阵 指示, 从第二对角化差分码本中查询得到差分预编码矩阵指示对应的码字; 第一预编码矩阵获取单元, 用于在码字获取单元得到差分预编码矩阵指 示对应的码字与参考预编码矩阵指示对应的码字后, 将差分预编码矩阵指示 对应的码字与参考预编码矩阵指示对应的码字进行相乘, 将乘积结果作为预 编码矩阵。
进一步地, 对于多用户多输多输出系统, 预编码矩阵获取模块 702具体 包括:
第二预编码矩阵获取单元, 用于在预编码矩阵指示接收模块 701接收到参 考预编码矩阵指示和所述差分预编码矩阵指示后, 根据同时配对的各个数据 接收端的参考预编码矩阵指示和差分预编码矩阵指示, 利用迫零波束赋形算 法或基于最大信泄噪比准则, 计算得到预编码矩阵。
本发明实施例所述的获取预编码矩阵的装置, 根据非差分码本和对角化 差分码本, 获取参考 PMI和差分 PMI, 可以节省反馈开销, 提高反馈精度, 从而进一步改进了反馈性能; 对角化差分码本中包括的码字为对角化矩阵, 可以保持非差分码本已具有的元素的幅度特性(如恒模特性, 有限字符集约 束特性)。 并且, 可以使用从旋转哈达马矩阵得到的非差分码本, 其中哈达马 矩阵的各列可以分别与均匀线阵发射天线配置和双极化发射天线配置下强相 关和低相关信道的特征向量匹配, 从而提高量化的准确性, 旋转矩阵为对角 化矩阵, 可以保持在哈达马矩阵量化的基础上对空间进行细致量化; 而且, 哈达马矩阵各个元素为 +1或者 -1, 可以保持非差分码本的恒模特性; 此外, 哈达马矩阵与其它矩阵或者矢量的乘法运算可以简化为加法或者减法运算, 从而可以大大降低基于 SINR (信干噪比)的 CQI (信道质量指示)计算、 PMI 选择和秩自适应的计算复杂度。 归一化哈达马矩阵的各个列相互正交, 可以 保持非差分码本中码字的酉矩阵特性, 可以保证发射的各个空间流功率分配 相同, 满秩情况下各个发射天线功率分配相同, 可以保证发射的各个空间流 功率分配相同, 满秩情况下各个发射天线功率分配相同; 进一步地, 从旋转 哈达马矩阵得到非差分码本的方法可以应用于 2、 4、 8、 16、 32、 64等个数 的发射天线且得到的非差分码本可以满足恒模特性、 酉矩阵特性以及降低基 于 SINR的 CQI计算、 PMI选择和秩自适应的计算复杂度。 实施例 10
参见图 8, 本发明实施例提供了一种获取预编码矩阵指示的方法, 包括: 801: 根据第一非差分码本, 基于预设的准则, 计算得到非差分预编码 矩阵指示。
其中, 第一非差分码本 ^)可以是本发明实施例设置的一种非差分码本, 第一非差分码本 ww中的码字从旋转哈达马矩阵中得到,具体地第一非差分码 本 ww中包括的码字 的结构可以如下:
其中, 表示第一非差分码本 ww中包括的非差分预编码矩阵指示, k = 0,l
Figure imgf000043_0001
表示第一非差分码本 Ww中包括的码字的数目, r表示第 一非差分码本中包括的码字的秩, 4表示与 对应的对角矩阵 R;t的索引, Rit表 示与 4对应的 n阶对角矩阵, A表示与 对应的归一化哈达马矩阵的索引, Η„ Λ 表示与 对应的 η阶归一化哈达马矩阵, η的取值与发射天线的数目相等, ( R; H„ , )(r)表示从旋转哈达马矩阵 R; Η„ ,中选择 r列构成的矩阵。
对角矩阵 .的结构具体可以表示为:
R. = diag{bik S'" ,bi 2eiS'k'2 ,...,bi teiS'k ,...,bi neiS'k'n),
bi f为实数, j2=- ,表示相移, = 1,2,...,;ί, n的取值与发射天线的数目 相等。从旋转哈达马矩阵 R4H„;t中选择 r列构成矩阵 (R4H„;tf),具体可以根据 实际应用状况,釆用相关的方法确定具体选择哪 r列,例如:对于非相关信道, 列的选择应使得从不同的矩阵 R; Η„ ,中分别抽取一个 r列子矩阵构成的集合
I中各个矩阵满足弦距最大; 对于相关信道, 列的选择应使得从不
Figure imgf000043_0002
R;H„ ,中分别抽取一个 r列子矩阵构成的集合 ί(Κ; Η„ , "
IV [ I中各个矩 阵满足在天线阵列空间投影的零方向增益最大, 还可以釆用其他任何可行的 方法进行选择, 对此不做具体限定。
或具体地, 第一非差分码本中包括的码字 的结构可以如下:
W = (R' "'W
其中, 表示第一非差分码本 ww中包括的预编码矩阵指示,
t = 0,l
Figure imgf000043_0003
表示第一非差分码本 Ww中包括的码字的数目, r表示第 一非差分码本 Ww中包括的码字的秩, 4表示与 对应的对角矩阵 的索引, Rit表示与 ik对应的 n阶对角矩阵, A表示与 对应的归一化哈达马矩阵的索引, Η„λ表示与 对应的 η阶归一化哈达马矩阵, η的取值与发射天线的数目相等,
Figure imgf000043_0004
表示从旋转哈达马矩阵^ H„ , R 中选择 r列构成的矩阵。
对角矩阵 .的结构具体可以表示为:
R. =^^{¾ 1,^ '2,.."^ '',"" ¾'"}, 为实数, f=-l 表示相移, t = l,2,...,n, n的取值与发射天线的数目相等。从旋转哈达马矩阵 R;H„ , Rf 中选 择 r列构成矩阵 (R H^R【)W, 具体可以根据实际应用状况, 釆用相关的方法 确定具体选择哪 r列, 例如: 对于非相关信道, 列的选择应使得从不同的矩阵 R; H„ , Rf 中分别抽取一个 r列子矩阵构成的集合 "f(R; H„ , Rf f)" [中各个矩阵满 足弦距最大; 对于相关信道, 列的选择应使得从不同的矩阵 R; H„, Rf 中分别 抽取一个 r列子矩阵构成的集合 i I( VR; H„ , Rf ι " [ I中各个矩阵满足在天线阵列空 间投影的零方向增益最大, 还可以釆用其他任何可行的方法进行选择, 对此 不做具体限定。
可选的, 所述第一非差分码本中包括的码字 的结构还可以如下:
) = R B( i
其中, 表示非差分码本 ww中包括的预编码矩阵指示, yt = o,i...|ww|- 1,
Ww|表示非差分码本 Ww中包括的码字的数目, r表示非差分码本 Ww中包括 的码字的秩, 4表示与 对应的对角矩阵 Rit的索引, R4表示与 4对应的 n阶 对角矩阵, B 表示与索引 对应的一个秩为 r的 n发射天线非差分码本中的 一个码字, 与 对应。 进一步地,参见图 8,获取第一预编码矩阵指示之后,该方法还可以包括: 802: 将非差分预编码矩阵指示反馈给数据发送端。
803: 数据发送端接收非差分预编码矩阵指示, 并利用其自身中预先存储 的第二非差分码本, 根据非差分预编码矩阵指示, 得到预编码矩阵。
其中, 第二非差分码本与第一非差分码本相一致。
本发明实施例所述的获取预编码矩阵指示的方法, 使用从旋转哈达马矩 配置和双极化发射天线配置下强相关和低相关信道的特征向量匹配, 从而提 高量化的准确性, 旋转矩阵为对角化矩阵, 可以保持在哈达马矩阵量化的基 础上对空间进行细致量化; 而且, 哈达马矩阵各个元素为 +1 或者 -1, 可以保 持非差分码本的恒模特性; 此外, 哈达马矩阵与其它矩阵或者矢量的乘法运 算可以简化为加法或者减法运算, 从而可以大大降低基于 SINR的 CQI计算、 PMI选择和秩自适应的计算复杂度。 归一化哈达马矩阵的各个列相互正交, 可以保持非差分码本中码字的酉矩阵特性, 可以保证发射的各个空间流功率 分配相同, 满秩情况下各个发射天线功率分配相同, 可以保证发射的各个空 间流功率分配相同, 满秩情况下各个发射天线功率分配相同; 进一步地, 从 旋转哈达马矩阵得到非差分码本的方法可以应用于 2、 4、 8、 16、 32、 64等 个数的发射天线且得到的非差分码本可以满足恒模特性、 酉矩阵特性以及降 低基于 SINR的 CQI计算、 PMI选择和秩自适应的计算复杂度。 实施例 11
参见图 9, 本发明实施例提供了一种获取预编码矩阵指示的装置, 所述装 置包括:
非差分预编码矩阵指示获取模块 901, 用于根据第一非差分码本,基于预 设的准则, 计算得到非差分预编码矩阵指示, 其中, 第一非差分码本中包括 的码字从旋转哈达马矩阵得到。
其中, 预设的准则与实施例 2中预设的准侧类似, 此处不再赘述。 第一 非差分码本 Ww可以是本发明实施例设置的一种非差分码本,第一非差分码本 w(r)中的码字从旋转哈达马矩阵中得到, 具体地第一非差分码本 ww中包括的 码字 的结构可以如下:
Figure imgf000045_0001
其中, 表示第一非差分码本 ww中包括的非差分预编码矩阵指示, k = 0,l
Figure imgf000045_0002
表示第一非差分码本 Ww中包括的码字的数目, r表示第 一非差分码本中包括的码字的秩, 4表示与 对应的对角矩阵 ¾的索引, Rit表 示与 4对应的 n阶对角矩阵, A表示与 对应的归一化哈达马矩阵的索引, Η„Λ 表示与 对应的 η阶归一化哈达马矩阵, η的取值与发射天线的数目相等, ( R; H„ , )(r)表示从旋转哈达马矩阵 R; Η„ ,中选择 r列构成的矩阵。
对角矩阵 Rit的结构具体可以表示为: R = ^g{¾y¾\^%...,¾y¾,,...,¾ ¾ , 为实数, = -1, ,表示相移, = 1,2,...,«, n的取值与发射天线的数目 相等。从旋转哈达马矩阵 RitH„A中选择 r列构成矩阵 (RitH„;t f),具体可以根据 实际应用状况,釆用相关的方法确定具体选择哪 r列,例如:对于非相关信道, 列的选择应使得从不同的矩阵 R; Η„ ,中分别抽取一个 r列子矩阵构成的集合
{(RitH„,;t }中各个矩阵满足弦距最大; 对于相关信道, 列的选择应使得从不 同的矩阵 Ι ; Η„,中分别抽取一个 r列子矩阵构成的集合 ί(Κ; Η„, " [中各个矩 阵满足在天线阵列空间投影的零方向增益最大, 还可以釆用其他任何可行的 方法进行选择, 对此不做具体限定。
或具体地, 第一非差分码本中包括的码字 的结构可以如下: W,(r) = ^iR H R 其中, 表示第一非差分码本 Ww中包括的预编码矩阵指示, fc = 0, l 1
Figure imgf000046_0001
Ww |表示第一非差分码本 Ww中包括的码字的数目, r表示第一非差分码本
Ww中包括的码字的秩, 4表示与 对应的对角矩阵 ¾的索引, R4表示与 4对 应的 n阶对角矩阵, 表示与 对应的哈达马矩阵的索引, Η„Λ表示与 对应 的 η阶哈达马矩阵, η的取值与发射天线的数目相等, (R; H„ , Rf f
V )表示从旋 转哈达马矩阵 R,. H„ ,. R 中选择 r列构成的矩阵。
对角矩阵 .的结构具体可以表示为:
R. 1
Figure imgf000046_0002
, ,表示相移, t = l, 2, ..., n , n的取值与发射天线的数目相等。从旋转哈达马矩阵 R; H„ , Rf 中选 择 r列构成矩阵 (R4H„;tR ), 具体可以根据实际应用状况, 釆用相关的方法 确定具体选择哪 r列, 例如: 对于非相关信道, 列的选择应使得从不同的矩阵 R; H„ , Rf 中分别抽取一个 r列子矩阵构成的集合 i(R; H„ , Rf "[中各个矩阵满 足弦距最大; 对于相关信道, 列的选择应使得从不同的矩阵 R; H„ , Rf 中分别 抽取一个 r列子矩阵构成的集合 i(R; H„ , Rf }
I V I中各个矩阵满足在天线阵列空 间投影的零方向增益最大, 还可以釆用其他任何可行的方法进行选择, 对此 不做具体限定。
进一步地, 参见图 9, 该装置还可以包括:
非差分预编码矩阵指示反馈模块 902,用于在非差分预编码矩阵指示获取 模块 901 得到非差分预编码矩阵指示后, 将非差分预编码矩阵指示反馈给数 据发送端, 使数据发送端利用其自身中预先存储的第二非差分码本, 根据非 差分预编码矩阵指示,得到预编码矩阵, 其中, 第二非差分码本与第一非差分 码本相一致。
本发明实施例所述的获取预编码矩阵指示的装置, 使用从旋转哈达马矩 配置和双极化发射天线配置下强相关和低相关信道的特征向量匹配, 从而提 高量化的准确性, 旋转矩阵为对角化矩阵, 可以保持在哈达马矩阵量化的基 础上对空间进行细致量化; 而且, 哈达马矩阵各个元素为 +1 或者 -1, 可以保 持非差分码本的恒模特性;, 此外, 哈达马矩阵与其它矩阵或者矢量的乘法运 算可以简化为加法或者减法运算, 从而可以大大降低基于 SINR的 CQI计算、 PMI选择和秩自适应的计算复杂度。 归一化哈达马矩阵的各个列相互正交, 可以保持非差分码本中码字的酉矩阵特性, 可以保证发射的各个空间流功率 分配相同, 满秩情况下各个发射天线功率分配相同, 可以保证发射的各个空 间流功率分配相同, 满秩情况下各个发射天线功率分配相同; 进一步地, 从 旋转哈达马矩阵得到非差分码本的方法可以应用于 2、 4、 8、 16、 32、 64等 个数的发射天线且得到的非差分码本可以满足恒模特性、 酉矩阵特性以及降 低基于 SINR的 CQI计算、 PMI选择和秩自适应的计算复杂度。 实施例 12
参见图 10, 本发明实施例提供了一种获取预编码矩阵的方法, 该方法包 括:
1001 : 接收非差分预编码矩阵指示; 其中, 非差分预编码矩阵指示由数 据接收端根据第一非差分码本得到并发送; 第一非差分码本中包括的码字从 旋转哈达马矩阵得到。
其中, 第一非差分码本 ( 可以是本发明实施例设置的一种非差分码本, 第一非差分码本 ww中的码字从旋转哈达马矩阵中得到,具体地第一非差分码 本 ww中包括的码字 的结构可以如下:
wir) = (R,W
其中, 表示第一非差分码本 ww中包括的非差分预编码矩阵指示, k = 0, l
Figure imgf000047_0001
表示第一非差分码本 Ww中包括的码字的数目, r表示第 一非差分码本中包括的码字的秩, 4表示与 对应的对角矩阵 ¾的索引, 表 示与 4对应的 n阶对角矩阵, A表示与 对应的归一化哈达马矩阵的索引, Η„Λ 表示与 对应的 η阶归一化哈达马矩阵, η的取值与发射天线的数目相等, ( R; H„ , )(r)表示从旋转哈达马矩阵 R; Η„ ,中选择 r列构成的矩阵。
对角矩阵 的结构具体可以表示为: R ^ { .', 2,..., ¾.',..., ¾."}, 为实数, = - 1, 表示相移, = 1,2_ , η , η的取值与发射天线的数目相 等。从旋转哈达马矩阵 RitH„A中选择 r列构成矩阵 (RitH„A f),具体可以根据实 际应用状况, 釆用相关的方法确定具体选择哪 r列, 例如: 对于非相关信道, 列的选择应使得从不同的矩阵 R; Η„ ,中分别抽取一个 r列子矩阵构成的集合 {(RitH„;t )Wj中各个矩阵满足弦距最大; 对于相关信道, 列的选择应使得从不 同的矩阵 R; H„,中分别抽取一个 r列子矩阵构成的集合 ί(Κ; Η„, " [中各个矩 阵满足在天线阵列空间投影的零方向增益最大, 还可以釆用其他任何可行的 方法进行选择, 对此不做具体限定。
或具体地, 第一非差分码本 构可以如下:
Figure imgf000048_0001
其中, 表示第一非差分码本 ww中包括的预编码矩阵指示,
k = 0, l
Figure imgf000048_0002
表示第一非差分码本 Ww中包括的码字的数目, r表示第 一非差分码本 Ww中包括的码字的秩, 4表示与 对应的对角矩阵 的索引, Rit表示与 ik对应的 n阶对角矩阵, A表示与 对应的归一化哈达马矩阵的索引, Η„λ表示与 Λ对应的 η阶归一化哈达马矩阵, η的取值与发射天线的数目相等, (R; H„ , Rf f)表示从旋转哈达马矩阵^ H„ , R 中选择 r列构成的矩阵。
对角矩阵 .的结构具体可以表示为:
R. =^^{¾ 1, , ¾'2,..., ¾',..., ¾'"}, 为实数, f = -l , 表示相移, t = l, 2 n的取值与发射天线的数目相等。从旋转哈达马矩阵 R; H„, Rf 中选 择 r列构成矩阵 (RitH„;tR )W, 具体可以根据实际应用状况, 釆用相关的方法 确定具体选择哪 r列, 例如: 对于非相关信道, 列的选择应使得从不同的矩阵 R; H„ , Rf 中分别抽取一个 r列子矩阵构成的集合 "f(R; H„ , Rf f)" [中各个矩阵满 足弦距最大; 对于相关信道, 列的选择应使得从不同的矩阵 R; H„, Rf 中分别 抽取一个 r列子矩阵构成的集合 i(R; H„ , Rf "
I V I [ I中各个矩阵满足在天线阵列空 间投影的零方向增益最大, 还可以釆用其他任何可行的方法进行选择, 对此 不做具体限定。
1002: 根据非差分预编码矩阵指示, 利用本地预先存储的第二非差分码 本, 得到预编码矩阵, 其中, 第二非差分码本与第一非差分码本相一致。
进一步地, 对于单用户多输入多输出系统, 根据非差分预编码矩阵指示, 利用本地预先存储的第二非差分码本, 得到预编码矩阵具体包括: 根据非差 分预编码矩阵指示, 从第二非差分码本中查询得到该非差分预编码矩阵指示 对应的码字; 将该非差分预编码矩阵指示对应的码字作为预编码矩阵。
进一步地, 对于多用户多输多输出系统, 根据非差分预编码矩阵指示, 利 用本地预先存储的第二非差分码本, 得到预编码矩阵具体包括: 根据同时配 对的各个数据接收端的非差分预编码矩阵指示, 利用迫零波束赋形算法或基 于最大信泄噪比准则, 计算得到预编码矩阵。 本发明实施例所述的获取预编码矩阵的方法, 根据非差分预编码矩阵指 示获取预编码矩阵, 非差分码本从旋转哈达马矩阵得到, 其中哈达马矩阵的 各列可以分别与均勾线阵发射天线配置和双极化发射天线配置下强相关和低 相关信道的特征向量匹配, 从而提高量化的准确性, 旋转矩阵为对角化矩阵, 可以保持在哈达马矩阵量化的基础上对空间进行细致量化; 而且, 哈达马矩 阵的各个元素为 +1或者 -1, 可以保持非差分码本的恒模特性; 此外, 哈达马 矩阵与其它矩阵或者矢量的乘法运算可以简化为加法或者减法运算, 从而可 以大大降低基于 SINR的 CQI计算、 PMI选择和秩自适应的计算复杂度。 归 一化哈达马矩阵的各个列相互正交, 可以保持非差分码本中码字的酉矩阵特 性, 可以保证发射的各个空间流功率分配相同, 满秩情况下各个发射天线功 率分配相同; 另外, 从旋转哈达马矩阵得到非差分码本的方法可以应用于 2、 4、 8、 16、 32、 64等个数的发射天线且得到的非差分码本可以满足恒模特性、 酉矩阵特性以及降低基于 SINR的 CQI计算、 PMI选择和秩自适应的计算复 杂度。 实施例 13
参见图 11, 本发明实施例提供了一种获取预编码矩阵的装置, 该装置包 括:
非差分预编码矩阵指示接收模块 1101,用于接收非差分预编码矩阵指示; 其中, 非差分预编码矩阵指示由数据接收端根据第一非差分码本得到并发送, 第一非差分码本中包括的码字从旋转哈达马矩阵得到。
其中, 第一非差分码本 ( 可以是本发明实施例设置的一种非差分码本, 第一非差分码本 Ww中的码字从旋转哈达马矩阵中得到,具体地第一非差分码 本 Ww中包括的码字 的结构可以如下:
wir) = (R,W
其中, 表示第一非差分码本 Ww中包括的非差分预编码矩阵指示, k = 0, l
Figure imgf000049_0001
表示第一非差分码本 Ww中包括的码字的数目, r表示第 一非差分码本中包括的码字的秩, 4表示与 对应的对角矩阵 R;t的索引, Rit表 示与 4对应的 n阶对角矩阵, 表示与 对应的哈达马矩阵的索引, H„ A表示 与 对应的 n阶哈达马矩阵, n的取值与发射天线的数目相等, (R; H„ , )W表 示从旋转哈达马矩阵 R; Η„ ,.中选择 r列构成的矩阵。 对角矩阵 R4的结构具体可以表示为:
b; i为实数, =-1, ,表示相移, = 1,2,...,/ί, n的取值与发射天线的数目相 等。从旋转哈达马矩阵 R4H„A中选择 r列构成矩阵 (R4H„;tf),具体可以根据实 际应用状况, 釆用相关的方法确定具体选择哪 r列, 例如: 对于非相关信道, 列的选择应使得从不同的矩阵 R; Η„ ,中分别抽取一个 r列子矩阵构成的集合 {(RitH„;t )}中各个矩阵满足弦距最大; 对于相关信道, 列的选择应使得从不 同的矩阵 R;H„,中分别抽取一个 r列子矩阵构成的集合 ί(Κ;Η„, " [中各个矩 阵满足在天线阵列空间投影的零方向增益最大, 还可以釆用其他任何可行的 方法进行选择, 对此不做具体限定。
或具体地, 第一非差分 的结构可以如下:
Figure imgf000050_0001
其中, 表示第一非差分码本 Ww中包括的预编码矩阵指示, fc = 0,l — 1
Figure imgf000050_0002
Ww|表示第一非差分码本 Ww中包括的码字的数目, r表示第一非差分码本
Ww中包括的码字的秩, 4表示与 对应的对角矩阵 ¾的索引, Rit表示与 4对 应的 n阶对角矩阵, 表示与 对应的哈达马矩阵的索引, Η„Λ表示与 对应 的 η阶哈达马矩阵, η的取值与发射天线的数目相等, (R; H„ , Rf f )表示从旋 转哈达马矩阵 R,. H„ ,. R 中选择 r列构成的矩阵。
对角矩阵 .的结构具体可以表示为:
R. = · { ¾1Α·,2^¾'2,···Α· '',···Α·,„Ζ }, t为实数, f =_l, 表示相移, t = l,2,...,n, n的取值与发射天线的数目相等。从旋转哈达马矩阵 R;H„,Rf 中选 择 r列构成矩阵 (R4H„;tR ), 具体可以根据实际应用状况, 釆用相关的方法 确定具体选择哪 r列, 例如: 对于非相关信道, 列的选择应使得从不同的矩阵 R; H„ , Rf 中分别抽取一个 r列子矩阵构成的集合 H„ , Rf " [中各个矩阵满 足弦距最大; 对于相关信道, 列的选择应使得从不同的矩阵 R;H„,Rf 中分别 抽取一个 r列子矩阵构成的集合 i I( VR; H„ , Rf } I中各个矩阵满足在天线阵列空 间投影的零方向增益最大, 还可以釆用其他任何可行的方法进行选择, 对此 不做具体限定。
非差分预编码矩阵处理模块 1102, 用于在非差分预编码矩阵指示接收模块 1101接收到非差分预编码矩阵指示后, 根据非差分预编码矩阵指示, 利用本 地预先存储的第二非差分码本, 计算得到预编码矩阵, 其中, 第二非差分码 本与第一非差分码本相一致。 进一步地, 对于单用户多输入多输出系统, 非差分预编码矩阵获取模块
1102, 具体用于在非差分预编码矩阵指示接收模块 1101接收到非差分预编码 矩阵指示后, 根据非差分预编码矩阵指示, 从第二非差分码本中查询得到该 非差分预编码矩阵指示对应的码字, 将该非差分预编码矩阵指示对应的码字 作为预编码矩阵。
进一步地, 对于多用户多输多输出系统, 预编码矩阵获取模块 1102, 具 体用于在非差分预编码矩阵指示接收模块 1101接收到非差分预编码矩阵指示 后, 根据同时配对的各个数据接收端的非差分预编码矩阵指示, 利用迫零波 束赋形算法或基于最大信泄噪比准则, 计算得到预编码矩阵。
本发明实施例所述的获取预编码矩阵的装置, 根据非差分预编码矩阵指 示获取预编码矩阵, 非差分码本从旋转哈达马矩阵得到, 其中哈达马矩阵的 各列可以分别与均勾线阵发射天线配置和双极化发射天线配置下强相关和低 相关信道的特征向量匹配, 从而提高量化的准确性, 旋转矩阵为对角化矩阵, 可以保持在哈达马矩阵量化的基础上对空间进行细致量化; 而且, 哈达马矩 阵的各个元素为 +1或者 -1, 可以保持非差分码本的恒模特性; 此外, 哈达马 矩阵与其它矩阵或者矢量的乘法运算可以简化为加法或者减法运算, 从而可 以大大降低基于 SINR的 CQI计算、 PMI选择和秩自适应的计算复杂度。 归 一化哈达马矩阵的各个列相互正交, 可以保持非差分码本中码字的酉矩阵特 性, 可以保证发射的各个空间流功率分配相同, 满秩情况下各个发射天线功 率分配相同; 另外, 从旋转哈达马矩阵得到非差分码本的方法可以应用于 2、 4、 8、 16、 32、 64等个数的发射天线且得到的非差分码本可以满足恒模特性、 酉矩阵特性以及降低基于 SINR的 CQI计算、 PMI选择和秩自适应的计算复 杂度。 以上实施例提供的技术方案中的全部或部分内容可以通过软件编程实 现, 其软件程序存储在可读取的存储介质中, 存储介质例如: 计算机中的硬 盘、 光盘或软盘。 以上所述仅为本发明的较佳实施例, 并不用以限制本发明, 凡在本发明 的精神和原则之内, 所作的任何修改、 等同替换、 改进等, 均应包含在本发 明的保护范围之内。

Claims

权利 要求 书
1、 一种获取预编码矩阵指示的方法, 其特征在于, 所述方法包括: 根据第一非差分码本和第一对角化差分码本, 获取参考预编码矩阵指示和 差分预编码矩阵指示, 其中, 所述第一对角化差分码本中包括的码字为对角化 矩阵。
2、 根据权利要求 1所述的获取预编码矩阵指示的方法, 其特征在于, 所述 根据第一非差分码本和第一对角化差分码本, 获取参考预编码矩阵指示和差分 预编码矩阵指示, 具体包括:
根据所述第一非差分码本, 基于预设的准则, 计算得到所述参考预编码矩 阵指示;
根据所述第一非差分码本、 所述第一对角化差分码本和所述参考预编码矩 阵指示, 基于所述预设的准则, 计算得到所述差分预编码矩阵指示; 或
所述根据第一非差分码本和第一对角化差分码本, 获取参考预编码矩阵指 示和差分预编码矩阵指示, 具体包括:
根据所述第一非差分码本和所述第一对角化差分码本, 基于预设的准则, 计算得到所述参考预编码矩阵指示和差分预编码矩阵指示。
3、 根据权利要求 1所述的获取预编码矩阵指示的方法, 其特征在于, 所述 第一对角化差分码本中包 Cfc的结构如下:
Figure imgf000052_0001
其中, 表示所述第一对角化差分码本 D 中包括的差分预编码矩阵指示, yt = 0, l...| | -l , |Z)|表示所述第一对角化差分码本 D中包括的码字的数目, = - 1, N表示发射天线的数目, N为正整数, 表示相移, i = l, 2』。
4、 根据权利要求 1所述的获取预编码矩阵指示的方法, 其特征在于, 所述 第一对角化差分码本
Figure imgf000052_0002
其中, 表示所述第一对角化差分码本 D 中包括的差分预编码矩阵指示, yt = 0, l...| | -l , |Z)|表示所述第一对角化差分码本 D中包括的码字的数目, = - 1, N表示发射天线的数目, N为正整数, *(m0)表示相移, i = \,l』, m0表示相邻 发射天线间的相移差。
5、 根据权利要求 1所述的获取预编码矩阵指示的方法, 其特征在于, 当包 含 N个发射天线的双极化发射天线阵的前 N/2个发射天线为一组同极化发射天 线, 后 N/2个发射天线为另外一组同极化发射天线时, 所述第一对角化差分码 本中包括的码字 Cfc的结构如下:
Ck =diag{Sm,e^Sm}
其中, sm =diag、ej ,ej ,...,ej ,...,e j( 'N' , 表示所述第一对角化差分码本 D 中包括的差分预编码矩阵指示, fc = 0,l...|Z)|- 1, |D|表示所述第一对角化差分码本 D中包括的码字的数目, f=-l , N表示发射天线的数目, N为偶数,
示相移, m和 "为自然数, = l,2..., N/2。
6、 根据权利要求 1所述的获取预编码矩阵指示的方法, 其特征在于, 所述 第一对角化差分码本中包括的码字 Cfc的结构如下:
C¾ = diag、ak leAi ,ak 2eA'2 ,...,ak ieA'' ,...,ak Ne≠kN
其中, 表示所述第一对角化差分码本 D中的差分预编码矩阵指示, k =
Figure imgf000053_0001
1, Ν表示发射天线的数目, Ν为正整数, i为实数, 表示相移, i = l,2』。
7、 根据权利要求 1所述的获取预编码矩阵指示的方法, 其特征在于, 所述 第一非差分码本中包括的码字 的结构如下:
wir)= (R,W
其中, 表示所述第一非差分码本 Ww中包括的非差分预编码矩阵指示, k = 0,l
Figure imgf000053_0002
表示所述第一非差分码本 Ww中包括的码字的数目, r表示 所述第一非差分码本 Ww中包括的码字的秩, 4表示与 对应的对角矩阵 的索 引, Rit表示与 4对应的 n阶对角矩阵, A表示与 对应的归一化哈达马矩阵的索 引, Η„ ,表示与 对应的 n阶归一化哈达马矩阵, n的取值与发射天线的数目相 等, (R; H„ , f)表示从旋转哈达马矩阵 . Η„ ,中选择 r列构成的矩阵; 或
所述第一非差分码本中包括的码字 w 的结构如下: W( ) = ^iR H R 其中, 表示所述第一非差分码本 Ww中包括的预编码矩阵指示, fc = 0,l — 1
Figure imgf000054_0001
Ww|表示所述第一非差分码本 Ww中包括的码字的数目, r表示所述第一非差分 码本 Ww中包括的码字的秩, 4表示与 对应的对角矩阵 R;t的索引, Rit表示与 4对 应的 n阶对角矩阵, A表示与 对应的归一化哈达马矩阵的索引, H„A表示与 对应的 n阶归一化哈达马矩阵, n的取值与发射天线的数目相等,(R; H„, Rf 表 示从旋转哈达马矩阵 R. H . RH中选择 r列构成的矩阵; 或
所述第一非差分码本中包括的码字 w^)的结构可以如下:
) = R B(i
其中, 表示非差分码本 ww中包括的预编码矩阵指示, yt = o,i...|ww| -i, |ww|表 示非差分码本 Ww中包括的码字的数目, r表示非差分码本 Ww中包括的码字的 秩, 4表示与 对应的对角矩阵 的索引, R4表示与 4对应的 n阶对角矩阵, 表示与索引 对应的一个秩为 r的 n发射天线非差分码本中的一个码字, 与 对应。
8、 根据权利要求 1-7中任意一项权利要求所述的获取预编码矩阵指示的方 法, 其特征在于, 所述获取参考预编码矩阵指示和差分预编码矩阵指示之后, 还包括:
将所述参考预编码矩阵指示和所述差分预编码矩阵指示反馈给数据发送 端, 使所述数据发送端利用其自身中预先存储的第二非差分码本和第二对角化 差分码本, 根据所述参考预编码矩阵指示和所述差分预编码矩阵指示, 得到预 编码矩阵, 其中, 所述第二非差分码本和所述第二对角化差分码本分别与所述 第一非差分码本和所述第一对角化差分码本相一致。
9、 根据权利要求 8所述的获取预编码矩阵指示的方法, 其特征在于, 所述 第一非差分码本为第一基码本, 所述第一对角化差分码本为第一变换码本; 所述第二非差分码本为第二基码本, 所述第二对角化差分码本为第二变换 码本。
10、 根据权利要求 1-7 中任一项权利要求所述的获取预编码矩阵指示的方 法, 其特征在于, 所述第一非差分码本为第一基码本, 所述第一对角化差分码 本为第一变换码本。
11、根据权利要求 1-7中任意一项权利要求所述的获取预编码矩阵指示的方 法, 其特征在于, 所述参考预编码矩阵指示和差分预编码矩阵指示用于指示预 编码矩阵 , 其中, 所述预编码矩阵 = (^ W«, 所述 ( ^为所述差分预编码矩阵 指示为 k时所对应的所述第一对角化差分码本中的码字, 所述 W«为所述参考预 编码矩阵指示为 n时所对应所述第一非差分码本中的码字。
12、 一种获取预编码矩阵的方法, 其特征在于, 所述方法包括:
接收参考预编码矩阵指示和差分预编码矩阵指示; 其中, 所述参考预编码 矩阵指示和所述差分预编码矩阵指示由数据接收端根据第一非差分码本和第一 对角化差分码本得到并发送; 所述第一对角化差分码本中包括的码字为对角化 矩阵;
根据所述参考预编码矩阵指示和所述差分预编码矩阵指示, 利用本地预先 存储的第二非差分码本和第二对角化差分码本, 得到预编码矩阵, 其中, 所述 第二非差分码本和所述第二对角化差分码本分别与所述第一非差分码本和所述 第一对角化差分码本相一致。
13、 根据权利要求 12所述的获取预编码矩阵的方法, 其特征在于, 对于单 用户多输入多输出系统, 所述根据所述参考预编码矩阵指示和所述差分预编码 矩阵指示, 利用本地预先存储的第二非差分码本和第二对角化差分码本, 得到 预编码矩阵, 具体包括:
根据所述参考预编码矩阵指示, 从所述第二非差分码本中查询得到所述参 考预编码矩阵指示对应的码字, 并根据所述差分预编码矩阵指示, 从所述第二 对角化差分码本中查询得到所述差分预编码矩阵指示对应的码字;
将所述差分预编码矩阵指示对应的码字与所述参考预编码矩阵指示对应的 码字进行相乘, 将乘积结果作为预编码矩阵。
14、 根据权利要求 11所述的获取预编码矩阵的方法, 其特征在于, 对于多 用户多输多输出系统, 所述根据所述参考预编码矩阵指示和所述差分预编码矩 阵指示, 利用本地预先存储的第二非差分码本和第二对角化差分码本, 得到预 编码矩阵, 具体包括:
根据同时配对的各个数据接收端的参考预编码矩阵指示和差分预编码矩阵 指示, 利用迫零波束赋形算法或基于最大信泄噪比准则, 计算得到预编码矩阵。
15、 根据权利要求 12所述的获取预编码矩阵的方法, 其特征在于, 所述参 考预编码矩阵指示和差分预编码矩阵指示用于指示预编码矩阵 , 其中, 所述 预编码矩阵 = (^ W«, 所述 ^为所述差分预编码矩阵指示为 k时所对应的所述 第一对角化差分码本中的码字, 所述 W "为所述参考预编码矩阵指示为 n时所对 应所述第一非差分码本中的码字。
16、 一种获取预编码矩阵指示的装置, 其特征在于, 所述装置包括: 预编码矩阵指示获取模块, 用于根据第一非差分码本和第一对角化差分码 本, 获取参考预编码矩阵指示和差分预编码矩阵指示, 其中, 所述第一对角化 差分码本中包括的码字为对角化矩阵。
17、 根据权利要求 16所述的获取预编码矩阵指示的装置, 其特征在于, 所 述预编码矩阵指示获取模块具体包括:
参考预编码矩阵指示计算单元, 用于根据所述第一非差分码本, 基于预设 的准则, 计算得到所述参考预编码矩阵指示;
差分预编码矩阵指示计算单元, 用于在所述参考预编码矩阵指示计算单元 得到参考预编码矩阵指示后, 根据所述第一非差分码本、 所述第一对角化差分 码本和所述参考预编码矩阵指示, 基于所述预设的准则, 计算得到所述差分预 编码矩阵指示; 或
所述预编码矩阵指示获取模块具体包括:
参考预编码矩阵指示和差分预编码矩阵指示计算单元, 用于根据所述第一 非差分码本和所述第一对角化差分码本, 基于预设的准则, 计算得到所述参考 预编码矩阵指示和差分预编码矩阵指示。
18、 根据权利要求 16-17中任意一项权利要求所述的获取预编码矩阵指示的 装置, 其特征在于, 所述装置还包括:
预编码矩阵指示反馈模块, 用于在所述预编码矩阵指示获取模块得到参考 预编码矩阵指示和差分预编码矩阵指示后, 将所述参考预编码矩阵指示和所述 差分预编码矩阵指示反馈给数据发送端, 使所述数据发送端利用其自身中预先 存储的第二非差分码本和第二对角化差分码本, 根据所述参考预编码矩阵指示 和所述差分预编码矩阵指示, 得到预编码矩阵, 其中, 所述第二非差分码本和 所述第二对角化差分码本分别与所述第一非差分码本和所述第一对角化差分码 本相一致。
19、 一种获取预编码矩阵的装置, 其特征在于, 所述装置包括:
预编码矩阵指示接收模块, 用于接收参考预编码矩阵指示和差分预编码矩 阵指示; 其中, 所述参考预编码矩阵指示和所述差分预编码矩阵指示由数据接 收端根据第一非差分码本和第一对角化差分码本得到并发送; 所述第一对角化 差分码本中包括的码字为对角化矩阵;
预编码矩阵获取模块, 用于在所述预编码矩阵指示接收模块接收到所述参 考预编码矩阵指示和所述差分预编码矩阵指示后, 根据所述参考预编码矩阵指 示和所述差分预编码矩阵指示, 利用本地预先存储的第二非差分码本和第二对 角化差分码本, 得到预编码矩阵, 其中, 所述第二非差分码本和所述第二对角 化差分码本分别与所述第一非差分码本和所述第一对角化差分码本相一致。
20、 根据权利要求 19所述的获取预编码矩阵的装置, 其特征在于, 对于单 用户多输入多输出系统, 所述预编码矩阵获取模块具体包括:
码字获取单元, 用于在所述预编码矩阵指示接收模块接收到所述参考预编 码矩阵指示和所述差分预编码矩阵指示后, 根据所述参考预编码矩阵指示, 从 所述第二非差分码本中查询得到所述参考预编码矩阵指示对应的码字, 并根据 所述差分预编码矩阵指示, 从所述第二对角化差分码本中查询得到所述差分预 编码矩阵指示对应的码字;
第一预编码矩阵获取单元, 用于在所述码字获取单元得到所述差分预编码 矩阵指示对应的码字与所述参考预编码矩阵指示对应的码字后, 将所述差分预 编码矩阵指示对应的码字与所述参考预编码矩阵指示对应的码字进行相乘, 将 乘积结果作为预编码矩阵。
21、 根据权利要求 19所述的获取预编码矩阵的装置, 其特征在于, 对于多 用户多输多输出系统, 所述预编码矩阵获取模块具体包括:
第二预编码矩阵获取单元, 用于在所述预编码矩阵指示接收模块接收到所述 参考预编码矩阵指示和所述差分预编码矩阵指示后, 根据同时配对的各个数据 接收端的参考预编码矩阵指示和差分预编码矩阵指示, 利用迫零波束赋形算法 或基于最大信泄噪比准则, 计算得到预编码矩阵。
22、 一种获取预编码矩阵指示的方法, 其特征在于, 所述方法包括: 根据第一非差分码本, 基于预设的准则, 计算得到非差分预编码矩阵指示; 其中, 所述第一非差分码本中包括的码字从旋转哈达马矩阵中得到。
23、 根据权利要求 22所述的获取预编码矩阵指示的方法, 其特征在于, 所述第一非差分码本中包括的码字 的结构如下:
Figure imgf000058_0001
其中, 表示所述第一非差分码本 ww中包括的非差分预编码矩阵指示, t = 0, l 表示所述第一非差分码本 Ww中包括的码字的数目, r表示
Figure imgf000058_0002
所述第一非差分码本 Ww中包括的码字的秩, 4表示与 对应的对角矩阵 的索 引, R4表示与 4对应的 n阶对角矩阵, A表示与 对应的归一化哈达马矩阵的索 引, Η„ ,表示与 对应的 n阶归一化哈达马矩阵, n的取值与发射天线的数目相 等, (R; H„ , f)表示从旋转哈达马矩阵 . Η„ ,中选择 r列构成的矩阵; 或
所述第一非差分码本中包括的码字 W^)的结构如下:
Figure imgf000058_0003
其中, 表示所述第一非差分码本 Ww中包括的预编码矩阵指示, fc = 0,l — 1
Figure imgf000058_0004
Ww |表示所述第一非差分码本 Ww中包括的码字的数目, r表示所述第一非差分 码本 Ww中包括的码字的秩, 4表示与 对应的对角矩阵 Rit的索引, Rit表示与 4对 应的 n阶对角矩阵, A表示与 对应的归一化哈达马矩阵的索引, H„A表示与 对应的 n阶归一化哈达马矩阵, n的取值与发射天线的数目相等,( VR; ¾H„, Rf 表 示从旋转哈达马矩阵 R; H ,. R 中选择 r列构成的矩阵。
24、根据权利要求 22-23中任意一项权利要求所述的获取预编码矩阵指示的 方法, 其特征在于, 所述计算得到非差分预编码矩阵指示之后, 还包括:
将所述非差分预编码矩阵指示反馈给数据发送端, 使所述数据发送端利用 其自身中预先存储的第二非差分码本, 根据所述非差分预编码矩阵指示, 得到 预编码矩阵, 其中, 所述第二非差分码本与所述第一非差分码本相一致。
25、 一种获取预编码矩阵的方法, 其特征在于, 所述方法包括:
接收非差分预编码矩阵指示; 其中, 所述非差分预编码矩阵指示由数据接 收端根据第一非差分码本得到并发送; 所述第一非差分码本中包括的码字从旋 转哈达马矩阵得到;
根据所述非差分预编码矩阵指示, 利用本地预先存储的第二非差分码本, 得到预编码矩阵, 其中, 所述第二非差分码本与所述第一非差分码本相一致。
26、 一种获取预编码矩阵指示的装置, 其特征在于, 所述装置包括: 非差分预编码矩阵指示获取模块, 用于根据第一非差分码本, 基于预设的 准侧, 计算得到非差分预编码矩阵指示, 其中, 所述第一非差分码本中包括的 码字从旋转哈达马矩阵得到。
27、 根据权利要求 26所述的获取预编码矩阵指示的装置, 其特征在于, 所 述装置还包括:
非差分预编码矩阵指示反馈模块, 用于在所述非差分预编码矩阵指示获取 模块得到非差分预编码矩阵指示后, 将所述非差分预编码矩阵指示反馈给数据 发送端, 使所述数据发送端利用其自身中预先存储的第二非差分码本, 根据所 述预编码矩阵指示, 得到预编码矩阵, 其中, 所述第二非差分码本与所述第一 非差分码本相一致。
28、 一种获取预编码矩阵的装置, 其特征在于, 所述装置包括:
非差分预编码矩阵指示接收模块, 用于接收非差分预编码矩阵指示; 其中, 所述非差分预编码矩阵指示由数据接收端根据第一非差分码本得到并发送; 所 述第一非差分码本中包括的码字从旋转哈达马矩阵得到;
非差分预编码矩阵处理模块, 用于在所述预编码矩阵指示接收模块接收到 所述非差分预编码矩阵指示后, 根据所述非差分预编码矩阵指示, 利用本地预 先存储的第二非差分码本, 得到预编码矩阵, 其中, 所述第二非差分码本与所 述第一非差分码本相一致。
PCT/CN2011/070323 2010-01-16 2011-01-17 获取预编码矩阵指示以及预编码矩阵的方法和装置 WO2011085695A1 (zh)

Priority Applications (6)

Application Number Priority Date Filing Date Title
EP11732685.0A EP2525505B1 (en) 2010-01-16 2011-01-17 Method and device for obtaining precoding matrix indicator (pmi) and precoding matrix
EP15182630.2A EP2985924B1 (en) 2010-01-16 2011-01-17 Method and apparatus for acquiring a precoding matrix indicator and a precoding matrix
US13/549,810 US8848768B2 (en) 2010-01-16 2012-07-16 Method and apparatus for acquiring a precoding matrix indicator and a precoding matrix
US13/730,044 US8625711B2 (en) 2010-01-16 2012-12-28 Method and apparatus for acquiring a Precoding Matrix Indicator and a Precoding Matrix
US14/340,971 US9054754B2 (en) 2010-01-16 2014-07-25 Method and apparatus for acquiring a precoding matrix indicator and a precoding matrix
US14/713,502 US9350432B2 (en) 2010-01-16 2015-05-15 Method and apparatus for acquiring a precoding matrix indicator and a precoding matrix

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN201010005350.9 2010-01-16
CN201010005350 2010-01-16
CN2010101115269A CN102130752B (zh) 2010-01-16 2010-02-11 获取预编码矩阵指示以及预编码矩阵的方法和装置
CN201010111526.9 2010-02-11

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US13/549,810 Continuation US8848768B2 (en) 2010-01-16 2012-07-16 Method and apparatus for acquiring a precoding matrix indicator and a precoding matrix

Publications (1)

Publication Number Publication Date
WO2011085695A1 true WO2011085695A1 (zh) 2011-07-21

Family

ID=44268653

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2011/070323 WO2011085695A1 (zh) 2010-01-16 2011-01-17 获取预编码矩阵指示以及预编码矩阵的方法和装置

Country Status (4)

Country Link
US (4) US8848768B2 (zh)
EP (2) EP2525505B1 (zh)
CN (3) CN102130752B (zh)
WO (1) WO2011085695A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9634748B2 (en) * 2013-06-29 2017-04-25 Huawei Technologies Co., Ltd. Method and apparatus for determining precoding matrix indicator, user equipment, and base station

Families Citing this family (35)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103378889A (zh) * 2012-04-24 2013-10-30 株式会社Ntt都科摩 码本生成方法、码本生成装置以及初始码本生成方法
WO2013185320A1 (zh) 2012-06-14 2013-12-19 华为技术有限公司 确定预编码矩阵指示的方法、用户设备、基站演进节点
CN103782533B (zh) 2012-07-02 2017-02-22 华为技术有限公司 确定预编码矩阵指示的方法、用户设备和基站
CN103580789B (zh) * 2012-07-30 2018-08-10 中兴通讯股份有限公司 多点协作传输预编码处理方法、装置及系统
CN103684657A (zh) * 2012-09-03 2014-03-26 夏普株式会社 预编码矩阵构造和索引值反馈方法及相关通信设备
CN103795489B (zh) * 2012-10-29 2017-05-24 电信科学技术研究院 传输编码指示信息和确定预编码矩阵的方法、系统及设备
CN108809388B (zh) 2013-04-03 2022-03-08 华为技术有限公司 信道状态信息上报方法、接收方法及设备
KR101819480B1 (ko) * 2013-04-28 2018-01-17 후아웨이 테크놀러지 컴퍼니 리미티드 프리코딩 행렬 인디케이터 피드백 방법, 수신단 및 송신단
CN112039566B (zh) 2013-05-10 2022-07-29 华为技术有限公司 确定预编码矩阵指示的方法、用户设备和基站
CN114257341B (zh) 2013-05-24 2023-08-29 中兴通讯股份有限公司 信道状态信息反馈方法和终端
US9432101B2 (en) * 2013-06-07 2016-08-30 Google Technology Holdings LLC Methods for codebook sub-sampling
US9509379B2 (en) 2013-06-17 2016-11-29 Huawei Technologies Co., Ltd. System and method for designing and using multidimensional constellations
EP3018852B1 (en) 2013-08-08 2019-10-16 Huawei Technologies Co., Ltd. Method for determining precoding matrix indicator, receiving device and transmitting device
US9178588B2 (en) * 2013-09-10 2015-11-03 Intel IP Corporation Method for determining a transmission direction for a communication, a method for determining a precoding matrix for a communication, and a device configured to do the same
US9800314B2 (en) * 2014-03-05 2017-10-24 Lg Electronics Inc. Method and apparatus for measuring channel variation for massive antenna array based beamforming in wireless communication system
CN106471752A (zh) * 2014-08-08 2017-03-01 富士通株式会社 基于波束的信息传输方法、装置以及通信系统
US10523383B2 (en) 2014-08-15 2019-12-31 Huawei Technologies Co., Ltd. System and method for generating waveforms and utilization thereof
US10084579B2 (en) 2014-11-17 2018-09-25 Samsung Electronics Co., Ltd. CSI feedback for MIMO wireless communication systems with polarized active antenna array
WO2016080742A1 (en) * 2014-11-17 2016-05-26 Samsung Electronics Co., Ltd. Csi feedback for mimo wireless communication systems with polarized active antenna array
WO2017071586A1 (en) 2015-10-30 2017-05-04 Huawei Technologies Co., Ltd. System and method for high-rate sparse code multiple access in downlink
WO2017190356A1 (en) * 2016-05-06 2017-11-09 Qualcomm Incorporated Enhancement of linear combination codebooks in fd-mimo
US10056941B2 (en) * 2016-06-20 2018-08-21 Qualcomm Incorporated Wireless communication impairments correction
CN107733559B (zh) * 2016-08-12 2021-06-22 北京华为数字技术有限公司 一种数据传输方法、发送设备及接收设备
CN106330277B (zh) * 2016-08-26 2019-06-21 西安电子科技大学 应用于gfdm通信系统的相关旋转预编码方法
CN108282203B (zh) * 2017-01-05 2021-08-10 中国移动通信集团四川有限公司 用于无线通信的方法和设备
CN108322283B (zh) * 2017-01-14 2021-09-07 华为技术有限公司 数据发送技术
WO2018129733A1 (zh) * 2017-01-16 2018-07-19 华为技术有限公司 确定信道状态信息的方法、接入网设备和终端设备
JP6423475B2 (ja) * 2017-03-30 2018-11-14 華為技術有限公司Huawei Technologies Co.,Ltd. プリコーディング行列インジケータを決定するための方法および装置、ユーザ機器、ならびに基地局
CN108809372B (zh) * 2017-04-26 2021-05-11 华为技术有限公司 一种指示及确定预编码向量的方法和设备
CN107302387B (zh) * 2017-06-30 2020-06-19 西安电子科技大学 一种高速飞行器中继双极化mimo信道建模方法
CN109495143B (zh) * 2017-09-11 2021-12-24 大唐移动通信设备有限公司 一种预编码矩阵的发送、接收方法及设备
CN110492916B (zh) * 2018-05-15 2021-08-03 华为技术有限公司 预编码矩阵指示方法及相关设备
US10404339B1 (en) 2018-08-09 2019-09-03 At&T Intellectual Property I, L.P. Precoding matrix indicator determination in wireless communication systems
CN114124176A (zh) * 2020-08-27 2022-03-01 华为技术有限公司 预编码方法及通信装置
CN115441911A (zh) * 2021-06-01 2022-12-06 华为技术有限公司 通信处理方法和通信处理装置

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101043298A (zh) * 2006-03-20 2007-09-26 华为技术有限公司 一种多天线通信中发射信号的方法及系统
CN101146078A (zh) * 2006-12-27 2008-03-19 中兴通讯股份有限公司 一种多输入多输出空间复用预编码矩阵的选择方法
CN101512929A (zh) * 2006-08-17 2009-08-19 交互数字技术公司 用于在mimo无线通信系统中提供有效预编码反馈的方法和设备

Family Cites Families (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6373859B1 (en) * 1996-09-10 2002-04-16 Hewlett-Packard Company Methods and apparatus for encoding and decoding data
US7885348B2 (en) 2006-02-09 2011-02-08 Intel Corporation MIMO communication system and method for beamforming using polar-cap codebooks
US7629902B2 (en) * 2007-06-08 2009-12-08 Samsung Electronics Co., Ltd. MIMO wireless precoding system robust to power imbalance
KR101293373B1 (ko) * 2007-06-25 2013-08-05 엘지전자 주식회사 다중안테나 시스템에서의 데이터 전송 방법
CN101335558B (zh) * 2007-06-29 2012-07-04 华为技术有限公司 多输入多输出信道的码本生成方法及装置
WO2009059039A2 (en) * 2007-11-02 2009-05-07 Interdigital Patent Holdings, Inc. Method and apparatus for generating channel quality indicator
KR101349826B1 (ko) * 2007-11-05 2014-01-09 엘지전자 주식회사 코드북 기반의 다중 입출력 시스템에 있어서, 피드백정보를 송신하는 방법 및 코드북 기반 프리코딩 기법을적용하는 다중 입출력 시스템에 있어서, 프리코딩 정보를송신하는 방법
KR101871707B1 (ko) * 2010-04-02 2018-06-27 엘지전자 주식회사 무선통신 시스템에서 채널상태정보 피드백 하는 단말 장치 및 그 방법
US8213873B2 (en) * 2010-04-09 2012-07-03 Telefonaktiebolaget Lm Ericsson (Publ) Method and apparatus for channel quality reporting in a wireless communication system
KR101835326B1 (ko) * 2010-09-26 2018-03-07 엘지전자 주식회사 다중 안테나 지원 무선 통신 시스템에서 효율적인 피드백 방법 및 장치
US8711907B2 (en) * 2010-10-01 2014-04-29 Intel Corporation PMI feedback with codebook interpolation
US20120314590A1 (en) * 2011-06-10 2012-12-13 Sharp Laboratories Of America, Inc. Enhanced precoding feedback for multiple-user multiple-input and multiple-output (mimo)

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101043298A (zh) * 2006-03-20 2007-09-26 华为技术有限公司 一种多天线通信中发射信号的方法及系统
CN101512929A (zh) * 2006-08-17 2009-08-19 交互数字技术公司 用于在mimo无线通信系统中提供有效预编码反馈的方法和设备
CN101146078A (zh) * 2006-12-27 2008-03-19 中兴通讯股份有限公司 一种多输入多输出空间复用预编码矩阵的选择方法

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
"Differential PMI Feedback for Overhead Reduction in LTE-A DL IMO, PANTECH&CURITEL, R1-093453", 3GPP TSG-RAN WG1 #58, 28 August 2009 (2009-08-28), XP050388251, Retrieved from the Internet <URL:http://ftp.3gpp.org/specs/html-info/TDocExMtg--R1-58--27293.htm> [retrieved on 20110401] *
See also references of EP2525505A4 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9634748B2 (en) * 2013-06-29 2017-04-25 Huawei Technologies Co., Ltd. Method and apparatus for determining precoding matrix indicator, user equipment, and base station
US9843370B2 (en) 2013-06-29 2017-12-12 Huawei Technologies Co., Ltd. Method and apparatus for determining precoding matrix indicator, user equipment, and base station
US10116363B2 (en) 2013-06-29 2018-10-30 Huawei Technologies Co., Ltd. Method and apparatus for determining precoding matrix indicator, user equipment, and base station
US10326505B2 (en) 2013-06-29 2019-06-18 Hauwei Technologies Co., Ltd. Method and apparatus for determining precoding matrix indicator, user equipment, and base station

Also Published As

Publication number Publication date
CN103220087A (zh) 2013-07-24
EP2525505A4 (en) 2012-11-21
EP2985924B1 (en) 2019-09-18
US8625711B2 (en) 2014-01-07
US9350432B2 (en) 2016-05-24
CN103248457A (zh) 2013-08-14
CN103220087B (zh) 2017-06-06
US20150256243A1 (en) 2015-09-10
US8848768B2 (en) 2014-09-30
US20120275500A1 (en) 2012-11-01
EP2985924A1 (en) 2016-02-17
CN103248457B (zh) 2016-12-07
EP2525505B1 (en) 2015-11-11
US9054754B2 (en) 2015-06-09
CN102130752A (zh) 2011-07-20
US20140334571A1 (en) 2014-11-13
CN102130752B (zh) 2013-04-24
EP2525505A1 (en) 2012-11-21
US20130121437A1 (en) 2013-05-16

Similar Documents

Publication Publication Date Title
WO2011085695A1 (zh) 获取预编码矩阵指示以及预编码矩阵的方法和装置
US11251843B2 (en) Methods and devices for determining precoder parameters in a wireless communication network
EP2901569B1 (en) Method for wifi beamforming, feedback, and sounding (wibeam)
US8553799B2 (en) Method and apparatus for obtaining precoding matrix indicator
JP5465339B2 (ja) マルチアンテナシステム中で情報を送受信するための方法および装置、ならびにそのマルチアンテナシステム
US10594372B2 (en) Codebook construction
EP2477342A1 (en) Method and system for acquiring channel state information
EP3171525A1 (en) System and method for channel information feedback in a wireless communications system
WO2011124021A1 (zh) 用于信息反馈以及预编码的方法和装置
WO2011137633A1 (zh) 反馈信道状态信息的方法及装置
WO2010069242A1 (en) Method for employing six-bit rank 1 and 2 codebooks for four transmit antennas
US9020061B2 (en) Codebook construction
JP2018201205A (ja) コードブック構造
JP5576558B2 (ja) 複数入力複数出力システムで高ランク適応コードブックを生成し、フィード・バックする方法および装置
WO2012022208A1 (zh) 信道信息的发送方法、终端、基站及lte-a系统

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 11732685

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

WWE Wipo information: entry into national phase

Ref document number: 2011732685

Country of ref document: EP