WO2020135573A1 - 信道状态信息csi反馈方法、设备及存储介质 - Google Patents

信道状态信息csi反馈方法、设备及存储介质 Download PDF

Info

Publication number
WO2020135573A1
WO2020135573A1 PCT/CN2019/128653 CN2019128653W WO2020135573A1 WO 2020135573 A1 WO2020135573 A1 WO 2020135573A1 CN 2019128653 W CN2019128653 W CN 2019128653W WO 2020135573 A1 WO2020135573 A1 WO 2020135573A1
Authority
WO
WIPO (PCT)
Prior art keywords
feedback
vector
coefficient
coefficients
base
Prior art date
Application number
PCT/CN2019/128653
Other languages
English (en)
French (fr)
Inventor
郑国增
吴昊
李永
李儒岳
鲁照华
陈艺戬
Original Assignee
中兴通讯股份有限公司
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 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2020135573A1 publication Critical patent/WO2020135573A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/0413—MIMO systems
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/0413—MIMO systems
    • H04B7/0456—Selection of precoding matrices or codebooks, e.g. using matrices antenna weighting
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B7/00—Radio transmission systems, i.e. using radiation field
    • H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
    • H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
    • H04B7/06—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station
    • H04B7/0613—Diversity 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
    • H04B7/0615—Diversity 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
    • H04B7/0619—Diversity 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
    • H04B7/0621—Feedback content
    • H04B7/0626—Channel coefficients, e.g. channel state information [CSI]

Definitions

  • Embodiments of the present invention relate to wireless communication technologies, for example, to a channel state information CSI feedback method, device, and storage medium.
  • MIMO Multiple-Input Multiple-Output
  • CSI feedback is a key technology for high-performance precoding or beamforming in MIMO systems.
  • the quantization feedback on the channel matrix will bring a relatively large feedback overhead, especially when supporting CSI feedback with a large bandwidth, the feedback overhead is an important issue that limits performance improvement.
  • an embodiment of the present invention provides a channel state information CSI feedback method, including: a terminal feeds back precoding matrix information, the precoding matrix information includes first base vector information, second base vector information, and second coefficient Amplitude and phase information;
  • the precoding vector in the feedback precoding subband is the linear combination of the first base vector, and the weighting coefficient used for the linear combination of the first base vector is the first coefficient; in the frequency domain units included in all CSI feedback bands, A vector composed of the first coefficients corresponding to the same first base vector is a linear combination of second base vectors, and the second coefficient is a weighting coefficient used for linear combination of the second base vectors.
  • An embodiment of the present invention also provides a channel state information CSI feedback method, including:
  • the base station receives the precoding matrix information fed back by the terminal, where the precoding matrix information includes first base vector information, second base vector information, amplitude and phase information of the second coefficient;
  • the precoding vector in the feedback precoding subband is the linear combination of the first base vector, and the weighting coefficient used for the linear combination of the first base vector is the first coefficient; in the frequency domain units included in all CSI feedback bands, A vector composed of the first coefficients corresponding to the same first base vector is a linear combination of second base vectors, and the second coefficient is a weighting coefficient used for linear combination of the second base vectors.
  • An embodiment of the present invention also provides a terminal, including:
  • a feedback unit for feeding back precoding matrix information includes first base vector information, second base vector information, amplitude and phase information of the second coefficient;
  • the precoding vector in the feedback precoding subband is the linear combination of the first base vector, and the weighting coefficient used for the linear combination of the first base vector is the first coefficient; in the frequency domain units included in all CSI feedback bands, A vector composed of the first coefficients corresponding to the same first base vector is a linear combination of second base vectors, and the second coefficient is a weighting coefficient used for linear combination of the second base vectors.
  • An embodiment of the present invention also provides a base station, including:
  • a receiving unit configured to receive precoding matrix information fed back by the terminal, where the precoding matrix information includes first base vector information, second base vector information, and amplitude and phase information of the second coefficient;
  • the precoding vector in the feedback precoding subband is the linear combination of the first base vector, and the weighting coefficient used for the linear combination of the first base vector is the first coefficient; in the frequency domain units included in all CSI feedback bands, A vector composed of the first coefficients corresponding to the same first base vector is a linear combination of second base vectors, and the second coefficient is a weighting coefficient used for linear combination of the second base vectors.
  • An embodiment of the present invention also provides a terminal, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and the computer program is executed by the processor to implement the terminal A method of performing channel state information CSI feedback.
  • An embodiment of the present invention also provides a base station, including a memory, a processor, and a computer program stored on the memory and executable on the processor.
  • the computer program is executed by the processor to implement the above base station A method of performing channel state information CSI feedback.
  • An embodiment of the present invention also provides a computer-readable storage medium, characterized in that an information processing program is stored on the computer-readable storage medium, and when the information processing program is executed by a processor, any of the above Channel state information CSI feedback method.
  • the embodiments of the present invention provide a channel state information CSI feedback method and device. By feeding back the first base vector information, the second base vector information, and the amplitude and phase information of the second coefficient, both can be reduced.
  • the CSI feedback overhead ensures high CSI feedback performance.
  • FIG. 1 is a schematic flowchart of a channel state information CSI feedback method according to Embodiment 1 of the present invention
  • FIG. 2 is a schematic flowchart of a channel state information CSI feedback method according to Embodiment 2 of the present invention
  • FIG. 3 is a schematic flowchart of a channel state information CSI feedback method according to Embodiment 1 of the present invention.
  • FIG. 4 is a schematic flowchart of a channel state information CSI feedback method according to Embodiment 2 of the present invention.
  • FIG. 5 is a schematic structural diagram of a terminal according to Embodiment 3 of the present invention.
  • FIG. 6 is a schematic structural diagram of a base station according to Embodiment 4 of the present invention.
  • CSI quantization feedback technology is an important part of MIMO technology.
  • DFT Discrete Fourier Transform
  • variations of DFT vectors are commonly used, such as the Kronecker product of multiple DFT vectors, or the cascaded form of DFT vectors, Or in the form of performing phase adjustment on the cascaded DFT vectors, the terminal reports the precoding instruction information in the above form to the base station through quantization feedback.
  • This type of precoding codebook can be classified as the first type of codebook. This codebook has a small overhead, but the CSI quantization accuracy is low and the performance is limited.
  • the other codebook uses linear weighted combination of the DFT vector or the Kronecker product of the DFT vector.
  • the weighted and combined vectors are called codebook base vectors.
  • the codebook base vector related information, the amplitude and phase information of the weighting coefficients As the precoding instruction information is fed back to the base station, such a precoding codebook can be classified as a second type codebook.
  • the number of columns of the precoding matrix fed back by the terminal is RI (Rank indicator), where the precoding vector of each layer is expressed as a set of codebook bases Linear combination of vectors, the group of codebook basis vectors can be called the first basis vector, the terminal calculates the weighting coefficient of the linear combination according to the first basis vector, and quantizes the feedback weighting coefficient amplitude and phase information, the weighting coefficient can be called The first coefficient.
  • the subband is a frequency domain granularity. For all RBs (Resource Blocks) included in the CSI feedback bandwidth, consecutive multiple RBs form a subband.
  • the precoding of a layer in a subband can be expressed as:
  • W 1 contains 2L first basis vectors, W 1 dimension is N t ⁇ 2L, its form is:
  • b 0 , b 1 ,..., b L-1 is composed of a set of orthogonal DFT vectors or the Kronecker product of DFT vectors, whose dimensions are N t represents the number of ports of CSI-RS (Channel State Information Reference Signal, Channel State Information Reference Signal).
  • This CSI feedback method uses the same DFT vector or Kronecker product of DFT vectors for antenna ports with different polarization directions, so W 1 contains two first basis vectors.
  • the information in W 1 is broadband feedback, that is, the information in W 1 is the same for different subbands in the entire CSI feedback band.
  • the dimension of w ⁇ is 2L ⁇ 1, which represents the weighting coefficient of 2L first basis vectors, which is called the first coefficient.
  • the terminal needs to feed back the phase and/or amplitude information of the weighting coefficient of each codebook basis vector for each subband. Therefore, when the number of subbands is large, the second type of codebook feedback will bring a large CSI feedback overhead. On the other hand, if only the information of the amplitude or phase of the weighting coefficients over the entire broadband is fed back, the high-performance gain that the second type of codebook feedback can bring cannot be sufficiently obtained. Therefore, how to obtain a higher feedback performance of the second type codebook through a smaller feedback overhead is a problem that the related art has not solved.
  • the embodiments of the present invention provide a new channel state information CSI feedback scheme, by which the CSI feedback overhead can be reduced, and at the same time, high CSI feedback performance can be ensured.
  • FIG. 1 is a schematic flowchart of a channel state information CSI feedback method according to Embodiment 1 of the present invention. As shown in FIG. 1, the method includes:
  • Step 101 The terminal feeds back precoding matrix information, where the precoding matrix information includes first base vector information, second base vector information, amplitude and phase information of the second coefficient; wherein, the precoding vector in the precoding subband is fed back Is the linear combination of the first basis vectors, and the weighting coefficient used in the linear combination of the first basis vectors is the first coefficient; in the frequency domain units included in all CSI feedback frequency bands, the weighting coefficients corresponding to the same first basis vector The vector composed of the first coefficients is the linear combination of the second basis vectors, and the second coefficient is the weighting coefficient used for the linear combination of the second basis vectors.
  • said second base vector represents the DFT matrix or DFT over the K base vectors of the DFT matrix selected sampling, the oversampling factor value over-sampling O f DFT matrix as one of the following: 1,2, 4, 8.
  • the method further includes: receiving a frequency domain range sent by the base station that needs to feed back CSI, and the frequency domain range of the CSI includes: a precoding subband that needs to feed back CSI or a comb feedback.
  • the second base vector is modified by one of the following methods: One: The base station configures a larger oversampling factor to increase the number of selectable DFT matrices; Method two: The terminal intercepts the DFT base vectors in a comb shape according to the precoding subband of the precoding information that needs to be fed back; or, when the base station configuration section does not When the precoding subband needs to be fed back, the corresponding position element of the precoding subband that does not require feedback of the precoding information in the DFT base vector is truncated; or, the base station configures whether to phase-rotate the second base vector to The configured middlemost precoding subband is used as a reference.
  • the CSI feedback frequency band refers to the frequency domain range configured by the base station that needs to feedback CSI.
  • the precoding subband is a frequency domain unit
  • the number of RBs included is determined by the number of RBs configured by the base station or included in the CSI feedback band and the number of second base vectors used, that is, the number of included RBs is RBNum is the number of RBs included in the CSI feedback frequency band, and K is the number of second basis vectors used in the CSI feedback frequency band.
  • the second coefficient is expressed as The dimension of is 2L ⁇ K, 2L represents the 2L first basis vectors contained in the matrix W 1 , and K represents the second basis vectors contained in the matrix W 3 .
  • the method further includes: receiving a feedback mode and a feedback subset indication of the second coefficient sent by the base station; the second coefficient in the precoding matrix information is fed back in one of the following ways : Method 1: The second coefficient is fed back according to the magnitude of the second coefficient as priority; Method 2: The second coefficient is fed back with the first basis vector as the priority; Method 3: The second basis vector is used as the priority Feedback the second coefficient; Method 4: Divide the first basis vector into two groups, and divide the first to Lth first basis vectors, that is, the 1st to Lth column vectors of W 1 , as the first group, the first L + 1 to a first group of two vectors, i.e. the first W L + 1 to 2L of 1 column vector, the second group, according to the priority of the packet inside a second feedback coefficient; five ways: second The coefficients are grouped, and the second coefficient is fed back according to the priority of the grouping.
  • the first method includes: sorting by the magnitude of the second coefficient, selecting partial coefficients from large to small, until the ratio of the power of the partial coefficients to the total power of the second coefficient is not less than ⁇ , the terminal feedbacks Partial coefficients, where ⁇ is a preset threshold; a bitmap of 2L ⁇ K is used to indicate the selected coefficient.
  • the second approach comprises: mode A: Select l 0 K coefficients in the optimal weighting coefficients in a first group of vectors, each of l 1 select one of the remaining K weighting coefficients in each of a first base vector Coefficients, l 1 ⁇ l 0 , use a bitmap of 2L ⁇ K size to indicate the selected coefficients; or, method B: select l 0 coefficients out of the K weighting coefficients of the optimal first base vector, and the remaining All the weighting coefficients of the first base vector, select a total of feedback coefficients Ml 0, the value of M is configured by a base station or a Where ⁇ M is the preset threshold; use one of the following sub-methods to feed back its selected subset: sub-method B-1: use a bitmap of 2L ⁇ K size to indicate its selected coefficient; sub-method B-2: use Bits indicate the selected optimal first basis vector, and reuse Instruct the remaining Ml 0 coefficients it chooses; or, method C: sort the two first basis vectors, and select a
  • said third approach comprises: D mode, select 2L l 0 coefficients in the optimal weighting coefficients in a second base vector, all remaining weighting coefficients in the second group of vectors, a total of select Ml 0 Coefficient, the value of M is configured by the base station or is Where ⁇ M is the preset threshold; use one of the following sub-modes to feed back the selected subset: sub-mode D-1: use a 2L ⁇ K-sized bitmap to indicate the coefficients selected; sub-mode D-2: use Indicate the selected optimal second basis vector, reuse bits indicate the remaining Ml 0 coefficients it selects; or, method E: sort the K second basis vectors, and select the weighting coefficients of the strongest and weakest second basis vectors l 0 , l 0 -1, respectively... ., l 0 -M+1 coefficients, using a bitmap of 2L ⁇ K size to indicate the selected coefficients.
  • the method 4 includes: Method F: In all weighting coefficients of the first basis vector in the two groups, each feedback Coefficients, The value of is configured by the base station or is Where ⁇ M is the preset threshold; through the bitmap or Indicate the selected coefficient; or, mode G: within two groups, feedback according to the priority of the first base vector and use a bitmap of 2L ⁇ K size to indicate the selected coefficient, including: Sub-mode G-1: Choose an optimal first basis vector for each of the two groups, and select l 0 coefficients for the weighting coefficients of the two optimal first basis vectors.
  • ⁇ M is a preset threshold
  • sub-mode G-2 select an optimal first basis vector from each of the two groups, and select l 0 coefficients from the weight coefficients of the two optimal first basis vectors, In the weighting coefficients of the remaining first basis vectors in the two groups, select l 1 coefficients, l 1 ⁇ l 0 ; or, mode H: in the two groups, feedback according to the priority of the second basis vector And use the 2L ⁇ K size bitmap to indicate its selected coefficients, including: Sub-mode H-1: Select an optimal second basis vector in each of the two groups, and the optimal second basis vector in the first group Choose l 0 coefficients among the weighting coefficients of the second group, select l 0 coefficients among the weighting coefficients of the optimal second basis vectors in the second group, and each of the weighting coefficients of the remaining second basis vectors in the two groups
  • the method 5 includes: method I: divide the second coefficient into GroupNumber combinations, the terminal selects the l most suitable combination for feedback, use Indicate the selected combination; or, method J: divide the second coefficient into GroupNumber combinations, select the most suitable combination, and then select l-1 combinations among the remaining combinations, and use log 2 GroupNumberbits to indicate the most appropriate selection Combination, use Indicate the remaining combinations.
  • the method further includes: receiving first signaling sent by the base station; the first signaling includes a first amplitude feedback penalty factor configured by the base station and a first indicating that power limitation is required Basis vector, used to perform power limitation on the weighting coefficient of the first basis vector requiring power limitation; the weighted coefficient amplitude of the first basis vector requiring power limitation fed back by the terminal cannot exceed the base station configuration after quantization Quantize the first amplitude feedback penalty factor, or multiply the amplitude of the weighting coefficient of the first base vector that requires power limitation by the first amplitude feedback penalty factor configured by the base station, and then perform quantization; and/or Second signaling; the second signaling includes a second amplitude feedback penalty factor configured by the base station and a second basis vector indicating that power limitation is required, and a weighting coefficient used for the second basis vector requiring power limitation Power limitation; the weighted coefficient amplitude of the second basis vector required for power limitation fed back by the terminal cannot exceed the second amplitude feedback penalty factor configured
  • the feedback precoding matrix information includes: when the feedback resources are insufficient and the number K of selected second basis vectors is configured by the terminal, the number of second basis vectors fed back is dynamically reduced; or, when the selected second basis When the number of vectors K is configured by the base station, if the feedback resources are sufficient, upload all the parameters configured by the base station, otherwise choose one of the following methods:
  • the precoding matrix information includes: when the feedback resources are insufficient and the number K of selected second basis vectors is configured by the terminal, the number of second basis vectors fed back is dynamically reduced; or, when the selected second basis When the number of vectors K is configured by the base station, if the feedback resources are sufficient, upload all the parameters configured by the base station, otherwise choose one of the following methods:
  • the terminal gives priority to the optimal feedback
  • the weighting coefficient of the first basis vector, and then the terminal dynamically informs the selected coefficient through the bitmap based on a certain criterion until the maximum feedback resource is reached; when the mode 2 includes the mode B, the precoding matrix information
  • the weighting coefficient of the optimal first basis vector in a group followed by feedback of the optimal weighting coefficient of the first basis vector in the second group, and then based on a certain criterion, dynamically notify the selected coefficient through the bitmap until the maximum feedback is reached Resources; when feeding back the precoding matrix information according to the sub-mode H-1 or sub-mode H-2 included in the mode H included in the fourth mode, if the feedback resources are insufficient, the optimal second base in the first group is preferentially fed back
  • the weighting factor of the vector which Feedback the weighting coefficient of the optimal second basis vector in the second group secondly, and then dynamically notify the selected coefficient through the bitmap based on a certain criterion until the maximum feedback resource is reached; or, when the method includes I feed back the precoding matrix information and use the bitmap to feed back the selected group.
  • the base station is dynamically notified according to a certain criterion according to the bitmap to the selected group until the maximum feedback resource is reached;
  • the precoding matrix information is fed back in mode J, if the feedback resources are insufficient, the weighting coefficient corresponding to the optimal group is preferentially fed back.
  • FIG. 2 is a schematic flowchart of a channel state information CSI feedback method according to Embodiment 2 of the present invention. As shown in FIG. 2, the method includes:
  • Step 201 The base station receives the precoding matrix information fed back by the terminal, where the precoding matrix information includes first base vector information, second base vector information, amplitude and phase information of the second coefficient; wherein, feedback within the precoding subband
  • the precoding vector is a linear combination of the first basis vectors, and the weighting coefficient used for the linear combination of the first basis vectors is the first coefficient; in the frequency domain units included in all CSI feedback frequency bands, it corresponds to the same first basis vector
  • the vector composed of the first coefficients is a linear combination of second base vectors, and the second coefficient is a weighting coefficient used for linear combination of the second base vectors.
  • the method further includes: the base station configures a frequency domain range that needs to be fed back CSI and sends it to the terminal, so that the terminal determines that feedback is needed according to the frequency domain range of the CSI Precoding subbands of precoding information.
  • the frequency domain range of the configured CSI includes: a precoding subband requiring CSI feedback or a precoding subband requiring comb feedback CSI and sparseness, or a precoding subband requiring no feedback CSI.
  • the method further includes: the base station configures the feedback mode and the feedback subset indication of the second coefficient, and sends it to the terminal, so that the terminal feedbacks according to the second coefficient
  • the mode and the feedback subset indicate the feedback of the second coefficient
  • the feedback mode of the second coefficient includes one of the following: Mode 1: The second coefficient is fed back according to the magnitude of the second coefficient as priority;
  • Method 2 Feedback the second coefficient with the first basis vector as the priority
  • Method 3 Feedback the second coefficient with the second basis vector as the priority
  • Method 4 Divide the first basis vector into two groups, Take the 1st to Lth first basis vectors, that is, the 1st to Lth column vectors of W 1 as the first group, and set the L+1 to 2Lth first basis vectors, that is, the Lth of W 1 +1 to the 2Lth column vector, as the second group, feedback the second coefficient according to the priority within the group;
  • Method 5 Group the second coefficient, feedback the second coefficient according to the priority of the group; the feedback subset Indicates a mode for instructing the terminal to feed back the subset of the second coefficient.
  • the method further includes: sending the first signaling and/or the second signaling to the terminal; wherein, the first signaling includes the first amplitude feedback configured by the base station A penalty factor and a first basis vector indicating that power limitation is required, used to perform power limitation on the weighting coefficient of the first basis vector that requires power limitation; the second signaling includes a second amplitude feedback penalty configured by the base station The factor and the second basis vector indicating that power limitation is required are used to perform power limitation on the weighting coefficient of the second basis vector requiring power limitation.
  • the first amplitude feedback penalty factor is or Or, the second amplitude feedback penalty factor is or
  • Embodiment 1 of the present invention provides a CSI feedback method for compressing channel coefficients in the spatial and frequency domains, that is, the precoding of all precoding subbands in a certain layer can be re-expressed as follows,
  • W 2 dimension is 2L ⁇ N s
  • N s represents the number of precoding subbands
  • S represents the frequency domain unit of the terminal feedback precoding, which can be adjusted according to the accuracy of the feedback.
  • S is called the size of the precoding subband, and it can be composed of one or more RBs, and N s is called the number of precoding subbands;
  • the 2L elements in each column correspond to a subband, which represents the 2L Coefficient of linear superposition of a basis vector.
  • the CSI feedback frequency band refers to the frequency domain range that the base station configures to need to feedback CSI, and the number of RBs included is RBNum.
  • the precoding subband is a frequency domain unit, and the number of RBs included in it may be configured by the base station or determined by the number of RBs in the feedback frequency band and the number of second base vectors used, that is, the number of included RBs is RBNum is the number of RBs included in the CSI feedback frequency band, and is the number of second base vectors used in the CSI feedback frequency band.
  • W 1 contains 2L first basis vectors, W 1 dimension is N t ⁇ 2L, and its form is: Among them, b 0 , b 1 ,..., b L-1 is composed of a set of orthogonal DFT (Discrete Fourier Transformation, discrete Fourier transform) vectors or the Kronecker product of DFT vectors, and its dimension is N t represents the number of CSI-RS ports.
  • DFT Discrete Fourier Transformation, discrete Fourier transform
  • the frequency domain compression matrix W 3 [w 0 w 1 ... w K-1 ], which represents a basis vector selected from the DFT matrix or the oversampled DFT matrix, and is called a second basis vector.
  • the dimension of the DFT matrix is N s ⁇ N s ; wherein, L can be one of the following: 2, 3, 4, 5, 6.
  • the K may take one of the following: 1, 2, ..., N s , that is, the maximum value that K can take is the number of precoding subbands N s .
  • the dimension is 2L ⁇ K, which represents the matrix after W 2 is compressed in the frequency domain, and represents the weighting coefficient of the second basis vector, which is called the second coefficient.
  • the weighting coefficient of the i-th first basis vector refers to The K coefficients in the i-th row; the weighting coefficient of the j-th second base vector refers to 2L coefficients of column j.
  • the weighting coefficient of the j-th second basis vector in the first group namely L coefficients in the jth column and rows 1 to L in the middle; the weighting coefficient of the jth second base vector in the second group, namely L coefficients in the jth column and L+1 to 2L rows in the middle.
  • FIG. 3 is a schematic flowchart of a channel state information CSI feedback method according to Embodiment 1 of the present invention. As shown in FIG. 3, the method includes:
  • step 301 the terminal receives the configuration information sent by the base station.
  • the configuration information includes at least one of the following: the frequency domain range configured by the base station that needs to feedback CSI, the feedback mode and the feedback subset indication of the configured second coefficient, the configured first amplitude feedback penalty factor and the indication that power is required Information such as the limited first base vector, the configured second amplitude feedback penalty factor, and the second base vector indicating that power limitation is required.
  • the frequency domain range of the CSI includes: a precoding subband requiring feedback CSI or a precoding subband requiring comb feedback CSI and sparseness, or a precoding subband requiring no feedback CSI.
  • Step 302 The terminal obtains precoding matrix information according to the channel estimation result and the configuration information of the base station.
  • Step 303 The terminal feeds back precoding matrix information, where the precoding matrix information includes first base vector information, second base vector information, and amplitude and phase information of the second coefficient.
  • the precoding vector in the feedback precoding subband is the linear combination of the first base vector, and the weighting coefficient used for the linear combination of the first base vector is the first coefficient; in the frequency domain units included in all CSI feedback bands, A vector composed of the first coefficients corresponding to the same first base vector is a linear combination of second base vectors, and the second coefficient is a weighting coefficient used for linear combination of the second base vectors.
  • said second base vector represents the DFT matrix or DFT over the K base vectors of the DFT matrix selected sampling, the oversampling factor value over-sampling O f DFT matrix as one of the following: 1,2, 4, 8.
  • the second base vector is a DFT base vector
  • the second base vector may be partially modified according to actual needs.
  • the selection of the second base vector, that is, the base vector in W 3 needs to consider some special cases, more Specifically, it includes the following special situations and solutions:
  • Method A-1 The base station may not need the terminal to upload the CSI of all precoding subbands.
  • the base station configuration needs to upload the CSI precoding subbands may be discontinuous, so the DFT base vector needs to be modified. Specifically, do the following several ways to improve:
  • the DFT matrix can be modified in the following two ways:
  • Sub-mode A-1-1-1 Oversampling DFT matrix is used.
  • the base station configures a larger oversampling factor and increases the number of selectable DFT matrices. Essentially, it increases the phase difference between adjacent elements of the DFT base vector to match Partial precoding subband missing causes phase shift to increase;
  • Sub-mode A-1-1-2 Partial DFT matrix is intercepted, that is, the DFT base vector only retains the first S 0 + ⁇ , S 0 +2 ⁇ , S 0 +3 ⁇ , ... elements.
  • Sub-mode A-1-2 The base station may configure only a part of subbands without uploading CSI.
  • a specific example is that the base station configures the S i th precoding subband without reporting CSI.
  • the base vector can be truncated DFT of S i-th element, to adapt precoding subband partial deletion.
  • the DFT base vector refers to the first precoding subband, that is, the first element of the DFT base vector is 1.
  • the second coefficient namely feedback of.
  • the 2L ⁇ K elements can select part of the subset or all of the feedback, and different subsets can also feedback different accuracy to improve performance.
  • the base station and the terminal need to agree on how the subset selected by the terminal needs to be notified to the base station, that is, the feedback subset indication of the second coefficient.
  • Method C-1 The second coefficient is fed back according to the magnitude of the second coefficient as a priority; the part with a larger amplitude of the second coefficient usually reflects the true channel characteristics more.
  • the terminal sorts the amplitude of the second coefficient from large to small, and selects partial coefficients from large to small, until the ratio of the sum of the power of the partial coefficients and the total power of the second coefficient is not less than ⁇ , and the feedback is selected.
  • ⁇ is the threshold, and typical values can be taken as 0.90, 0.95, 0.99, etc.
  • the terminal uses a bitmap of 2L ⁇ K size to indicate its selected coefficient.
  • Method C-2 Select coefficients from the perspective of the first basis vector; where the first basis vector corresponds to the beam in space, usually the received signal will be concentrated in some specific beam directions, so the priority and feedback of the specific beam are increased Accuracy can improve system performance.
  • one of the following sub-modes may be selected:
  • Sub-mode C-2-1 the terminal selects the optimal first basis vector based on certain criteria (such as the K coefficient power and maximum corresponding to a certain first basis vector), then the K of the optimal first basis vector Weighting factors (corresponding to A certain line) feedback l 0 coefficients; the quantization precision of the weighting coefficient of this first base vector can be appropriately increased, for example, the amplitude and phase are quantized to (3, 4) bits or (4, 4) bits; in the remaining
  • Each of the K weighting coefficients of a basis vector selects l 1 (l 1 ⁇ l 0 ) coefficients, and the amplitude and phase are quantized as (2, 3) bits or (3, 3) bits, where the amplitude can be directly quantized or according to the most The amplitude of the optimal first base vector is differentially quantized.
  • the terminal uses a 2L ⁇ K size bitmap to indicate its selected coefficients.
  • the base station can know the feedback subset and its optimal first basis vector according to the bitmap information fed back by the terminal, and can further calculate it according to the configured quantization accuracy The total cost of feedback of the second coefficient.
  • Submode C-2-2 terminal based on certain criteria, a first group of selecting the optimal vector, K is the optimum weighting coefficients of the first base vector feedback coefficient l 0, this can be increased the The quantization accuracy of the weighting coefficient of the optimal first basis vector, for example, the amplitude and phase are quantized to (3, 4) bits or (4, 4) bits; in the remaining first basis except the optimal first basis vector A total of Ml 0 coefficients are selected from all the weighting coefficients of the vector.
  • the value of M can be configured by the base station or Where ⁇ M is a preset threshold; its amplitude and phase quantization accuracy is (2, 3) bits or (3, 3) bits, where the amplitude can be directly quantized or differentially quantized according to the amplitude of the optimal first base vector.
  • the part of the second coefficient that the terminal indicates to its feedback may be implemented in one of the following sub-modes:
  • Sub-mode C-2-2-1 The terminal uses a 2L ⁇ K size bitmap to indicate its selected coefficients.
  • the base station can know the feedback subset and the optimal first basis vector according to the bitmap information fed back by the terminal. With the configured quantization accuracy, the total cost of feeding back the second coefficient can be calculated. Note that the number of weighting coefficients selected in the non-optimal first basis vector must be less than l 0 , otherwise the base station cannot implicitly know the position of the optimal first basis vector through the bitmap.
  • Sub mode C-2-2-2 terminal use Indicate the selected optimal first basis vector, reuse Indicates that selects the remaining coefficients Ml 0, the base station further configuration according to the quantization accuracy, a feedback overhead can be calculated second coefficients.
  • Sub-mode C-2-3 The terminal sorts the strength of 2L beams based on certain criteria, then the corresponding beams from strong to weak are selected l 0 , l 0 -1, ..., l 0 -M With +1 coefficient, the base station can configure different feedback accuracy according to the strength of its beam.
  • the terminal uses a 2L ⁇ K size bitmap to indicate its selected coefficients.
  • the base station can know the order of the feedback subset and the first basis vector according to the bitmap information fed back by the terminal. Further, according to the configured quantization accuracy, the feedback number can be calculated. The total cost of the second coefficient.
  • Method C-3 Considered from the perspective of the second basis vector; where a second basis vector corresponds to a delay path of the channel, if a certain strongest delay path has higher feedback accuracy, system performance can be improved. Specifically, to feed the second coefficient with the second base vector as the priority, one of the following sub-modes may be selected:
  • Sub-mode C-3-1 The terminal selects the optimal second basis vector based on certain criteria (for example, the sum of the power of 2L coefficients corresponding to a certain second basis vector is the largest), and the optimal second basis vector 2 weighting factors ((corresponding to )) In a column l 0 feedback coefficients.
  • the quantization accuracy of the weighting coefficient of this base vector can be appropriately increased, for example, the amplitude and phase quantization accuracy is (3,4)bits or (4,4)bits; among all the weighting coefficients of the remaining second base vectors, the amplitude is selected again (Ml 0 ) coefficients with the largest coefficients, the value of M can be configured by the base station or Where ⁇ M is a preset threshold; its amplitude and phase quantization accuracy is (2, 3) bits or (3, 3) bits.
  • the amplitude can be directly quantized or differentially quantized according to the amplitude of the optimal second basis vector.
  • the part of the second coefficient that the terminal indicates to its feedback may be implemented in one of the following sub-modes:
  • Sub-mode C-3-1-1 The terminal uses a bitmap of 2L ⁇ K size to indicate its selected coefficients.
  • the base station can know the feedback subset and the optimal second basis vector according to the bitmap information fed back by the terminal. With the configured quantization accuracy, the total cost of feeding back the second coefficient can be calculated. Note that the number of coefficients selected in the non-optimal second basis vector must be less than l 0 , otherwise the base station cannot implicitly know the position of the optimal second basis vector through the bitmap.
  • Sub mode C-3-1-2 terminal use Indicate the selected optimal second basis vector, reuse Indicates that selects the remaining coefficients Ml 0, the base station further configuration according to the quantization accuracy, a feedback overhead can be calculated second coefficients.
  • Sub-mode C-3-2 The terminal sorts the strengths of the K second basis vectors based on certain criteria, then the weighting coefficients of the corresponding second basis vectors from strong to weak are selected to be l 0 , l 0- 1,...,l 0 -M+1 coefficients, the base station can configure different amplitude and phase quantization accuracy according to the strength of its second base vector.
  • the terminal uses a 2L ⁇ K size bitmap to indicate its selected coefficients.
  • the base station can know the feedback subset and the order of the second base vector according to the bitmap information fed back by the terminal. Further, according to the configured quantization accuracy, the feedback number can be calculated. The total cost of the second coefficient.
  • Method C-4 Divide the first basis vector into two groups, and divide the 1st to Lth first basis vectors, that is, the 1st to Lth column vectors of W 1 , as the first group, divide the L+1 to the 2L first group of vectors, i.e., the first W L + 1 to 2L of 1 column vector, the second group, according to the priority of the packets inside, a second feedback coefficient; wherein, L is a first basis vector 1st to Precoding of CSI-RS ports, and L first base vectors are Precoding to N t CSI-RS ports. Select within the group according to certain rules Part of the subset, and improve feedback accuracy, will improve system performance.
  • At least one of the following sub-modes can be selected:
  • Sub-mode C-4-1 Based on certain criteria (such as the ordering of the magnitude of the second coefficient), the weighting coefficient of the first base vector contained in the first group (ie The first to the L) Coefficients,
  • the value of can be configured by the base station or Where ⁇ M is a preset threshold; the weighting coefficient of the first basis vector contained in the second group (ie In line L+1 to 2L) and select Coefficients.
  • the terminal uses a bitmap of 2L ⁇ K size or Indicate the coefficient of its choice.
  • Sub-mode C-4-2 The terminal selects an optimal first basis vector from each of the two groups of first basis vectors based on certain criteria.
  • the corresponding coefficients of the two optimal first basis vectors can feed back more elements and increase the quantization accuracy. Specifically, one of the following sub-methods can be adopted.
  • Sub-mode C-4-2-1 The terminal selects an optimal first basis vector from each of the two groups of first basis vectors based on certain criteria.
  • Optimal weighting coefficients of the first two groups in each of the feedback vector l 0 coefficients, quantization of these coefficients can be increased precision, such as amplitude and phase quantization (3,4) bits or (4,4) bits.
  • amplitude and phase quantization 3,4) bits or (4,4) bits.
  • ⁇ M is a preset threshold
  • its amplitude and phase quantization accuracy is (2, 3) bits or (3, 3) bits, where the amplitude can be directly quantized or differentially quantized according to the amplitudes of the two optimal first basis vectors.
  • the terminal uses a 2L ⁇ K-sized bitmap to indicate its selected coefficients.
  • the base station can know the feedback subset and the optimal first basis vector in each of the two groups according to the bitmap information fed back by the terminal, and further according to the configured quantization accuracy, The total cost of feeding back the second coefficient can be calculated. Note that the number of coefficients selected in the non-optimal first basis vector must be less than l 0 , otherwise the base station cannot implicitly know the position of the optimal first basis vector through the bitmap.
  • Sub-mode C-4-2-2 The terminal selects an optimal first basis vector from each of the two groups of first basis vectors based on certain criteria.
  • Optimal weighting coefficients of the first two groups in each of the feedback vector l 0 coefficients, quantization of these coefficients can be increased precision, such as amplitude and phase quantization (3,4) bits or (4,4) bits.
  • the two groups respectively select l 1 (l 1 ⁇ l 0 ) coefficients (ie Select 1 coefficient for each line of ), the amplitude and phase quantization accuracy is (2, 3) bits or (3, 3) bits, where the amplitude can be directly quantized or differentially quantized according to the amplitude of the two optimal first basis vectors .
  • the terminal uses a 2L ⁇ K-sized bitmap to indicate its selected coefficients.
  • the base station can know the feedback subset and the optimal first basis vectors in the two groups according to the bitmap information fed back by the terminal, and further according to the configured quantization accuracy, The total cost of feeding back the second coefficient can be calculated.
  • Sub-mode C-4-3 Based on certain criteria, the terminal selects an optimal second basis vector from each of the two groups of first basis vectors.
  • the corresponding coefficients of these two second basis vectors can feed back more elements and increase the quantization accuracy.
  • one of the following sub-methods can be adopted:
  • Sub-mode C-4-3-1 Based on certain criteria, the terminal selects an optimal second basis vector from each of the two groups of first basis vectors. Choose l 0 coefficients from the weighting coefficients of the optimal second basis vector in the first group, and choose l 0 coefficients from the weighting coefficients of the optimal second basis vector in the second group. These coefficients can be increased appropriately Quantization accuracy, for example, quantization of amplitude and phase to (3, 4) bits or (4, 4) bits.
  • the value of can be configured by the base station or Where ⁇ M is a preset threshold; the amplitude and phase quantization accuracy is (2, 3) bits or (3, 3) bits, where the amplitude can be directly quantized or differentially quantized according to the amplitudes of the two optimal second basis vectors.
  • the terminal uses a bitmap of 2L ⁇ K size to indicate its selected coefficients.
  • the base station can know the feedback subset and the optimal second basis vector of each of the two groups according to the bitmap information fed back by the terminal, and further according to the configured quantization accuracy, The total cost of feeding back the second coefficient can be calculated. Note that the number of coefficients selected in the non-optimal first basis vector must be less than l 0 , otherwise the base station cannot implicitly know the position of the optimal second basis vector through the bitmap.
  • Sub-mode C-4-3-2 Based on certain criteria, the terminal selects an optimal second basis vector from each of the two groups of first basis vectors. Choose l 0 coefficients from the weighting coefficients of the optimal second basis vector in the first group, and choose l 0 coefficients from the weighting coefficients of the optimal second basis vector in the second group. These coefficients can be appropriate Increase the quantization accuracy, such as quantizing the amplitude and phase to (3, 4) bits or (4, 4) bits.
  • the two groups respectively select l 1 (l 1 ⁇ l 0 ) coefficients from the weighting coefficients of the remaining second basis vectors (ie Select 1 coefficient for each column of ), and its amplitude and phase quantization accuracy is (2, 3) bits or (3, 3) bits, where the amplitude can be directly quantized or differentiated according to the amplitudes of the two optimal second basis vectors Quantify.
  • the terminal uses a 2L ⁇ K-sized bitmap to indicate its selected coefficients.
  • the base station can know the feedback subset and the optimal second basis vectors in the two groups according to the bitmap information fed back by the terminal, and further according to the configured quantization accuracy, The total cost of feeding back the second coefficient can be calculated.
  • Method C-5 The second coefficients will be grouped and feedback will be based on the priority of the grouping; the beam and corresponding delay of the received signal will be relatively concentrated, which is specifically reflected in The larger part of will be concentrated in the second coefficient A certain location. Therefore, you can Grouping, the terminal feedback the most suitable combination, increase the quantization accuracy, can improve system performance. Specifically, at least one of the following sub-modes can be selected:
  • Method C-5-1 Group according to the size of L and K, and divide them into Combinations, where L s and K s respectively represent the grouping granularity of the first basis vector and the second basis vector, and the grouping granularity can be selected according to the values of L and K and the feedback accuracy.
  • the terminal selects the l most suitable combination for feedback based on certain criteria. End use Bitmap or The bits indicate the selected combination, and the base station can calculate the total cost of feeding back the second coefficient according to the combination selected by the terminal and further according to the configured quantization accuracy.
  • Method C-5-2 Group according to the size of L and K, and divide them into Combinations, where L s and K s respectively represent the grouping granularity of the first basis vector and the second basis vector, and the grouping granularity can be selected according to the values of L and K and the feedback accuracy. Based on certain criteria, the terminal selects the most suitable combination. The coefficients in this combination can appropriately increase the quantization accuracy. For example, the amplitude and phase are quantized to (3, 4) bits or (4, 4) bits. The terminal then selects (M-1) combinations. The coefficients in these combinations can have amplitude and phase quantization accuracy of (2, 3) bits or (3, 3) bits. End use Indicate the optimal combination of its choice, in addition need Instructing the other combinations selected, and further according to the configured quantization accuracy, the total cost of feeding back the second coefficient can be calculated.
  • the first base vector corresponds to a beam in space. Because the beam has directivity, some beams in the signal of one terminal are too strong, which will cause interference to other users. Therefore, the base station usually limits certain beams, such as limiting the broadband amplitude power of certain beams. In addition, according to the reciprocity of the Time Division Duplex (TDD) system, the base station can also limit unnecessary second basis vectors to reduce multipath interference of the signal. Specifically, the following methods can be adopted:
  • Method D-1 Group all optional first basis vectors, that is, the base station selects a group, and if the beam fed back by the terminal is located in the group configured by the base station, the power limitation of the subband amplitude is performed, that is, Some rows (corresponding to some beams) do power limitation.
  • the i-th first basis vector fed back by the terminal happens to be the first basis vector that the base station needs to do power limiting, you can choose one of the following two sub-modes:
  • Sub-mode D-1-1 The amplitude of the i-th and i+L lines of the quantization cannot exceed the first amplitude feedback penalty factor configured by the base station.
  • the first amplitude feedback penalty factor that can be configured by the base station is or
  • Sub mode D-1-2 The amplitude of the i-th and i+L rows of the multiplied by a certain first amplitude feedback penalty factor and then quantified.
  • the first amplitude feedback penalty factor is desirable or
  • Method D-2 In the TDD system, based on the principle of reciprocity, the uplink channel estimated by the base station through the uplink reference signal can reciprocate to the downlink channel. Therefore, the base station can roughly estimate the delay of the downlink channel.
  • a second basis vector essentially reflects a delay path, and the base station can avoid interference caused by some unnecessary delay components by limiting the second basis vector selected by the terminal.
  • Sub-mode D-2-1 Group all the optional second base vectors.
  • the base station selects a group through high-level parameter configuration. If a second base vector selected by the terminal overlaps with the group configured by the base station, you can use the following two Seed mode for power limitation:
  • Sub-mode D-2-1-1 All weighting coefficients corresponding to the second base vector cannot exceed the second amplitude feedback penalty factor configured by the base station after amplitude quantization.
  • the second amplitude feedback penalty factor configured by the base station is or
  • Sub-mode D-2-1-2 All weighting coefficients of the second base vector, the amplitude as a whole is multiplied by a certain second amplitude feedback penalty factor and then quantified, the second amplitude feedback penalty factor may be desirable or
  • the terminal estimates the downlink channel by measuring the CSI-RS, and uses other reference signals to estimate the interference. According to the channel conditions, the terminal will feedback the RI (Rank indicator), CQI (Channel quality indicator) and channel quality indicator (CQI) according to the instructions of the base station. PMI (Precoding matrix indicator). Since the base station cannot predict the RI fed back by the terminal or the base station triggers multiple CSI reports at the same time, it may cause insufficient resources for uplink CSI transmission. Therefore, it is necessary to define which part of the resource is given priority when the feedback resource is insufficient. According to the improved way of implementing the second basis vector and the feedback way of the second coefficient and various ways indicated by the feedback subset, one of the following ways can be selected:
  • Manner E-1 The selected number K of second basis vectors is fed back by the terminal. If the CSI resources fed back are insufficient, the terminal can dynamically reduce the number of second basis vectors fed back to reduce feedback overhead.
  • Method E-2 The number K of the selected second base vector is configured by the base station. If the feedback CSI resource is insufficient, specifically one of the following sub-modes can be selected:
  • Sub-mode E-2-1 feedback CSI according to mode C-1, and if the feedback resources are sufficient, upload all the parameters configured by the base station. If the feedback resources are insufficient, then the terminal can dynamically inform the selected coefficients through the bitmap based on certain criteria until the maximum feedback resources are reached.
  • Sub-mode E-2-2 feedback CSI according to mode C-2. If the feedback resources are sufficient, upload all the parameters configured by the base station. Specifically, when CSI is fed back according to sub-mode C-2-1 and sub-mode C-2-3, if the feedback resources are insufficient, the terminal preferentially feeds back the coefficient corresponding to the optimal first basis vector, and then the terminal may dynamically pass based on a certain criterion The bitmap informs the selected coefficients until the maximum feedback resource is reached; the CSI is fed back according to sub-mode C-2-2, and the feedback subset is indicated according to sub-mode C-2-2-1.
  • the terminal If the feedback resources are insufficient, the terminal gives priority to the most feedback Optimizing the weighting coefficient corresponding to the first basis vector, and then the terminal can dynamically notify the selected coefficient through the bitmap based on certain criteria until the maximum feedback resource is reached; feedback CSI according to sub-mode C-2-2, and according to sub-mode C -2-2-2
  • the feedback subset indicates that if the feedback resources are insufficient, the terminal only feeds back the weighting coefficient of the optimal first basis vector.
  • Sub-mode E-2-3 feedback CSI according to mode C-3. If the feedback resources are sufficient, upload all the parameters configured by the base station. Specifically, the CSI is fed back according to sub-mode C-3-1, and according to the feedback subset indication of sub-mode C-3-1-1, if the feedback resources are insufficient, the terminal preferentially feeds back the weighting coefficient of the optimal second basis vector, and then the terminal Based on certain criteria, the selected coefficients can be dynamically notified through the bitmap until the maximum feedback resource is reached; the CSI is fed back according to sub-mode C-3-1, and the feedback subset is indicated according to sub-mode C-3-1-2, if If the feedback resources are insufficient, the terminal only feeds back the weighting coefficient of the optimal second basis vector; when CSI is fed back according to sub-mode C-3-2, if the feedback resources are insufficient, the terminal gives priority to the weighting coefficient of the optimal second basis vector, and then the terminal can Based on certain criteria, the selected coefficients are dynamically notified through the bitmap until the maximum feedback resource is
  • Sub-mode E-2-4 feedback CSI according to mode C-4. If the feedback resources are sufficient, upload all the parameters configured by the base station.
  • the CSI is fed back according to sub-mode C-4-1 and the selected coefficient is indicated according to the bitmap mode. If the feedback resources are insufficient, the selected coefficient is dynamically notified through the bitmap until the maximum feedback resource is reached; according to sub-mode C-4-1 4-2-1 and sub-mode C-4-2-2 feedback CSI, if the feedback resources are insufficient, the terminal gives priority to the weighting coefficient of the optimal first basis vector in the first group, and secondly the optimal first on the second group Based on the weighting coefficient of the base vector, the terminal can dynamically notify the selected coefficient through the bitmap based on certain criteria until the maximum feedback resource is reached; feedback according to sub-mode C-4-3-1 and sub-mode C-4-3-2 CSI, if the feedback resources are insufficient, the terminal gives priority to the weighting coefficient of the optimal second basis vector in the first group, and secondly feedbacks the weighting coefficient of the optimal second basis vector in the second group, and then the terminal can dynamically based on certain criteria
  • the bitmap is used to notify the selected coefficient until the maximum feedback resource is reached
  • Sub-mode E-2-5 CSI is fed back according to mode C-5-1, and the selected packet is fed back using the bitmap. If the feedback resources are insufficient, the terminal can dynamically notify the base station of the selected packet based on the bitmap based on a certain criterion until Until the maximum feedback resource is reached; the CSI is fed back according to method C-5-2, and if the feedback resource is sufficient, all the parameters configured by the base station are uploaded. If the feedback resources are insufficient, the terminal gives priority to the weighting coefficient corresponding to the optimal group.
  • FIG. 4 is a schematic flowchart of the channel state information CSI feedback method provided in the second embodiment of the present invention. As shown in FIG. 4, the method includes:
  • Step 401 The base station configures the frequency domain range that needs to feedback CSI and sends it to the terminal.
  • the frequency domain range of the configured CSI includes: a precoding subband requiring CSI feedback or a precoding subband requiring comb feedback CSI and sparseness, or a precoding subband requiring no feedback CSI.
  • Step 402 The base station configures the feedback mode and feedback subset indication of the second coefficient, and sends it to the terminal.
  • the feedback method of the second coefficient includes one of the following: Method 1: Feedback the second coefficient according to the magnitude of the second coefficient as the priority; Method 2: Feedback the second coefficient using the first base vector as the priority ; Method 3: Use the second basis vector as the priority to feed back the second coefficient; Method 4: Divide the first basis vector into two groups, and divide the 1st to Lth first basis vectors, that is, the W 1st 1 to L-th column vector, as the first group, the first L + 1 to 2L th first group of vectors, i.e., the first W L + 1 to 2L of 1 column vector, as the second set, in accordance with the internal packet The second coefficient of the priority is fed back; Method 5: Group the second coefficient and feed back the second coefficient according to the priority of the group.
  • the feedback subset indication is used to instruct the terminal to feed back the subset of the second coefficient.
  • step 403 the terminal feeds back the precoding matrix information according to the frequency domain range of the CSI and the feedback mode of the second coefficient and the feedback subset indication required by the base station configuration.
  • the precoding matrix information includes first base vector information, second base vector information, amplitude and phase information of the second coefficient.
  • the precoding vector in the feedback precoding subband is the linear combination of the first base vector, and the weighting coefficient used for the linear combination of the first base vector is the first coefficient; in the frequency domain units included in all CSI feedback bands, A vector composed of the first coefficients corresponding to the same first base vector is a linear combination of second base vectors, and the second coefficient is a weighting coefficient used for linear combination of the second base vectors.
  • the CSI feedback frequency band refers to the frequency domain range configured by the base station that needs to feedback CSI.
  • the precoding subband is a frequency domain unit
  • the number of RBs included is determined by the number of RBs configured by the base station or included in the CSI feedback band and the number of second base vectors used, that is, the number of included RBs is RBNum is the number of RBs included in the CSI feedback frequency band, and K is the number of second basis vectors used in the CSI feedback frequency band.
  • the second coefficient is expressed as The dimension of is 2L ⁇ K, 2L represents the 2L first basis vectors contained in the matrix W 1 , and K represents the K second basis vectors contained in the matrix W 3 .
  • the second base vector can be partially modified according to actual needs, including the following three methods:
  • the second base vector can be adjusted in one of the following sub-modes: Sub-mode 1: Configure a larger oversampling factor and increase the number of selectable DFT matrices; Sub-method 2: The second basis vector is intercepted by a comb.
  • the second base vector part element is truncated.
  • the second base vector can be configured to perform phase rotation, with the middlemost precoding subband configured as a reference.
  • the terminal may feed back a partial subset, and indicate the selected subset by feedback. According to the selected subset and the feedback indication, the second coefficient is fed back according to one of the following methods agreed with the base station:
  • Method one includes: sorting by the magnitude of the second coefficient, selecting partial coefficients from large to small, until the ratio of the power of the partial coefficients to the total power of the second coefficient is not less than ⁇ , and the terminal feeds back the partial Coefficient, where ⁇ is a preset threshold; use a bitmap of 2L ⁇ K size to indicate its selected coefficient.
  • the second way includes: feeding back the second coefficient with the first base vector as the priority, specifically one of the following sub-ways can be used:
  • Mode A l 0 coefficients selected in weighting coefficients K optimized first base vector, the quantization accuracy increases. Among the remaining K weighting coefficients of the first base vectors, select l 1 (l 1 ⁇ l 0 ) coefficients to appropriately reduce the quantization accuracy.
  • the terminal uses a bitmap of 2L ⁇ K size to indicate its selected coefficient.
  • Mode B Select l 0 K coefficients in the optimal weighting coefficients of the first base vector, the quantization accuracy increases. All the remaining weighting coefficients of the first base vector, and then select a total feedback coefficients -l 0, the value of M may be configured by a base station or a Where ⁇ M is a preset threshold, and the quantization precision is appropriately reduced.
  • the terminal may use one of the following sub-modes to feed back its selected subset.
  • Sub-mode B-1 The terminal uses a 2L ⁇ K bitmap to indicate its selected coefficient.
  • Sub-mode B-2 terminal use Indicate the selected optimal first basis vector, reuse Indicates that selects the remaining coefficients Ml 0.
  • Method C The terminal sorts the 2L first basis vectors, and the weighting coefficients of the first basis vectors from strong to weak are selected l 0 , l 0 -1, ..., l 0 -M+1 coefficients for feedback .
  • the terminal uses a bitmap of 2L ⁇ K size to indicate its selected coefficient.
  • Mode three includes: feeding back the second coefficient with the second base vector as the priority, specifically one of the following sub-modes can be used:
  • Method D Select 10 coefficients among the 2L weighting coefficients of the optimal second basis vector to increase the quantization accuracy. All remaining weighting coefficients in the second group vector, select a total of feedback coefficients Ml 0, the value of M is configured by a base station or a Where ⁇ M is a preset threshold, and the quantization precision is appropriately reduced.
  • the terminal may use one of the following sub-modes to feed back its selected subset.
  • Sub-mode D-1 The terminal uses a bitmap of 2L ⁇ K size to indicate its selected coefficient.
  • Sub-mode D-2 terminal use Indicate the selected optimal second basis vector, reuse Indicates that selects the remaining coefficients Ml 0.
  • Method E The terminal sorts the K second basis vectors, and the weighting coefficients of the second basis vectors from strong to weak are selected to be l 0 , l 0 -1, ..., l 0 -M+1 coefficients, respectively.
  • the terminal uses a bitmap of 2L ⁇ K size to indicate its selected coefficient.
  • Method 4 includes: dividing the first basis vector into two groups, and feeding back the second coefficient according to the priority. Specifically, one of the following sub-modes can be used:
  • Method F In all weighting coefficients of the first basis vector in the two groups, each feedback Coefficients, The value of is configured by the base station or is Where ⁇ M is a preset threshold; the terminal uses the bitmap or Indicates the selected coefficient.
  • Method G Within the two groups, the feedback is based on the priority of the first basis vector. Specifically, one of the following sub-methods can be used:
  • Sub-mode G-1 select an optimal first basis vector in each of the two groups.
  • a first weighting coefficient optimal two groups each selected vector l 0 coefficients, the quantization accuracy increases. Reselect each weighting coefficient of the remaining first basis vectors in the two groups Coefficients,
  • the value of is configured by the base station or is Where ⁇ M is a preset threshold, and the quantization precision is appropriately reduced.
  • the terminal uses a bitmap of 2L ⁇ K size to indicate its selected coefficient.
  • Sub-mode G-2 select an optimal first basis vector in each of the two groups.
  • a first weighting coefficient optimal two groups each selected vector l 0 coefficients, the quantization accuracy increases.
  • the terminal uses a bitmap of 2L ⁇ K size to indicate its selected coefficient.
  • Method H Within two groups, the feedback is based on the priority of the second base vector, and one of the following sub-modes can be used specifically:
  • Sub-mode H-1 select an optimal second basis vector in each of the two groups. Select 10 coefficients among the weighting coefficients of the optimal second basis vector in the first group, and select 10 coefficients among the weighting coefficients of the optimal second basis vector in the second group to increase the quantization accuracy. Reselect each weighting coefficient of the remaining second basis vectors in the two groups Coefficients, The value of is configured by the base station or is Where ⁇ M is a preset threshold, and the quantization precision is appropriately reduced.
  • the terminal uses a bitmap of 2L ⁇ K size to indicate its selected coefficient.
  • Sub-mode H-2 select an optimal second basis vector in each of the two groups. Select l 0 coefficients from the weighting coefficients of the optimal second basis vector in the first group, and select l 0 coefficients from the weighting coefficients of the optimal second basis vector in the second group to increase the quantization accuracy . Among the weighting coefficients of the remaining second base vectors in the two groups, select l 1 (l 1 ⁇ l 0 ) coefficients to appropriately reduce the quantization accuracy.
  • the terminal uses a bitmap of 2L ⁇ K size to indicate its selected coefficient.
  • Method 5 includes: grouping the second coefficients and feeding back according to the priority of the grouping. Specifically, one of the following sub-modes can be used:
  • Method I divide the second coefficient into GroupNumber combinations, the terminal selects the l most suitable combination for feedback, use Indicate its selected combination.
  • Method J The second coefficient is divided into GroupNumber combinations, and the terminal selects the most suitable combination to increase the quantization accuracy. Among the remaining combinations, 1-1 combinations are selected to appropriately reduce the quantization accuracy. The terminal uses log 2 GroupNumberbits to indicate the most suitable combination selected, use Indicate the remaining combinations.
  • the base station may also limit the selection of the first basis vector and the second basis vector, thereby avoiding interference.
  • the following methods can be used:
  • Method 1 The base station groups all the optional first basis vectors, and notifies the terminal of the first basis vector that needs to be restricted by signaling to limit the feedback power. Specifically, one of the following sub-methods may be adopted: 1. The base station configures power limitation on certain first base vectors, and the weighted coefficient amplitude of the power-limited first base vector cannot exceed the first amplitude feedback configured by the base station after quantization Penalty factor; 2. The base station configures power limitation on some first base vectors, and the amplitude of the weighting coefficient of the first base vector for power limitation is multiplied by the first amplitude feedback penalty factor configured by the base station before quantization.
  • Method 2 The base station groups all the optional second base vectors, and notifies the terminal of the second base vector that needs to be restricted by signaling to limit the feedback power. Specifically, one of the following sub-methods may be adopted: 1. The base station configures power limitation on some second base vectors, and the weighted coefficient amplitude of the power-limited second base vector cannot exceed the second amplitude feedback configured by the base station after quantization Penalty factor; 2. The base station configures power limitation on some second base vectors, and the amplitude of the weighting coefficient of the second base vector used for power limitation is multiplied by the second amplitude feedback penalty factor configured by the base station before quantization.
  • the terminal may discard part of the feedback content according to the priority. Specifically, one of the following methods can be used:
  • Method 2 When the selected number of second base vectors K is configured by the base station, if the feedback resources are sufficient, upload all the parameters configured by the base station, otherwise choose one of the following sub-modes: 1. When feedback the precoding matrix according to the method 1 Information, if the feedback resources are insufficient, based on a certain criterion, dynamically notify the selected coefficients through the bitmap until the maximum feedback resource is reached; 2.
  • the terminal When the precoding matrix information is fed back in the manner A or C included in the second way If the feedback resources are insufficient, the terminal gives priority to the weighted coefficients of the optimal first basis vector, and then the terminal can dynamically notify the selected coefficients through the bitmap based on certain criteria until the maximum feedback resource is reached; when the second method includes Way B feeds back precoding matrix information and uses sub-mode B-1 feedback subset indication, if the feedback resources are insufficient, the terminal preferentially feeds back the weighting coefficient of the optimal first basis vector, and then the terminal can dynamically pass the bits based on certain criteria
  • the graph notifies the selected coefficients until the maximum feedback resource is reached; when the precoding matrix information is fed back in the manner B included in the second way and indicated by the sub-mode B-2 feedback subset, if the feedback resources are insufficient, the terminal only feeds back the most Optimizing the weighting coefficient of the first base vector; 3.
  • the terminal When the precoding matrix information is fed back in the manner D included in the third way and indicated by the sub-mode D-1 feedback subset, if the feedback resources are insufficient, the terminal will give the best feedback first.
  • the weighted coefficients of the two basis vectors, and then the terminal can dynamically notify the selected coefficients through the bitmap based on certain criteria until the maximum feedback resource is reached; when the precoding matrix information is fed back in the manner included in the third way and the sub-way is used D-2
  • the feedback subset is indicated, if the feedback resources are insufficient, the terminal only feeds back the weighting coefficient of the optimal second basis vector; when the precoding matrix information is fed back in the manner E included in the third way, if the feedback resources are insufficient, the terminal Priority feedback of the weighting coefficient of the optimal second basis vector, and then the terminal can dynamically notify the selected coefficient through the bitmap based on a certain criterion until the maximum feedback resource is reached;
  • the feedback resources if the feedback resources are insufficient, dynamically inform the selected coefficient
  • the base station When feeding back the precoding matrix information according to the method I included in the fifth method and using the bitmap to feedback the selected group, if the feedback resources are insufficient, the base station is dynamically notified according to a certain criterion according to the bitmap of its selection Grouping until the maximum feedback resource is reached; when the precoding matrix information is fed back in the manner J included in the fifth way, if the feedback resources are insufficient, the weighting coefficient corresponding to the optimal group is preferentially fed back.
  • the precoding matrix information fed back by the terminal includes first base vector information, second base vector information, amplitude and phase information of the second coefficient; wherein, within one precoding subband
  • the weighting coefficient of the first base vector is the first coefficient, and the correlation of the precoding vectors of different precoding subbands is used.
  • the matrix composed of the first coefficients of all precoding subbands can be compressed in the frequency domain using the second base vector ,
  • the weighting coefficient of the second basis vector is the second coefficient. In this way, not only can the CSI feedback overhead be reduced, but also a high CSI feedback performance is guaranteed.
  • FIG. 5 is a schematic structural diagram of a terminal according to Embodiment 3 of the present invention.
  • the terminal includes: a feedback unit for feeding back precoding matrix information.
  • the precoding matrix information includes first base vector information and Two base vector information, the amplitude and phase information of the second coefficient; wherein, the precoding vector in the feedback precoding subband is the linear combination of the first base vector, and the weighting coefficient used for the linear combination of the first base vector is the first coefficient ;
  • the vector composed of the first coefficients corresponding to the same first base vector is a linear combination of second base vectors, and the second coefficient is the The weighting coefficient used for the linear combination of the second basis vector.
  • said second base vector represents the DFT matrix or DFT over the K base vectors of the DFT matrix selected sampling, the oversampling factor value over-sampling O f DFT matrix as one of the following: 1,2, 4, 8.
  • the terminal Before the feedback unit feeds back the precoding matrix information, the terminal further includes: a processing unit configured to obtain precoding matrix information according to the channel estimation result and the configuration information sent by the base station; the configuration information sent by the base station includes at least the following One: the frequency domain range of the CSI that needs to be fed back by the base station, the feedback mode and feedback subset indication of the configured second coefficient, the configured first amplitude feedback penalty factor and the first basis vector indicating that power limitation is required, the configured The second amplitude feedback penalty factor and the second basis vector indicating that power limitation is required.
  • the specific process is as follows: the specific process is as follows:
  • the terminal further includes: a receiving unit, configured to receive a frequency domain range requiring CSI feedback sent by the base station before the terminal feeds back precoding matrix information, and the frequency domain range of the CSI includes: a precoding subband requiring CSI feedback Either a comb-like feedback CSI precoding subband and sparseness are required, or a CSI precoding subband is not required; the processing unit is used when the base station comb configures a subband requiring feedback precoding, by one of the following ways Modification of the second base vector: Method 1: The base station configures a larger oversampling factor to increase the number of selectable DFT matrices; Method 2: The terminal intercepts the DFT base vector in a comb shape according to the precoding subbands of the precoding information that needs to be fed back; Or, when the configuration part of the base station does not need to feed back the precoded subband, the corresponding position element of the precoding subband that does not require feedback of the precoding information in the DFT base vector
  • the CSI feedback frequency band refers to the frequency domain range configured by the base station that needs to feedback CSI.
  • the precoding subband is a frequency domain unit
  • the number of RBs included is determined by the number of RBs configured by the base station or included in the CSI feedback band and the number of second base vectors used, that is, the number of included RBs is RBNum is the number of RBs included in the CSI feedback frequency band, and K is the number of second basis vectors used in the CSI feedback frequency band.
  • the second coefficient is expressed as The dimension of is 2L ⁇ K, 2L represents the 2L first basis vectors contained in the matrix W 1 , and K represents the K second basis vectors contained in the matrix W 3 .
  • the receiving unit is also used to receive the feedback mode and the feedback subset indication of the second coefficient sent by the base station before the terminal feeds back the precoding matrix information;
  • the processing unit is further configured to select the second coefficient to be fed back in one of the following ways: method 1: feedback the second coefficient according to the magnitude of the second coefficient as a priority; method 2: using the first basis vector The second coefficient is fed back for priority.
  • Method 5 Group the second coefficient, and feedback the second coefficient according to the priority of the group.
  • the first method includes: sorting by the magnitude of the second coefficient, selecting partial coefficients from large to small, until the ratio of the power of the partial coefficients to the total power of the second coefficient is not less than ⁇ , the terminal feedbacks Partial coefficients, where ⁇ is a preset threshold; a bitmap of 2L ⁇ K is used to indicate the selected coefficient.
  • the second approach comprises: mode A: Select l 0 K coefficients in the optimal weighting coefficients in a first group of vectors, each of l 1 select one of the remaining K weighting coefficients in each of a first base vector Coefficient, l 1 ⁇ l 0 , use a bitmap of 2L ⁇ K size to indicate the selected coefficient; or, method B: select l 0 coefficients out of the K weighting coefficients of the optimal first basis vector, and the remaining All the weighting coefficients of the first base vector, select a total of feedback coefficients Ml 0, the value of M is configured by a base station or a Where ⁇ M is the preset threshold; use one of the following sub-methods to feed back its selected subset: sub-method B-1: use a bitmap of 2L ⁇ K size to indicate its selected coefficient; sub-method B-2: use Indicate the selected optimal first basis vector, reuse Instruct the remaining Ml 0 coefficients it chooses; or, method C: sort the 2L first basis vectors, and select l 0
  • said third approach comprises: D mode, select 2L l 0 coefficients in the optimal weighting coefficients in a second base vector, all remaining weighting coefficients in the second group of vectors, a total of select Ml 0 Coefficient, the value of M is configured by the base station or is Where ⁇ M is the preset threshold; use one of the following sub-modes to feed back the selected subset: sub-mode D-1: use a 2L ⁇ K-sized bitmap to indicate the coefficients selected; sub-mode D-2: use Indicate the selected optimal second basis vector, reuse bits indicate the remaining Ml 0 coefficients it selects; or, method E: sort the K second basis vectors, and select the weighting coefficients of the strongest and weakest second basis vectors l 0 , l 0 -1, respectively... ., l 0 -M+1 coefficients, using a bitmap of 2L ⁇ K size to indicate the selected coefficients.
  • the method 4 includes: method F: method F: in all weighting coefficients of the first basis vector in the two groups, each feedback Coefficients, The value of is configured by the base station or is Where ⁇ M is the preset threshold; through the bitmap or Indicate the selected coefficient; or, mode G: within two groups, feedback according to the priority of the first base vector and use a bitmap of 2L ⁇ K size to indicate the selected coefficient, including: Sub-mode G-1: Choose an optimal first basis vector for each of the two groups, and select l 0 coefficients for the weighting coefficients of the two optimal first basis vectors.
  • ⁇ M is a preset threshold
  • sub-mode G-2 select an optimal first basis vector from each of the two groups, and select l 0 coefficients from the weight coefficients of the two optimal first basis vectors, In the weighting coefficients of the remaining first basis vectors in the two groups, select l 1 coefficients, l 1 ⁇ l 0 ; or, mode H: in the two groups, feedback according to the priority of the second basis vector And use the 2L ⁇ K size bitmap to indicate its selected coefficients, including: Sub-mode H-1: Select an optimal second basis vector in each of the two groups, and the optimal second basis vector in the first group Choose l 0 coefficients among the weighting coefficients of the second group, select l 0 coefficients among the weighting coefficients of the optimal second basis vectors in the second group, and each of the weighting coefficients of the remaining second basis vectors in the two groups
  • the method 5 includes: method I: divide the second coefficient into GroupNumber combinations, the terminal selects the l most suitable combination for feedback, use Indicate the selected combination; or, method J: divide the second coefficient into a combination, select the most suitable combination, and then select l-1 combinations among the remaining combinations, and use log 2 GroupNumber bits to indicate the most suitable combination selected ,use Indicate the remaining combinations.
  • the receiving unit is also used to receive the first signaling sent by the base station before the terminal feeds back the precoding matrix information; the first signaling includes the first amplitude feedback penalty factor configured by the base station and indicates that power limitation is required
  • the first basis vector is used to perform power limitation on the weighting coefficient of the first basis vector that needs power limitation; the processing unit is also used to limit the feedback of the first basis vector that needs power limitation After the quantization of the weighting coefficient amplitude cannot exceed the first amplitude feedback penalty factor configured by the base station, or multiply the amplitude of the weighting coefficient of the first base vector that requires power limitation by the first amplitude feedback penalty factor configured by the base station, and then perform quantization ;
  • the receiving unit is also used to receive the second signaling sent by the base station before the terminal feeds back the precoding matrix information; the second signaling includes the second amplitude feedback penalty factor configured by the base station and indicating the need
  • the second base vector for power limitation is used for power limiting the weighting coefficient of the second base vector for power limitation; the processing
  • the processing unit is also used to dynamically reduce the number of feedback second basis vectors when the feedback resources are insufficient and the number of selected second basis vectors K is configured by the terminal; or, when the number of selected second basis vectors When K is configured by the base station, if the feedback resources are sufficient, upload all the parameters configured by the base station, otherwise choose one of the following methods:
  • dynamic The bitmap is used to notify the selected coefficients until the maximum feedback resource is reached; or, when the precoding matrix information is fed back in the manner A or C included in the second way, if the feedback resources are insufficient, the terminal gives priority to the optimal first feedback
  • the weighting coefficient of the base vector and then the terminal dynamically informs the selected coefficient through the bitmap based on a certain criterion until the maximum feedback resource is reached; when the mode 2 includes the mode B, the precoding matrix information is fed back and the sub-mode B-1 is used When the feedback subset indicates that
  • the weighting coefficient of the optimal first basis vector in the group followed by feedback of the weighting coefficient of the optimal first basis vector in the second group, and then based on a certain criterion, dynamically notify the selected coefficient through the bitmap until the maximum feedback resource is reached ;
  • the optimal second base vector in the first group is preferentially fed back Weighting factor, second feedback
  • the weighting coefficient of the optimal second basis vector within the group and then based on a certain criterion, dynamically notify the selected coefficients through the bitmap until the maximum feedback resource is reached; or, when feedback feedback precoding is performed in the manner I included in the fifth way
  • the base station is dynamically notified according to a certain criterion according to the bitmap to the selected group until the maximum feedback resource is reached;
  • the base station includes: a receiving unit, configured to receive precoding matrix information fed back by a terminal, and the precoding matrix information includes a first basis vector Information, second base vector information, amplitude and phase information of the second coefficient; where the precoding vector in the feedback precoding subband is the linear combination of the first base vector, and the weighting coefficient used for the linear combination of the first base vector is The first coefficient; in the frequency domain unit included in all CSI feedback frequency bands, the vector composed of the first coefficients corresponding to the same first base vector is a linear combination of the second base vectors, and the second coefficient Weighting coefficients used for linear combination of the second basis vector.
  • the base station further includes: a configuration unit for configuring the frequency domain range requiring CSI feedback before the base station receives the precoding matrix information fed back by the terminal; a transmission unit for transmitting the configured frequency domain range requiring CSI feedback to the The terminal, so that the terminal determines a precoding subband that needs to feed back precoding information according to the frequency domain range of the CSI.
  • the frequency domain range of the configured CSI includes: a precoding subband requiring CSI feedback or a precoding subband requiring comb feedback CSI and sparseness, or a precoding subband requiring no feedback CSI.
  • the configuration unit is also used to configure the feedback mode and the feedback subset indication of the second coefficient before the base station receives the precoding matrix information fed back by the terminal;
  • the sending unit is also used to configure the feedback mode of the configured second coefficient And the feedback subset indication are sent to the terminal, so that the terminal feeds back the second coefficient according to the feedback mode of the second coefficient and the feedback subset indication;
  • the feedback mode of the second coefficient includes one of the following: Mode 1: According to the second coefficient Amplitude magnitude of the second feedback coefficient;
  • Method 2 The first base vector is used as the priority to feedback the second coefficient;
  • Method 3 The second base vector is used as the priority to feedback the second coefficient;
  • Method four The The first basis vectors are divided into two groups, and the 1st to Lth first basis vectors, that is, the 1st to Lth column vectors of W 1 , are used as the first group, and the L+1th to 2Lth first basis vector, i.e.
  • the second group according to the priority of the packet inside a second feedback coefficient; five ways: second coefficient group, the first priority in accordance with the feedback packet Two coefficients; the feedback subset indication is used to instruct the terminal to feed back the subset of the second coefficient.
  • the second coefficient is expressed as The dimension of is 2L ⁇ K, 2L represents the 2L first basis vectors contained in the matrix W 1 , and K represents the K second basis vectors contained in the matrix W 3 .
  • the first method includes: sorting by the magnitude of the second coefficient, selecting the partial coefficient power from large to small, until the ratio of the power of the partial coefficient to the total power of the second coefficient is not less than ⁇ , the terminal feeds back The partial coefficients, where ⁇ is a preset threshold; a bitmap of 2L ⁇ K size is used to indicate its selected coefficient.
  • the second approach comprises: mode A: l 0 selected coefficients K coefficients in the optimal vector corresponding to the first group, each of l 1 selected one in each of the remaining K coefficients vectors corresponding to the first group coefficient, l 1 ⁇ l 0, using 2L ⁇ K bitmap size indicating which coefficients selected; or, mode B: l 0 selected coefficients K coefficients in the optimal vector corresponding to a first group, the remaining Among all the coefficients corresponding to the first base vector, a total of -l 0 coefficients are selected for feedback.
  • the value of M is configured by the base station or is Where ⁇ M is the preset threshold; use one of the following sub-methods to feed back its selected subset: sub-method B-1: use a bitmap of 2L ⁇ K size to indicate its selected coefficient; sub-method B-2: use Indicate the selected optimal first basis vector, reuse Instruct the remaining Ml 0 coefficients it chooses; or, method C: sort the 2L first base vectors, and select l 0 , l 0 -1, respectively, from the coefficients corresponding to the strong first base vectors.. .,l 0 -M+1 coefficients are used for feedback, and a bitmap of 2L ⁇ K size is used to indicate the selected coefficients.
  • said third approach comprises: mode D, l 0 coefficients selected in 2L optimum coefficients corresponding to a second base vector, the coefficients in all of the remaining base vector corresponding to the second, a total of select Ml 0 Coefficient, the value of M is configured by the base station or is Where ⁇ M is the preset threshold; use one of the following sub-modes to feed back the selected subset: sub-mode D-1: use a 2L ⁇ K-sized bitmap to indicate the coefficients selected; sub-mode D-2: use Indicate the selected optimal second basis vector, reuse bits indicate the remaining Ml 0 coefficients it chooses; or, method E: sort the K second basis vectors, and select the coefficients corresponding to the strongest to weakest second basis vectors l 0 , l 0 -1, .. ., l 0 -M+1 coefficients, using a bitmap of 2L ⁇ K size to indicate the selected coefficients.
  • the method 4 includes: method F: within two groups, each feedback Coefficients, The value of is configured by the base station or is Where ⁇ M is the preset threshold; through the bitmap or Indicate the selected coefficient; or, mode G: within two groups, feedback according to the priority of the first base vector and use a bitmap of 2L ⁇ K size to indicate the selected coefficient, including: Sub-mode G-1: Choose an optimal first basis vector for each of the two groups, and select 10 coefficients for the corresponding coefficients of the two optimal first basis vectors. The remaining first basis vectors in the two groups correspond to all the coefficients.
  • ⁇ M is a preset threshold
  • sub-embodiment G-2 in each of two groups of a first group select an optimum vector corresponding to the optimal coefficients of the two first base vector of each selected coefficients l 0, In each of the remaining first base vector corresponding coefficients in the two groups, select l 1 coefficients, l 1 ⁇ l 0 ; or, mode H: in the two groups, feedback according to the priority of the second base vector and Use a 2L ⁇ K bitmap to indicate the selected coefficients, including: Sub-mode H-1: Select an optimal second basis vector in each of the two groups, and the optimal second basis vector in the first group Select l 0 coefficients among the corresponding coefficients, select l 0 coefficients among the corresponding coefficients of the optimal second basis vector in the second group, and the remaining second basis vectors within the two groups correspond to all the coefficients.
  • select Coefficients The value of is configured by the base station or is Where ⁇ M is a preset threshold; or sub-mode H-2: select an optimal second basis vector from each of the two groups, and select l 0 from the corresponding coefficient of the optimal second basis vector within the first group Coefficients, select l 0 coefficients from the corresponding coefficients of the optimal second base vector in the second group, and then select l 1 coefficients from the remaining corresponding coefficients of the second base vector in the two groups, l 1 ⁇ l 0 .
  • the method 5 includes: method I: divide the second coefficient into GroupNumber combinations, the terminal selects the l most suitable combination for feedback, use Indicate its selected combination; Or, Method J: Divide the second coefficient into GroupNumber combinations, select 1 most suitable combination, and then select l-1 combinations among the remaining combinations, use log 2 GroupNumberbits to indicate the most suitable selection Combination, use Indicate the remaining combinations.
  • the configuration unit is also used to configure the first amplitude feedback penalty factor and the first basis vector indicating that power limitation is required; the sending unit is also used to send the precoding matrix information fed back by the terminal to the base station The terminal sends first signaling; wherein the first signaling includes a first amplitude feedback penalty factor configured by the base station and a first basis vector indicating that power limitation is required, and is used for the first basis for power limitation The power of the vector corresponding coefficient is limited; and/or, the configuration unit is also used to configure a second amplitude feedback penalty factor and a second base vector indicating that power limitation is required; the sending unit is also used to receive at the base station Before the precoding matrix information fed back by the terminal, second signaling is sent to the terminal; wherein, the second signaling includes a second amplitude feedback penalty factor configured by the base station and a second basis vector indicating that power limitation is required, which is used for The second base vector that needs to be power-limited corresponds to the power of the coefficient.
  • the first amplitude feedback penalty factor is or Or, the second amplitude feedback penalty factor is or
  • An embodiment of the present invention also provides a terminal, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the computer program is executed by the processor, the foregoing tasks are realized.
  • a method for CSI feedback of channel state information performed by a terminal.
  • An embodiment of the present invention also provides a base station, including a memory, a processor, and a computer program stored on the memory and executable on the processor.
  • the computer program is executed by the processor to implement any of the above tasks.
  • a method for CSI feedback of channel state information performed by a base station.
  • An embodiment of the present invention also provides a computer-readable storage medium having an information processing program stored on the computer-readable storage medium.
  • the information processing program is executed by a processor to implement any of the above channel state information CSI Feedback method.
  • computer storage medium includes both volatile and nonvolatile implemented in any method or technology for storing information such as computer readable instructions, data structures, program modules, or other data Sex, removable and non-removable media.
  • Computer storage media include but are not limited to Random Access Memory (RAM), Read-Only Memory (ROM), Erasable Programmable Read-Only Memory (Electrically Programmable Read-Only Memory, EEPROM) , Flash memory or other memory technology, compact disc read-only memory (Compact Disc Read-Only Memory, CD-ROM), digital versatile disc (Digital Video Disc, DVD) or other optical disc storage, magnetic box, magnetic tape, magnetic disk storage or other magnetic A storage device, or any other medium that can be used to store desired information and can be accessed by a computer.
  • the communication medium generally contains computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transmission mechanism, and may include any information delivery medium .

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

本发明实施例公开了一种信道状态信息CSI反馈方法、设备及存储介质，所述方法包括：终端反馈预编码矩阵信息，所述预编码矩阵信息包括第一基矢量信息、第二基矢量信息、第二系数的幅度和相位信息；其中，反馈预编码子带内的预编码矢量为一组第一基矢量的线性合并，一组第一基矢量线性合并所使用的加权系数为第一系数；在所有CSI反馈频带所包含的频域单位内，对应于同一组所述第一基矢量的第一系数所组成的矢量为一组第二基矢量的线性合并，所述第二系数为一组第二基矢量进行线性合并所使用的加权系数。

Description

信道状态信息CSI反馈方法、设备及存储介质
本申请要求在2018年12月28日提交中国专利局、申请号为201811625788.X的中国专利申请的优先权，该申请的全部内容通过引用结合在本申请中。
技术领域
本发明实施例涉及无线通信技术，例如涉及一种信道状态信息CSI反馈方法、设备及存储介质。
背景技术
在MIMO(Multiple-Input Multiple-Output，多输入多输出)无线通信系统中，通过对多根发送天线进行预编码或波束成型，可以达到提升传输效率和可靠性的目的。为了实现高性能的预编码或波束成型，预编码矩阵或波束成型矢量需要比较好的匹配信道，这就需要发射端能较好的获得信道状态信息(Channel State Information，CSI)。因此，CSI反馈是在MIMO系统中实现高性能预编码或波束成型的关键技术。然而，在进行CSI反馈的时候，对信道矩阵的量化反馈会带来比较大的反馈开销，特别是支持大带宽的CSI反馈时，反馈开销是限制性能提升的重要问题。
发明内容
有鉴于此，本发明实施例提供了一种信道状态信息CSI反馈方法，包括：终端反馈预编码矩阵信息，所述预编码矩阵信息包括第一基矢量信息、第二基矢量信息、第二系数的幅度和相位信息；
其中，反馈预编码子带内的预编码矢量为第一基矢量的线性合并，第一基矢量线性合并所使用的加权系数为第一系数；在所有CSI反馈频带所包含的频域单位内，对应于同一所述第一基矢量的所述第一系数所组成的矢量为第二基矢量的线性合并，所述第二系数为所述第二基矢量进行线性合并所使用的加权系数。
本发明实施例还提供了一种信道状态信息CSI反馈方法，包括：
基站接收终端反馈的预编码矩阵信息，所述预编码矩阵信息包括第一基矢量信息、第二基矢量信息、第二系数的幅度和相位信息；
其中，反馈预编码子带内的预编码矢量为第一基矢量的线性合并，第一基矢量线性合并所使用的加权系数为第一系数；在所有CSI反馈频带所包含的频 域单位内，对应于同一所述第一基矢量的所述第一系数所组成的矢量为第二基矢量的线性合并，所述第二系数为所述第二基矢量进行线性合并所使用的加权系数。
本发明实施例还提供了一种终端，包括：
反馈单元，用于反馈预编码矩阵信息，所述预编码矩阵信息包括第一基矢量信息、第二基矢量信息、第二系数的幅度和相位信息；
其中，反馈预编码子带内的预编码矢量为第一基矢量的线性合并，第一基矢量线性合并所使用的加权系数为第一系数；在所有CSI反馈频带所包含的频域单位内，对应于同一所述第一基矢量的所述第一系数所组成的矢量为第二基矢量的线性合并，所述第二系数为所述第二基矢量进行线性合并所使用的加权系数。
本发明实施例还提供了一种基站，包括：
接收单元，用于接收终端反馈的预编码矩阵信息，所述预编码矩阵信息包括第一基矢量信息、第二基矢量信息、第二系数的幅度和相位信息；
其中，反馈预编码子带内的预编码矢量为第一基矢量的线性合并，第一基矢量线性合并所使用的加权系数为第一系数；在所有CSI反馈频带所包含的频域单位内，对应于同一所述第一基矢量的所述第一系数所组成的矢量为第二基矢量的线性合并，所述第二系数为所述第二基矢量进行线性合并所使用的加权系数。
本发明实施例还提供了一种终端，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述计算机程序被所述处理器执行时实现上述终端执行的信道状态信息CSI反馈的方法。
本发明实施例还提供了一种基站，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述计算机程序被所述处理器执行时实现上述基站执行的信道状态信息CSI反馈的方法。
本发明实施例还提供了一种计算机可读存储介质，其特征在于，所述计算机可读存储介质上存储有信息处理程序，所述信息处理程序被处理器执行时实现上述任一项所述信道状态信息CSI反馈的方法。
与相关技术相比，本发明实施例提供了一种信道状态信息CSI反馈方法及设备，通过反馈第一基矢量信息、第二基矢量信息和第二系数的幅度和相位信息，既能够减小CSI反馈开销，又保证了较高的CSI反馈性能。
附图说明
图1为本发明实施方式一提供的信道状态信息CSI反馈方法的流程示意图；
图2为本发明实施方式二提供的信道状态信息CSI反馈方法的流程示意图；
图3为本发明实施例一提供的信道状态信息CSI反馈方法的流程示意图；
图4为本发明实施例二提供的信道状态信息CSI反馈方法的流程示意图；
图5为本发明实施方式三提供的终端的结构示意图；
图6为本发明实施方式四提供的基站的结构示意图。
具体实施方式
为使本申请的目的、技术方案和优点更加清楚明白，下文中将结合附图对本申请的实施例进行详细说明。
在附图的流程图示出的步骤可以在诸如一组计算机可执行指令的计算机系统中执行。并且，虽然在流程图中示出了逻辑顺序，但是在某些情况下，可以以不同于此处的顺序执行所示出或描述的步骤。
CSI量化反馈技术是MIMO技术的一个重要组成部分。在传统的无线通信系统中，通常使用DFT(Discrete Fourier Transform，离散傅里叶变换)矢量或者DFT矢量的变化形式，例如多个DFT矢量的克罗内克积，或者DFT矢量的级联形式，或者对级联的DFT矢量进行相位调整的形式，终端通过量化反馈，将上述形式的预编码指示信息上报给基站。这种类型的预编码码本可以归类为第一类码本，这种码本开销较小，但是CSI量化精度较低，性能较为受限。另一种码本通过对DFT矢量或者DFT矢量的克罗内克积进行线性加权合并，被加权合并的矢量称为码本基矢量，将码本基矢量相关信息、加权系数的幅度和相位信息作为预编码指示信息反馈给基站，这样的预编码码本可归类为第二类码本。具体的，第二类码本技术中，终端反馈的预编码矩阵的列数，即信道秩，为RI(Rank indicator，秩指示)，其中每一层的预编码矢量表示为一组码本基矢量的线性合并，该组码本基矢量可称为第一基矢量，终端根据第一基矢量计算出线性合并的加权系数，并量化反馈加权系数的幅度和相位信息，该加权系数可以称为第一系数。为了提升反馈的性能，通常需要将第一系数的幅度和相位信息按子带进行上报。所述子带为一种频域粒度，对于CSI反馈带宽中包含的所有RB(Resource block，资源块)，将连续的多个RB构成一个子带。
对于上述的CSI反馈方法，某一层在一个子带的预编码可以表示为：
Figure PCTCN2019128653-appb-000001
其中，W 1包含了2L个第一基矢量，W 1维度为N t×2L，其形式为：
Figure PCTCN2019128653-appb-000002
其中b 0,b 1,...,b L-1由一组正交的DFT矢量或DFT矢量的克罗内克积构成，其维度为
Figure PCTCN2019128653-appb-000003
其中N t表示CSI-RS(Channel State Information Reference Signal，信道状态信息参考信号)的端口数。这种CSI反馈方式，对不同极化方向的天线端口采用相同的DFT矢量或DFT矢量的克罗内克积，所以W 1包含了2个第一基矢量。一般来说，W 1中的信息是宽带反馈的，即对于整个CSI反馈频带中不同的子带，W 1中的信息是一样的。w～的维度为2L×1，表示2L个第一基矢量的加权系数，称为第一系数。
一般来说，为提升第二类码本的性能，终端需要针对每个子带反馈各个码本基矢量的加权系数的相位和/或幅度信息。因此，当子带个数较多时第二类码本反馈会带来较大的CSI反馈开销。另一方面，如果仅反馈加权系数的幅度或相位在整个宽带上的信息，无法充分得到第二类码本反馈能够带来的高性能增益。因此，如何通过较小的反馈开销，得到较高的第二类码本反馈性能，是相关技术未能解决的问题。
为此，本发明实施例提供了一种新的信道状态信息CSI反馈方案，通过该方案，即能够减小CSI反馈开销，同时又能够保证较高的CSI反馈性能。
实施方式一
图1为本发明实施方式一提供的信道状态信息CSI反馈方法的流程示意图，如图1所示，该方法包括：
步骤101，终端反馈预编码矩阵信息，所述预编码矩阵信息包括第一基矢量信息、第二基矢量信息、第二系数的幅度和相位信息；其中，反馈预编码子带内的预编码矢量为第一基矢量的线性合并，第一基矢量线性合并所使用的加权系数为第一系数；在所有CSI反馈频带所包含的频域单位内，对应于同一所述第一基矢量的所述第一系数所组成的矢量为第二基矢量的线性合并，所述第二系数为所述第二基矢量进行线性合并所使用的加权系数。
其中，所述第二基矢量表示从DFT矩阵或者过采样DFT矩阵中选择的K个DFT基矢量，所述过采样DFT矩阵的过采样因子O f的取值为以下之一：1、2、4、8。
其中，在终端反馈预编码矩阵信息之前，该方法还包括：接收基站发送的需要反馈CSI的频域范围，所述CSI的频域范围包括：需要反馈CSI的预编码子带或者需要梳状反馈CSI的预编码子带及稀疏程度，或者不需要反馈CSI的预编码子带；当基站梳状配置需要反馈预编码信息的预编码子带时，通过以下 方式之一修正第二基矢量：方式一：基站配置较大的过采样因子，增加可选择的DFT矩阵数目；方式二：终端根据需要反馈的预编码信息的预编码子带，梳状截取DFT基矢量；或者，当基站配置部分不需要反馈预编码的子带时，截去DFT基矢量中所述不需要反馈预编码信息的预编码子带的对应位置元素；或者，基站配置是否将所述第二基矢量做相位旋转，以配置的最中间的预编码子带为参考。
其中，所述CSI反馈频带是指基站配置需要反馈CSI的频域范围。
其中，所述预编码子带是一个频域单位，其包含的RB个数由基站配置或者由CSI反馈频带包含的RB数目和所使用的第二基矢量数目确定，即包含RB数目为
Figure PCTCN2019128653-appb-000004
RBNum为所述CSI反馈频带内包含的RB数目，K为所述CSI反馈频带使用的第二基矢量数目。
其中，所述第二系数表示为
Figure PCTCN2019128653-appb-000005
的维度为2L×K，2L表示矩阵W 1中包含的2L个第一基矢量，K表示矩阵W 3中包含的个第二基矢量。
其中，在终端反馈预编码矩阵信息之前，该方法还包括：接收基站发送的第二系数的反馈方式和反馈子集指示；所述预编码矩阵信息中的第二系数通过以下方式之一进行反馈：方式一：按照第二系数的幅度大小为优先级反馈第二系数；方式二：以所述第一基矢量为优先级反馈第二系数；方式三：以所述第二基矢量为优先级反馈第二系数；方式四：将所述第一基矢量分成两组，将第1到第L个第一基矢量，即W 1的第1到第L个列向量，作为第一组，将第L+1到第2个第一基矢量，即W 1的第L+1到第2L个列向量，作为第二组，根据分组内部的优先级反馈第二系数；方式五：对第二系数进行分组，根据分组的优先级反馈第二系数。
其中，所述方式一包括：以第二系数的幅度大小进行排序，从大到小选取部分系数，直到所述部分系数的功率之和与第二系数总功率的比值不小于δ，终端反馈所述部分系数，其中δ为预设阈值；使用2L·K大小的比特图指示其选择的系数。
其中，所述方式二包括：方式A：在最优的第一基矢量的K个加权系数中选择l 0个系数，在其余的各第一基矢量的K个加权系数中各选择l 1个系数，l 1＜l 0，使用2L·K大小的比特图指示其选择的系数；或者，方式B：在最优的第一基矢量的K个加权系数中选择l 0个系数，在其余的第一基矢量的所有加权系数中，总共再选择M-l 0个系数进行反馈，M的值由基站配置或者为
Figure PCTCN2019128653-appb-000006
其中δ M为预设阈值；使用以下子方式之一反馈其选择的子集：子方式B-1：使用2L·K大小的比特图指示其选择的系数；子方式B-2：使用
Figure PCTCN2019128653-appb-000007
比特(bits)指示选 择的最优第一基矢量，再用
Figure PCTCN2019128653-appb-000008
指示其选择的其余M-l 0个系数；或者，方式C：对2个第一基矢量进行排序，从强到弱的第一基矢量的加权系数中分别选取l 0,l 0-1，...,l 0-M+1个系数进行反馈，使用2L·K大小的比特图指示其选择的系数。
其中，所述方式三包括：方式D，在最优的第二基矢量的2L个加权系数中选择l 0个系数，在其余的第二基矢量的所有加权系数中，总共再选择M-l 0个系数，M的值由基站配置或者为
Figure PCTCN2019128653-appb-000009
其中δ M为预设阈值；使用以下子方式之一反馈其选择的子集：子方式D-1：使用2L·K大小的比特图指示其选择的系数；子方式D-2：使用
Figure PCTCN2019128653-appb-000010
指示选择的最优第二基矢量，再用
Figure PCTCN2019128653-appb-000011
bits指示其选择的其余M-l 0个系数；或者，方式E：对K个第二基矢量进行排序，从强到弱的第二基矢量的加权系数分别选取l 0,l 0-1，...,l 0-M+1个系数，使用2L·K大小的比特图指示其选择的系数。
其中，所述方式四包括：方式F：在两个分组内第一基矢量的所有加权系数中，各反馈
Figure PCTCN2019128653-appb-000012
个系数，
Figure PCTCN2019128653-appb-000013
的值由基站配置或者为
Figure PCTCN2019128653-appb-000014
其中δ M为预设阈值；通过比特图或者
Figure PCTCN2019128653-appb-000015
指示选择的系数；或者，方式G：在两个分组内，根据第一基矢量的优先级进行反馈且使用2L·K大小的比特图指示其选择的系数，包括：子方式G-1：在两个分组各选择一个最优的第一基矢量，两个最优的第一基矢量的加权系数中各选择l 0个系数，在两个分组内的其余第一基矢量的所有加权系数中各再选择
Figure PCTCN2019128653-appb-000016
个系数，
Figure PCTCN2019128653-appb-000017
的值由基站配置或者为
Figure PCTCN2019128653-appb-000018
其中δ M为预设阈值；或者子方式G-2：在两个分组各选择一个最优的第一基矢量，两个最优的第一基矢量的加权系数中各选择l 0个系数，在两个分组内的其余第一基矢量的加权系数中各再选择l 1个系数，l 1＜l 0；或者，方式H：在两个分组内，根据第二基矢量的优先级进行反馈且使用2L·K大小的比特图指示其选择的系数，包括：子方式H-1：在两个分组各选择一个最优的第二基矢量，第1个分组内最优的第二基矢量的加权系数中选择l 0个系数，第2个分组内的最优的第二基矢量的加权系数中选择l 0个系数，在两个分组内的其余第二基矢量的所有加权系数中各再选择
Figure PCTCN2019128653-appb-000019
个系数，
Figure PCTCN2019128653-appb-000020
的值由基站配置或者为
Figure PCTCN2019128653-appb-000021
其中δ M为预设阈值；或者子方式H-2：在两个分组各选择一个最优的第二基矢量，第1个分组内的最优的第二基矢量的加权系数中选择l 0个系数，第2个分组内的最优的第二基矢量的加权系数中选择l 0个系数，在两个分组内的其余第二基矢量的加权系数中各再选择l 1个系数，l 1＜l 0。
其中，所述方式五包括：方式I：将第二系数分成GroupNumber个组合，终端 选择l个最合适的组合进行反馈，使用
Figure PCTCN2019128653-appb-000022
指示其选择的组合；或者，方式J：将第二系数分成GroupNumber个组合，选择1个最合适的组合，在其余组合中再选择l-1个组合，使用log 2GroupNumberbits指示选择的最合适的组合，使用
Figure PCTCN2019128653-appb-000023
指示其余的组合。
其中，在终端反馈预编码矩阵信息之前，该方法还包括：接收基站发送的第一信令；所述第一信令包括基站配置的第一幅度反馈惩罚因子及指示需要做功率限制的第一基矢量，用于对所述需要做功率限制的第一基矢量的加权系数做功率限制；所述终端反馈的所述需要做功率限制的第一基矢量的加权系数幅度量化之后不能超过基站配置的第一幅度反馈惩罚因子，或者将所述需要做功率限制的第一基矢量的加权系数的幅度乘上基站配置的第一幅度反馈惩罚因子之后再进行量化；和/或，接收基站发送的第二信令；所述第二信令包括基站配置的第二幅度反馈惩罚因子及指示需要做功率限制的第二基矢量，用于对所述需要做功率限制的第二基矢量的加权系数做功率限制；所述终端反馈的所述需要做功率限制的第二基矢量的加权系数幅度量化之后不能超过基站配置的第二幅度反馈惩罚因子，或者将所述需要做功率限制的第二基矢量的加权系数的幅度乘上基站配置的第二幅度反馈惩罚因子之后再进行量化。
其中，所述反馈预编码矩阵信息，包括：当反馈资源不足且选择的第二基矢量数目K由终端配置时，动态地减小反馈的第二基矢量数目；或者，当选择的第二基矢量数目K由基站配置时，如果反馈资源充足，则上传所有基站配置的参数，否则选择以下方式之一：当按照所述方式一反馈预编码矩阵信息时，如果反馈资源不足，则基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；或者，当按照所述方式二包括的方式A或方式C反馈预编码矩阵信息时，如果反馈资源不足，终端优先反馈最优第一基矢量的加权系数，接着终端基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；当所述方式二包括的方式B反馈预编码矩阵信息且使用子方式B-1反馈子集指示时，如果反馈资源不足，终端优先反馈最优第一基矢量的加权系数，接着终端基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；当按照所述方式二包括的方式B反馈预编码矩阵信息且使用子方式B-2反馈子集指示时，如果反馈资源不足，终端只反馈最优第一基矢量的加权系数；或者，当按照所述方式三包括的方式D反馈预编码矩阵信息且使用子方式D-1反馈子集指示时，如果反馈资源不足，终端优先反馈最优第二基矢量的加权系数，接着终端基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；当按照所述方式三包括的方式D反 馈预编码矩阵信息且使用子方式D-2反馈子集指示时，如果反馈资源不足，终端只反馈最优第二基矢量的加权系数；当按照所述方式三包括的方式E反馈预编码矩阵信息时，如果反馈资源不足，终端优先反馈最优第二基矢量的加权系数，接着终端基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；或者，当按照所述方式四包括的方式F反馈预编码矩阵信息时，如果反馈资源不足，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；当按照所述方式四包括的方式G中包括的子方式G-1或子方式G-2反馈预编码矩阵信息时，如果反馈资源不足，优先反馈第一个分组中最优第一基矢量的加权系数，其次反馈第二分组内最优的第一基矢量的加权系数，接着基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；当按照所述方式四包括的方式H中包括的子方式H-1或子方式H-2反馈预编码矩阵信息时，如果反馈资源不足，优先反馈第一分组内最优第二基矢量的加权系数，其次反馈第二分组内最优的第二基矢量的加权系数，接着基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；或者，当按照所述方式五包括的方式I反馈预编码矩阵信息且使用比特图反馈选择的分组时，如果反馈资源不足，基于一定准则动态地根据比特图通知基站其选择的分组直至到达最大反馈资源为止；当按照所述方式五包括的方式J反馈预编码矩阵信息时，如果反馈资源不足，优先反馈最优分组对应的加权系数。
实施方式二
图2为本发明实施方式二提供的信道状态信息CSI反馈方法的流程示意图，如图2所示，该方法包括：
步骤201，基站接收终端反馈的预编码矩阵信息，所述预编码矩阵信息包括第一基矢量信息、第二基矢量信息、第二系数的幅度和相位信息；其中，反馈预编码子带内的预编码矢量为第一基矢量的线性合并，第一基矢量线性合并所使用的加权系数为第一系数；在所有CSI反馈频带所包含的频域单位内，对应于同一所述第一基矢量的所述第一系数所组成的矢量为第二基矢量的线性合并，所述第二系数为所述第二基矢量进行线性合并所使用的加权系数。
其中，在基站接收终端反馈的预编码矩阵信息之前，该方法还包括：所述基站配置需要反馈CSI的频域范围并发送给终端，以便所述终端根据所述CSI的频域范围确定需要反馈预编码信息的预编码子带。
其中，所述配置的CSI的频域范围包括：需要反馈CSI的预编码子带或者需要梳状反馈CSI的预编码子带及稀疏程度，或者不需要反馈CSI的预编码子带。
其中，在基站接收终端反馈的预编码矩阵信息之前，该方法还包括：所述 基站配置第二系数的反馈方式和反馈子集指示，并发送给终端，以便终端根据所述第二系数的反馈方式和反馈子集指示反馈第二系数；所述第二系数的反馈方式包括以下之一：方式一：按照第二系数的幅度大小为优先级反馈第二系数；
方式二：以所述第一基矢量为优先级反馈第二系数；方式三：以所述第二基矢量为优先级反馈第二系数；方式四：将所述第一基矢量分成两组，将第1到第L个第一基矢量，即W 1的第1到第L个列向量，作为第一组，将第L+1到第2L个第一基矢量，即W 1的第L+1到第2L个列向量，作为第二组，根据分组内部的优先级反馈第二系数；方式五：对第二系数进行分组，根据分组的优先级反馈第二系数；所述反馈子集指示用于指示终端反馈所述第二系数的子集的方式。
其中，在基站接收终端反馈的预编码矩阵信息之前，该方法还包括：向终端发送第一信令和/或第二信令；其中，所述第一信令包括基站配置的第一幅度反馈惩罚因子及指示需要做功率限制的第一基矢量，用于对所述需要做功率限制的第一基矢量的加权系数做功率限制；所述第二信令包括基站配置的第二幅度反馈惩罚因子及指示需要做功率限制的第二基矢量，用于对所述需要做功率限制的第二基矢量的加权系数做功率限制。
其中，所述第一幅度反馈惩罚因子为
Figure PCTCN2019128653-appb-000024
或者
Figure PCTCN2019128653-appb-000025
或者，所述第二幅度反馈惩罚因子为
Figure PCTCN2019128653-appb-000026
或者
Figure PCTCN2019128653-appb-000027
下面通过两个具体的实施例详细阐述本发明实施方式一、二提供的技术方案。
实施例一
本发明实施例一提出了一种对空域、频域信道系数压缩的方法进行CSI反馈，即某一层在所有预编码子带的预编码可以重新表示为如下形式，
Figure PCTCN2019128653-appb-000028
其中，对于该层的预编码矢量，将各个预编码子带所有第一系数组成矩阵W 2。W 2维度为2L×N s，N s表示预编码子带的个数
Figure PCTCN2019128653-appb-000029
S表示终端反馈预编码的频域单位，可以根据反馈的精度进行调整。S称为预编码子带大小，它可以由1个或者多个RB组成，N s称为预编码子带个数；其每一列的2L个元素， 对应于一个子带，表示上述2L个第一基矢量线性叠加的系数。CSI反馈频带是指基站配置需要反馈CSI的频域范围，其包含的RB个数为RBNum。所述的预编码子带是一个频域单位，其包含的RB个数可以由基站配置或者由反馈频带内的RB数目和所使用的第二基矢量数目确定，即包含RB数目为
Figure PCTCN2019128653-appb-000030
RBNum为所述CSI反馈频带包含的RB数目，为所述CSI反馈频带使用的第二基矢量数目。
W 1包含2L个第一基矢量，W 1维度为N t×2L，其形式为：
Figure PCTCN2019128653-appb-000031
其中，b 0,b 1,...,b L-1由一组正交的DFT(Discrete Fourier Transformation，离散傅里叶变换)矢量或DFT矢量的克罗内克积构成，其维度为
Figure PCTCN2019128653-appb-000032
N t表示CSI-RS的端口数。
其中，频域压缩矩阵W 3＝[w 0 w 1 ... w K-1]，表示从DFT矩阵或者过采样DFT矩阵中选择的个基矢量，称为第二基矢量。DFT矩阵的维度是N s×N s；其中，所述L可取以下之一：2、3、4、5、6。所述K可取以下之一：1、2、……，N s，即K最大能取的值是预编码子带的个数N s。
其中，所述的DFT矩阵，其列向量为，
Figure PCTCN2019128653-appb-000033
其中，m＝m 0,m 0+O f,m 0+2O f,...,m 0+(N s-1)O f；(m 0＝1,2,...,O f)，O f为过采样因子，过采样因子的取值可以为1、或2、或4、或8等。
其中，
Figure PCTCN2019128653-appb-000034
维度为2L×K，表示将W 2经过频域压缩后的矩阵，表示第二基矢量的加权系数，称之为第二系数。在本实施例中，除方式C-4外，第i个第一基矢量的加权系数，指的是
Figure PCTCN2019128653-appb-000035
第i行的K个系数；第j个第二基矢量的加权系数，指的是
Figure PCTCN2019128653-appb-000036
的j列的2L个系数。在方式C-4中，第一个分组内，第j个第二基矢量的加权系数，即
Figure PCTCN2019128653-appb-000037
中第j列、第1到L行的L个系数；第二个分组内，第j个第二基矢量的加权系数，即
Figure PCTCN2019128653-appb-000038
中第j列、第L+1到2L行的L个系数。
在上述新的子带的预编码表示形式基础上，图3为本发明实施例一提供的信道状态信息CSI反馈方法的流程示意图，如图3所示，该方法包括：
步骤301，终端接收基站发送的配置信息。
其中，所述配置信息包括以下至少之一：基站配置的需要反馈CSI的频域范围、配置的第二系数的反馈方式和反馈子集指示、配置的第一幅度反馈惩罚因子及指示需要做功率限制的第一基矢量、配置的第二幅度反馈惩罚因子及指示需要做功率限制的第二基矢量等信息。
其中，所述CSI的频域范围包括：需要反馈CSI的预编码子带或者需要梳 状反馈CSI的预编码子带及稀疏程度，或者不需要反馈CSI的预编码子带。
步骤302，终端根据信道估计的结果及基站的配置信息，得到预编码矩阵信息。
步骤303，终端反馈预编码矩阵信息，所述预编码矩阵信息包括第一基矢量信息、第二基矢量信息、第二系数的幅度和相位信息。
其中，反馈预编码子带内的预编码矢量为第一基矢量的线性合并，第一基矢量线性合并所使用的加权系数为第一系数；在所有CSI反馈频带所包含的频域单位内，对应于同一所述第一基矢量的所述第一系数所组成的矢量为第二基矢量的线性合并，所述第二系数为所述第二基矢量进行线性合并所使用的加权系数。
其中，所述第二基矢量表示从DFT矩阵或者过采样DFT矩阵中选择的K个DFT基矢量，所述过采样DFT矩阵的过采样因子O f的取值为以下之一：1、2、4、8。
其中，所述第二基矢量为DFT基矢量，所述第二基矢量可以根据实际需要进行部分修正，例如，第二基矢量，即W 3中基矢量的选取需要考虑一些特殊的情况，更具体地，包括以下几种特殊情况及解决方案：
方式A-1：基站有可能不需要终端上传所有预编码子带的CSI，基站配置需要上传CSI的预编码子带可能是不连续的，因此需要对DFT基矢量做出修正。具体地，做如下几种子方式进行改进：
子方式A-1-1：基站采用梳状配置需要上传CSI的预编码子带，即基站指定第S 0+Δ,S 0+2Δ,S 0+3Δ,...(Δ＝1,2,3,4...表示配置的稀疏程度)个预编码子带的CSI需要反馈，此时DFT矩阵可以采用如下两种方式进行修正：
子方式A-1-1-1：采用过采样的DFT矩阵，基站配置较大的过采样因子，增加可选择的DFT矩阵数目，本质上是增大DFT基矢量相邻元素相位差，以匹配部分预编码子带缺失造成相位偏移增大；子方式A-1-1-2：截取部分DFT矩阵，即DFT基矢量只保留第S 0+Δ,S 0+2Δ,S 0+3Δ,...个元素。此时，DFT矩阵的不同基矢量仍正交，但是同一个基矢量的相邻元素的相位差变大，可以匹配部分预编码子带缺失造成相位偏移增大；子方式A-1-2：基站可能只配置一部分子带不需要上传CSI，一个具体的例子是，基站配置第S i个预编码子带无需上报CSI。则进行频域压缩时，可以通过截去DFT基矢量的第S i个元素，以适应部分预编码子带缺失的情况。
方式B-1：所述DFT基矢量，是以第一个预编码子带为参考，即DFT基矢量的第一个元素为1。可以通过改变参考的预编码子带，将其位于所有预编码子带的中间，可以减小累积的量化误差。具体地，对原来的DFT基矢量做如下相 位旋转，
Figure PCTCN2019128653-appb-000039
其中，对于频域压缩反馈方法，另一个重要的问题是第二系数，即
Figure PCTCN2019128653-appb-000040
的反馈。按照第二系数的幅度大小、第一基矢量、第二基矢量、第一基矢量分组或者第一基矢量和第二基矢量联合分组的优先级，
Figure PCTCN2019128653-appb-000041
的2L·K个元素可以选取部分子集或者全部反馈，不同的子集还可以反馈不同的精度，以提高性能。此外，基站与终端需要约定终端选择的子集需要以什么方式通知基站，即第二系数的反馈子集指示。具体的反馈方式及子集指示方式，可以选择以下方式之一：
方式C-1：按照第二系数的幅度大小为优先级反馈第二系数；其中，第二系数幅度较大的部分，通常更能反映真实的信道特征。一个具体的例子是，终端对第二系数的幅度从大到小进行排序，从大到小选取部分系数，直到所述部分系数的功率之和与第二系数总功率比值不小于δ，反馈选取的这部分系数。其中，δ为阈值，典型值可取0.90、0.95、0.99等。终端使用2L·K大小的比特图指示其选择的系数。
方式C-2：从第一基矢量的角度选取系数；其中，第一基矢量对应于空间中的波束，通常是接收信号会集中于一些特定的波束方向，因此提高特定波束的优先级及反馈精度，能够提升系统性能。具体地，以所述第一基矢量为优先级反馈第二系数，可以选择以下子方式之一：
子方式C-2-1：终端基于一定准则(比如某个第一基矢量对应的K个系数功率和最大)，选择最优的第一基矢量，则这个最优的第一基矢量的K个加权系数(对应于
Figure PCTCN2019128653-appb-000042
的某一行)中反馈l 0个系数；可以适当增加这个第一基矢量的加权系数的量化精度，比如幅度和相位量化为(3，4)bits或者(4，4)bits；在其余的第一基矢量的K个加权系数中各选择l 1(l 1＜l 0)个系数，幅度和相位量化为(2，3)bits或者(3，3)bits，其中幅度可以直接量化或者根据最优第一基矢量的幅度进行差分量化。终端使用2L·K大小的比特图指示其选择的系数，基站根据终端反馈的比特图信息可以知道反馈的子集及其反馈最优的第一基矢量，进一步根据配置的量化精度，可以计算出反馈第二系数的总开销。
子方式C-2-2：终端基于一定准则，选择最优的第一基矢量，则所述最优的第一基矢量的K个加权系数中反馈l 0个系数，可以适当增加这个所述最优的第一基矢量的加权系数的量化精度，比如幅度和相位量化为(3，4)bits或者(4，4)bits；在除了最优的第一基矢量之外的其余第一基矢量的所有加权系数中再选择 总共M-l 0个系数，M的值可以由基站配置或者为
Figure PCTCN2019128653-appb-000043
其中δ M为预设阈值；其幅度和相位量化精度为(2，3)bits或者(3，3)bits，其中幅度可以直接量化或者根据最优第一基矢量的幅度进行差分量化。
终端指示其反馈的部分第二系数，可以通过以下子方式之一实现：
子方式C-2-2-1：终端使用2L·K大小的比特图指示其选择的系数，基站根据终端反馈的比特图信息可以知道反馈的子集及最优的第一基矢量，进一步根据配置的量化精度，可以计算出反馈第二系数的总开销。注意，在非最优第一基矢量选择的加权系数个数必须少于l 0，否则基站无法通过比特图隐含知道最优第一基矢量的位置。
子方式C-2-2-2：终端使用
Figure PCTCN2019128653-appb-000044
指示选择的最优第一基矢量，再用
Figure PCTCN2019128653-appb-000045
指示其选择的其余M-l 0个系数，基站进一步根据配置的量化精度，可以计算出反馈第二系数的总开销。
子方式C-2-3：终端基于一定准则，对2L个波束的强弱进行排序，则对应的从强到弱的波束分别选取l 0,l 0-1，...,l 0-M+1个系数，基站可以根据其波束的强弱不同，配置不同的反馈精度。终端使用2L·K大小的比特图指示其选择的系数，基站根据终端反馈的比特图信息可以知道反馈的子集及第一基矢量的排序情况，进一步根据配置的量化精度，可以计算出反馈第二系数的总开销。
方式C-3：从第二基矢量角度考虑；其中，一个第二基矢量对应于信道的一个时延路径，如果某个最强的时延路径反馈精度较高，可以提高系统性能。具体地，以所述第二基矢量为优先级反馈第二系数，可以选择以下子方式之一：
子方式C-3-1：终端基于一定准则(比如某个第二基矢量对应的2L个系数的功率之和最大)，选择最优的第二基矢量，该最优的第二基矢量的2个加权系数((对应于
Figure PCTCN2019128653-appb-000046
的某一列))中反馈l 0个系数。可以适当增加这个基矢量的加权系数的量化精度，比如幅度和相位量化精度为(3，4)bits或者(4，4)bits；在其余第二基矢量的所有加权系数中，总共再选择幅度系数最大的(M-l 0)个系数，M的值可以由基站配置或者为
Figure PCTCN2019128653-appb-000047
其中δ M为预设阈值；其幅度和相位量化精度为(2，3)bits或者(3，3)bits其中幅度可以直接量化或者根据最优第二基矢量的幅度进行差分量化。
终端指示其反馈的部分第二系数，可以通过以下子方式之一实现：
子方式C-3-1-1：终端使用2L·K大小的比特图指示其选择的系数，基站根据终端反馈的比特图信息可以知道反馈的子集及最优的第二基矢量，进一步根据配置的量化精度，可以计算出反馈第二系数的总开销。注意，在非最优第二基矢量选择的系数个数必须少于l 0，否则基站无法通过比特图隐含知道最优第二基矢量的位置。
子方式C-3-1-2：终端使用
Figure PCTCN2019128653-appb-000048
指示选择的最优第二基矢量，再用
Figure PCTCN2019128653-appb-000049
指示其选择的其余M-l 0个系数，基站进一步根据配置的量化精度，可以计算出反馈第二系数的总开销。
子方式C-3-2：终端基于一定准则，对K个第二基矢量的强弱进行排序，则对应的从强到弱的第二基矢量的加权系数中分别选取l 0,l 0-1，...,l 0-M+1个系数，基站可以根据其第二基矢量的强弱不同，配置不同的幅度和相位量化精度。终端使用2L·K大小的比特图指示其选择的系数，基站根据终端反馈的比特图信息可以知道反馈的子集及第二基矢量的排序情况，进一步根据配置的量化精度，可以计算出反馈第二系数的总开销。
方式C-4：将第一基矢量分成两组，将第1到第L个第一基矢量，即W 1的第1到第L个列向量，作为第一组，将第L+1到第2L个第一基矢量，即W 1的第L+1到第2L个列向量，作为第二组，根据分组内部的优先级，反馈第二系数；其中，L个第一基矢量是对第1到
Figure PCTCN2019128653-appb-000050
个CSI-RS端口的预编码，另外L个第一基矢量是对第
Figure PCTCN2019128653-appb-000051
到N t个CSI-RS端口的预编码。在组内按照一定规则选取
Figure PCTCN2019128653-appb-000052
的部分子集，并提高反馈精度，将会提高系统性能。
具体地，可以选择以下子方式至少之一：
子方式C-4-1：终端基于一定准则(比如第二系数的幅度大小排序)，在第一组包含的第一基矢量的加权系数(即
Figure PCTCN2019128653-appb-000053
的第1到L行)中选择
Figure PCTCN2019128653-appb-000054
个系数，
Figure PCTCN2019128653-appb-000055
的值可以由基站配置或者为
Figure PCTCN2019128653-appb-000056
其中δ M为预设阈值；在第二组包含的第一基矢量的加权系数(即
Figure PCTCN2019128653-appb-000057
的第L+1到2L行)中再选择
Figure PCTCN2019128653-appb-000058
个系数。终端使用2L·K大小的比特图或者
Figure PCTCN2019128653-appb-000059
指示其选择的系数。
子方式C-4-2：终端基于一定准则，两组第一基矢量中各选择一个最优的第一基矢量。这两个最优第一基矢量的对应系数，可以反馈更多元素，并且增加量化精度。具体地，可以采取以下子方式之一。
子方式C-4-2-1：终端基于一定准则，两组第一基矢量中各选择一个最优的第一基矢量。两个最优第一基矢量的加权系数中各反馈l 0个系数，这些系数可以适当增加量化精度，比如幅度和相位量化为(3，4)bits或者(4，4)bits。在每一个分组内的其余第一基矢量的所有加权系数中，总共各再选择
Figure PCTCN2019128653-appb-000060
个系数，
Figure PCTCN2019128653-appb-000061
的值可以由基站配置或者为
Figure PCTCN2019128653-appb-000062
其中δ M为预设阈值；其幅度和相位量化精度为(2，3)bits或者(3，3)bits，其中幅度可以直接量化或者根据两个最优第一基矢量的幅度进行差分量化。终端使用2L·K大小的比特图指示其 选择的系数，基站根据终端反馈的比特图信息可以知道反馈的子集及两个分组中各自最优的第一基矢量，进一步根据配置的量化精度，可以计算出反馈第二系数的总开销。注意，在非最优第一基矢量选择的系数个数必须少于l 0，否则基站无法通过比特图隐含知道最优第一基矢量的位置。
子方式C-4-2-2：终端基于一定准则，两组第一基矢量中各选择一个最优的第一基矢量。两个最优第一基矢量的加权系数中各反馈l 0个系数，这些系数可以适当增加量化精度，比如幅度和相位量化为(3，4)bits或者(4，4)bits。两个组分别在其余第一基矢量的加权系数中各再选择l 1(l 1＜l 0)个系数(即
Figure PCTCN2019128653-appb-000063
的每行选l 1个系数)，其幅度和相位量化精度为(2，3)bits或者(3，3)bits其中幅度可以直接量化或者根据两个最优第一基矢量的幅度进行差分量化。终端使用2L·K大小的比特图指示其选择的系数，基站根据终端反馈的比特图信息可以知道反馈的子集及两个分组内各自最优的第一基矢量，进一步根据配置的量化精度，可以计算出反馈第二系数的总开销。
子方式C-4-3：终端基于一定准则，两组第一基矢量中各选择一个最优的第二基矢量。这两个第二基矢量的对应系数，可以反馈更多元素，并且增加量化精度。具体地，可以采取以下子方式之一：
子方式C-4-3-1：终端基于一定准则，两组第一基矢量中各选择一个最优的第二基矢量。第1个分组内最优的第二基矢量的加权系数中选择l 0个系数，第2个分组内的最优的第二基矢量的加权系数中选择l 0个系数，这些系数可以适当增加量化精度，比如幅度和相位量化为(3，4)bits或者(4，4)bits。在每一个分组内的其余第二基矢量的所有加权系数中，总共各再选择
Figure PCTCN2019128653-appb-000064
个系数，
Figure PCTCN2019128653-appb-000065
的值可以由基站配置或者为
Figure PCTCN2019128653-appb-000066
其中δ M为预设阈值；其幅度和相位量化精度为(2，3)bits或者(3，3)bits，其中幅度可以直接量化或者根据两个最优第二基矢量的幅度进行差分量化。终端使用2L·K大小的比特图指示其选择的系数，基站根据终端反馈的比特图信息可以知道反馈的子集及两个分组中各自最优的第二基矢量，进一步根据配置的量化精度，可以计算出反馈第二系数的总开销。注意，在非最优第一基矢量选择的系数个数必须少于l 0，否则基站无法通过比特图隐含知道最优第二基矢量的位置。
子方式C-4-3-2：终端基于一定准则，两组第一基矢量中各选择一个最优的第二基矢量。第1个分组内的最优的第二基矢量的加权系数中选择l 0个系数，第2个分组内的最优的第二基矢量的加权系数中选择l 0个系数，这些系数可以适当增加量化精度，比如幅度和相位量化为(3，4)bits或者(4，4)bits。两个组分别在其余第二基矢量的加权系数中各再选择l 1(l 1＜l 0)个系数(即
Figure PCTCN2019128653-appb-000067
的每列选l 1个系数)，其幅度和相位量化精度为(2，3)bits或者(3，3)bits，其中幅度可以直接量化或者根据两个最优第二基矢量的幅度进行差分量化。终端使用2L·K大小的比特图指示其选择的系数，基站根据终端反馈的比特图信息可以知 道反馈的子集及两个分组内各自最优的第二基矢量，进一步根据配置的量化精度，可以计算出反馈第二系数的总开销。
方式C-5：将对第二系数进行分组，根据分组的优先级进行反馈；接收信号的波束及对应的时延会相对地集中，具体体现在
Figure PCTCN2019128653-appb-000068
的幅度较大的部分，会集中分布在第二系数
Figure PCTCN2019128653-appb-000069
的某一个位置。因此，可以对
Figure PCTCN2019128653-appb-000070
进行分组，终端反馈最合适的组合，增加量化精度，可以提高系统性能。具体地，可以选择以下子方式至少之一：
方式C-5-1：根据L和K的大小进行分组，将其分成
Figure PCTCN2019128653-appb-000071
个组合，其中L s和K s分别表示第一基矢量和第二基矢量的分组粒度，分组粒度可以根据L和K的值及反馈精度进行选择。终端基于一定准则，选择l个最合适的组合进行反馈。终端使用
Figure PCTCN2019128653-appb-000072
比特图或者
Figure PCTCN2019128653-appb-000073
bits指示其选择的组合，基站根据终端选择的组合，进一步根据配置的量化精度，可以计算出反馈第二系数的总开销。
方式C-5-2：根据L和K的大小进行分组，将其分成
Figure PCTCN2019128653-appb-000074
个组合，其中L s和K s分别表示第一基矢量和第二基矢量的分组粒度，分组粒度可以根据L和K的值及反馈精度进行选择。终端基于一定准则，选择1个最合适的组合，这个组合内系数可以适当增加量化精度，比如幅度和相位量化为(3，4)bits或者(4，4)bits。终端再选择(M-1)个组合，这些组合内的系数，其幅度和相位量化精度可以为(2，3)bits或者(3，3)bits。终端使用
Figure PCTCN2019128653-appb-000075
Figure PCTCN2019128653-appb-000076
指示其选择的最优组合，此外还需要
Figure PCTCN2019128653-appb-000077
指示选择的其它组合，进一步根据配置的量化精度，可以计算出反馈第二系数的总开销。
其中，在实际通信系统中，第一基矢量对应于空间中的波束，由于波束具有方向性，一个终端的信号中的一些波束信号强度太大，将会给其它用户带来干扰。因此，基站通常会对特定的波束做限制，比如限制特定波束的宽带幅度功率。此外，基站根据时分双工(Time Division Duplex，TDD)系统的互易性，基站也可以限制不必要的第二基矢量，减小信号的多径干扰。具体地，可以采取以下几种方式：
方式D-1：将全部可选的第一基矢量进行分组，即基站选择一个分组，如果终端反馈的波束位于基站配置的分组中，则对子带幅度做功率限制，即对
Figure PCTCN2019128653-appb-000078
的某些行(对应于一些波束)做功率限制。
比如，终端反馈的第i个第一基矢量，恰好是基站需要做功率限制的第一基矢量，则可以选择以下两种子方式之一：
子方式D-1-1：
Figure PCTCN2019128653-appb-000079
的第i、i+L行的幅度量化后不能超过基站配置的第一幅度反馈惩罚因子，基站可以配置的第一幅度反馈惩罚因子为
Figure PCTCN2019128653-appb-000080
或者
Figure PCTCN2019128653-appb-000081
子方式D-1-2：
Figure PCTCN2019128653-appb-000082
的第i、i+L行的幅度整体乘以一定的第一幅度反馈惩罚因子之后再进行量化，第一幅度反馈惩罚因子可取
Figure PCTCN2019128653-appb-000083
或者
Figure PCTCN2019128653-appb-000084
方式D-2：在TDD系统中，基于互易性原理，基站通过上行的参考信号估计的上行信道，可以互易到下行信道。因此，基站能够大致估计下行信道的时延情况。而一个第二基矢量本质上反映的是一条时延路径，基站可以通过限制终端选择的第二基矢量，进而避免一些非必要的时延分量带来的干扰。
具体地，可以通过以下子方式之一实现：
子方式D-2-1：将所有可选的第二基矢量进行分组，基站通过高层参数配置选择一个分组，如果终端选择的一个第二基矢量与基站配置的分组重叠，则可以通过如下两种子方式进行功率限制：
子方式D-2-1-1：该第二基矢量对应的所有加权系数，幅度量化后不能超过基站配置的第二幅度反馈惩罚因子，基站可以配置的第二幅度反馈惩罚因子为
Figure PCTCN2019128653-appb-000085
或者
Figure PCTCN2019128653-appb-000086
子方式D-2-1-2：该第二基矢量的所有加权系数，幅度整体乘以一定的第二幅度反馈惩罚因子之后再进行量化，第二幅度反馈惩罚因子可取
Figure PCTCN2019128653-appb-000087
或者
Figure PCTCN2019128653-appb-000088
需要指出的是，上述方式D-1和D-2并不是互斥的，两者可以结合使用，同时避免干扰和减小不必要的第二基矢量的反馈。
其中，终端通过测量CSI-RS估计下行信道，利用其它参考信号估计干扰，根据信道条件，终端会根据基站指示，反馈RI(Rank indicator，秩指示)、CQI(Channel quality indicator，信道质量指示)和PMI(Precoding matrix indicator， 预编码指示)。由于基站无法预知终端反馈的RI或者基站同时触发了多个CSI报告，有可能造成上行传输CSI的资源不足。因此，有必要定义反馈资源不足时，优先反馈哪一部分的资源。根据实现第二基矢量的改进方式以及第二系数的反馈方式及反馈子集指示的各种方式，可以选择以下方式之一：
方式E-1：选择的第二基矢量数目K由终端反馈，如果反馈的CSI资源不足，终端可以动态地减小反馈的第二基矢量数目，减小反馈开销。
方式E-2：选择的第二基矢量数目K由基站配置，如果反馈的CSI资源不足，具体地可以选择以下子方式之一：
子方式E-2-1：按照方式C-1反馈CSI，如果反馈资源充足，则上传所有基站配置的参数。如果反馈资源不足，接着终端可以基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止。
子方式E-2-2：按照方式C-2反馈CSI，如果反馈资源充足，则上传所有基站配置的参数。具体地，按照子方式C-2-1和子方式C-2-3反馈CSI时，如果反馈资源不足，终端优先反馈最优第一基矢量对应的系数，接着终端可以基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；按照子方式C-2-2反馈CSI，且按照子方式C-2-2-1反馈子集指示，如果反馈资源不足，终端优先反馈最优第一基矢量对应的加权系数，接着终端可以基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；按照子方式C-2-2反馈CSI，且按照子方式C-2-2-2反馈子集指示，如果反馈资源不足，终端只反馈最优第一基矢量的加权系数。
子方式E-2-3：按照方式C-3反馈CSI，如果反馈资源充足，则上传所有基站配置的参数。具体地，按照子方式C-3-1反馈CSI，且按照子方式C-3-1-1反馈子集指示，如果反馈资源不足，终端优先反馈最优第二基矢量的加权系数，接着终端可以基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；按照子方式C-3-1反馈CSI，且按照子方式C-3-1-2反馈子集指示，如果反馈资源不足，终端只反馈最优第二基矢量的加权系数；按照子方式C-3-2反馈CSI时，如果反馈资源不足，终端优先反馈最优第二基矢量的加权系数，接着终端可以基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止。
子方式E-2-4：按照方式C-4反馈CSI，如果反馈资源充足，则上传所有基站配置的参数。
具体地，按照子方式C-4-1反馈CSI且按照比特图方式指示选择的系数，如果反馈资源不足，动态地通过比特图通知其选择的系数直至到达最大反馈资源 为止；按照子方式C-4-2-1和子方式C-4-2-2反馈CSI，如果反馈资源不足，终端优先反馈第一分组中最优第一基矢量的加权系数，其次反馈第二分组上最优的第一基矢量的加权系数，接着终端可以基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；按照子方式C-4-3-1和子方式C-4-3-2反馈CSI，如果反馈资源不足，终端优先反馈第一分组中最优第二基矢量的加权系数，其次反馈第二分组内最优的第二基矢量的加权系数，接着终端可以基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止。
子方式E-2-5：按照方式C-5-1反馈CSI，且使用比特图反馈选择的分组，如果反馈资源不足，终端可以基于一定准则，动态地根据比特图通知基站其选择的分组直至到达最大反馈资源为止；按照方式C-5-2反馈CSI，如果反馈资源充足，则上传所有基站配置的参数。如果反馈资源不足，终端优先反馈最优分组对应的加权系数。
需要指出的是，以上方案对终端反馈的每一层的CSI都适用。
实施例二
在上述实施例一中新的子带的预编码表示形式基础上，图4为本发明实施例二提供的信道状态信息CSI反馈方法的流程示意图，如图4所示，该方法包括：
步骤401，基站配置需要反馈CSI的频域范围并发送给终端。
其中，所述配置的CSI的频域范围包括：需要反馈CSI的预编码子带或者需要梳状反馈CSI的预编码子带及稀疏程度，或者不需要反馈CSI的预编码子带。
步骤402，所述基站配置第二系数的反馈方式和反馈子集指示，并发送给终端。
其中，所述第二系数的反馈方式包括以下之一：方式一：按照第二系数的幅度大小为优先级反馈第二系数；方式二：以所述第一基矢量为优先级反馈第二系数；方式三：以所述第二基矢量为优先级反馈第二系数；方式四：将所述第一基矢量分成两组，将第1到第L个第一基矢量，即W 1的第1到第L个列向量，作为第一组，将第L+1到第2L个第一基矢量，即W 1的第L+1到第2L个列向量，作为第二组，根据分组内部的优先级反馈第二系数；方式五：对第二系数进行分组，根据分组的优先级反馈第二系数。
其中，所述反馈子集指示用于指示终端反馈所述第二系数的子集的方式。
步骤403，终端根据基站配置的需要反馈CSI的频域范围和第二系数的反馈 方式及反馈子集指示反馈预编码矩阵信息。
其中，所述预编码矩阵信息包括第一基矢量信息、第二基矢量信息、第二系数的幅度和相位信息。
其中，反馈预编码子带内的预编码矢量为第一基矢量的线性合并，第一基矢量线性合并所使用的加权系数为第一系数；在所有CSI反馈频带所包含的频域单位内，对应于同一所述第一基矢量的所述第一系数所组成的矢量为第二基矢量的线性合并，所述第二系数为所述第二基矢量进行线性合并所使用的加权系数。
其中，所述CSI反馈频带是指基站配置需要反馈CSI的频域范围。
其中，所述预编码子带是一个频域单位，其包含的RB个数由基站配置或者由CSI反馈频带包含的RB数目和所使用的第二基矢量数目确定，即包含RB数目为
Figure PCTCN2019128653-appb-000089
RBNum为所述CSI反馈频带内包含的RB数目，K为所述CSI反馈频带使用的第二基矢量数目。
其中，所述第二系数表示为
Figure PCTCN2019128653-appb-000090
的维度为2L×K，2L表示矩阵W 1中包含的2L个第一基矢量，K表示矩阵W 3中包含的K个第二基矢量。
其中，所述的第二基矢量可以根据实际需要进行部分修正，包括如下三种方式：
1、当基站梳状配置需要反馈预编码的子带时，第二基矢量可以通过以下子方式之一进行调整：子方式一：配置较大的过采样因子，增加可选择的DFT矩阵数目；子方式二：梳状截取第二基矢量。
2、当基站配置部分不需要反馈预编码的子带时，截去第二基矢量部分元素。
3、配置第二基矢量可以做相位旋转，以配置的最中间的预编码子带为参考。
其中，所述第二系数，终端可以反馈部分子集，通过反馈指示其所选的子集。所选子集及反馈指示，根据跟基站约定的以下方式之一反馈第二系数：
方式一包括：以所述第二系数的幅度大小进行排序，从大到小选取部分系数，直到所述部分系数的功率之和与第二系数总功率的比值不小于δ，终端反馈所述部分系数，其中δ为预设阈值；使用2L·K大小的比特图指示其选择的系数。
方式二包括：以所述第一基矢量为优先级反馈第二系数，具体可以使用以下子方式之一：
方式A：在最优的第一基矢量的K个加权系数中选择l 0个系数，增加量化精度。在其余的各第一基矢量的K个加权系数中各选择l 1(l 1＜l 0)个系数，适当减小量化精度。终端使用2L·K大小的比特图指示其选择的系数。
方式B：在最优的第一基矢量的K个加权系数中选择l 0个系数，增加量化精度。在其余的第一基矢量的所有加权系数中，再选择总共-l 0个系数进行反馈，M的值可以由基站配置或者为
Figure PCTCN2019128653-appb-000091
其中δ M为预设阈值，适当减小量化精度。终端可以使用以下子方式之一，反馈其选择的子集。
子方式B-1：终端使用2L·K大小的比特图指示其选择的系数。
子方式B-2：终端使用
Figure PCTCN2019128653-appb-000092
指示选择的最优第一基矢量，再用
Figure PCTCN2019128653-appb-000093
指示其选择的其余M-l 0个系数。
方式C：终端对2L个第一基矢量进行排序，从强到弱的第一基矢量的加权系数分别选取l 0,l 0-1，...,l 0-M+1个系数进行反馈。终端使用2L·K大小的比特图指示其选择的系数。
方式三包括：以所述第二基矢量为优先级反馈第二系数，具体可以使用以下子方式之一：
方式D：在最优的第二基矢量的2L个加权系数中选择l 0个系数，增加量化精度。在其余的第二基矢量的所有加权系数中，再选择总共M-l 0个系数进行反馈，M的值由基站配置或者为
Figure PCTCN2019128653-appb-000094
其中δ M为预设阈值，适当减小量化精度。终端可以使用以下子方式之一，反馈其选择的子集。
子方式D-1：终端使用2L·K大小的比特图指示其选择的系数。
子方式D-2：终端使用
Figure PCTCN2019128653-appb-000095
指示选择的最优第二基矢量，再用
Figure PCTCN2019128653-appb-000096
指示其选择的其余M-l 0个系数。
方式E：终端对K个第二基矢量进行排序，从强到弱的第二基矢量的加权系数分别选取l 0,l 0-1，...,l 0-M+1个系数。终端使用2L·K大小的比特图指示其选择的系数。
方式四包括：将所述第一基矢量进分成两组，根据优先级，反馈第二系数，具体可以使用以下子方式之一：
方式F：在两个分组内第一基矢量的所有加权系数中，各反馈
Figure PCTCN2019128653-appb-000097
个系数，
Figure PCTCN2019128653-appb-000098
的值由基站配置或者为
Figure PCTCN2019128653-appb-000099
其中δ M为预设阈值；终端通过比特图或者
Figure PCTCN2019128653-appb-000100
指示选择的系数。
方式G：在两个分组内，根据第一基矢量的优先级进行反馈，具体可以使用以下子方式之一：
子方式G-1：在两个分组各选择一个最优的第一基矢量。两个最优的第一基矢量的加权系数中各选择l 0个系数，增加量化精度。在两个分组内其余第一基矢 量的所有加权系数中各再选择
Figure PCTCN2019128653-appb-000101
个系数，
Figure PCTCN2019128653-appb-000102
的值由基站配置或者为
Figure PCTCN2019128653-appb-000103
其中δ M为预设阈值，适当减小量化精度。终端使用2L·K大小的比特图指示其选择的系数。
子方式G-2：在两个分组各选择一个最优的第一基矢量。两个最优的第一基矢量的加权系数中各选择l 0个系数，增加量化精度。在两个分组内的其余第一基矢量的加权系数中各再选择l 1(l 1＜l 0)个系数，适当减小量化精度。终端使用2L·K大小的比特图指示其选择的系数。
方式H：在两个分组内，根据第二基矢量的优先级进行反馈，具体可以使用以下子方式之一：
子方式H-1：在两个分组各选择一个最优的第二基矢量。第1个分组内最优的第二基矢量的加权系数中选择l 0个系数，第2个分组内的最优的第二基矢量的加权系数中选择l 0个系数，，增加量化精度。在两个分组内的其余第二基矢量的所有加权系数中各再选择
Figure PCTCN2019128653-appb-000104
个系数，
Figure PCTCN2019128653-appb-000105
的值由基站配置或者为
Figure PCTCN2019128653-appb-000106
其中δ M为预设阈值，适当减小量化精度。终端使用2L·K大小的比特图指示其选择的系数。
子方式H-2：在两个分组各选择一个最优的第二基矢量。第1个分组内的最优的第二基矢量的加权系数中选择l 0个系数，第2个分组内的最优的第二基矢量的加权系数中选择l 0个系数，，增加量化精度。在两个分组内的其余第二基矢量的加权系数中各再选择l 1(l 1＜l 0)个系数，适当减小量化精度。终端使用2L·K大小的比特图指示其选择的系数。
方式五包括：对第二系数进行分组，根据分组的优先级进行反馈，具体可以使用以下子方式之一：
方式I：将第二系数分成GroupNumber个组合，终端选择l个最合适的组合进行反馈，使用
Figure PCTCN2019128653-appb-000107
指示其选择的组合。
方式J：将第二系数分成GroupNumber个组合，终端选择1个最合适的组合，增加量化精度。在其余组合中在选择l-1个组合，适当减小量化精度。终端使用log 2GroupNumberbits指示选择的最合适的组合，使用
Figure PCTCN2019128653-appb-000108
指示其余的组合。
其中，基站还可以限制所述第一基矢量、第二基矢量的选择，进而避免干扰。具体可以采用以下方式：
方式一：基站对所有可选的第一基矢量进行分组，对需要进行限制的第一 基矢量通过信令通知终端，对反馈的功率做限制。具体地，可以采用以下子方式之一：1、基站配置对某些第一基矢量做功率限制，该做功率限制的第一基矢量的加权系数幅度量化之后不能超过基站配置的第一幅度反馈惩罚因子；2、基站配置对某些第一基矢量做功率限制，该做功率限制的第一基矢量的加权系数的幅度乘上基站配置的第一幅度反馈惩罚因子之后再进行量化。
方式二：基站对所有可选的第二基矢量进行分组，对需要进行限制的第二基矢量通过信令通知终端，对反馈的功率做限制。具体地，可以采用以下子方式之一：1、基站配置对某些第二基矢量做功率限制，该做功率限制的第二基矢量的加权系数幅度量化之后不能超过基站配置的第二幅度反馈惩罚因子；2、基站配置对某些第二基矢量做功率限制，该做功率限制的第二基矢量的加权系数的幅度乘上基站配置的第二幅度反馈惩罚因子之后再进行量化。
其中，当上行反馈CSI的资源不足时，终端可以根据优先级丢弃部分反馈内容。具体地，可以采用以下方式之一：
方式一：当反馈资源不足且选择的第二基矢量数目K由终端配置时，动态地减小反馈的第二基矢量数目。
方式二，当选择的第二基矢量数目K由基站配置时，如果反馈资源充足，则上传所有基站配置的参数，否则选择以下子方式之一：1、当按照所述方式一反馈预编码矩阵信息时，如果反馈资源不足，则基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；2、当按照所述方式二包括的方式A或方式C反馈预编码矩阵信息时，如果反馈资源不足，终端优先反馈最优第一基矢量的加权系数，接着终端可以基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；当所述方式二包括的方式B反馈预编码矩阵信息且使用子方式B-1反馈子集指示时，如果反馈资源不足，终端优先反馈最优第一基矢量的加权系数，接着终端可以基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；当按照所述方式二包括的方式B反馈预编码矩阵信息且使用子方式B-2反馈子集指示时，如果反馈资源不足，终端只反馈最优第一基矢量的加权系数；3、当按照所述方式三包括的方式D反馈预编码矩阵信息且使用子方式D-1反馈子集指示时，如果反馈资源不足，终端优先反馈最优第二基矢量的加权系数，接着终端可以基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；当按照所述方式三包括的方式D反馈预编码矩阵信息且使用子方式D-2反馈子集指示时，如果反馈资源不足，终端只反馈最优第二基矢量的加权系数；当按照所述方式三包括的方式E反馈预编码矩阵信息时，如果反馈资源不足，终端优先反馈最优第二基矢量的加权系数，接着终端可以基于一定准则，动态地通过比 特图通知其选择的系数直至到达最大反馈资源为止；4、当按照所述方式四包括的方式F反馈预编码矩阵信息时，如果反馈资源不足，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；当按照所述方式四包括的方式G中包括的子方式G-1或子方式G-2反馈预编码矩阵信息时，如果反馈资源不足，优先反馈第一分组内最优第一基矢量的加权系数，其次反馈第二分组内最优的第一基矢量的加权系数，接着基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；当按照所述方式四包括的方式H中包括的子方式H-1或子方式H-2反馈预编码矩阵信息时，如果反馈资源不足，优先反馈第一分组内最优第二基矢量的加权系数，其次反馈第二分组内最优的第二基矢量的加权系数，接着基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；5、当按照所述方式五包括的方式I反馈预编码矩阵信息且使用比特图反馈选择的分组时，如果反馈资源不足，基于一定准则动态地根据比特图通知基站其选择的分组直至到达最大反馈资源为止；当按照所述方式五包括的方式J反馈预编码矩阵信息时，如果反馈资源不足，优先反馈最优分组对应的加权系数。
本发明实施例一、二提供的技术方案，终端反馈的预编码矩阵信息包括第一基矢量信息、第二基矢量信息、第二系数的幅度和相位信息；其中，在一个预编码子带内的第一基矢量的加权系数为第一系数，利用不同预编码子带预编码矢量的相关性，所有预编码子带的第一系数构成的矩阵可以在频域上使用第二基矢量进行压缩，第二基矢量的加权系数为第二系数。如此，既能够减小CSI反馈开销，又保证了较高的CSI反馈性能。
上述基站的配置操作并不存在固定的前后顺序，配置的信息也可以合并一起发送给终端。
实施方式三
图5为本发明实施方式三提供的终端的结构示意图，如图5所示，该终端包括：反馈单元，用于反馈预编码矩阵信息，所述预编码矩阵信息包括第一基矢量信息、第二基矢量信息、第二系数的幅度和相位信息；其中，反馈预编码子带内的预编码矢量为第一基矢量的线性合并，第一基矢量线性合并所使用的加权系数为第一系数；在所有CSI反馈频带所包含的频域单位内，对应于同一所述第一基矢量的所述第一系数所组成的矢量为第二基矢量的线性合并，所述第二系数为所述第二基矢量进行线性合并所使用的加权系数。
其中，所述第二基矢量表示从DFT矩阵或者过采样DFT矩阵中选择的K个DFT基矢量，所述过采样DFT矩阵的过采样因子O f的取值为以下之一：1、2、4、8。
其中，在反馈单元反馈预编码矩阵信息之前，该终端还包括：处理单元，用于根据信道估计的结果及基站发送的配置信息，得到预编码矩阵信息；所述基站发送的配置信息包括以下至少之一：基站配置的需要反馈CSI的频域范围、配置的第二系数的反馈方式和反馈子集指示、配置的第一幅度反馈惩罚因子及指示需要做功率限制的第一基矢量、配置的第二幅度反馈惩罚因子及指示需要做功率限制的第二基矢量等信息。具体过程如下：
其中，该终端还包括：接收单元，用于在终端反馈预编码矩阵信息之前，接收基站发送的需要反馈CSI的频域范围，所述CSI的频域范围包括：需要反馈CSI的预编码子带或者需要梳状反馈CSI的预编码子带及稀疏程度，或者不需要反馈CSI的预编码子带；处理单元，用于当基站梳状配置需要反馈预编码的子带时，通过以下方式之一修正第二基矢量：方式一：基站配置较大的过采样因子，增加可选择的DFT矩阵数目；方式二：终端根据需要反馈的预编码信息的预编码子带，梳状截取DFT基矢量；或者，当基站配置部分不需要反馈预编码的子带时，截去DFT基矢量中所述不需要反馈预编码信息的预编码子带的对应位置元素；或者，基站配置是否将所述第二基矢量做相位旋转，以配置的最中间的预编码子带为参考。
其中，所述CSI反馈频带是指基站配置需要反馈CSI的频域范围。
其中，所述预编码子带是一个频域单位，其包含的RB个数由基站配置或者由CSI反馈频带包含的RB数目和所使用的第二基矢量数目确定，即包含RB数目为
Figure PCTCN2019128653-appb-000109
RBNum为所述CSI反馈频带内包含的RB数目，K为所述CSI反馈频带使用的第二基矢量数目。
其中，所述第二系数表示为
Figure PCTCN2019128653-appb-000110
的维度为2L×K，2L表示矩阵W 1中包含的2L个第一基矢量，K表示矩阵W 3中包含的K个第二基矢量。
其中，所述接收单元，还用于在终端反馈预编码矩阵信息之前，接收基站发送的第二系数的反馈方式和反馈子集指示；
所述处理单元，还用于通过以下方式之一选择反馈的所述第二系数：方式一：按照第二系数的幅度大小为优先级反馈第二系数；方式二：以所述第一基矢量为优先级反馈第二系数方式三：以所述第二基矢量为优先级反馈第二系数；方式四：将所述第一基矢量分成两组，将第1到第L个第一基矢量，即W 1的第1到第L个列向量，作为第一组，将第L+1到第2L个第一基矢量，即W 1的第L+1到第2L个列向量，作为第二组，根据分组内部的优先级反馈第二系数；方式五：对第二系数进行分组，根据分组的优先级反馈第二系数。
其中，所述方式一包括：以第二系数的幅度大小进行排序，从大到小选取 部分系数，直到所述部分系数的功率之和与第二系数总功率的比值不小于δ，终端反馈所述部分系数，其中δ为预设阈值；使用2L·K大小的比特图指示其选择的系数。
其中，所述方式二包括：方式A：在最优的第一基矢量的K个加权系数中选择l 0个系数，在其余的各第一基矢量的K个加权系数中各选择l 1个系数，l 1＜l 0，使用2L·K大小的比特图指示其选择的系数；或者，方式B：在最优的第一基矢量的K个加权系数中选择l 0个系数，在其余的第一基矢量的所有加权系数中，总共再选择M-l 0个系数进行反馈，M的值由基站配置或者为
Figure PCTCN2019128653-appb-000111
其中δ M为预设阈值；使用以下子方式之一反馈其选择的子集：子方式B-1：使用2L·K大小的比特图指示其选择的系数；子方式B-2：使用
Figure PCTCN2019128653-appb-000112
指示选择的最优第一基矢量，再用
Figure PCTCN2019128653-appb-000113
指示其选择的其余M-l 0个系数；或者，方式C：对2L个第一基矢量进行排序，从强到弱的第一基矢量的加权系数中分别选取l 0,l 0-1，...,l 0-M+1个系数进行反馈，使用2L·K大小的比特图指示其选择的系数。
其中，所述方式三包括：方式D，在最优的第二基矢量的2L个加权系数中选择l 0个系数，在其余的第二基矢量的所有加权系数中，总共再选择M-l 0个系数，M的值由基站配置或者为
Figure PCTCN2019128653-appb-000114
其中δ M为预设阈值；使用以下子方式之一反馈其选择的子集：子方式D-1：使用2L·K大小的比特图指示其选择的系数；子方式D-2：使用
Figure PCTCN2019128653-appb-000115
指示选择的最优第二基矢量，再用
Figure PCTCN2019128653-appb-000116
bits指示其选择的其余M-l 0个系数；或者，方式E：对K个第二基矢量进行排序，从强到弱的第二基矢量的加权系数分别选取l 0,l 0-1，...,l 0-M+1个系数，使用2L·K大小的比特图指示其选择的系数。
其中，所述方式四包括：方式F：方式F：在两个分组内第一基矢量的所有加权系数中，各反馈
Figure PCTCN2019128653-appb-000117
个系数，
Figure PCTCN2019128653-appb-000118
的值由基站配置或者为
Figure PCTCN2019128653-appb-000119
其中δ M为预设阈值；通过比特图或者
Figure PCTCN2019128653-appb-000120
指示选择的系数；或者，方式G：在两个分组内，根据第一基矢量的优先级进行反馈且使用2L·K大小的比特图指示其选择的系数，包括：子方式G-1：在两个分组各选择一个最优的第一基矢量，两个最优的第一基矢量的加权系数中各选择l 0个系数，在两个分组内的其余第一基矢量的所有加权系数中各再选择
Figure PCTCN2019128653-appb-000121
个系数，
Figure PCTCN2019128653-appb-000122
的值由基站配置或者为
Figure PCTCN2019128653-appb-000123
其中δ M为预设阈值；或者子方式G-2：在两个分组各选择一个最优的第一基矢量，两个最优的第一基矢量的加权系数中各选择l 0个系数，在两个分组内的其余第一基矢量的加权系数中各再选择l 1个系数，l 1＜l 0；或者，方式H：在两个分组内，根据第二基矢量的优先级进行反馈且使用2L·K大小的比特图指 示其选择的系数，包括：子方式H-1：在两个分组各选择一个最优的第二基矢量，第1个分组内最优的第二基矢量的加权系数中选择l 0个系数，第2个分组内的最优的第二基矢量的加权系数中选择l 0个系数，在两个分组内的其余第二基矢量的所有加权系数中各再选择
Figure PCTCN2019128653-appb-000124
个系数，
Figure PCTCN2019128653-appb-000125
的值由基站配置或者为
Figure PCTCN2019128653-appb-000126
其中δ M为预设阈值；或者子方式H-2：在两个分组各选择一个最优的第二基矢量，第1个分组内的最优的第二基矢量的加权系数中选择l 0个系数，第2个分组内的最优的第二基矢量的加权系数中选择l 0个系数，在两个分组内的其余第二基矢量的加权系数中各再选择l 1个系数，l 1＜l 0。
其中，所述方式五包括：方式I：将第二系数分成GroupNumber个组合，终端选择l个最合适的组合进行反馈，使用
Figure PCTCN2019128653-appb-000127
指示其选择的组合；或者，方式J：将第二系数分成个组合，选择1个最合适的组合，在其余组合中再选择l-1个组合，使用log 2GroupNumberbits指示选择的最合适的组合，使用
Figure PCTCN2019128653-appb-000128
指示其余的组合。
其中，所述接收单元，还用于在终端反馈预编码矩阵信息之前，接收基站发送的第一信令；所述第一信令包括基站配置的第一幅度反馈惩罚因子及指示需要做功率限制的第一基矢量，用于对所述需要做功率限制的第一基矢量的加权系数做功率限制；所述处理单元，还用于限制反馈的所述需要做功率限制的第一基矢量的加权系数幅度量化之后不能超过基站配置的第一幅度反馈惩罚因子，或者将所述需要做功率限制的第一基矢量的加权系数的幅度乘上基站配置的第一幅度反馈惩罚因子之后再进行量化；和/或，所述接收单元，还用于在终端反馈预编码矩阵信息之前，接收基站发送的第二信令；所述第二信令包括基站配置的第二幅度反馈惩罚因子及指示需要做功率限制的第二基矢量，用于对所述需要做功率限制的第二基矢量的加权系数做功率限制；所述处理单元，还用于限制反馈的所述需要做功率限制的第二基矢量的加权系数幅度量化之后不能超过基站配置的第二幅度反馈惩罚因子，或者将所述需要做功率限制的第二基矢量的加权系数的幅度乘上基站配置的第二幅度反馈惩罚因子之后再进行量化。
其中，所述处理单元，还用于当反馈资源不足且选择的第二基矢量数目K由终端配置时，动态地减小反馈的第二基矢量数目；或者，当选择的第二基矢量数目K由基站配置时，如果反馈资源充足，则上传所有基站配置的参数，否则选择以下方式之一：当按照所述方式一反馈预编码矩阵信息时，如果反馈资源不足，则基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；或者，当按照所述方式二包括的方式A或方式C反馈预编码矩 阵信息时，如果反馈资源不足，终端优先反馈最优第一基矢量的加权系数，接着终端基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；当所述方式二包括的方式B反馈预编码矩阵信息且使用子方式B-1反馈子集指示时，如果反馈资源不足，终端优先反馈最优第一基矢量的加权系数，接着终端基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；当按照所述方式二包括的方式B反馈预编码矩阵信息且使用子方式B-2反馈子集指示时，如果反馈资源不足，终端只反馈最优第一基矢量的加权系数；或者，当按照所述方式三包括的方式D反馈预编码矩阵信息且使用子方式D-1反馈子集指示时，如果反馈资源不足，终端优先反馈最优第二基矢量的加权系数，接着终端基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；当按照所述方式三包括的方式D反馈预编码矩阵信息且使用子方式D-2反馈子集指示时，如果反馈资源不足，终端只反馈最优第二基矢量的加权系数；当按照所述方式三包括的方式E反馈预编码矩阵信息时，如果反馈资源不足，终端优先反馈最优第二基矢量的加权系数，接着终端基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；或者，当按照所述方式四包括的方式F反馈预编码矩阵信息时，如果反馈资源不足，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；当按照所述方式四包括的方式G中包括的子方式G-1或子方式G-2反馈预编码矩阵信息时，如果反馈资源不足，优先反馈第一个分组中最优第一基矢量的加权系数，其次反馈第二分组内最优的第一基矢量的加权系数，接着基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；当按照所述方式四包括的方式H中包括的子方式H-1或子方式H-2反馈预编码矩阵信息时，如果反馈资源不足，优先反馈第一分组内最优第二基矢量的加权系数，其次反馈第二分组内最优的第二基矢量的加权系数，接着基于一定准则，动态地通过比特图通知其选择的系数直至到达最大反馈资源为止；或者，当按照所述方式五包括的方式I反馈预编码矩阵信息且使用比特图反馈选择的分组时，如果反馈资源不足，基于一定准则动态地根据比特图通知基站其选择的分组直至到达最大反馈资源为止；当按照所述方式五包括的方式J反馈预编码矩阵信息时，如果反馈资源不足，优先反馈最优分组对应的加权系数。
实施方式四
图6为本发明实施方式四提供的基站的结构示意图，如图6所示，该基站包括：接收单元，用于接收终端反馈的预编码矩阵信息，所述预编码矩阵信息包括第一基矢量信息、第二基矢量信息、第二系数的幅度和相位信息；其中，反馈预编码子带内的预编码矢量为第一基矢量的线性合并，第一基矢量线性合并所使用的加权系数为第一系数；在所有CSI反馈频带所包含的频域单位内， 对应于同一所述第一基矢量的所述第一系数所组成的矢量为第二基矢量的线性合并，所述第二系数为所述第二基矢量进行线性合并所使用的加权系数。
其中，该基站还包括：配置单元，用于在基站接收终端反馈的预编码矩阵信息之前，配置需要反馈CSI的频域范围；发送单元，用于将配置的需要反馈CSI的频域范围发送给终端，以便所述终端根据所述CSI的频域范围确定需要反馈预编码信息的预编码子带。
其中，所述配置的CSI的频域范围包括：需要反馈CSI的预编码子带或者需要梳状反馈CSI的预编码子带及稀疏程度，或者不需要反馈CSI的预编码子带。
其中，配置单元，还用于在基站接收终端反馈的预编码矩阵信息之前，配置第二系数的反馈方式和反馈子集指示；所述发送单元，还用于将配置的第二系数的反馈方式和反馈子集指示发送给终端，以便终端根据所述第二系数的反馈方式和反馈子集指示反馈第二系数；所述第二系数的反馈方式包括以下之一：方式一：按照第二系数的幅度大小反馈第二系数；方式二：以所述第一基矢量为优先级反馈第二系数；方式三：以所述第二基矢量为优先级反馈第二系数；方式四：将所述第一基矢量分成两组，将第1到第L个第一基矢量，即W 1的第1到第L个列向量，作为第一组，将第L+1到第2L个第一基矢量，即W 1的第L+1到第2L个列向量，作为第二组，根据分组内部的优先级反馈第二系数；方式五：对第二系数进行分组，根据分组的优先级反馈第二系数；所述反馈子集指示用于指示终端反馈所述第二系数的子集的方式。
其中，其中，所述第二系数表示为
Figure PCTCN2019128653-appb-000129
的维度为2L×K，2L表示矩阵W 1中包含的2L个第一基矢量，K表示矩阵W 3中包含的K个第二基矢量。
其中，所述方式一包括：以第二系数的幅度大小进行排序，从大到小选取部分系数功率，直到所述部分系数的功率之和与第二系数总功率的比值不小于δ，终端反馈所述部分系数，其中δ为预设阈值；使用2L·K大小的比特图指示其选择的系数。
其中，所述方式二包括：方式A：在最优的第一基矢量对应的K个系数中选择l 0个系数，在其余的各第一基矢量对应的K个系数中各选择l 1个系数，l 1＜l 0，使用2L·K大小的比特图指示其选择的系数；或者，方式B：在最优的第一基矢量对应的K个系数中选择l 0个系数，在其余的第一基矢量对应的所有系数中，总共再选择-l 0个系数进行反馈，M的值由基站配置或者为
Figure PCTCN2019128653-appb-000130
其中δ M为预设阈值；使用以下子方式之一反馈其选择的子集：子方式B-1：使用2L·K大小的比特图指示其选择的系数；子方式B-2：使用
Figure PCTCN2019128653-appb-000131
指示选择的最优第一基矢量，再用
Figure PCTCN2019128653-appb-000132
指示其选择的其余M-l 0个系数；或者，方式 C：对2L个第一基矢量进行排序，从强到弱的第一基矢量对应的系数中分别选取l 0,l 0-1，...,l 0-M+1个系数进行反馈，使用2L·K大小的比特图指示其选择的系数。
其中，所述方式三包括：方式D，在最优的第二基矢量对应的2L个系数中选择l 0个系数，在其余的第二基矢量对应的所有系数中，总共再选择M-l 0个系数，M的值由基站配置或者为
Figure PCTCN2019128653-appb-000133
其中δ M为预设阈值；使用以下子方式之一反馈其选择的子集：子方式D-1：使用2L·K大小的比特图指示其选择的系数；子方式D-2：使用
Figure PCTCN2019128653-appb-000134
指示选择的最优第二基矢量，再用
Figure PCTCN2019128653-appb-000135
bits指示其选择的其余M-l 0个系数；或者，方式E：对K个第二基矢量进行排序，从强到弱的第二基矢量对应的系数分别选取l 0,l 0-1，...,l 0-M+1个系数，使用2L·K大小的比特图指示其选择的系数。
其中，所述方式四包括：方式F：在两个分组内，各反馈
Figure PCTCN2019128653-appb-000136
个系数，
Figure PCTCN2019128653-appb-000137
的值由基站配置或者为
Figure PCTCN2019128653-appb-000138
其中δ M为预设阈值；通过比特图或者
Figure PCTCN2019128653-appb-000139
指示选择的系数；或者，方式G：在两个分组内，根据第一基矢量的优先级进行反馈且使用2L·K大小的比特图指示其选择的系数，包括：子方式G-1：在两个分组各选择一个最优的第一基矢量，两个最优的第一基矢量的对应系数中各选择l 0个系数，在两个分组内的其余第一基矢量对应所有系数中各再选择
Figure PCTCN2019128653-appb-000140
个系数，
Figure PCTCN2019128653-appb-000141
的值由基站配置或者为
Figure PCTCN2019128653-appb-000142
其中δ M为预设阈值；或者子方式G-2：在两个分组各选择一个最优的第一基矢量，两个最优的第一基矢量的对应系数中各选择l 0个系数，在两个分组内的其余第一基矢量对应系数中各再选择l 1个系数，l 1＜l 0；或者，方式H：在两个分组内，根据第二基矢量的优先级进行反馈且使用2L·K大小的比特图指示其选择的系数，包括：子方式H-1：在两个分组各选择一个最优的第二基矢量，第1个分组内最优的第二基矢量的对应的系数中选择l 0个系数，第2个分组内的最优的第二基矢量的对应系数中选择l 0个系数，在两个分组内的其余第二基矢量对应所有系数中各再选择
Figure PCTCN2019128653-appb-000143
个系数，
Figure PCTCN2019128653-appb-000144
的值由基站配置或者为
Figure PCTCN2019128653-appb-000145
其中δ M为预设阈值；或者子方式H-2：在两个分组各选择一个最优的第二基矢量，第1个分组内的最优的第二基矢量的对应系数中选择l 0个系数，第2个分组内的最优的第二基矢量的对应系数中选择l 0个系数，在两个分组内的其余第二基矢量对应系数中各再选择l 1个系数，l 1＜l 0。
其中，所述方式五包括：方式I：将第二系数分成GroupNumber个组合，终端选择l个最合适的组合进行反馈，使用
Figure PCTCN2019128653-appb-000146
指示其选择的组合； 或者，方式J：将第二系数分成GroupNumber个组合，选择1个最合适的组合，在其余组合中再选择l-1个组合，使用log 2GroupNumberbits指示选择的最合适的组合，使用
Figure PCTCN2019128653-appb-000147
指示其余的组合。
其中，所述配置单元，还用于配置第一幅度反馈惩罚因子及指示需要做功率限制的第一基矢量；所述发送单元，还用于在基站接收终端反馈的预编码矩阵信息之前，向终端发送第一信令；其中，所述第一信令包括基站配置的第一幅度反馈惩罚因子及指示需要做功率限制的第一基矢量，用于对所述需要做功率限制的第一基矢量对应系数的功率做限制；和/或，所述配置单元，还用于配置第二幅度反馈惩罚因子及指示需要做功率限制的第二基矢量；所述发送单元，还用于在基站接收终端反馈的预编码矩阵信息之前，向终端发送第二信令；其中，所述第二信令包括基站配置的第二幅度反馈惩罚因子及指示需要做功率限制的第二基矢量，用于对所述需要做功率限制的第二基矢量对应系数的功率做限制。
其中，所述第一幅度反馈惩罚因子为
Figure PCTCN2019128653-appb-000148
或者
Figure PCTCN2019128653-appb-000149
或者，所述第二幅度反馈惩罚因子为
Figure PCTCN2019128653-appb-000150
或者
Figure PCTCN2019128653-appb-000151
本发明实施例还提供了一种终端，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述计算机程序被所述处理器执行时实现上述任一项终端执行的信道状态信息CSI反馈的方法。
本发明实施例还提供了一种基站，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述计算机程序被所述处理器执行时实现上述任一项基站执行的信道状态信息CSI反馈的方法。
本发明实施例还提供了一种计算机可读存储介质，所述计算机可读存储介质上存储有信息处理程序，所述信息处理程序被处理器执行时实现上述任一项所述信道状态信息CSI反馈的方法。
本领域普通技术人员可以理解，上文中所公开方法中的全部或某些步骤、系统、装置中的功能模块/单元可以被实施为软件、固件、硬件及其适当的组合。在硬件实施方式中，在以上描述中提及的功能模块/单元之间的划分不一定对应于物理组件的划分；例如，一个物理组件可以具有多个功能，或者一个功能或步骤可以由若干物理组件合作执行。某些组件或所有组件可以被实施为由处理 器，如数字信号处理器或微处理器执行的软件，或者被实施为硬件，或者被实施为集成电路，如专用集成电路。这样的软件可以分布在计算机可读介质上，计算机可读介质可以包括计算机存储介质(或非暂时性介质)和通信介质(或暂时性介质)。如本领域普通技术人员公知的，术语计算机存储介质包括在用于存储信息(诸如计算机可读指令、数据结构、程序模块或其他数据)的任何方法或技术中实施的易失性和非易失性、可移除和不可移除介质。计算机存储介质包括但不限于随机存取存储器(Random Access Memory，RAM)、只读存储器(Read-Only Memory，ROM)、带电可擦可编程只读存储器(Electrically Erasable Programmable Read-Only Memory，EEPROM)、闪存或其他存储器技术、光盘只读存储器(Compact Disc Read-Only Memory，CD-ROM)、数字多功能盘(Digital Video Disc，DVD)或其他光盘存储、磁盒、磁带、磁盘存储或其他磁存储装置、或者可以用于存储期望的信息并且可以被计算机访问的任何其他的介质。此外，本领域普通技术人员公知的是，通信介质通常包含计算机可读指令、数据结构、程序模块或者诸如载波或其他传输机制之类的调制数据信号中的其他数据，并且可包括任何信息递送介质。
虽然本申请所揭露的实施方式如上，但所述的内容仅为便于理解本申请而采用的实施方式，并非用以限定本申请。任何本申请所属领域内的技术人员，在不脱离本申请所揭露的精神和范围的前提下，可以在实施的形式及细节上进行任何的修改与变化。

Claims (25)

  1. 一种信道状态信息CSI反馈方法，包括：
    终端反馈预编码矩阵信息，所述预编码矩阵信息包括第一基矢量信息、第二基矢量信息、第二系数的幅度和相位信息；
    其中，反馈预编码子带内的预编码矢量为一组第一基矢量的线性合并，一组第一基矢量线性合并所使用的加权系数为第一系数；在所有CSI反馈频带所包含的频域单位内，对应于同一组第一基矢量的第一系数所组成的矢量为一组第二基矢量的线性合并，所述第二系数为一组第二基矢量进行线性合并所使用的加权系数。
  2. 根据权利要求1所述的方法，其中，所述一组第二基矢量表示从离散傅里叶变换DFT矩阵或者过采样DFT矩阵中选择的K个DFT基矢量，所述过采样DFT矩阵的过采样因子O f的取值为以下之一：1、2、4、8。
  3. 根据权利要求2所述的方法，在所述终端反馈预编码矩阵信息之前，还包括：
    接收基站发送的需要反馈CSI的频域范围，所述需要反馈CSI的频域范围包括：需要反馈CSI的预编码子带，或者需要梳状反馈CSI的预编码子带及稀疏程度，或者不需要反馈CSI的预编码子带；
    在所述基站梳状配置需要反馈预编码信息的预编码子带的情况下，通过以下方式之一修正第二基矢量：
    方式一：所述基站配置过采样因子大于1；
    方式二：所述终端根据所述需要反馈的预编码信息的预编码子带，梳状截取DFT基矢量；
    或者，在所述基站配置部分不需要反馈预编码的子带的情况下，截去DFT基矢量中所述不需要反馈预编码信息的预编码子带的对应位置元素；
    或者，所述基站配置将DFT基矢量做相位旋转，以配置的中间的预编码子带为参考。
  4. 根据权利要求1所述的方法，其中，所述所有CSI反馈频带是指基站配置需要反馈CSI的频域范围。
  5. 根据权利要求1所述的方法，其中，所述预编码子带是一个频域单位，所述预编码子带包含的资源块RB个数由基站配置或者由CSI反馈频带包含的RB的数目和所使用的第二基矢量的数目确定，所述预编码子带包含的RB数目的计算公式为
    Figure PCTCN2019128653-appb-100001
    其中，S为所述预编码子带包含的RB数目，RBNum 为所述CSI反馈频带内包含的RB的数目，K为所述CSI反馈频带使用的第二基矢量的数目。
  6. 根据权利要求1所述的方法，其中，所述第二系数表示为
    Figure PCTCN2019128653-appb-100002
    的维度为2L×K，2L表示矩阵W 1中包含的第一基矢量的数目，K表示矩阵W 3中包含的K个第二基矢量的数目，其中，W 1为一组第一基矢量，W 3为一组第二基矢量。
  7. 根据权利要求6所述的方法，在所述终端反馈预编码矩阵信息之前，还包括：
    接收基站发送的第二系数的反馈方式和反馈子集指示；
    所述预编码矩阵信息中的第二系数通过以下方式之一进行反馈：
    方式一：按照第二系数的幅度大小为优先级反馈第二系数；
    方式二：以第一基矢量为优先级反馈第二系数；
    方式三：以第二基矢量为优先级反馈第二系数；
    方式四：将2L个第一基矢量分成两组，将第1个第一基矢量到第L个第一基矢量，作为第一组，将第L+1个第一基矢量到第2L个第一基矢量，作为第二组，根据分组内部的优先级反馈第二系数，其中，所述第1个第一基矢量到所述第L个第一基矢量为W 1的第1个列向量到第L个列向量，所述第L+1个第一基矢量到所述第2L个第一基矢量为W 1的第L+1个列向量到第2L个列向量；
    方式五：对第二系数进行分组，根据分组的优先级反馈第二系数。
  8. 根据权利要求7所述的方法，其中，所述方式一包括：
    以第二系数的幅度从大到小的顺序进行排序，从排序结果中选取部分系数，直到所述部分系数的功率之和与所述第二系数总功率的比值不小于δ，所述终端反馈所述部分系数，其中，δ为预设阈值；
    使用2L·K大小的比特图指示选择的系数。
  9. 根据权利要求7所述的方法，其中，所述方式二包括：
    方式A：在选择的第一基矢量的K个加权系数中选择l 0个系数，在除所述选择的第一基矢量外的第一基矢量中的每个第一基矢量的K个加权系数中选择l 1个系数，l 0大于l 1且l 0小于K，使用2L·K大小的比特图指示选择的系数；
    或者，方式B：在选择的第一基矢量的K个加权系数中选择l 0个系数，在除所述选择的第一基矢量外的第一基矢量的所有加权系数中，总共选择M-l 0个系数进行反馈，M由所述基站配置或者为
    Figure PCTCN2019128653-appb-100003
    其中，l 0小于K，δ M为预设阈值；
    使用以下子方式之一反馈选择的子集：
    子方式B-1：使用2L·K大小的比特图指示选择的系数；
    子方式B-2：使用
    Figure PCTCN2019128653-appb-100004
    比特指示所述选择的第一基矢量，用
    Figure PCTCN2019128653-appb-100005
    比特指示选择的所述M-l 0个系数；
    或者，方式C：根据2L个第一基矢量对应的波束从强到弱的顺序对所述2L个第一基矢量进行排序，从排序后的2L个第一基矢量的加权系数中分别选取l 0,l 0-1，...,l 0-M+1个系数进行反馈，使用2L·K大小的比特图指示选择的系数。
  10. 根据权利要求7所述的方法，其中，所述方式三包括：
    方式D，在选择的第二基矢量的2L个加权系数中选择l 0个系数，在除所述选择的第二基矢量外的第二基矢量的所有加权系数中，总共选择M-l 0个系数，M由所述基站配置或者为
    Figure PCTCN2019128653-appb-100006
    其中，l 0小于K，δ M为预设阈值；
    使用以下子方式之一反馈选择的子集：
    子方式D-1：使用2L·K大小的比特图指示选择的系数；
    子方式D-2：使用
    Figure PCTCN2019128653-appb-100007
    比特指示所述选择的第二基矢量，用
    Figure PCTCN2019128653-appb-100008
    比特指示选择的所述M-l 0个系数；
    或者，方式E：根据K个第二基矢量对应的时延路径从强到弱的顺序对所述K个第二基矢量进行排序，从排序后的K个第二基矢量的加权系数分别选取l 0,l 0-1，...,l 0-M+1个系数，使用2L·K大小的比特图指示选择的系数。
  11. 根据权利要求7所述的方法，其中，所述根据分组内部的优先级反馈第二系数，包括：
    方式F：在每个分组内第一基矢量的所有加权系数中，反馈
    Figure PCTCN2019128653-appb-100009
    个系数，
    Figure PCTCN2019128653-appb-100010
    由基站配置或者为
    Figure PCTCN2019128653-appb-100011
    其中，δ M为预设阈值；通过比特图或者
    Figure PCTCN2019128653-appb-100012
    bits比特指示选择的系数；
    或者，方式G：在两个分组内，根据第一基矢量的优先级进行反馈且使用2L·K大小的比特图指示选择的系数，包括：
    子方式G-1：在每个分组中选择一个第一基矢量，在每个选择的第一基矢量的加权系数中选择l 0个系数，在每个分组内除选择的第一基矢量外的第一基矢量的所有加权系数中选择
    Figure PCTCN2019128653-appb-100013
    个系数，
    Figure PCTCN2019128653-appb-100014
    由所述基站配置或者为
    Figure PCTCN2019128653-appb-100015
    其中，δ M为预设阈值；或者，子方式G-2：在每个分组中选择一个第一基矢量，在每个选择的的第一基矢量的加权系数中选择l 0个系数，在每个分组内除选择的第一基矢量外的第一基矢量中的每个第一基矢量的加权系数中选择l 1个系数， l 1＜l 0；
    或者，方式H：在两个分组内，根据第二基矢量的优先级进行反馈且使用2L·K大小的比特图指示其选择的系数，包括：
    子方式H-1：在每个分组中选择一个第二基矢量，在所述第一分组内的选择的第二基矢量的加权系数中选择l 0个系数，在所述第二分组内的选择的第二基矢量的加权系数中选择l 0个系数，在每个分组内除选择的第二基矢量外的第二基矢量的所有加权系数中选择
    Figure PCTCN2019128653-appb-100016
    个系数，
    Figure PCTCN2019128653-appb-100017
    由所述基站配置或者为
    Figure PCTCN2019128653-appb-100018
    其中，δ M为预设阈值；或者，子方式H-2：在每个分组中选择一个第二基矢量，在所述第一分组内的选择的第二基矢量的加权系数中选择l 0个系数，在所述第二分组内的选择的第二基矢量的加权系数中选择l 0个系数，在每个分组内除选择的第二基矢量外的第二基矢量中的每个第二基矢量的加权系数中选择l 1个系数，l 1＜l 0。
  12. 根据权利要求7所述的方法，其中，所述方式五包括：
    方式I：将第二系数分成GroupNumber个组合，所述终端选择l个组合进行反馈，使用
    Figure PCTCN2019128653-appb-100019
    比特指示选择的组合，其中，l小于GroupNumber；
    或者，方式J：将第二系数分成GroupNumber个组合，选择1个组合，在除选择的组合外的组合中选择l-1个组合，使用log 2GroupNumber比特指示所述选择的组合，使用
    Figure PCTCN2019128653-appb-100020
    比特指示除选择的组合外的的组合，其中，l小于GroupNumber。
  13. 根据权利要求1所述的方法，在所述终端反馈预编码矩阵信息之前，还包括以下至少之一：
    接收基站发送的第一信令；所述第一信令包括所述基站配置的第一幅度反馈惩罚因子及指示需要做功率限制的第一基矢量，其中，所述第一信令用于对所述需要做功率限制的第一基矢量的加权系数做功率限制；所述终端反馈的所述需要做功率限制的第一基矢量的加权系数幅度量化之后不能超过所述基站配置的所述第一幅度反馈惩罚因子，或者将所述需要做功率限制的第一基矢量的加权系数的幅度乘以所述基站配置的所述第一幅度反馈惩罚因子之后再进行量化；
    接收基站发送的第二信令；所述第二信令包括所述基站配置的第二幅度反馈惩罚因子及指示需要做功率限制的第二基矢量，其中，所述第二信令用于对所述需要做功率限制的第二基矢量的加权系数做功率限制；所述终端反馈的所 述需要做功率限制的第二基矢量的加权系数幅度量化之后不能超过所述基站配置的所述第二幅度反馈惩罚因子，或者将所述需要做功率限制的第二基矢量的加权系数的幅度乘以所述基站配置的所述第二幅度反馈惩罚因子之后再进行量化。
  14. 根据权利要求7-12任一项所述的方法，其中，所述反馈预编码矩阵信息，包括：
    在反馈资源不足且选择的第二基矢量数目K由所述终端配置的情况下，减小反馈的第二基矢量数目；
    或者，在选择的第二基矢量数目K由所述基站配置，且反馈资源充足的情况下，上传所有基站配置的参数，在不是选择的第二基矢量数目K由所述基站配置，且反馈资源充足的情况下，选择以下方式之一：
    在按照所述方式一反馈预编码矩阵信息，且反馈资源不足的情况下，基于准则，通过比特图通知选择的系数直至到达最大反馈资源为止；
    在按照所述方式二包括的方式A或方式C反馈预编码矩阵信息，且反馈资源不足的情况下，所述终端反馈选择的第一基矢量的加权系数，所述终端基于准则，通过比特图通知选择的系数直至到达最大反馈资源为止；
    在按照所述方式二包括的方式B反馈预编码矩阵信息且使用子方式B-1反馈子集指示，且反馈资源不足的情况下，所述终端反馈选择的第一基矢量的加权系数，所述终端基于一定准则，通过比特图通知选择的系数直至到达最大反馈资源为止；
    在按照所述方式二包括的方式B反馈预编码矩阵信息且使用子方式B-2反馈子集指示，且反馈资源不足的情况下，所述终端只反馈选择的第一基矢量的加权系数；
    在按照所述方式三包括的方式D反馈预编码矩阵信息且使用子方式D-1反馈子集指示，且反馈资源不足的情况下，所述终端反馈选择的第二基矢量的加权系数，所述终端基于准则，通过比特图通知选择的系数直至到达最大反馈资源为止；
    在按照所述方式三包括的方式D反馈预编码矩阵信息且使用子方式D-2反馈子集指示，且反馈资源不足的情况下，所述终端只反馈选择的第二基矢量的加权系数；
    在按照所述方式三包括的方式E反馈预编码矩阵信息，且反馈资源不足的情况下，所述终端反馈选择的第二基矢量的加权系数，所述终端基于准则，通过比特图通知选择的系数直至到达最大反馈资源为止；
    在按照所述方式四包括的方式F反馈预编码矩阵信息，且反馈资源不足的情况下，通过比特图通知选择的系数直至到达最大反馈资源为止；
    在按照所述方式四包括的方式G中包括的子方式G-1或子方式G-2反馈预编码矩阵信息，且反馈资源不足的情况下，依次反馈所述第一个分组内的选择的第一基矢量的加权系数和所述第二分组内的选择的第一基矢量的加权系数，基于准则，通过比特图通知选择的系数直至到达最大反馈资源为止；
    在按照所述方式四包括的方式H中包括的子方式H-1或子方式H-2反馈预编码矩阵信息，且反馈资源不足的情况下，依次反馈所述第一分组内的选择的第二基矢量的加权系数和所述第二分组内的选择的第二基矢量的加权系数，基于准则，通过比特图通知选择的系数直至到达最大反馈资源为止；
    在按照所述方式五包括的方式I反馈预编码矩阵信息且使用比特图反馈选择的分组，且反馈资源不足的情况下，基于准则，根据比特图通知基站选择的分组直至到达最大反馈资源为止；
    在按照所述方式五包括的方式J反馈预编码矩阵信息，且反馈资源不足的情况下，反馈选择的分组对应的加权系数。
  15. 一种信道状态信息CSI反馈方法，包括：
    基站接收终端反馈的预编码矩阵信息，所述预编码矩阵信息包括第一基矢量信息、第二基矢量信息、第二系数的幅度和相位信息；
    其中，反馈预编码子带内的预编码矢量为一组第一基矢量的线性合并，一组第一基矢量线性合并所使用的加权系数为第一系数；在所有CSI反馈频带所包含的频域单位内，对应于同一组第一基矢量的第一系数所组成的矢量为一组第二基矢量的线性合并，所述第二系数为一组第二基矢量进行线性合并所使用的加权系数。
  16. 根据权利要求15所述的方法，在所述基站接收终端反馈的预编码矩阵信息之前，还包括：
    所述基站配置需要反馈CSI的频域范围并发送给所述终端，以使所述终端根据所述CSI的频域范围确定需要反馈预编码信息的预编码子带。
  17. 根据权利要求16所述的方法，其中，所述配置的需要反馈CSI的频域范围包括：需要反馈CSI的预编码子带或者需要梳状反馈CSI的预编码子带及稀疏程度，或者不需要反馈CSI的预编码子带。
  18. 根据权利要求15所述的方法，在所述基站接收终端反馈的预编码矩阵信息之前，还包括：
    所述基站配置第二系数的反馈方式和反馈子集指示，并发送给所述终端，以使所述终端根据所述第二系数的反馈方式和所述反馈子集指示反馈第二系数；
    所述第二系数的反馈方式包括以下之一：
    方式一：按照第二系数的幅度大小为优先级反馈第二系数；
    方式二：以第一基矢量为优先级反馈第二系数；
    方式三：以第二基矢量为优先级反馈第二系数；
    方式四：将2L个第一基矢量分成两组，将第1个第一基矢量到第L个第一基矢量，作为第一组，将第L+1个第一基矢量到第2L个第一基矢量，作为第二组，根据分组内部的优先级反馈第二系数，其中，2L表示矩阵W 1中包含的第一基矢量的数目，W 1为一组第一基矢量，所述第1个第一基矢量到所述第L个第一基矢量为W 1的第1个列向量到第L个列向量，所述第L+1个第一基矢量到所述第2L个第一基矢量为W 1的第L+1个列向量到第2L个列向量；
    方式五：对第二系数进行分组，根据分组的优先级反馈第二系数；
    所述反馈子集指示用于指示终端反馈所述第二系数的子集的方式。
  19. 根据权利要求15所述的方法，在所述基站接收终端反馈的预编码矩阵信息之前，还包括：
    向所述终端发送第一信令和第二信令中的至少之一；
    其中，所述第一信令包括所述基站配置的第一幅度反馈惩罚因子及指示需要做功率限制的第一基矢量，其中，所述第一信令用于对所述需要做功率限制的第一基矢量的加权系数做功率限制；所述第二信令包括所述基站配置的第二幅度反馈惩罚因子及指示需要做功率限制的第二基矢量，其中，所述第二信令用于对所述需要做功率限制的第二基矢量的加权系数做功率限制。
  20. 根据权利要求19所述的方法，其中，
    所述第一幅度反馈惩罚因子为
    Figure PCTCN2019128653-appb-100021
    或者
    Figure PCTCN2019128653-appb-100022
    或者，所述第二幅度反馈惩罚因子为
    Figure PCTCN2019128653-appb-100023
    或者
    Figure PCTCN2019128653-appb-100024
  21. 一种终端，包括：
    反馈单元，设置为反馈预编码矩阵信息，所述预编码矩阵信息包括第一基矢量信息、第二基矢量信息、第二系数的幅度和相位信息；
    其中，反馈预编码子带内的预编码矢量为一组第一基矢量的线性合并，一组第一基矢量线性合并所使用的加权系数为第一系数；在所有信道状态信息CSI反馈频带所包含的频域单位内，对应于同一组第一基矢量的第一系数所组成的矢量为一组第二基矢量的线性合并，所述第二系数为一组第二基矢量进行线性合并所使用的加权系数。
  22. 一种基站，包括：
    接收单元，设置为接收终端反馈的预编码矩阵信息，所述预编码矩阵信息包括第一基矢量信息、第二基矢量信息、第二系数的幅度和相位信息；
    其中，反馈预编码子带内的预编码矢量为一组第一基矢量的线性合并，一组第一基矢量线性合并所使用的加权系数为第一系数；在所有信道状态信息CSI反馈频带所包含的频域单位内，对应于同一组第一基矢量的第一系数所组成的矢量为一组第二基矢量的线性合并，所述第二系数为一组第二基矢量进行线性合并所使用的加权系数。
  23. 一种终端，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述计算机程序被所述处理器执行时实现如权利要求1至14中任一项所述信道状态信息CSI反馈方法。
  24. 一种基站，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述计算机程序被所述处理器执行时实现如权利要求15至20中任一项所述信道状态信息CSI反馈方法。
  25. 一种计算机可读存储介质，存储有信息处理程序，所述信息处理程序被处理器执行时实现如权利要求1至20中任一项所述信道状态信息CSI反馈方法。
PCT/CN2019/128653 2018-12-28 2019-12-26 信道状态信息csi反馈方法、设备及存储介质 WO2020135573A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811625788.X 2018-12-28
CN201811625788.XA CN110535498B (zh) 2018-12-28 2018-12-28 信道状态信息csi反馈方法及设备

Publications (1)

Publication Number Publication Date
WO2020135573A1 true WO2020135573A1 (zh) 2020-07-02

Family

ID=68659809

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/128653 WO2020135573A1 (zh) 2018-12-28 2019-12-26 信道状态信息csi反馈方法、设备及存储介质

Country Status (2)

Country Link
CN (1) CN110535498B (zh)
WO (1) WO2020135573A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2024061263A1 (zh) * 2022-09-21 2024-03-28 维沃移动通信有限公司 Pmi组合系数的发送方法、装置及终端

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110535498B (zh) * 2018-12-28 2022-10-18 中兴通讯股份有限公司 信道状态信息csi反馈方法及设备
CN110535504A (zh) * 2019-03-29 2019-12-03 中兴通讯股份有限公司 信息映射方法、获取方法及相关设备
US20230066978A1 (en) * 2019-12-27 2023-03-02 Qiaoyu Li Uplink grant downlink control information for frequency domain compressed uplink precoding
CN113508538B (zh) * 2020-02-05 2023-06-23 上海诺基亚贝尔股份有限公司 描绘每路径角度和延迟信息的信道状态信息（csi）反馈增强
CN113783589B (zh) * 2020-06-09 2023-03-10 华为技术有限公司 信道状态信息的传输方法及相关装置

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108271265A (zh) * 2017-01-03 2018-07-10 华为技术有限公司 通信方法、基站和终端设备
WO2018223794A1 (zh) * 2017-06-08 2018-12-13 索尼公司 无线通信系统中的电子设备、通信方法和存储介质
CN110535498A (zh) * 2018-12-28 2019-12-03 中兴通讯股份有限公司 信道状态信息csi反馈方法及设备
CN110535513A (zh) * 2018-11-02 2019-12-03 中兴通讯股份有限公司 一种csi反馈及接收方法、装置、存储介质

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5426023B2 (ja) * 2009-06-23 2014-02-26 アルカテル−ルーセント チャネル状態情報のための方法および装置
CN106160926B (zh) * 2015-04-08 2019-12-24 中兴通讯股份有限公司 在多输入多输出系统中反馈信道状态信息的方法和装置
WO2018029644A2 (en) * 2016-08-12 2018-02-15 Telefonaktiebolaget Lm Ericsson (Publ) Progressive advanced csi feedback
CN107888265B (zh) * 2016-09-30 2019-08-23 中兴通讯股份有限公司 信道信息的反馈方法及装置、终端、基站

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108271265A (zh) * 2017-01-03 2018-07-10 华为技术有限公司 通信方法、基站和终端设备
WO2018223794A1 (zh) * 2017-06-08 2018-12-13 索尼公司 无线通信系统中的电子设备、通信方法和存储介质
CN110535513A (zh) * 2018-11-02 2019-12-03 中兴通讯股份有限公司 一种csi反馈及接收方法、装置、存储介质
CN110535498A (zh) * 2018-12-28 2019-12-03 中兴通讯股份有限公司 信道状态信息csi反馈方法及设备

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2024061263A1 (zh) * 2022-09-21 2024-03-28 维沃移动通信有限公司 Pmi组合系数的发送方法、装置及终端

Also Published As

Publication number Publication date
CN110535498B (zh) 2022-10-18
CN110535498A (zh) 2019-12-03

Similar Documents

Publication Publication Date Title
WO2020135573A1 (zh) 信道状态信息csi反馈方法、设备及存储介质
JP7336489B2 (ja) 更に最適化されたオーバーヘッドを有するマルチビームコードブック
WO2018228308A1 (zh) 一种信道状态信息反馈的方法、装置和存储介质
US10541735B2 (en) Beamforming using an antenna array
WO2017157082A1 (zh) 一种csi反馈方法、预编码方法、终端及基站
JP7450625B2 (ja) 無線通信ネットワークにおけるフィードバック報告のための方法および装置
US20220286261A1 (en) Method for Compressing Wireless Channel State Information Feedback
US8917789B2 (en) Precoding for time or frequency correlated MIMO channels
CN105897320B (zh) 使用因数化预编码的方法和设备
CN101969363B (zh) 信道状态信息反馈方法及终端
WO2021102952A1 (en) Method for wireless channel reference signal transmission and channel state information feedback
IL261686A (en) Methods and Devices for Determining Pre-Encoded Data on a Wireless Network
WO2018082701A1 (zh) 信道信息量化反馈的方法和装置、电子设备、存储介质
CN102823155A (zh) 具有用于预编码mimo传送的子集限定的参数化码本
WO2017152785A1 (zh) 一种csi反馈方法、预编码方法及装置
CN108111211B (zh) 信道状态信息的反馈方法、装置及管理设备
WO2011123977A1 (zh) 针对天线阵列的相关矩阵反馈方法和系统

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: 19906501

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 17/11/2021)

122 Ep: pct application non-entry in european phase

Ref document number: 19906501

Country of ref document: EP

Kind code of ref document: A1