WO2014089992A1 - 协作消除干扰的方法、装置及系统 - Google Patents

协作消除干扰的方法、装置及系统 Download PDF

Info

Publication number
WO2014089992A1
WO2014089992A1 PCT/CN2013/082942 CN2013082942W WO2014089992A1 WO 2014089992 A1 WO2014089992 A1 WO 2014089992A1 CN 2013082942 W CN2013082942 W CN 2013082942W WO 2014089992 A1 WO2014089992 A1 WO 2014089992A1
Authority
WO
WIPO (PCT)
Prior art keywords
interference
vector
space
legacy
base station
Prior art date
Application number
PCT/CN2013/082942
Other languages
English (en)
French (fr)
Inventor
陈诗军
姚珂
刘娟
谢赛锦
郁光辉
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 WO2014089992A1 publication Critical patent/WO2014089992A1/zh

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04JMULTIPLEX COMMUNICATION
    • H04J11/00Orthogonal multiplex systems, e.g. using WALSH codes
    • H04J11/0023Interference mitigation or co-ordination
    • H04J11/005Interference mitigation or co-ordination of intercell interference
    • H04J11/0053Interference mitigation or co-ordination of intercell interference using co-ordinated multipoint transmission/reception

Definitions

  • the present invention relates to the field of communications, and in particular, to a method, apparatus, and system for cooperatively eliminating interference.
  • Background Art As a standard of the fourth generation communication system (4th Generation, referred to as 4G), advanced international mobile communication
  • COMP Coordinated Multi-Point Transmission/Reception
  • the so-called COMP is the plurality of base stations cooperative transmission, multiple cooperating base stations which serve one or more user terminal (User Equipment, abbreviated as UE) 0 Third Generation Partnership Project (3rd Generation Partnership Project, referred to as 3GPP) defined
  • the COMP includes two scenarios: one is multi-point cooperative scheduling, that is, the scheduling information exchanged between adjacent nodes, so that the interference between the transmission signals of each cell is coordinated; the other is multi-point joint processing, that is, multiple The coordinated nodes provide services to the target UE by sharing data, Channel Situation Information (CSI), and scheduling information.
  • CSI Channel Situation Information
  • the advantage of multi-point cooperative scheduling is that there is no need to exchange too much information between the nodes, which has no effect on the wireless interface, and the disadvantage is that the cooperative transmission gain cannot be obtained, so that the spectrum utilization efficiency cannot be improved;
  • the advantage of the multi-point joint processing is that Macro diversity and cooperative transmission gain and processing gain of high-order antennas, but the disadvantage is that a large amount of data information and CSI information need to be exchanged, and the implementation complexity is high.
  • Signal interference occurs between multiple cells. As shown in FIG. 1, three adjacent cells are cell 1, cell 2, and cell 3.
  • the first UE (UE1), the second UE (UE2), and the third UE (UE3) belong to the cell 1, the cell 2, and the cell 3, respectively.
  • the base station of the cell 2 and the cell 3 (here, it can be represented by BBU2 and BBU3, wherein the BBU is the indoor baseband processing unit, the signal sent by the building base band unit) interferes with the UE in the cell 1, and the base station of the cell 1 and the cell 3
  • the signals sent by (BBU1 and BBU3) may cause interference to the UE in the cell 2, and the signals sent by the base stations (BBU1 and BBU2) of the cell 1 and the cell 2 may cause interference to the UE in the cell 3.
  • the inter-cell multi-antenna cooperation technology used to reduce signal interference between cells includes the following steps: Step S202: The UE measures channel information of the main signal and the interference signal, and obtains the measured channel information.
  • Step S204 The base station to which the UE belongs sends the channel information of the interference signal to the adjacent base station; and Step S206: The neighboring base station determines the space with less interference according to the interference channel information, and determines the signal of the local station. Precoding, the base station to which the UE belongs sends a signal in a space where the determined interference is relatively small. It can be seen from the above description of the steps of the multi-cooperation technology that the small-area multi-antenna cooperation technology includes the step of the UE feeding back channel information to the associated base station.
  • the UE since the amount of data of the channel information is too large, in order to reduce the amount of data transmission, the UE usually performs codebook processing on the channel information instead of directly feeding back all the channel information.
  • the above codebook processing mainly refers to: matching the channel information obtained by the estimated channel with the codeword vector (ie, the codebook) in the codebook table, and then matching the codeword that is the most matching (ie, the most relevant).
  • the index corresponding to the vector is used as feedback information.
  • the above feedback information includes a primary channel information codeword (ie, a Precoding Matrix Indicator, abbreviated as PMI) and two types of interference channel codewords, where the two types of interference channel codewords are: The Worst Codeword Index (WCI), one is the Best Codeword Index (BCI), which corresponds to the maximum interference codeword and the minimum interference codeword, respectively.
  • WCI The Worst Codeword Index
  • BCI Best Codeword Index
  • the UE side feedback WCI or BCI
  • the BBU side sends the WCI to the cooperative base station, and the cooperative base station determines the precoding according to the fed back WCI or BCI, thereby reducing interference to the neighbor UE. In this way, interference suppression is performed.
  • the embodiments of the present invention provide a method, an apparatus, and a system for cooperatively eliminating interference, so as to at least solve the technical problem that it is difficult to effectively reduce interference to neighbor UEs caused by the difficulty in determining the size of legacy interference in the prior art.
  • a method for cooperatively eliminating interference including: generating, by a UE, multi-dimensional code localization information according to channel information; and determining, by the UE, legacy interference to obtain a legacy interference flag;
  • the UE sends the multi-dimensional code localization information and the legacy interference flag to the cooperative base station by using the primary base station, and is configured to instruct the cooperative base station to obtain a precoding matrix according to the multi-dimensional code localization information and the legacy interference flag.
  • the multi-dimensional code localization information includes: a PMI of the UE main signal channel, and/or a quantization result of a vector in an interference channel value space.
  • the quantization result of the vector in the interference value space includes at least one of the following: a codeword dimension vector, a component coefficient amplitude quantization index, and a component coefficient phase quantization index.
  • the generating, by the UE, the multi-dimensional code localization information according to the channel information includes: performing SVD decomposition on an equivalent interference channel obtained by multiplying a detection factor and an interference channel; and performing singular value correspondence that is greater than a threshold after performing the SVD decomposition.
  • the UE estimating the legacy interference to obtain the legacy interference flag includes: performing the following operations on each component in the interference value space: a component of a vector of other dimensions other than the above-described Q dimension in the orthogonal group selected as the above-mentioned legacy interference; Performing residual interference estimation on the components of the vectors of the other dimensions mentioned above to obtain the above-mentioned legacy interference target
  • performing the legacy interference estimation to obtain the legacy interference flag includes: if the legacy interference estimate of the component of the vector of the one dimension is less than a predetermined legacy interference flag threshold, The vector number of the component in the orthogonal group is used as the legacy interference flag; and/or in the case where the legacy interference is a component of a vector of a plurality of dimensions, performing the legacy interference estimation to obtain the legacy interference flag includes: If the smallest residual interference estimate of the residual interference estimate of the components of the vector of the plurality of dimensions is smaller than the predetermined legacy interference flag threshold, the vector number of the component corresponding to the minimum residual interference estimate in the orthogonal group is used as the legacy interference. Sign.
  • a method for cooperatively canceling interference including: a cooperative base station receiving multi-dimensional code localization information and a legacy interference flag sent by a primary base station; and the cooperative base station calculating, according to the multi-dimensional code localization information, Obtaining interference null space; the cooperative base station according to the interference zero space and the legacy interference indicator The residual interference space is obtained; the cooperative base station obtains a precoding matrix according to the interference null space and the legacy interference space.
  • the cooperative base station calculates the interference null space according to the multi-dimensional code localization information and the legacy interference flag, where: the cooperative base station is configured according to multiple dimensions of the interference value space carried in the multi-dimensional code localization information.
  • the quantization result recovers the interference space; the SVD decomposition is performed jointly on the interference space; the vector of the right singular value matrix corresponding to the eigenvalue of the threshold of the zero space vector selection threshold is obtained as the interference zero space.
  • the cooperative base station obtains a legacy interference space according to the interference null space and the legacy interference flag, including: projecting a codeword vector corresponding to the legacy interference flag into the interference null space to obtain a legacy interference vector; and using the legacy interference vector A vector whose modulus value is greater than or equal to the residual interference space vector threshold is used as the legacy interference space.
  • the obtaining, by the cooperative base station, the precoding matrix according to the interference null space and the legacy interference space comprises: calculating a normalization of each of the legacy interference spaces and a PMI of the primary signal channel of the cooperative base station in the interference null space.
  • the re-determining the PMI includes: using a codeword as the determined PMI, where the codeword is a codebook in which a PMI of a main signal channel of the multi-cooperating base station is related to a correlation norm of the codeword is greater than a main signal.
  • the obtaining, by the cooperative base station, the precoding matrix according to the interference null space and the legacy interference space comprises: calculating a normalization of each of the legacy interference spaces and a PMI of the primary signal channel of the cooperative base station in the interference null space.
  • the correlation norm of the projection vector if the calculated correlation norm is less than or equal to the residual interference vector and the main signal correlation threshold, the interference null space is used as the precoding vector space; otherwise, the precoding vector space is re-determined;
  • the codebook vector corresponding to the PMI of the primary signal channel of the cooperative base station is projected onto the determined precoding vector space to obtain the precoding matrix.
  • the re-determining the precoding vector space comprises: determining whether the dimension N of the legacy interference space is smaller than the dimension M of the interference null space, wherein M and N are natural numbers; if less, according to the legacy interference space structure (MN) a vector of the legacy interference space of the above dimension N is represented by a vector of the above-mentioned interference null space as a base vector; and a vector represented by the above-mentioned interference null space is used as a base vector
  • MN legacy interference space structure
  • the vector of the interference space is sorted according to the order of the modulus values, and the M (linear) independent vector is constructed with the constructed (MN) vectors; the M linear independent vectors are started from the residual interference space vector with the largest modulus value.
  • the orthogonalization processing is performed; the (MN) or more (MN) vectors from the backward to the forward in the orthogonalization result obtained by the orthogonalization processing are used as the above-mentioned re-determined precoding vector space.
  • the re-determining the precoding vector space comprises: determining whether the dimension N of the legacy interference space is greater than or equal to the dimension M of the interference null space, where M and N are natural numbers; if greater than or equal to, the legacy interference Spatial SVD decomposition; the vector in the right singular value matrix corresponding to the eigenvalue of the eigenvalues of the residual interference space vector threshold obtained by decomposing the SVD and the sum of the squares of all the eigenvalues is used as the above-mentioned re-determined pre-coding vector space.
  • a device for cooperatively canceling interference which is located in a UE, and includes: a multi-dimensional code localization module, configured to generate multi-dimensional code localization information according to channel information; a legacy interference estimation module, setting In order to estimate the legacy interference, the legacy interference flag is obtained.
  • the feedback module is configured to send the multi-dimensional code localization information and the legacy interference flag to the cooperative base station by using the primary base station, to indicate that the cooperative base station performs the multi-dimensional code localization information according to the foregoing
  • the above legacy interference flag is obtained by a precoding matrix.
  • the multi-dimensional code localization information includes: a PMI of the UE main signal channel, and/or a quantization result of a vector in an interference channel value space, where the quantization result of the vector in the interference value space includes at least one of the following: Codeword dimension vector, component coefficient magnitude quantization index, component coefficient phase quantization index.
  • the multi-dimensional code localization module includes: a decomposition unit configured to perform SVD decomposition on an equivalent interference channel obtained by multiplying a detection factor by an interference channel; and an interference value space determining unit configured to perform the SVD decomposition a space composed of a vector of a right singular value matrix corresponding to a singular value greater than a threshold as the interference value space; a quantization unit configured to perform the following operations on each of the interference value spaces: Projecting in each orthogonal group; in the case where the feedback dimension is Q-dimensional, calculating the sum of the squares of the Q components whose projection component size is located in the front Q in each orthogonal group, and selecting the orthogonal group having the largest sum of squares, wherein Q is a positive integer; quantizing the Q components of the current vector in the selected orthogonal group to obtain the quantization result; the legacy interference estimation module is configured to perform the following operations on each of the interference value spaces Determining the above-mentioned legacy interference flag: the current component is in the above
  • the legacy interference estimation module is further configured to: if the legacy interference is a component of a vector of one dimension, if the legacy interference estimate of the component of the vector of the one dimension is less than a predetermined legacy interference flag threshold, then the component is The vector number in the above orthogonal group as the above-mentioned legacy interference flag; and/or the above legacy
  • the interference estimation module is further configured to: if the legacy interference is a component of a vector of multiple dimensions, if the minimum legacy interference estimate of the legacy interference estimates of the components of the plurality of dimensions of the vector is less than the predetermined legacy interference flag threshold, Then, the vector number of the component corresponding to the minimum residual interference estimate described above in the orthogonal base vector group is used as the legacy interference flag.
  • a device for cooperatively canceling interference is located in a cooperative base station, and includes: a receiving module, configured to receive multi-dimensional code localization information and a legacy interference flag sent by the primary base station;
  • the calculation module is configured to calculate the interference null space according to the multi-dimensional code localization information;
  • the legacy interference space calculation module is configured to obtain a legacy interference space according to the interference zero space and the legacy interference flag;
  • the precoding calculation module is set according to the foregoing The interference null space and the left interference space get the precoding matrix.
  • the interference zero space calculation module includes: a recovery unit, configured to recover the interference space according to the quantization result of the vector of the plurality of dimensions in the interference value space carried in the multi-dimensional code localization information; the joint decomposition unit, setting In order to jointly perform the SVD decomposition on the interference space, the interference null space determining unit is configured to set, as the interference zero space, a vector of the right singular value matrix corresponding to the eigenvalue of the threshold of the zero space vector selection threshold obtained by decomposing the SVD.
  • the legacy interference space calculation module includes: a legacy interference vector determining unit, configured to project a codeword vector corresponding to the legacy interference flag into the interference null space to obtain a legacy interference vector; and a legacy interference space determining unit, configured to A vector in the above-mentioned legacy interference vector whose modulus value is greater than or equal to the residual interference space vector threshold is used as a legacy interference space.
  • the precoding calculation module includes: a first correlation norm calculation unit configured to calculate a normalized projection of each of the legacy interference spaces and a PMI of the primary signal channel of the cooperative base station in the interference null space; a correlation norm of the vector; the PMI re-determining unit is configured to determine the PMI of the main signal channel of the multi-cooperating base station as the calculated correlation norm is less than or equal to the legacy interference vector and the main signal correlation threshold PMI, otherwise, re-determining the PMI; the first pre-coding calculation unit is configured to project the codebook vector corresponding to the PMI determined above to the interference null space to obtain the pre-coding matrix.
  • the precoding calculation module includes: a second correlation norm calculation unit configured to calculate a normalized projection of each of the legacy interference spaces and a PMI of the primary signal channel of the cooperative base station in the interference null space a correlation norm of the vector; a precoding vector space redetermining unit, configured to use the interference null space as a precoding vector space if the calculated correlation norm is less than or equal to the legacy interference vector and the main signal correlation threshold, Otherwise, the precoding vector space is re-determined; the second precoding calculation unit is configured to project the codebook vector corresponding to the PMI of the primary signal channel of the cooperative base station to the determined precoding vector space to obtain the precoding matrix.
  • a system for cooperatively canceling interference comprising the UE of the apparatus for cooperatively eliminating interference described above, and a cooperative base station provided with the apparatus for cooperatively eliminating interference described above.
  • the UE carries the multi-dimensional code localization information and the self-estimated legacy interference flag in the information fed back to the primary base station, and after receiving the feedback information, the cooperative base station may perform the multi-dimensional code localization according to the information.
  • the information and legacy interference flags calculate the precoding matrix because the legacy interference flag is carried therein so that the cooperative base station further controls the correlation between the precoding matrix and the legacy interference space, so that the resulting precoding matrix further reduces interference to neighboring stations.
  • FIG. 2 is a preferred flowchart of an inter-cell multi-antenna cooperation technique according to the related art
  • FIG. 3 is a cooperative cancellation according to an embodiment of the present invention.
  • a preferred flowchart of a method for interference
  • FIG. 4 is another preferred flowchart of a method for cooperatively eliminating interference according to an embodiment of the present invention
  • FIG. 5 is a diagram of a UE generating multi-dimensional code localization information according to an embodiment of the present invention.
  • a preferred flow chart of legacy interference estimation is a preferred flow chart for calculating a interference null space by a cooperative base station according to an embodiment of the present invention
  • FIG. 7 is a preferred embodiment of a system for cooperative interference cancellation according to an embodiment of the present invention.
  • FIG. 1 is a preferred flowchart of an inter-cell multi-antenna cooperation technique according to the related art
  • FIG. 3 is a cooperative cancellation according to an embodiment of the present invention.
  • a preferred flowchart of a method for interference
  • FIG. 4 is
  • FIG. 8 is a block diagram showing a preferred configuration of an apparatus for cooperatively eliminating interference according to an embodiment of the present invention
  • FIG. 9 is a block diagram showing another preferred configuration of an apparatus for cooperatively canceling interference according to an embodiment of the present invention
  • FIG. 11 is an interference zero space calculation module according to an embodiment of the present invention
  • a block diagram of a preferred species
  • FIG. 12 is a block diagram showing a preferred structure of a legacy interference space calculation module according to an embodiment of the present invention
  • FIG. 13 is a block diagram showing a preferred structure of a precoding calculation module according to an embodiment of the present invention.
  • An embodiment of the present invention provides a method for cooperatively canceling interference, which is described from the UE side. As shown in FIG. 3, the method includes the following steps: Step S302: The UE generates multi-dimensional code localization information according to the channel information; Step S304 The UE estimates the legacy interference to obtain the legacy interference flag.
  • Step S306 The UE sends the multi-dimensional code localization information and the legacy interference flag to the cooperative base station by using the primary base station, and is used to indicate the cooperative base station according to the multi-dimensional code localization information and the legacy interference flag. Get a precoding matrix.
  • the embodiment of the present invention provides a method for cooperatively canceling interference, which is described from the cooperative base station side. As shown in FIG. 4, the method includes the following steps: Step S402: The cooperative base station receives the multi-dimensional code localization information sent by the primary base station.
  • step S404 the cooperative base station calculates the interference null space according to the multi-dimensional code localization information
  • step S406 the cooperative base station obtains the legacy interference space according to the interference null space and the legacy interference flag
  • step S408 the cooperative base station according to the interference zero
  • the spatial and legacy interference spaces are precoded.
  • the UE carries the multi-dimensional code localization information and the self-estimated legacy interference flag in the information fed back to the primary base station, and after receiving the feedback information, the cooperative base station may perform the localization information and the legacy according to the multi-dimensional code.
  • the interference flag calculates the precoding matrix because it carries the legacy interference flag so that the cooperative base station can further control the correlation between the precoding matrix and the legacy interference space, so that the finally obtained precoding matrix further reduces the interference to the neighboring station.
  • the multi-dimensional code localization information generated by the UE includes, but is not limited to, the following information: a Precoding Matrix Indicator (PMI) of the primary signal channel of the UE, and/or a dry
  • the quantized result of the vector in the scrambling space By transmitting this information to the cooperative base station, the cooperative base station can be effectively restored to the original channel information.
  • the quantization result may include, but is not limited to, at least one of the following: a codeword, a component coefficient amplitude quantization index, and a component coefficient phase quantization index.
  • the cooperating base station can recover the interference space according to the amplitude and phase therein.
  • Step S502 Perform singular value decomposition (Singular Value Decomposition, S for short) on the result obtained by multiplying the detection factor by the interference channel.
  • Step S504 taking a space composed of vectors of the right singular value matrix corresponding to the singular value larger than the threshold after performing the SVD decomposition as the interference value space;
  • Step S506 performing the following operations on each of the interference value spaces respectively Determining the legacy interference flag: Projecting the current vector to each orthogonal group in the codebook; in the case where the feedback dimension is Q-dimensional, calculating the square of the Q components whose projection component size is located in the front Q in each orthogonal group And selecting a quadrature group having the largest sum of squares; quantizing the Q components of the current vector in the selected orthogonal group to obtain the quantization result; and placing the current component in the selected orthogonal group a component of a vector other than the Q dimension as the legacy interference; a residual interference estimate for a component of the vector of the other dimension To the legacy interference flag, where Q is a positive integer.
  • the R8 codebook table as shown in Table 1 and the codebook correlation table as shown in Table 2 may be
  • step S506 there are mainly two cases, one is that the legacy interference is a vector of one dimension, and the other is that the legacy interference is a vector of more than one dimension -
  • performing the legacy interference estimation to obtain the legacy interference flag includes: if the legacy interference estimate of the component of the vector of the one dimension is less than a predetermined legacy interference flag threshold, then the component is a vector number within an orthogonal group as the legacy interference flag; and/or 2)
  • performing the legacy interference estimation to obtain the legacy interference flag includes: if the minimum residual interference estimate of the residual interference estimate of the components of the vector of the plurality of dimensions is smaller than a predetermined one If the residual interference flag threshold is left, the vector number of the component corresponding to the smallest residual interference estimate in the orthogonal group is used as the legacy interference flag.
  • the residual interference estimate for a vector of one or more dimensions can be calculated according to the following formula:
  • SINR main signal energy / ( (G*H*W) * (G*H*W) H ); where SINR represents the legacy interference estimate, G represents the detection factor, H represents an interference signal, and W represents the current The component of the calculated dimensional vector, ( ⁇ ) H represents transpose, and ((G*H*W) * (G*H*W) H ) represents the interfering signal energy.
  • Step S404 the calculation of the interference null space may be implemented according to the steps shown in FIG. 6, including: Step S602: The cooperative base station according to the interference value carried in the multi-dimensional code localization information The quantization result of the vector of the multiple dimensions in the space recovers the interference space.
  • Step S604 Perform SVD decomposition on the interference space jointly;
  • Step S606 The vector of the right singular value matrix corresponding to the eigenvalue of the threshold of the zero space vector selection threshold obtained by decomposing the SVD is used as the interference null space.
  • calculating the legacy interference space may be implemented according to the following steps, including: Step 1: Projecting the codeword vector corresponding to the legacy interference flag to the interference null space to obtain a residual interference direction
  • Step 2 The vector of the legacy interference vector whose modulus value is greater than or equal to the residual interference space vector threshold is used as the legacy interference space.
  • the correlation norm is less than or If the threshold is equal to the predetermined threshold, the interference null space is directly used as the precoding vector space, and the codebook vector corresponding to the PMI is projected to the interference null space.
  • the codeword is a codebook in which a correlation norm of a PMI of the primary signal channel of the multi-cooperating base station and the codeword is greater than the legacy interference vector and a main signal correlation a threshold, and a codeword that minimizes a correlation norm of the legacy interference space and a normalized projection vector of a PMI of the primary signal channel of the multi-cooperating base station in the interference null space.
  • a codeword is determined, and the codeword is used as the PMI, and the corresponding determined codeword needs to satisfy the above two conditions: 1) making the PMI of the main signal channel of the multi-cooperating base station and the codeword related to the codeword The number is greater than the legacy interference vector and the main signal correlation threshold; 2) selecting, from the codewords satisfying condition 1, a correlation criterion of the normalized projection vectors of the PMI of the primary signal channel of the multi-cooperating base station in the interference null space The codeword of the smallest codeword as the final choice, ie as the re-determined PMI.
  • Method 1) includes the following steps:
  • S1 determining whether the dimension N of the legacy interference space is smaller than the dimension M of the interference null space, where M and N are natural numbers; S2: if less, constructing (M-N) vectors according to the legacy interference space;
  • S3 The vector of the left interference space of dimension N is represented by a vector of interference null space as a base vector;
  • S4 The vector of the residual interference space represented by the vector with the interference null space as the base vector is sorted in order of the modulus values, and the M (N-N) vectors are constructed to constitute M linear independent vectors;
  • (MN) or more (MN) vectors from the backward to the forward in the orthogonalization result obtained by the orthogonalization process are used as the re-determined precoding vector space.
  • (MN) or more (MN) vectors can be selected from the back to the front as needed.
  • Mode 2) mainly includes the following steps:
  • S2 If greater than or equal to, the SVD decomposition is performed on the left interference space; S3: the ratio between the square of the eigenvalues obtained by decomposing the SVD and the sum of squares of all eigenvalues is smaller than the eigenvalue of the residual interference space vector threshold.
  • the vector in the singular value matrix acts as a precoding vector space.
  • a system for cooperating to eliminate interference is also provided.
  • the device is used to implement the foregoing embodiments and preferred embodiments, and details are not described herein.
  • the term "unit” or “module” may implement a combination of software and/or hardware for a predetermined function.
  • FIG. 7 is a block diagram of a preferred structure of a system for cooperative interference cancellation according to an embodiment of the present invention.
  • the method includes: a UE 702, a primary base station 704, and a cooperative base station 706.
  • the UE is provided with means for cooperatively canceling interference
  • the cooperative base station is also provided with means for cooperatively canceling interference.
  • the cooperative interference cancellation device located in the UE includes: a multidimensional code localization module 802, a legacy interference estimation module 804, and a feedback module 806. The structure will be described below.
  • the multi-dimensional code localization module 802 is configured to generate multi-dimensional code localization information according to the channel information; preferably, the UE selects the most relevant codeword according to the correlation between the channel and the codebook codeword according to the channel information, and In the orthogonal group in which the codeword is located, the component coefficients of other dimensions are quantized;
  • the legacy interference estimation module 804 is coupled to the multi-dimensional code localization module 802, and is configured to estimate the legacy interference to obtain a legacy interference flag.
  • the UE estimates the interference of the multi-dimensional code localization discarding dimension. If the threshold is exceeded, The largest interfering dimensional codeword is used as a legacy interference flag.
  • the feedback module 806 is coupled to the legacy interference estimation module 804, and configured to send the multi-dimensional code localization information and the legacy interference flag to the cooperative base station by using the primary base station, to indicate that the cooperative base station is configured according to the multi-dimensional code
  • the localization information and the legacy interference flag are obtained by a precoding matrix.
  • the multi-dimensional code localization module 802 includes: a decomposition unit 902 configured to perform SVD decomposition on a result obtained by multiplying a detection factor by an interference channel; and an interference value space determining unit 904 configured to perform SVD decomposition.
  • each orthogonal group in the codebook is projected; in the case where the feedback dimension is Q-dimensional, the calculation is performed at each positive
  • the sum of the projection components in the intersection group is located in the sum of the squares of the Q components of the front Q, and the orthogonal group having the largest sum of squares is selected, wherein
  • the legacy interference estimation module 804 is set to each of the interference value spaces
  • the component performs the following operations to determine a legacy interference flag: a component of a vector of other dimensions other than the Q dimension of the current component in the selected orthogonal group as the legacy interference; a vector for the other dimension
  • the component performs residual interference estimation to obtain the legacy interference flag.
  • the legacy interference estimation module 804 is further configured to: if the legacy interference is a component of a vector of one dimension, if the legacy interference estimate of the component of the vector of the one dimension is less than a predetermined legacy interference flag threshold, then The vector number of the component in the orthogonal group is used as the legacy interference flag; and/or the legacy interference estimation module 804 is further configured to, if the residual interference is a component of a vector of multiple dimensions, if the multiple The smallest residual interference estimate of the residual interference estimate of the components of the vector of the dimensions is smaller than the predetermined legacy interference flag threshold, and the vector number of the component corresponding to the smallest legacy interference estimate is within the orthogonal basis vector group As the legacy interference sign.
  • the precoding apparatus located in the cooperative base station includes: a receiving module 1002, an interference null space calculating module 1004, a legacy interference space calculating module 1006, and a precoding calculating module 1008. The structure will be described below.
  • the receiving module 1002 is configured to receive the multi-dimensional code localization information and the legacy interference flag sent by the primary base station;
  • the interference null space calculation module 1004 is coupled to the receiving module 1002 and configured to calculate the interference null space according to the multi-dimensional code localization information; preferably, the interference null space calculation module 1004 can multi-dimensionalize all received interference channels.
  • the codeword is restored to the interference channel vector, and the joint decomposition is performed to obtain the interference null space.
  • the interference channel is subjected to joint analysis to obtain a residual interference space.
  • the precoding calculation module 1008 is coupled to the legacy interference space calculation module 1006 and configured to obtain a precoding matrix according to the interference null space and the legacy interference space.
  • the interference null space calculation module 1004 includes: a recovery unit 1102, configured to set a vector according to multiple dimensions in the interference value space carried in the multi-dimensional code localization information.
  • the quantized result recovers the interference space;
  • the joint decomposition unit 1104 is configured to perform SVD decomposition on the interference space joint;
  • the interference null space determining unit 1106 is configured to set the eigenvalue of the eigenvalue obtained by decomposing the SVD to be smaller than the eigenvalue of the zero space vector selection threshold
  • the vector of the corresponding right singular value matrix acts as an interference null space.
  • the legacy interference space calculation module 1006 includes: a legacy interference vector determining unit 1202, configured to project a codeword vector corresponding to the legacy interference flag into the interference null space to obtain a legacy interference vector;
  • the legacy interference space determining unit 1204 is configured to set, as the legacy interference space, a vector whose modulus value in the legacy interference vector is greater than or equal to the residual interference space vector threshold. In a preferred embodiment, as shown in FIG.
  • the precoding calculation module 1008 includes: a first correlation norm calculation unit 1302, configured to calculate each of the legacy interference spaces and a main signal of the cooperative base station a correlation norm of the normalized projection vector of the PMI of the channel in the interference null space; the PMI re-determining unit 1304, set to the case where the calculated correlation norm is less than or equal to the residual interference vector and the main signal correlation threshold
  • the PMI of the primary signal channel of the multi-cooperating base station is used as the determined PMI, otherwise, the PMI is re-determined;
  • the first pre-coding calculation unit 1306 is configured to project the codebook vector corresponding to the determined PMI into the interference null space. Encoding matrix.
  • the PMI re-determining unit 1304 is further configured to re-determine the PMI in the following manner: using a codeword as the determined PMI, where the codeword is a PMI in a codebook that causes a primary signal channel of the multi-cooperating base station to The correlation norm of the codeword is greater than the primary signal correlation threshold, and the codeword of the legacy interference space and the PMI of the primary signal channel of the multi-cooperating base station having the smallest correlation norm of the normalized projection vector of the interference null space.
  • the precoding calculation module 1008 may further include: a second correlation norm calculation unit 10082, configured to calculate each of the legacy interference spaces and a primary signal channel of the cooperative base station a correlation norm of the normalized projection vector of the PMI in the interference null space; the precoding vector space redetermining unit 10084 is set to be less than or equal to the residual interference vector and the main signal correlation threshold in the calculated correlation norm In this case, the interference null space is used as the precoding vector space, otherwise, the precoding vector space is re-determined; the second precoding calculation unit 10086 is set to the codebook vector corresponding to the PMI of the main signal channel of the cooperative base station. Projecting the determined precoding vector space to obtain the precoding matrix.
  • a second correlation norm calculation unit 10082 configured to calculate each of the legacy interference spaces and a primary signal channel of the cooperative base station a correlation norm of the normalized projection vector of the PMI in the interference null space
  • the precoding vector space redetermining unit 10084 is set to be less than or equal to the residual interference vector and the
  • the preferred embodiment further provides a method for generating a precoding vector space, if the vector of the legacy interference space and the PMI (PMI represents the main signal vector) corresponding to the codeword vector in the interference vector of zero space If they are less than or equal to the threshold C, all vectors are removed from the legacy interference space B. If the last legacy interference space B is not empty, the following processing is performed:
  • the vector of the interference space B is constructed from the maximum modulus value to construct a linear independent vector group, and then starts from the interference space vector of the maximum modulus value , Perform orthogonalization. Then, starting from the orthogonal vector obtained at the beginning, the orthogonal vector is removed, and finally (jk) vectors are reserved to form a precoding vector space.
  • the dimension of the left interference space is greater than or equal to the dimension of the interference zero space: SVD decomposition is performed on the left interference space, starting from the minimum eigenvalue, selecting the eigenvalue square/the sum of all the eigenvalues and the sum of the eigenvalues corresponding to the threshold e
  • the vectors in the singular value matrix constitute the precoding vector space.
  • the main signal vector X may be selected according to the following formula, where X may be any codeword in the codebook, or may be other non-codebook vectors, as long as it can satisfy
  • X may be any codeword in the codebook, or may be other non-codebook vectors, as long as it can satisfy
  • ) is the smallest, and X can be used as the main signal vector, where AO represents the precoding space.
  • AO represents the precoding space.
  • the manner provided by the foregoing embodiment of the present invention can further control the size of the legacy interference, and can reduce the abnormal point of the SINR
  • a legacy interference control method in a 3-cell 3 UE scenario is taken as an example for description.
  • the threshold c (the residual interference vector and the main signal correlation threshold) is: 0.6;
  • the threshold a zero space vector selection threshold) is: 0.001;
  • the threshold u (the UE legacy interference flag threshold): 35;
  • the threshold f (the legacy interference space) Vector Threshold): 0.3;
  • Threshold e main signal precoding vector selection threshold
  • the components of the largest component and the included dimensional vector are selected for quantization, and the multidimensional quantization results for the interference value space are respectively - the first dimension : The code word is 12, the coefficient is quantized to 1; the second dimension: the code word is 10, and the coefficient is quantized to 0.2704 + 0.2650i.
  • the amplitude quantization index is 0, and the phase quantization index is 5;
  • the legacy interference estimate SINR is:
  • the legacy interference estimate SINR is:
  • Step S2 The base station receives the multi-dimensional code localization channel information and the legacy interference flag, and sends the information to the cooperative base station.
  • the base station 1 receives the multi-dimensional code localization channel information of the UE1 (the codewords and coefficients of the three dimensions of the cell 2 to the UE1 in step S1 and the codewords and coefficients of the three dimensions of the cell 3 to the UE1 in the step S1), and respectively sends them to Cooperative base station 2 and cooperative base station 3.
  • Step S3 After receiving the multi-dimensional code localization channel information, the cooperative base station recovers the interference channel space A, and performs SVD decomposition on A to obtain its zero space as AO.
  • Base station 2 and base station 3 are based on the received (pair of cells 2 in step one)
  • the codewords and coefficients of the three dimensions of UE1 and the codewords and coefficients of the three dimensions of UE3 to UE1) respectively obtain the interference interference space.
  • the angle quantization based on the recovery is shown in Table 3, and the amplitude quantization is shown in Table 4. table 3
  • the interference space recovered from the amplitude and angle tables of Tables 3 and 4 above is:
  • the interference space recovered by base station 3 is: [ 0.3598 - 0.403 H
  • Step S4 The cooperative base station projects the codeword vector corresponding to all received legacy interference codeword flags into the interference null space to obtain a legacy interference vector, and retains the vector of the legacy interference vector whose modulus value is greater than or equal to the threshold f to form a legacy interference space B.
  • the base station 2 does not receive the legacy interference flag; the codeword vector corresponding to the vector flag in the legacy interference flag of the base station 3, and the residual interference vector obtained by projecting into the interference null space is:
  • the residual interference vector has a modulus of 1, which is greater than the threshold f, so the legacy interference space B of the base station 3 is:
  • Step S5 The cooperative base station calculates the correlation norm of the normalized projection vector of the residual interference space B and the PMI in the space AO. If the relevant norm is greater than the threshold c, the PMI is reselected, otherwise the original PMI of the base station is used. In the preferred embodiment, base station 2 does not receive the legacy interference flag, so base station 2 uses the original PMI (base station 2 has a PMI of 6).
  • the legacy interference flag received by the base station 3 is 2, and the corresponding codebook index is 14 codebook space, and the correlation norm of the normalized projection vector of the PMI in the space A0 is 1, which is greater than the threshold c, so the base station 3 To reselect PMI (base station 3 original PMI is 14).
  • the codebook space of PMI is [0.5 0.5 -0.5 0.5] T
  • the normalized projection vector of AO is -
  • the correlation norm of the normalized projection vector of the PMI in the spatial AO is greater than the threshold c
  • the correlation norm of any codeword in the PMI and the codebook is calculated, if the correlation norm is greater than the threshold e, and the residual interference space
  • the correlation norm of the normalized projection vector of B and PMI in space AO is the smallest.
  • the correlation norm calculated by the base station 3 is greater than the threshold c
  • the correlation norm of the PMI and the 16 codewords in the codebook is calculated, and the 16 related norms are 0.5, 0.5, 0.5, 0.5.
  • the threshold e is 0.5, 0.5, 0.5, 0.2706, 0.6533, 0.6533, 0.2706, 0.5, 0.5, 0.5, 1, 0, 0, wherein the threshold e is 0.5, 0.5, 0.5, 0.5, 0.2706, 0.6533, 0.6533, 0.2706, 0.5, 0.5 , 0.5, 1, the codeword corresponding to the norm of the threshold e is 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 14, wherein the residual interference
  • the norm of the normalized projection vector of space B and PMI in space AO is 0.5143, 0.5368, 0.5942, 0.7223 0.3014, 0.9022, 0.7195, 0.2943 0.6147, 0.5725 0.6041, 0.5298, 1.0000, and the minimum correlation norm is 0.2943, corresponding to The code word is 8, so the base station 3 finally selects a PMI of 8.
  • Step S6 The precoding matrix W is obtained by projecting the codebook vector corresponding to the PMI to the zero space of the interference channel space A.
  • the PMI of the base station 2 is 6, and the PMI of the base station 3 is 8.
  • the codewords corresponding to the respective PMIs are respectively used to project the null space of the corresponding interference channel space A to obtain a precoding matrix W.
  • the precoding matrix of the base station 2 is: [0.4300 + 0.2296 ⁇ -0.2134 + 0.3843 ⁇ -0.0620 - 0.3554 ⁇ 0.5004 + 0.4343 ⁇ ];
  • the precoding matrix of the base station 3 is: [0.5057 - 0.2280 ⁇ 0.2115 - 0.2182 ⁇ 0.0356 - 0.6374 ⁇ -0.1377 - 0.4165i Preferred Embodiment 2
  • the legacy interference control method in the 2-cell 2ue scenario is taken as an example, where the cell 2 performs COMP for the UE1.
  • BBU1 and BBU2 respectively represent two base stations; UE1 and UE2 respectively represent two terminals.
  • the UE1 belongs to the BBU1 and finds the downlink signal interference of the BBU2.
  • the UE2 belongs to the BBU2 and finds the downlink signal interference of the BBU1.
  • Threshold c residual interference vector and main signal correlation threshold
  • threshold a zero space vector selection threshold
  • threshold u UE legacy interference flag threshold
  • 15 15 (dB)
  • threshold f legacy interference
  • Space vector threshold 0.7.
  • the interference channel H of cell 2 to cell 1 is:
  • the quantization scheme of the next largest coefficient 0.5075 is less than 0.54, the quantization is lbit, the value is 0, and the actual value is 0.44; the phase is -153.9273 degrees between -180 degrees and -135, and the quantization is 3bit, and the value is 000, the actual representative - 157 degrees.
  • the quantized coefficients are: -0.4050 - 0.1719io
  • SINR12 16.35 dB (corresponding codebook number 10); Since SINR11 is the smallest and less than the threshold u, the legacy interference flag is: Group number 3, quantized to 2 bits (10).
  • the UE1 feeds back the multi-dimensional code localization result and the legacy interference flag to the serving base station BBU1; the BBU1 forwards the multi-dimensional code localization result and the legacy interference flag to the cooperative base station 2.
  • the cooperative base station 2 After receiving the cooperation request, the cooperative base station 2 recovers the multi-dimensional codewords of BBU1 according to Table 4 and Table 5: Table 4
  • the codeword vector corresponding to the flag forwarded by BBU1 is projected to the interference null space to obtain B1 as: [0.3239 - 0.0104 ⁇
  • the residual interference space is: [0.5000 -0.5000
  • the projection vector PI of the PMI to the interference null space is:
  • the residual interference space B is:
  • a legacy interference control method in a 3-cell 3 UE scenario is taken as an example for description.
  • BBU1, BBU2, and BBU3 respectively represent three base stations, and UE1, UE2, and UE3 respectively represent three terminals.
  • UE1 belongs to BBU1 and finds downlink signal interference of BBU3.
  • UE2 belongs to BBU2 and finds
  • the threshold c (the residual interference vector and the main signal correlation threshold) is: 0.7; the threshold a (zero space vector selection threshold) is: 0.001; the threshold u (the UE legacy interference flag threshold): 31; the threshold f (the legacy interference space vector threshold) ): 0.3; Threshold e (main signal precoding vector selection threshold): 0.3.
  • H slg the UE1 main signal channel
  • the interference channel H mtf of cell 3 to UE2 is:
  • G*H mtf performs the SVD decomposition value space vector: [-0.5386 + 0.2049 ⁇
  • the quantified results obtained by multidimensional quantization of the interference value space are: Maximum dimension: codeword 3; sub-large dimension: codeword 2; the vector number in the orthogonal vector basis group is: 1 ; the residual interference estimate SINR of the second largest dimension vector is: maximum dimension number 3, sub-large dimension number 2 .
  • SINR 1.6780; since the SINR is less than the threshold, the residual interference flag is: 2. Assume that the UE3 main signal channel is -
  • UE1 feeds the multi-dimensional code localization result and the legacy interference flag to the serving base station BBU1;
  • UE2 feeds back the multi-dimensional code localization result and the legacy interference flag to the serving base station BBU2;
  • the BBU 1 forwards the codebook information and the legacy interference flag to the cooperative base station 3, and the BBU 2 forwards the codebook information and the legacy interference flag to the cooperative base station 3.
  • the codeword vector corresponding to the flag forwarded by BBU1 is projected onto the interference null space to obtain B1 as: [0.5, 0.5, -0.5, -0.5] ⁇ ;
  • the codeword vector corresponding to the flag forwarded by BBU2 is projected to the interference null space to obtain ⁇ 2 as: [0.5, 0 + 0.5 ⁇ , -0.5, 0 - 0.5i] T .
  • the residual interference space is:
  • the projection vector PI of the PMI to the interference null space is:
  • the legacy interference space B is:
  • Embodiment 4 a legacy interference control method of three UEs in three cells is taken as an example for description.
  • the cell 3 performs COMP for UE1 and UE2.
  • BBU1, BBU2, and BBU3 respectively represent three base stations, and UE1, UE2, and UE3 respectively represent three terminals.
  • the UE1 belongs to the BBU1 and finds the downlink signal interference of the BBU3.
  • the UE2 belongs to the BBU2 and finds the downlink signal interference of the BBU3.
  • the UE3 belongs to the BBU3 and no downlink signal interference is found.
  • the threshold c (the residual interference vector and the main signal correlation threshold) is: 0.7; the threshold a (zero space vector selection threshold) is: 0.001; the threshold u (the UE legacy interference flag threshold): 31; the threshold e (the residual interference space vector threshold) ): 0.2.
  • the UE1 main signal channel H slg is:
  • the component of the largest component and the included dimensional vector is selected to perform more on the interference value space.
  • the quantized results obtained by dimension quantization are: maximum dimension: codeword 3; sub-large dimension: codeword 1; third dimension: codeword 4; the vector numbers in the orthogonal vector basis group are: 3,1, 4 ;
  • the quantized results obtained by multi-dimensional quantization of the interference value space are: maximum dimension: codeword 14; sub-large dimension: codeword 13; third largest dimension: codeword 16; vector number in the orthogonal vector base group For: 2,1,4 ;
  • the UE3 main signal channel is - [0.6988 - 0.1689 ⁇ , -0.3181 + 0.1407 ⁇ , 1.6536 - 0.2026 ⁇ , 1.2404 + 0.3565 ⁇ 0.1272 - 0.6056 ⁇ , -1.3569 - 0.5173 ⁇ , 1.1567 + 0.5585 ⁇ , -1.3527 - 0.8050i] .
  • the obtained PMI 14.
  • UE1 feeds back the multi-dimensional code localization result and the legacy interference flag to the serving base station BBU1; UE2 feeds back the multi-dimensional code localization result and the legacy interference flag to the serving base station BBU2; BBU1 forwards the codebook information and the legacy interference flag to the cooperative base station 3.
  • the BBU 2 forwards the coded information and the legacy interference flag to the cooperative base station 3.
  • the cooperative base station 3 recovers the multi-dimensional codewords of BBU1 and BBU2 according to Tables 3 and 4.
  • [0.3905 - 0.3515 ⁇
  • the vector of the right singular value matrix corresponding to the singular value whose eigenvalue is less than the threshold a constitutes the interference null space, and the interference zero space is:
  • BBUl will project the codeword vector corresponding to the forwarded flag and project it into the interference null space to obtain B1:
  • BBU2 will project the codeword vector corresponding to the forwarded flag and project it into the interference null space to obtain B2:
  • the resulting residual interference space is:
  • the projection vector PI of the PMI to the interference null space is:
  • a software is also provided for performing the technical solutions described in the above embodiments and preferred embodiments.
  • a storage medium is provided, the software being stored, including but not limited to: an optical disk, a floppy disk, a hard disk, a rewritable memory, and the like. From the above description, it can be seen that the present invention achieves the following technical effects: the UE carries the multi-dimensional code localization information and the self-estimated legacy interference flag in the information fed back to the primary base station, and the cooperative base station receives the feedback information.
  • the precoding matrix can be calculated according to the multi-dimensional code localization information and the legacy interference flag, because the legacy interference flag is carried therein, so that the cooperative base station further controls the correlation between the precoding matrix and the legacy interference space, so that the finally obtained precoding matrix
  • the interference to the neighboring stations is further reduced.
  • modules or steps of the present invention can be implemented by a general-purpose computing device, which can be concentrated on a single computing device or distributed over a network composed of multiple computing devices. Alternatively, they may be implemented by program code executable by the computing device, such that they may be stored in the storage device by the computing device and, in some cases, may be different from the order herein.
  • the steps shown or described are performed, or they are separately fabricated into individual integrated circuit modules, or a plurality of modules or steps are fabricated as a single integrated circuit module.
  • the invention is not limited to any specific combination of hardware and software.

Landscapes

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

Abstract

本发明公开了一种协作消除干扰的方法、装置及系统,其中,该方法包括:用户终端(UE)根据信道信息生成多维码本化信息;上述UE对遗留干扰进行估计得到遗留干扰标志;上述UE通过主基站将上述多维码本化信息和上述遗留干扰标志发送给协作基站,用于指示上述协作基站根据上述多维码本化信息和上述遗留干扰标志得到预编码矩阵。本发明解决了现有技术中难以确定遗留干扰的大小而导致的难以有效降低对邻站UE的干扰的技术问题,达到了有效降低对邻站UE的干扰,提高协作多点传输(COMP)的增益的技术效果。

Description

协作消除干扰的方法、 装置及系统
技术领域 本发明涉及通信领域, 具体而言, 涉及一种协作消除干扰的方法、 装置及系统。 背景技术 作为第四代通信系统 (4th Generation, 简称为 4G) 的标准, 高级国际移动通信
(International Mobile Telecommunications- Advanced,简禾尔为 IMT- Advanced )对于系统 的性能提出了更高的要求, 尤其是对上下行的频率效率有更高的要求。 协作多点传输 ( Coordinated Multi-Point transmission/reception , 简称为 COMP ) 近年来一直作为 IMT-Advanced的热点技术被广泛关注和研究。 COMP技术是一种提高高速率传输覆盖范围、 小区边缘服务质量和吞吐量、 以及 系统吞吐量的技术, 也是提高系统频谱利用率的重要技术。 所谓 COMP, 就是多个基 站协作传输, 这多个协作的基站服务于一个或多个用户终端(User Equipment, 简称为 UE)0 第三代合作伙伴项目 (3rd Generation Partnership Project, 简称为 3GPP)定义的 COMP包括两种场景: 一种是多点协作调度, 即通过相邻节点之间交互的调度信息, 使各个小区传输信号之间的干扰得到协调; 另一种是多点联合处理, 即多个协作节点 之间通过共享数据、 信道状态信息 (Channel Situation Information, 简称为 CSI) 以及 调度信息等为目标 UE提供服务。 多点协作调度的优点是各个节点之间不需要交互太多的信息, 对无线接口没有影 响, 而缺点是无法获得协作传输增益, 以致无法提高频谱利用效率; 多点联合处理的 优点是可以获得宏分集及协作传输增益以及高阶天线的处理增益, 而缺点是需要进行 大量的数据信息和 CSI信息的交互, 实现复杂度较高。 多小区之间会产生信号干扰, 如图 1所示, 相邻的三个小区分别是小区 1、 小区 2 以及小区 3。 其中, 第一 UE (UE1 )、 第二 UE (UE2)、 第三 UE (UE3 ) 分别归属于 小区 1、 小区 2和小区 3。 小区 2和小区 3的基站 (这里可以用 BBU2和 BBU3表示, 其中, BBU为室内基带处理单元, Building Base Band Unit)发出的信号会对小区 1中 的 UE产生干扰, 小区 1和小区 3的基站(BBU1和 BBU3 )发出的信号会对小区 2中 的 UE产生干扰, 小区 1和小区 2的基站(BBU1和 BBU2)发出的信号会对小区 3中 的 UE产生干扰。 目前, 为了降低小区间的信号干扰所采用的小区间多天线协作技术, 如图 2所示, 包括以下步骤: 步骤 S202: UE测量主信号和干扰信号的信道信息, 并把测量得到的信道信息反 馈给 UE所属的基站; 步骤 S204: UE所属的基站把干扰信号的信道信息发给相邻的基站; 以及 步骤 S206: 相邻的基站根据干扰信道信息确定干扰比较小的空间, 确定本站信号 的预编码, UE所属的基站在确定的干扰比较小的空间中发送信号。 由上述对多协作技术的步骤的介绍可知, 小区间多天线协作技术中包括 UE向所 属的基站反馈信道信息的步骤。 然而, 由于信道信息的数据量太大, 为了减少数据传 输量, 通常 UE要对信道信息进行码本化处理, 而不是直接反馈全部的信道信息。 其 中,上述的码本化处理主要指:将估计信道得到的信道信息与码本表中的码字向量(即, 码本) 进行匹配, 然后将最匹配 (即, 相关性最大) 的码字向量对应的索引作为反馈 信息。上述反馈信息有主信道信息码字(即预编码矩阵指示 Precoding Matrix Indicator, 简称为 PMI) 和两种类型的干扰信道码字, 其中, 这两种类型的干扰信道码字为: 一 种是最坏码字索引(Worst Codeword Index, 简称为 WCI), 一种是最好码字索引(Best Codeword Index, 简称为 BCI), 它们分别对应最大干扰码字和最小干扰码字。 相关技术中主要采用的是 UE侧反馈 WCI (或者 BCI), BBU侧把 WCI发给协作 基站, 协作基站根据反馈的 WCI或者 BCI确定预编码, 从而减少对邻站 UE的干扰。 通过这种方式进行干扰的抑制, 因为码本化而造成的信道损失, 造成协作基站找到的 邻站 UE的低干扰空间并不是很准确, 部分干扰会遗留到预编码空间内, 从而导致不 能够有效降低对邻站 UE的干扰, 因此降低了 COMP的增益, 也使得 COMP增益水平 波动范围增大。 针对上述的问题, 目前尚未提出有效的解决方案。 发明内容 本发明实施例提供了一种协作消除干扰的方法、 装置及系统, 以至少解决现有技 术中难以确定遗留干扰的大小而导致的难以有效降低对邻站 UE的干扰的技术问题。 根据本发明实施例的一个方面, 提供了一种协作消除干扰的方法, 包括: UE根据 信道信息生成多维码本化信息; 上述 UE对遗留干扰进行估计得到遗留干扰标志; 上 述 UE通过主基站将上述多维码本化信息和上述遗留干扰标志发送给协作基站, 用于 指示上述协作基站根据上述多维码本化信息和上述遗留干扰标志得到预编码矩阵。 优选地, 上述多维码本化信息包括: 上述 UE主信号信道的 PMI、 和 /或干扰信道 值空间中的向量的量化结果。 优选地, 上述干扰值空间中向量的量化结果包括以下至少之一: 码字维度向量、 分量系数幅度量化索引、 分量系数相位量化索引。 优选地, 上述 UE根据上述信道信息生成上述多维码本化信息包括: 对检测因子 与干扰信道相乘后得到的等效干扰信道进行 SVD分解; 取进行上述 SVD分解后的大 于门限的奇异值对应的右奇异值矩阵的向量组成的空间作为上述干扰值空间; 对上述 干扰值空间中的每个向量分别执行以下操作: 将当前向量向码本中的各个正交组进行 投影; 在反馈维度为 Q维的情况下, 计算在每个正交组中投影分量大小位于前 Q的 Q 个分量的平方和, 选择平方和最大的正交组, 其中, Q为正整数; 对上述当前向量在 上述选择的正交组中的上述 Q个分量进行量化得到上述量化结果; 上述 UE对上述遗 留干扰进行估计得到遗留干扰标志包括: 对上述干扰值空间中的每个分量执行以下操 作: 将当前分量在上述选择的正交组中除上述 Q维之外的其他维度的向量的分量作为 上述遗留干扰; 对上述其他维度的向量的分量进行遗留干扰估计得到上述遗留干扰标
优选地, 在上述遗留干扰为一个维度的向量的分量的情况下, 进行遗留干扰估计 得到上述遗留干扰标志包括: 如果上述一个维度的向量的分量的遗留干扰估计小于预 定的遗留干扰标志门限, 则将该分量在上述正交组内的向量编号作为上述遗留干扰标 志; 和 /或在上述遗留干扰为多个维度的向量的分量的情况下, 进行遗留干扰估计得到 上述遗留干扰标志包括: 如果上述多个维度的向量的分量的遗留干扰估计中最小的遗 留干扰估计小于上述预定的遗留干扰标志门限, 则将上述最小的遗留干扰估计对应的 分量在上述正交组内的向量编号作为上述遗留干扰标志。 优选地, 按照以下公式计算上述遗留干扰估计: SINR=主信号能量 / ( ( GHW)
(GHW) H ) ; 其中, SINR表示上述遗留干扰估计, G表示上述检测因子, H表示干 扰信号信道, W表示当前计算的维度向量的分量, ( · ) H表示转置。 根据本发明实施例的另一方面, 提供了一种协作消除干扰的方法, 包括: 协作基 站接收主基站发送的多维码本化信息和遗留干扰标志; 上述协作基站根据上述多维码 本化信息计算得到干扰零空间; 上述协作基站根据上述干扰零空间和上述遗留干扰标 志得到遗留干扰空间; 上述协作基站根据上述干扰零空间和遗留干扰空间得到预编码 矩阵。 优选地, 上述协作基站根据上述多维码本化信息和上述遗留干扰标志计算上述干 扰零空间包括: 上述协作基站根据上述多维码本化信息中携带的上述干扰值空间中的 多个维度的向量的量化结果恢复出干扰空间; 对上述干扰空间联合进行 SVD分解; 将 SVD 分解后得到的特征值中小于零空间向量选取门限的特征值对应的右奇异值矩阵 的向量作为上述干扰零空间。 优选地, 上述协作基站根据上述干扰零空间和上述遗留干扰标志得到遗留干扰空 间包括: 对上述遗留干扰标志对应的码字向量向上述干扰零空间进行投影得到遗留干 扰向量; 将上述遗留干扰向量中模值大于或等于遗留干扰空间向量门限的向量作为遗 留干扰空间。 优选地,上述协作基站根据上述干扰零空间和遗留干扰空间得到预编码矩阵包括: 计算上述遗留干扰空间中的每个向量与上述协作基站的主信号信道的 PMI在上述干扰 零空间的归一化投影向量的相关范数; 如果计算得到的相关范数均小于或等于遗留干 扰向量和主信号相关性门限, 则将上述多协作基站的主信号信道的 PMI 作为确定的 PMI, 否则, 重新确定 PMI; 将上述确定的 PMI对应的码本向量向上述干扰零空间进 行投影得到上述预编码矩阵。 优选地, 上述重新确定 PMI包括: 将码字作为上述确定的 PMI, 其中, 上述码字 是码本中使得上述多协作基站的主信号信道的 PMI与该码字的相关范数大于主信号相 关性门限, 且使得上述遗留干扰空间与上述多协作基站的主信号信道的 PMI在上述干 扰零空间的归一化投影向量的相关范数最小的码字。 优选地,上述协作基站根据上述干扰零空间和遗留干扰空间得到预编码矩阵包括: 计算上述遗留干扰空间中的每个向量与上述协作基站的主信号信道的 PMI在上述干扰 零空间的归一化投影向量的相关范数; 如果计算得到的相关范数均小于或等于遗留干 扰向量和主信号相关性门限, 则将干扰零空间作为预编码向量空间, 否则, 重新确定 上述预编码向量空间; 将上述协作基站的主信号信道的 PMI对应的码本向量向上述确 定的预编码向量空间进行投影得到上述预编码矩阵。 优选地, 重新确定上述预编码向量空间包括: 确定上述遗留干扰空间的维度 N是 否小于上述干扰零空间的维度 M, 其中, M和 N为自然数; 如果小于, 则根据遗留干 扰空间构造 (M-N) 个向量; 将上述维度为 N的遗留干扰空间的向量用上述干扰零空 间的向量作为基向量进行表示; 将用上述干扰零空间的向量作为基向量进行表示的遗 留干扰空间的向量按照模值从大到小的顺序进行排序后, 与构造的 (M-N) 个向量构 成 M个线性无关向量; 从模值最大的遗留干扰空间向量开始对上述 M个线性无关向 量进行正交化处理; 将正交化处理后得到的正交化结果中从后向前的 (M-N) 个或多 于 (M-N) 个向量作为上述重新确定的预编码向量空间。 优选地, 重新确定上述预编码向量空间包括: 确定上述遗留干扰空间的维度 N是 否大于或等于上述干扰零空间的维度 M, 其中, M和 N为自然数; 如果大于或等于, 则对上述遗留干扰空间进行 SVD分解; 将 SVD分解后得到的特征值平方与所有特征 值平方和之间的比值小于遗留干扰空间向量门限的特征值对应的右奇异值矩阵中的向 量作为上述重新确定的预编码向量空间。 根据本发明实施例的又一方面, 提供了一种协作消除干扰的装置, 位于 UE中, 包括: 多维码本化模块, 设置为根据信道信息生成多维码本化信息; 遗留干扰估计模 块, 设置为对遗留干扰进行估计得到遗留干扰标志; 反馈模块, 设置为通过主基站将 上述多维码本化信息和上述遗留干扰标志发送给协作基站, 用于指示上述协作基站根 据上述多维码本化信息和上述遗留干扰标志得到预编码矩阵。 优选地, 上述多维码本化信息包括: 上述 UE主信号信道的 PMI、 和 /或干扰信道 值空间中的向量的量化结果, 其中, 上述干扰值空间中向量的量化结果包括以下至少 之一: 码字维度向量、 分量系数幅度量化索引、 分量系数相位量化索引。 优选地, 上述多维码本化模块包括: 分解单元, 设置为对检测因子与干扰信道相 乘后得到的等效干扰信道进行 SVD 分解; 干扰值空间确定单元, 设置为取进行上述 SVD 分解后的大于门限的奇异值对应的右奇异值矩阵的向量组成的空间作为上述干 扰值空间; 量化单元, 设置为对上述干扰值空间中的每个向量分别执行以下操作: 将 当前向量向码本中的各个正交组进行投影; 在反馈维度为 Q维的情况下, 计算在每个 正交组中投影分量大小位于前 Q的 Q个分量的平方和, 选择平方和最大的正交组, 其 中, Q为正整数; 对上述当前向量在上述选择的正交组中的上述 Q个分量进行量化得 到上述量化结果; 上述遗留干扰估计模块设置为对上述干扰值空间中的每个分量执行 以下操作以确定上述遗留干扰标志: 将当前分量在上述选择的正交组中除上述 Q维之 外的其他维度的向量的分量作为上述遗留干扰; 对上述其他维度的向量的分量进行遗 留干扰估计得到上述遗留干扰标志。 优选地, 上述遗留干扰估计模块还设置为在上述遗留干扰为一个维度的向量的分 量的情况下, 如果该一个维度的向量的分量的遗留干扰估计小于预定的遗留干扰标志 门限, 则将该分量在上述正交组内的向量编号作为上述遗留干扰标志; 和 /或上述遗留 干扰估计模块还设置为在上述遗留干扰为多个维度的向量的分量的情况下, 如果上述 多个维度的向量的分量的遗留干扰估计中最小的遗留干扰估计小于上述预定的遗留干 扰标志门限, 则将上述最小的遗留干扰估计对应的分量在上述正交基向量组内的向量 编号作为上述遗留干扰标志。 根据本发明实施例的又一方面, 提供了一种协作消除干扰的装置, 位于协作基站 中, 包括: 接收模块, 设置为接收主基站发送的多维码本化信息和遗留干扰标志; 干 扰零空间计算模块, 设置为根据上述多维码本化信息计算得到干扰零空间; 遗留干扰 空间计算模块, 设置为根据上述干扰零空间和上述遗留干扰标志得到遗留干扰空间; 预编码计算模块, 设置为根据上述干扰零空间和遗留干扰空间得到预编码矩阵。 优选地, 上述干扰零空间计算模块包括: 恢复单元, 设置为根据上述多维码本化 信息中携带的上述干扰值空间中的多个维度的向量的量化结果恢复出干扰空间; 联合 分解单元, 设置为对上述干扰空间联合进行 SVD分解; 干扰零空间确定单元, 设置为 将 SVD 分解后得到的特征值中小于零空间向量选取门限的特征值对应的右奇异值矩 阵的向量作为上述干扰零空间。 优选地, 上述遗留干扰空间计算模块包括: 遗留干扰向量确定单元, 设置为对上 述遗留干扰标志对应的码字向量向上述干扰零空间进行投影得到遗留干扰向量; 遗留 干扰空间确定单元, 设置为将上述遗留干扰向量中模值大于或等于遗留干扰空间向量 门限的向量作为遗留干扰空间。 优选地, 上述预编码计算模块包括: 第一相关范数计算单元, 设置为计算上述遗 留干扰空间中的每个向量与上述协作基站的主信号信道的 PMI在上述干扰零空间的归 一化投影向量的相关范数; PMI重新确定单元, 设置为在计算得到的相关范数均小于 或等于遗留干扰向量和主信号相关性门限的情况下, 将上述多协作基站的主信号信道 的 PMI作为确定的 PMI, 否则, 重新确定 PMI; 第一预编码计算单元, 设置为将上述 确定的 PMI对应的码本向量向上述干扰零空间进行投影得到上述预编码矩阵。 优选地, 上述预编码计算模块包括: 第二相关范数计算单元, 设置为计算上述遗 留干扰空间中的每个向量与上述协作基站的主信号信道的 PMI在上述干扰零空间的归 一化投影向量的相关范数; 预编码向量空间重新确定单元, 设置为在计算得到的相关 范数均小于或等于遗留干扰向量和主信号相关性门限的情况下, 将干扰零空间作为预 编码向量空间, 否则, 重新确定上述预编码向量空间; 第二预编码计算单元, 设置为 将上述协作基站的主信号信道的 PMI对应的码本向量向上述确定的预编码向量空间进 行投影得到上述预编码矩阵。 根据本发明实施例的又一方面, 提供了一种协作消除干扰的系统, 包括上述的协 作消除干扰的装置的 UE, 和设置有上述的协作消除干扰的装置的协作基站。 在本发明实施例中, UE在向主基站反馈的信息中携带有多维码本化信息和自身估 计的遗留干扰标志, 协作基站在接收到反馈的这些信息后, 可以根据其中的多维码本 化信息和遗留干扰标志计算预编码矩阵, 因为其中携带有遗留干扰标志使得协作基站 进一步控制预编码矩阵和遗留干扰空间的相关性, 使得最终得到的预编码矩阵进一步 降低了对邻站的干扰。 通过上述方式解决了现有技术中难以确定遗留干扰的大小而导 致的难以有效降低对邻站 UE的干扰的技术问题, 达到了有效降低对邻站 UE的干扰, 提高 COMP的增益的技术效果。 附图说明 此处所说明的附图用来提供对本发明的进一步理解, 构成本申请的一部分, 本发 明的示意性实施例及其说明用于解释本发明, 并不构成对本发明的不当限定。 在附图 中- 图 1是根据相关技术的多小区的网络架构示意图; 图 2是根据相关技术的小区间多天线协作技术的一种优选流程图; 图 3是根据本发明实施例的协作消除干扰的方法的一种优选流程图; 图 4是根据本发明实施例的协作消除干扰的方法的另一种优选流程图; 图 5是根据本发明实施例的 UE生成多维码本化信息并进行遗留干扰估计的一种 优选流程图; 图 6是根据本发明实施例的协作基站计算干扰零空间的一种优选流程图; 图 7是根据本发明实施例的协作消除干扰的系统的一种优选结构框图; 图 8是根据本发明实施例的协作消除干扰的装置的一种优选结构框图; 图 9是根据本发明实施例的协作消除干扰的装置的另一种优选结构框图; 图 10是根据本发明实施例的协作消除干扰的装置的又一种优选结构框图; 图 11是根据本发明实施例的干扰零空间计算模块的一种优选结构框图; 图 12是根据本发明实施例的遗留干扰空间计算模块的一种优选结构框图; 图 13是根据本发明实施例的预编码计算模块的一种优选结构框图。 具体实施方式 下文中将参考附图并结合实施例来详细说明本发明。 需要说明的是, 在不冲突的 情况下, 本申请中的实施例及实施例中的特征可以相互组合。 本发明实施例提供了一种优选的协作消除干扰的方法,从 UE侧进行描述,如图 3 所示, 该方法包括以下步骤: 步骤 S302: UE根据信道信息生成多维码本化信息; 步骤 S304: UE对遗留干扰进行估计得到遗留干扰标志; 步骤 S306: UE通过主基站将多维码本化信息和遗留干扰标志发送给协作基站, 用于指示协作基站根据该多维码本化信息和遗留干扰标志得到预编码矩阵。 本发明实施例提供了一种优选的协作消除干扰的方法, 从协作基站侧进行描述, 如图 4所示, 该方法包括以下步骤: 步骤 S402: 协作基站接收主基站发送的多维码本化信息和遗留干扰标志; 步骤 S404: 协作基站根据多维码本化信息计算得到干扰零空间; 步骤 S406: 协作基站根据上述干扰零空间和遗留干扰标志得到遗留干扰空间; 步骤 S408: 协作基站根据上述干扰零空间和遗留干扰空间得到预编码矩阵。 在上述优选实施方式中, UE在向主基站反馈的信息中携带多维码本化信息和自身 估计的遗留干扰标志, 协作基站在接收到反馈的这些信息后, 可以根据多维码本化信 息和遗留干扰标志计算预编码矩阵, 因为其中携带有遗留干扰标志使得协作基站可以 进一步控制预编码矩阵和遗留干扰空间的相关性, 使得最终得到的预编码矩阵进一步 降低了对邻站的干扰。 通过上述方式解决了现有技术中难以确定遗留干扰的大小而导 致的难以有效降低对邻站 UE的干扰的技术问题, 达到了有效降低对邻站 UE的干扰, 提高 COMP的增益的技术效果。 在上述步骤 S302 中, UE所生成的多维码本化信息包括但不限于以下信息: UE 的主信号信道的预编码矩阵指示 (Precoding Matrix Indicator, 简称为 PMI)、 和 /或干 扰值空间中的向量的量化结果。 通过将这些信息传送给协作基站, 可以使得协作基站 有效恢复出原本的信道信息。其中, 量化结果可以包括但不限于以下至少之一: 码字、 分量系数幅度量化索引、 以及分量系数相位量化索引。 协作基站可以根据其中的幅度 和相位恢复出干扰空间。 下面对上述的步骤 S302和上述步骤 S304的实现方式进行具体描述, SP, 具体描 述 UE是如何生成多维码本化信息和估计遗留干扰标志的。 如图 5所示, UE可以按照以下步骤生成多维码本化信息并进行遗留干扰估计: 步骤 S502: 对检测因子与干扰信道相乘后得到的结果进行奇异值分解 (Singular Value Decomposition, 简称为 S VD ); 步骤 S504: 取进行 SVD分解后的大于门限的奇异值对应的右奇异值矩阵的向量 组成的空间作为干扰值空间; 步骤 S506: 对干扰值空间中的每个向量分别执行以下操作以确定遗留干扰标志: 将当前向量向码本中的各个正交组进行投影; 在反馈维度为 Q维的情况下, 计算在每 个正交组中投影分量大小位于前 Q的 Q个分量的平方和, 选择平方和最大的正交组; 对所述当前向量在所述选择的正交组中的所述 Q个分量进行量化得到所述量化结果; 将当前分量在所述选择的正交组中除所述 Q维之外的其他维度的向量的分量作为所述 遗留干扰; 对所述其他维度的向量的分量进行遗留干扰估计得到所述遗留干扰标志, 其中, Q为正整数。 优选地, 在上述步骤 S506中, 可以采用如表 1所示的 R8码本表和如表 2所示的 码本相关性表进行量化。 表 1
Figure imgf000011_0001
-0.5 0+0.5i 0.5 0-0.5i -0.5 0.5 0.5 -0.5
表 2
Figure imgf000012_0001
优选地,在步骤 S506中,主要存在两种情况,一种是遗留干扰是一个维度的向量, 一种是遗留干扰是多于一个维度的向量-
1 )在遗留干扰为一个维度的向量的分量的情况下,进行遗留干扰估计得到遗留干 扰标志包括: 如果该一个维度的向量的分量的遗留干扰估计小于预定的遗留干扰标志 门限, 则将该分量在正交组内的向量编号作为所述遗留干扰标志; 和 /或 2)在遗留干扰为多个维度的向量的分量的情况下,进行遗留干扰估计得到所述遗 留干扰标志包括: 如果多个维度的向量的分量的遗留干扰估计中最小的遗留干扰估计 小于预定的遗留干扰标志门限, 则将最小的遗留干扰估计对应的分量在正交组内的向 量编号作为遗留干扰标志。 可以按照以下公式计算一个或多个维度的向量的遗留干扰估计:
SINR=主信号能量 / ( (G*H*W) * (G*H*W) H); 其中, SINR表示所述遗留干扰估计, G表示所述检测因子, H表示干扰信号, W 表示当前计算的维度向量的分量, ( ·) H表示转置, ((G*H*W) * (G*H*W) H ) 表 示干扰信号能量。 下面对协作基站所执行的步骤进行具体描述, 上述步骤 S404, 计算干扰零空间可 以按照如图 6所示的步骤实现, 包括: 步骤 S602:协作基站根据多维码本化信息中携带的干扰值空间中的多个维度的向 量的量化结果恢复出干扰空间; 优选地, 可以按照以下方式进行恢复出干扰空间: 根据码本表, 取出码字对应的 码字向量 Ci; 根据幅度表和角度量化表, 取出幅度 Ai和角度 6i量化表, 恢复出维度 系数 ai, 其中, = A* (cosW) + sinW)), 恢复出的多维化信道向量 = C * ', 其中, i为维度编号。 步骤 S604: 对干扰空间联合进行 SVD分解; 步骤 S606: 将 SVD分解后得到的特征值中小于零空间向量选取门限的特征值对 应的右奇异值矩阵的向量作为干扰零空间。 上述步骤 S404, 计算遗留干扰空间可以按照以下步骤实现, 包括: 步骤 1 : 对遗留干扰标志对应的码字向量向干扰零空间进行投影得到遗留干扰向
步骤 2: 将该遗留干扰向量中模值大于或等于遗留干扰空间向量门限的向量作为 遗留干扰空间。 在计算的得到预编码矩阵之前, 需要预先计算遗留干扰空间中的每个向量与协作 基站的主信号信道的 PMI在干扰零空间的归一化投影向量的相关范数, 如果相关范数 小于或等于预定门限, 则直接将干扰零空间作为预编码向量空间, 将 PMI对应的码本 向量向干扰零空间进行投影, 如果相关范数大于预定门限, 则表明遗留干扰的影响较 大, 因此, 需要重新确定 PMI或者重新确定干扰零空间, 从而达到降低干扰的目的。 下面对如何重新确定 PMI和干扰零空间进行具体描述:
1 )重新确定 PMI。将码字作为所述确定的 PMI, 其中, 该码字是码本中使得所述 多协作基站的主信号信道的 PMI与该码字的相关范数大于所述遗留干扰向量和主信号 相关性门限, 且使得所述遗留干扰空间与所述多协作基站的主信号信道的 PMI在所述 干扰零空间的归一化投影向量的相关范数最小的码字。 即, 确定一个码字, 将给码字 作为 PMI, 相应的确定的这个码字需要满足上述的两个条件: 1 )使得所述多协作基站 的主信号信道的 PMI与该码字的相关范数大于遗留干扰向量和主信号相关性门限; 2) 从满足条件 1的码字中选择使得所述多协作基站的主信号信道的 PMI在所述干扰零空 间的归一化投影向量的相关范数最小的码字的作为最终选择的码字, 即作为重新确定 的 PMI。
2) 在本实施例中提供了两种重新确定干扰零空间的方式: 方式 1 ) 包括以下步骤:
S1 : 确定遗留干扰空间的维度 N是否小于干扰零空间的维度 M, 其中, M和 N 为自然数; S2: 如果小于, 则根据遗留干扰空间构造 (M-N) 个向量;
S3:将维度为 N的遗留干扰空间的向量用干扰零空间的向量作为基向量进行表示;
S4: 将用干扰零空间的向量作为基向量进行表示的遗留干扰空间的向量按照模值 从大到小的顺序进行排序后, 与构造的 (M-N) 个向量构成 M个线性无关向量;
S5: 从模值最大的遗留干扰空间向量开始对该 M 个线性无关向量进行正交化处 理;
S6: 将正交化处理后得到的正交化结果中从后向前的 (M-N) 个或多于 (M-N) 个向量作为重新确定的预编码向量空间。优选的, 可以根据需要从后向前选择(M-N) 个或多于 (M-N) 个向量。 方式 2), 主要包括以下步骤:
S1 : 确定遗留干扰空间的维度 N是否大于或等于干扰零空间的维度 M, 其中, M 和 N为自然数;
S2: 如果大于或等于, 则对遗留干扰空间进行 SVD分解; S3:将 SVD分解后得到的特征值中平方与所有特征值平方和之间的比值小于遗留 干扰空间向量门限的特征值对应的右奇异值矩阵中的向量作为预编码向量空间。 在本实施例中还提供了一种协作消除干扰的系统, 该装置用于实现上述实施例及 优选实施方式, 已经进行过说明的不再赘述。如以下所使用的,术语"单元"或者"模块" 可以实现预定功能的软件和 /或硬件的组合。尽管以下实施例所描述的装置较佳地以软 件来实现, 但是硬件, 或者软件和硬件的组合的实现也是可能并被构想的。 图 7是根 据本发明实施例的协作消除干扰的系统的一种优选结构框图, 如图 7 所示, 包括: UE702、 主基站 704、 协作基站 706。 其中, UE中设置有协作消除干扰的装置, 协作 基站中也设置有协作消除干扰的装置, 下面对该结构进行说明。 如图 8所示, 位于 UE中的协作消除干扰的装置包括: 多维码本化模块 802、遗留 干扰估计模块 804以及反馈模块 806。 下面对该结构进行说明。
1 )多维码本化模块 802, 设置为根据信道信息生成多维码本化信息; 优选地, UE 对信道信息根据信道和码本码字的相关性, 选择相关性最大的码字, 并在该码字所在 的正交组里, 对其他维度的分量系数进行量化;
2)遗留干扰估计模块 804, 与多维码本化模块 802耦合, 设置为对遗留干扰进行 估计得到遗留干扰标志; 优选地, UE对多维码本化丢弃维度的干扰进行估计, 如果超 过门限, 则把最大的干扰的维度码字作为遗留干扰标志。
3 ) 反馈模块 806, 与遗留干扰估计模块 804耦合, 设置为通过主基站将所述多维 码本化信息和所述遗留干扰标志发送给协作基站, 用于指示所述协作基站根据所述多 维码本化信息和所述遗留干扰标志得到预编码矩阵。 如图 9所示, 多维码本化模块 802包括: 分解单元 902, 设置为对检测因子与干 扰信道相乘后得到的结果进行 SVD分解;干扰值空间确定单元 904,设置为取进行 SVD 分解后的大于门限的奇异值对应的右奇异值矩阵的向量组成的空间作为所述干扰值空 间; 量化单元 906, 设置为对所述干扰值空间中的每个向量分别执行以下操作: 将当 前向量向码本中的各个正交组进行投影; 在反馈维度为 Q维的情况下, 计算在每个正 交组中投影分量大小位于前 Q的 Q个分量的平方和,选择平方和最大的正交组,其中,
Q为正整数; 对所述当前向量在所述选择的正交组中的所述 Q个分量进行量化得到所 述量化结果; 遗留干扰估计模块 804设置为对所述干扰值空间中的每个分量执行以下 操作以确定遗留干扰标志: 将当前分量在所述选择的正交组中除所述 Q维之外的其他 维度的向量的分量作为所述遗留干扰; 对所述其他维度的向量的分量进行遗留干扰估 计得到所述遗留干扰标志。 在一个优选实施方式中, 遗留干扰估计模块 804还设置为在遗留干扰为一个维度 的向量的分量的情况下, 如果该一个维度的向量的分量的遗留干扰估计小于预定的遗 留干扰标志门限, 则将该分量在所述正交组内的向量编号作为所述遗留干扰标志; 和 / 或遗留干扰估计模块 804还设置为在遗留干扰为多个维度的向量的分量的情况下, 如 果所述多个维度的向量的分量的遗留干扰估计中最小的遗留干扰估计小于所述预定的 遗留干扰标志门限, 则将所述最小的遗留干扰估计对应的分量在所述正交基向量组内 的向量编号作为所述遗留干扰标志。 如图 10所示, 位于协作基站中的预编码装置包括: 接收模块 1002、 干扰零空间 计算模块 1004、 遗留干扰空间计算模块 1006、 以及预编码计算模块 1008。 下面对该 结构进行说明。
1 ) 接收模块 1002, 设置为接收主基站发送的多维码本化信息和遗留干扰标志;
2)干扰零空间计算模块 1004, 与接收模块 1002耦合, 设置为根据所述多维码本 化信息计算得到干扰零空间; 优选地,干扰零空间计算模块 1004可以把所有收到的干 扰信道多维化码字恢复成干扰信道向量, 并进行联合分解, 得到干扰零空间。
3 )遗留干扰空间计算模块 1006, 与干扰零空间计算模块 1004耦合, 设置为根据 所述干扰零空间和所述遗留干扰标志得到遗留干扰空间; 优选地, 遗留干扰空间计算 模块 1106对所有遗留标志的干扰信道进行联合分析, 得到遗留干扰空间。
4)预编码计算模块 1008, 与遗留干扰空间计算模块 1006耦合, 设置为根据所述 干扰零空间和遗留干扰空间得到预编码矩阵。 在一个优选实施方式中, 如图 11所示, 干扰零空间计算模块 1004包括: 恢复单 元 1102, 设置为根据所述多维码本化信息中携带的所述干扰值空间中的多个维度的向 量的量化结果恢复出干扰空间; 联合分解单元 1104, 设置为对干扰空间联合进行 SVD 分解; 干扰零空间确定单元 1106, 设置为将 SVD分解后得到的特征值中小于零空间 向量选取门限的特征值对应的右奇异值矩阵的向量作为干扰零空间。 在一个优选实施方式中, 如图 12所示, 遗留干扰空间计算模块 1006包括: 遗留 干扰向量确定单元 1202, 设置为对遗留干扰标志对应的码字向量向干扰零空间进行投 影得到遗留干扰向量; 遗留干扰空间确定单元 1204, 设置为将所述遗留干扰向量中模 值大于或等于遗留干扰空间向量门限的向量作为遗留干扰空间。 在一个优选实施方式中, 如图 13所示, 预编码计算模块 1008包括: 第一相关范 数计算单元 1302, 设置为计算所述遗留干扰空间中的每个向量与所述协作基站的主信 号信道的 PMI 在所述干扰零空间的归一化投影向量的相关范数; PMI 重新确定单元 1304, 设置为在计算得到的相关范数均小于或等于遗留干扰向量和主信号相关性门限 的情况下,将多协作基站的主信号信道的 PMI作为确定的 PMI,否则,重新确定 PMI; 第一预编码计算单元 1306,设置为将确定的 PMI对应的码本向量向干扰零空间进行投 影得到预编码矩阵。 优选地, 上述 PMI重新确定单元 1304还设置为按照以下方式重新确定 PMI: 将 码字作为所述确定的 PMI,其中,该码字是码本中使得多协作基站的主信号信道的 PMI 与该码字的相关范数大于主信号相关性门限, 且使得遗留干扰空间与多协作基站的主 信号信道的 PMI在干扰零空间的归一化投影向量的相关范数最小的码字。 在一个优选实施方式中,所述预编码计算模块 1008还可以包括: 第二相关范数计 算单元 10082, 设置为计算所述遗留干扰空间中的每个向量与所述协作基站的主信号 信道的 PMI在所述干扰零空间的归一化投影向量的相关范数; 预编码向量空间重新确 定单元 10084, 设置为在计算得到的相关范数均小于或等于遗留干扰向量和主信号相 关性门限的情况下, 将干扰零空间作为预编码向量空间, 否则, 重新确定所述预编码 向量空间; 第二预编码计算单元 10086, 设置为将所述协作基站的主信号信道的 PMI 对应的码本向量向所述确定的预编码向量空间进行投影得到所述预编码矩阵。 优选地, 本优选实施例还提供了一种生成预编码向量空间的方式, 如果遗留干扰 空间的向量和 PMI (PMI表示主信号向量) 对应的码字向量在干扰零空间的投影向量 的相关性都小于等于门限 C, 则从遗留干扰空间 B中去除所有向量, 如果最后的遗留 干扰空间 B如果不为空, 则进行以下处理:
1 ) 如果遗留干扰空间 B的维度 (k) 小于干扰零空间的维度 (j ): 对干扰空间 B的向量从最大模值开始进行构造线性无关向量组, 然后从最大模值 的干扰空间向量开始, 进行正交化。 然后从最开始得到的正交向量开始, 去掉正交向 量, 最后保留 (j-k) 个向量, 组成预编码向量空间。 2) 如果遗留干扰空间的维度大于等于干扰零空间的维度: 对遗留干扰空间进行 SVD分解, 从最小特征值开始, 选择特征值平方 /所有特征 值平方和小于门限 e 的包含特征值对应的右奇异值矩阵中的向量组成预编码向量空 间。 优选地, 可以按照以下公式选择主信号向量 X, 其中, X可以是码本中的任一码 字,也可以是其他非码本向量,只要可以满足 |PMI*X|大于等于门限 D,且 maX ( |Bi*(X 在空间 AO的归一化投影向量) | ) 最小, 就可以将该 X作为主信号向量, 其中, AO表 示预编码空间。 通过本发明实施例上述提供的方式可以进一步控制遗留干扰的大小, 同时可以减 少 COMP后 SINR陡变异常点, 有利于编码方式和速率的自适应调整。 下面结合几个优选实施例对本发明实施例进行具体描述。 优选实施例 1 在本优选实施例中, 以 3小区 3个 UE场景的遗留干扰控制方法为例进行描述。 其中, 门限 c (遗留干扰向量和主信号相关性门限) 为: 0.6; 门限 a (零空间向量选取门限) 为: 0.001 ; 门限 u (UE遗留干扰标志门限): 35; 门限 f (遗留干扰空间向量门限): 0.3; 门限 e (主信号预编码向量选取门限): 0.2。 该三个小区三个用户场景情形下的 CoMP 的遗留干扰控制方法可以包括如下步 骤: 步骤 SI : UE侧估计信道和计算并反馈多维码本化信道信息。 假设 UE1主信号信道为 HI 1 : H11 =
[ 0.3131 + 0.2794Ϊ -0.3569 - 0.4109Ϊ -0.0625 - 0.2399Ϊ 1.8446 - 0.0288Ϊ 0.5752 - 0.0294Ϊ -0.2097 + 0.7725Ϊ -0.6426 - 0.1475Ϊ 0.8445 - 0.6428Ϊ]; 主信号信道码本化后, 得到的 PMI=4; 其中, 预先设定的检测因子 G = [-0.0008 - 0.3464i 0.3352 - 0.1492i]。 小区 2对小区 1的干扰信道为 H12: H12 =
[ -0.3136 - 0.0537Ϊ 0.3010 - 0.0164Ϊ -0.0771 - 0.0953Ϊ -1.1493 + 0.077H 0.8876 + 0.7816Ϊ -0.6232 + 0.0735Ϊ 0.3503 + 0.6433Ϊ -0.2748 - 0.8774i]o 对 G*H12进行 SVD分解 (这里 *表示相乘), 取分解后的值空间为: [ 0.6396 - 0.3848Ϊ -0.3294 - 0.0216Ϊ
0.2916 - 0.3072Ϊ -0.3157 - 0.2342Ϊ]; 在确定维度数量是 3维的情况下, 选择最大分量和包含的维度向量的分量进行量 化, 对干扰值空间多维度量化结果分别为- 第一个维度: 码字为 12, 系数量化为 1 ; 第二个维度: 码字为 10, 系数量化为 0.2704 + 0.2650i, 根据幅度和角度量化表可 得: 幅度量化索引为 0, 相位量化索引为 5; 第三个维度: 码字为 11, 系数量化为: 0.2557 + 0.0436i, 根据幅度和角度量化表 可得: 幅度量化索引为 0, 相位量化索引为 5。 遗留干扰估计 SINR为:
SINR = 315.7029; 由于最小 SINR大于门限 所以没有遗留干扰标志。 小区 3对小区 1的干扰信道为 H13: H13 =
[ 0.1800 + 0.0064Ϊ -0.5820 + 1.9147Ϊ 0.0238 - 0.7970Ϊ -0.5242 - 1.4639Ϊ
0.4784 - 1.1865Ϊ 0.7670 + 0.3493Ϊ -0.3136 - 0.555H -0.7843 + 0.1310i]; 对 G*H13进行 SVD分解后, 取分解后的值空间为: [ -0.0099 + 0.3593Ϊ
0.6578 - 0.1370Ϊ -0.3137 + 0.0993Ϊ -0.5070 - 0.2324Ϊ]; 在确定维度数量为 3维的情况下, 选择最大分量和包含的维度向量的分量进行量 化, 对干扰值空间多维度量化结果分别为: 第一个维度: 码字为 15, 系数量化为 1 ; 第二个维度: 码字为 13, 系数量化为 -0.4477 - 0.4814i, 根据幅度和角度量化表可 得: 幅度量化索引为 1, 相位量化索引为 2; 第三个维度: 码字为 16, 系数量化为 0.0008 - 0.4224i, 根据幅度和角度量化表可 得: 幅度量化索引为 1, 相位量化索引为 3。 遗留干扰估计 SINR为:
SINR = 32.8548; 由于 SINR大于门限 所以遗留干扰在正交向量基组内向量编号为: 2。 步骤 S2: 基站收到多维码本化信道信息和遗留干扰标志后发给协作基站。 基站 1收到 UE1的多维码本化信道信息(步骤 S1中的小区 2对 UE1的三个维度 的码字和系数和小区 3对 UE1 的三个维度的码字和系数), 并分别发送给协作基站 2 和协作基站 3。 步骤 S3: 协作基站收到多维码本化信道信息后, 恢复干扰信道空间 A, 对 A进行 SVD分解后得到其零空间为 AO。 基站 2和基站 3根据收到的 (步骤一中的小区 2对 UE1的三个维度的码字和系数和小区 3对 UE1的三个维度的码字和系数), 分别得到 恢复干扰空间。其中, 在恢复时所依据的角度量化如表 3所示, 幅度量化如表 4所示。 表 3
Figure imgf000021_0001
表 4
Figure imgf000021_0002
在本优选实施例中, 根据上述表 3和表 4的幅度和角度表格恢复的干扰空间为:
[ 0.6587 + 0.0736Ϊ
-0.1205 - 0.4163Ϊ
0.4987 + 0.0074Ϊ
0.0395 - 0.350H] ; 对基站 2恢复的干扰空间进行 SVD分解,取特征值小于门限 a的奇异值对应的右 奇异值矩阵的向量组成的干扰零空间为 -
[ 0.1569 - 0.4040Ϊ -0.4974 - 0.0369Ϊ -0.0083 - 0.3522Ϊ
0.8870 - 0.0050Ϊ 0.0325 + 0.126H -0.0864 + 0.0315Ϊ
0.0435 - 0.1227Ϊ 0.8503 - 0.0066Ϊ -0.0056 - 0.1057Ϊ
-0.0833 - 0.0390Ϊ -0.0150 + 0.1048Ϊ 0.9253 - 0.0033Ϊ] ; 根据上述表 3和表 4基站 3恢复的干扰空间为: [ 0.3598 - 0.403 H
-0.5566 - 0.0720Ϊ 0.2226 - 0.0720Ϊ
0.4194 + 0.403H]; 对基站 3恢复的干扰空间进行 SVD分解,取特征值小于门限 a的奇异值对应的右 奇异值矩阵的向量组成的干扰零空间为 -
[ 0.4274 - 0.3637Ϊ -0.2259 + 0.0610Ϊ -0.1319 + 0.5666Ϊ
0.7871 + 0.063H 0.0915 + 0.014H 0.1385 - 0.1839Ϊ 0.0690 - 0.0617Ϊ 0.9630 + O.OllOi -0.0195 + 0.0940Ϊ 0.2163 + 0.0787Ϊ -0.0675 - 0.0682Ϊ 0.7712 + 0.0678Ϊ]。 步骤 S4:协作基站对所有收到的遗留干扰码字标志对应的码字向量向干扰零空间 投影得到遗留干扰向量, 保留遗留干扰向量中模值大于等于门限 f 的向量形成遗留干 扰空间 B。 在本优选实施例中, 基站 2没有收到遗留干扰标志; 基站 3的遗留干扰标志中的 向量标志对应的码字向量, 向干扰零空间投影得到遗留干扰向量为:
[0.5, 0.5i, -0.5, 0.5] τ; 遗留干扰向量的模值为 1, 大于门限 f, 因此基站 3的遗留干扰空间 B为:
[0.5, 0.5i, -0.5, 0.5] τ; 步骤 S5:协作基站计算遗留干扰空间 B与 PMI在空间 AO的归一化投影向量的相 关范数。 如果此相关范数大于门限 c, 则重新选择 PMI, 否则用本基站原来的 PMI。 在本优选实施例中,基站 2没有收到遗留干扰标志,因此基站 2用原来的 PMI (基 站 2的 PMI为 6)。 基站 3收到的遗留干扰标志为 2, 其对应的码本索引为 14的码本空间, 其与 PMI 在空间 A0的归一化投影向量的相关范数为 1,大于门限 c,因此基站 3要重新选择 PMI (基站 3原有的 PMI为 14)。其中 PMI的码本空间为 [0.5 0.5 -0.5 0.5] T, AO的归 一化投影向量为-
[0.5 0.5 -0.5 0.5 其中,如果 PMI在空间 AO的归一化投影向量的相关范数大于门限 c,则计算 PMI 与码本中的任一码字的相关范数,如果相关范数大于门限 e,且遗留干扰空间 B与 PMI 在空间 AO的归一化投影向量的相关范数最小。 在本优选实施例中, 基站 3计算出的相关范数大于门限 c, 则计算 PMI与码本中 的 16个码字的相关范数, 这 16个相关范数为 0.5、 0.5、 0.5、 0.5、 0.2706、 0.6533、 0.6533、 0.2706、 0.5、 0.5、 0.5、 0.5、 0、 1、 0、 0, 其中大于门限 e的为 0.5、 0.5、 0.5、 0.5、 0.2706、 0.6533、 0.6533、 0.2706、 0.5、 0.5、 0.5、 0.5、 1, 大于门限 e 的范数所 对应的码字分别为 1、 2、 3、 4、 5、 6、 7、 8、 9、 10、 11、 12、 14, 其中, 遗留干扰 空间 B与 PMI在空间 AO的归一化投影向量的相关范数为 0.5143、 0.5368、 0.5942、 0.7223 0.3014、 0.9022、 0.7195、 0.2943 0.6147、 0.5725 0.6041、 0.5298、 1.0000, 最小相关范数为 0.2943, 对应的码字为 8, 所以基站 3最终选择的 PMI为 8。 步骤 S6: 用 PMI对应的码本向量向干扰信道空间 A的零空间投影得到预编码矩 阵 W。 在本优选实施例中,基站 2的 PMI为 6,基站 3的 PMI为 8,分别采用各自的 PMI 对应的码字向对应的干扰信道空间 A的零空间进行投影得到预编码矩阵 W。 因此, 基站 2的预编码矩阵是: [ 0.4300 + 0.2296Ϊ -0.2134 + 0.3843Ϊ -0.0620 - 0.3554Ϊ 0.5004 + 0.4343Ϊ]; 基站 3的预编码矩阵是: [ 0.5057 - 0.2280Ϊ 0.2115 - 0.2182Ϊ 0.0356 - 0.6374Ϊ -0.1377 - 0.4165i 优选实施例 2 在本优选实施例中, 以 2小区 2ue场景的遗留干扰控制方法为例进行描述, 其中, 小区 2给 UE1做 COMP。 其中, BBU1, BBU2分别表示两个基站; UE1, UE2分别 表示两个终端。
UE1归属在 BBU1下, 且发现 BBU2的下行信号干扰; UE2归属在 BBU2下, 且 发现 BBU1的下行信号干扰。 门限 c (遗留干扰向量和主信号相关性门限) 为: 0.6; 门限 a (零空间向量选取门限) 为: 0.1 ; 门限 u (UE遗留干扰标志门限): 15(dB); 门限 f (遗留干扰空间向量门限): 0.7。 假设 UE1的主信号信道为-
[-0.7264 + 0.2697Ϊ -0.6658 - 1.277H -0.0972 - 0.3957Ϊ 0.4820 + 0.1496Ϊ
2.0465 + 0.865H 0.4859 + 1.5874Ϊ 1.0333 - 0.4792Ϊ 0.3115 + 0.6340Ϊ]; 主信号信道码本化后, 得到的 PMI=12; 其中, 预先设定的检测因子 G= [-0.9755 - 0.0289Ϊ 2.6506 + 0.2058i]。 小区 2对小区 1的干扰信道 H为:
[0.0772 - 0.4964Ϊ 0.2142 - 0.4225Ϊ -0.1221 - 0.1337Ϊ 0.4899 - 0.7944Ϊ -0.7083 - 0.3485Ϊ -0.1310 + 0.4070Ϊ 0.9815 - 0.8919Ϊ 0.4952 - 0.4300Ϊ]; G*H进行 SVD分解取右奇异矩阵的第一列 (即为干扰值空间) 为: [ -0.4249 + 0.1317i -0.1462 - 0.3268Ϊ
0.6501 + 0.4546Ϊ 0.2018 + 0.062H]; 在确定维度数量为 2维的情况下, 选择最大分量和包含的维度向量的分量进行量 化, 对干扰值空间进行多维度量化或得到的量化结果分别为: 最大维度: 码字编号 9, 量化为 4比特 (1000), 系数为: 1, 不需要上报; 次大维度:码字编号 12,所在正交向量基组内向量编号为: 4,量化为 2比特(11 ), 系数为: -0.4558 - 0.2230i = 0.5075*eA(j* (-2.6865)), 相位为 -153.9273度。 根据次大系数的量化方案, 0.5075小于 0.54,量化为 lbit取值为 0,实际代表 0.44; 相位 -153.9273度在 -180度和 -135之间, 量化为 3bit, 取值为 000, 实际代表 -157度。 量化后的系数为: -0.4050 - 0.1719io 对各个遗留维度向量的遗留干扰估计 SINR分别为: SINR11= 13.72 dB (对应码本编号 11, 正交基组内编号为 3 );
SINR12= 16.35 dB (对应码本编号 10); 由于 SINR11最小且小于小于门限 u, 因此遗留干扰标志为: 组内编号 3, 量化为 2比特 (10)。
UE1将多维度码本化结果以及遗留干扰标志反馈给服务基站 BBU1 ; BBU1把多维度码本化结果以及遗留干扰标志转发给协作基站 2。 协作基站 2收到协作请求后, 对 BBU1的多维度码字按照表 4和表 5进行恢复: 表 4
Figure imgf000025_0001
表 5
Figure imgf000025_0002
Figure imgf000026_0001
αΐ =
[ 0.4309 - 0.1452Ϊ
0.1694 + 0.2435Ϊ
-0.7725 + 0.0002Ϊ
-0.3144 + 0.0982Ϊ] ; 对 [alf进行 SVD分解 ([ · ]H表示转置, 即对 od转置后进行 SVD分解), 结果
特征值: [1 0 0 0]; 取其特征值小于门限 a 的奇异值对应的右奇异值矩阵的向量组成的干扰零空间 为:
[ 0.2653 - 0.1326Ϊ -0.6780 - 0.3702Ϊ -0.2290 - 0.2368Ϊ
0.8548 - 0.0370Ϊ 0.1366 + 0.3654Ϊ 0.0092 + 0.166H
0.2950 - 0.2553Ϊ 0.0155 - 0.2512Ϊ -0.3688 - 0.2272Ϊ
0.0876 - 0.1414Ϊ -0.4326 + 0.0228Ϊ 0.8210 - 0.0457Ϊ];
BBU1转发来的标志对应的码字向量, 向干扰零空间投影得到 B1为: [ 0.3239 - 0.0104Ϊ
-0.5255 - 0.1122Ϊ
-0.2220 + 0.1122i
0.6274 + 0.0104Ϊ]; 该向量长度为: 0.9216; 根据门限 f 判断, 遗留干扰空间为: [ 0.5000 -0.5000
-0.5000
0.5000];
PMI向干扰零空间的投影向量 PI为:
[ 0.5635 + 0.0798Ϊ
-0.3920 - 0.2992Ϊ
-0.0590 + 0.3370Ϊ
0.3089 - 0.4124Ϊ];
B1*P1/( |B1|*|P1|)= 0.7051, 根据门限 c进行判断, 应保留 Bl。 遗留干扰空间 B为:
[ 0.3239 - 0.0104Ϊ
-0.5255 - 0.1122Ϊ
-0.2220 + 0.1122i
0.6274 + 0.0104i
由于 B的维度 K=l, 小于干扰零空间的维度 J =3, 因此需要构造
B=
[ 0.3239 - 0.0104Ϊ
-0.5255 - 0.1122Ϊ
-0.2220 + 0.1122i
0.6274 + 0.0104Ϊ];
Bl=
[ -0.3542 - 0.3806Ϊ
-0.3889 + 0.2532Ϊ -0.2065 - 0.1389Ϊ
0.1948 + 0.0332Ϊ];
B2=
[ 0.0949 - 0.2472Ϊ
-0.5163 + 0.0538Ϊ
-0.5909 - 0.1150Ϊ
1.4484 - 0.0353i]。 对上述构造的 B、 Bl、 B2进行正交化处理, 正交化的结果为:
[ 0.3239 - 0.0104Ϊ -0.4345 - 0.2899Ϊ 0.0225 - 0.1847i
-0.5255 - 0.1122Ϊ -0.2235 + 0.1392Ϊ 0.4099 - 0.0368Ϊ
-0.2220 + 0.1122i -0.1800 - 0.228H -0.0314 - 0.2427Ϊ
0.6274 + 0.0104Ϊ 0.0309 + 0.2010Ϊ 0.3560 - 0.0948Ϊ] ; 对上述正交化的结果从后向前取 J-K个向量,将取的 J-K个向量作为预编码空间:
[ -0.4345 - 0.2899Ϊ 0.0225 - 0.1847Ϊ
-0.2235 + 0.1392Ϊ 0.4099 - 0.0368Ϊ
-0.1800 - 0.228H -0.0314 - 0.2427Ϊ
0.0309 + 0.2010Ϊ 0.3560 - 0.0948Ϊ] ; 将 PMI对应的码字向量向预编码向量空间投影得到投影向量, 对投影向量做归一 化后得到预编码矩阵: [ 0.4995 + 0.2526Ϊ
0.0667 - 0.520H
0.1426 + 0.3206Ϊ
-0.2903 - 0.4521i 优选实施例 3 在本优选实施例中,以 3小区 3UE场景的遗留干扰控制方法为例进行描述。其中, 小区 3 ^ UE1禾口 UE2 j¾ COMP。
BBUl , BBU2, BBU3分别表示三个基站, UE1, UE2, UE3分别表示三个终端。 UE1归属在 BBU1下, 发现 BBU3的下行信号干扰; UE2归属在 BBU2下, 发现
BBU3的下行信号干扰; UE3归属在 BBU3下, 未发现下行信号干扰。 门限 c (遗留干扰向量和主信号相关性门限) 为: 0.7; 门限 a (零空间向量选取门限) 为: 0.001; 门限 u (UE遗留干扰标志门限): 31; 门限 f (遗留干扰空间向量门限): 0.3; 门限 e (主信号预编码向量选取门限): 0.3。 假设 UE1主信号信道 Hslg为:
[-0.7086 - 0.1596Ϊ, -0.2563 + 0.3978Ϊ, 0.1030 + 1.1413Ϊ, -0.5756 - 0.9590Ϊ 0.1364 + 1.2499Ϊ ,1.6319 + 0.1339Ϊ ,-1.0330 - 0.2442Ϊ, -0.0549 + 1.0216i]; 主信号信道码本化后, 得到的 PMI=6; 其中, 预先设定的检测因子 G = [0.3019 + 0.9052i -1.0590 - 2.0130i] o 小区 3对 UEl的干扰信道 Hmtf为:
[0.1837 + 0.2300Ϊ, -0.0307 + 0.886H, 0.1168 - 0.4964Ϊ, 1.2073 + 0.3406Ϊ -0.7757 + 0.2183Ϊ, -0.1480 - 0.7176Ϊ, -0.1838 - 0.0916Ϊ, 0.5855 - 0.1312Ϊ]; G*Hmtf进行 SVD分解取值空间向量为:
[0.3355 - 0.4742Ϊ -0.6357 - 0.3930Ϊ 0.1499 - 0.128H -0.2507 - 0.0472Ϊ]; 在确定维度数量为 2维的情况下, 选择最大分量和包含的维度向量的分量进行量 化, 对干扰值空间进行多维度量化后得到的量化结果分别为: 最大维度: 码字 12; 次大维度: 码字 9; 所在正交向量基组内的向量编号为: 最大维度编号 4, 次大维度编号 1 ; 次大维度向量的遗留干扰估计 SINR为: SINR = 27.5316。 由于最小 SINR小于门限 因此遗留干扰标志为: 1 (改为遗留干扰向量在组内 编号) 假设 UE2主信号信道 Hslg为:
[0.3246 + 0.1434Ϊ0, 3047 + 0.1287Ϊ, -0.7787 + 0.2328Ϊ, 1.0716 - 0.4984Ϊ
-0.5840 - 1.4254Ϊ, -1.0533 + 0.1778Ϊ, -0.0612 - 0.5368Ϊ, 0.9195 - 1.2835Ϊ]; 主信号信道码本化后, 得到的 PMI=15 ; 检测因子 G = [0.1564 + 0.1254Ϊ 0.6638 + 1.7118i]。 小区 3对 UE2的干扰信道 Hmtf为:
[-1.1894 + 0.2548Ϊ, -0.2402 + 0.1227Ϊ, -0.0347 + 0.6238Ϊ, 0.6497 + 0.1689Ϊ
-0.9537 + 1.0782Ϊ, -0.7289 - 0.5774Ϊ, 0.9038 + 0.5232Ϊ, -0.6957 - 1.5622Ϊ];
G*Hmtf进行 SVD分解取值空间向量为: [-0.5386 + 0.2049Ϊ
0.0902 + 0.3279Ϊ
-0.0758 - 0.3970Ϊ
0.4579 + 0.4234Ϊ]; 对干扰值空间进行多维度量化后得到的量化结果分别为: 最大维度: 码字 3; 次大维度: 码字 2; 所在正交向量基组内向量编号为: 1 ; 次大维度向量的遗留干扰估计 SINR分别为: 最大维度编号 3, 次大维度编号 2。 其中, SINR = 1.6780; 由于 SINR小于门限 因此遗留干扰标志为: 2。 假设 UE3主信号信道为-
[0.6988 - 0.1689Ϊ, -0.3181 + 0.1407Ϊ, 1.6536 - 0.2026Ϊ, 1.2404 + 0.3565Ϊ
0.1272 - 0.6056Ϊ, -1.3569 - 0.5173Ϊ, 1.1567 + 0.5585Ϊ, -1.3527 - 0.8050Ϊ]; 主信号信道进行码本化后, 得到的 PMI=14。
UEl将多维度码本化结果以及遗留干扰标志反馈给服务基站 BBUl ; UE2把多维 度码本化结果以及遗留干扰标志反馈给服务基站 BBU2;
BBU1把码本化信息和遗留干扰标志转发给协作基站 3, BBU2把码本化信息和遗 留干扰标志转发给协作基站 3。 协作基站 3在收到协作请求后,对 BBU1和 BBU2的值空间向量进行恢复,得到: αΐ = [0.5000, 0 - 0.5000Ϊ, 0.5000, 0 - 0.5000i]H α2 = [0.5000, -0.5000, 0.5000, -0.5000]Η 由于值空间向量的反馈向量只采用 1维, 因此直接取码本作为值空间向量。 对 [αΐ, α2]进行 SVD分解, 结果为: U = [-0.7071 + 0.0000Ϊ 0.7071 + 0.0000Ϊ
-0.5000 - 0.5000Ϊ -0.5000 - 0.5000Ϊ] D = [1.3066 0 0 0
0 0.5412 0 0] v =
[-0.4619 - 0.1913Ϊ, 0.1913 - 0.4619Ϊ, -0.6986 - 0.0165Ϊ, -0.0329 - 0.1027Ϊ
0.1913 + 0.4619Ϊ, 0.4619 - 0.1913Ϊ, 0.0494 - 0.0959Ϊ, -0.6918 - 0.0987Ϊ
-0.4619 - 0.1913Ϊ, 0.1913 - 0.4619Ϊ, 0.6986 + 0.0165Ϊ, 0.0329 + 0.1027Ϊ
0.1913 + 0.4619Ϊ, 0.4619 - 0.1913Ϊ, -0.0494 + 0.0959Ϊ, 0.6918 + 0.0987i]。 取其中特征值小于门限 a的奇异值对应的右奇异值矩阵的向量组成干扰零空间, 得到的干扰零空间为:
[-0.6986 - 0.0165Ϊ -0.0329 - 0.1027Ϊ
0.0494 - 0.0959Ϊ -0.6918 - 0.0987Ϊ
0.6986 + 0.0165Ϊ 0.0329 + 0.1027Ϊ
-0.0494 + 0.0959Ϊ 0.6918 + 0.0987Ϊ] ;
BBU1转发来的标志对应的码字向量, 向干扰零空间投影得到 B1为: [0.5, 0.5,-0.5,-0.5] τ;
BBU2转发来的标志对应的码字向量, 向干扰零空间投影得到 Β2为: [0.5, 0 + 0.5Ϊ, -0.5, 0 - 0.5i] T。 根据门限 f进行判断, 遗留干扰空间为:
[0.5 0.5
0.5 0+ 0.5i
-0.5 -0.5000
-0.5 0 - 0.5i]。
PMI向干扰零空间的投影向量 PI为:
[ 0.5000 + 0.0000Ϊ
-0.0000 - 0.0000Ϊ -0.5000 + O.OOOOi
-0.0000 + 0.0000Ϊ] ;
Bl *Pl/( |B1 |*|P1 |)=0.7071 ;
B2*Pl/( |B2|*|P1 |)= 0.7071; 根据门限 c进行判断可知需要保留 Bl, B2。 因此, 遗留干扰空间 B为:
[0.5000 + O.OOOOi 0.5000 + O.OOOOi
0.5000 - O.OOOOi -0.0000 + 0.5000Ϊ
-0.5000 + O.OOOOi -0.5000
-0.5000 + O.OOOOi -0.0000 - 0.5000Ϊ]; 由于 B的维度 K=2等于干扰零空间维度的为 J=2。因此。对遗留干扰空间进行 SVD 分解, 分解得到的结果为-
U = [-0.7071 0.7071
-0.5000 + 0.5000Ϊ -0.5000 + 0.5000Ϊ]
D = [1.3066 0 0 0
0 0.5412 0 0]
V =
[-0.4619 + 0.1913Ϊ, 0.1913 + 0.4619Ϊ, 0.6986 + 0.0165Ϊ, 0.1027 - 0.0329Ϊ
-0.4619 - 0.1913Ϊ, 0.1913 - 0.4619Ϊ, -0.0959 - 0.0494Ϊ, 0.6918 + 0.0987Ϊ
0.4619 - 0.1913Ϊ, -0.1913 - 0.4619Ϊ, 0.6986 + 0.0165Ϊ, 0.1027 - 0.0329Ϊ
0.4619 + 0.1913Ϊ, -0.1913 + 0.4619Ϊ, -0.0959 - 0.0494Ϊ, 0.6918 + 0.0987Ϊ]; 选择特征值平方 /所有特征值平方和小于门限 e的特征值对应的右奇异值矩阵中的 向量作为预编码向量空间, 得到的预编码向量空间为:
[ 0.1913 + 0.4619Ϊ 0.1913 - 0.4619Ϊ -0.1913 - 0.4619Ϊ -0.1913 + 0.4619i]。 将 PMI对应的码字向量向预编码向量空间进行投影得到预编码矩阵, 该预编码矩 阵为-
[0.2500 - 0.0000Ϊ
-0.1768 - 0.1768Ϊ
-0.2500 - 0.0000Ϊ
0.1768 + 0.1768i 优选实施例 4 在本优选实施例中, 以 3小区 3个 UE场景的遗留干扰控制方法为例进行描述。 其中, 小区 3给 UE1和 UE2做 COMP。
BBUl , BBU2, BBU3分别表示三个基站, UE1, UE2, UE3分别表示三个终端。
UE1归属在 BBU1下, 发现 BBU3的下行信号干扰; UE2归属在 BBU2下, 发现 BBU3的下行信号干扰; UE3归属在 BBU3下, 未发现下行信号干扰。 门限 c (遗留干扰向量和主信号相关性门限) 为: 0.7; 门限 a (零空间向量选取门限) 为: 0.001; 门限 u (UE遗留干扰标志门限): 31; 门限 e (遗留干扰空间向量门限): 0.2。 假设 UE1主信号信道 Hslg为:
[-0.7086 - 0.1596Ϊ, -0.2563 + 0.3978Ϊ, 0.1030 + 1.1413Ϊ, -0.5756 - 0.9590Ϊ
0.1364 + 1.2499Ϊ ,1.6319 + 0.1339Ϊ ,-1.0330 - 0.2442Ϊ, -0.0549 + 1.0216i]; 主信号信道码本化后, 得到的 PMI=6; 其中, 预先设定的 G = [0.3019 + 0.9052i -1.0590 - 2.0130i] o 小区 3对 UE1的干扰信道 Hmtf为:
[0.1837 + 0.2300Ϊ, -0.0307 + 0.886H, 0.1168 - 0.4964Ϊ, 1.2073 + 0.3406Ϊ -0.7757 + 0.2183Ϊ, -0.1480 - 0.7176Ϊ, -0.1838 - 0.0916Ϊ, 0.5855 - 0.1312Ϊ]; G*Hmtf进行 SVD分解取值空间向量为:
[0.3355 - 0.4742Ϊ -0.6357 - 0.3930Ϊ 0.1499 - 0.128H -0.2507 - 0.0472Ϊ]; 在确定维度数量为 3维的情况下, 选择最大分量和包含的维度向量的分量进行量 对干扰值空间进行多维度量化后得到的量化结果分别为: 最大维度: 码字 3; 次大维度: 码字 1 ; 第 3大维度: 码字 4; 所在正交向量基组内向量编号分别为: 3,1,4; 第 3大维度向量的遗留干扰估计 SINR为: SINR = 29.5921。 因最小 SINR小于门限 U, 所以遗留干扰标志为: 4。 假设 UE2主信号信道 Hslg为:
[0.3246 + 0.1434Ϊ0, 3047 + 0.1287Ϊ, -0.7787 + 0.2328Ϊ, 1.0716 - 0.4984Ϊ -0.5840 - 1.4254Ϊ, -1.0533 + 0.1778Ϊ, -0.0612 - 0.5368Ϊ, 0.9195 - 1.2835Ϊ]; 主信号信道码本化后, 得到的 PMI=15; 其中, 预先设定的检测因子 G = [0.1564 + 0.1254i 0.6638 + 1.7118i]。 小区 3对 UE2的干扰信道 Hmtf为:
[-1.1894 + 0.2548Ϊ, -0.2402 + 0.1227Ϊ, -0.0347 + 0.6238Ϊ, 0.6497 + 0.1689Ϊ -0.9537 + 1.0782Ϊ, -0.7289 - 0.5774Ϊ, 0.9038 + 0.5232Ϊ, -0.6957 - 1.5622Ϊ];
G*Hmtf进行其中, 预先设定的分解取值空间向量为: [-0.5386 + 0.2049Ϊ
0.0902 + 0.3279Ϊ -0.0758 - 0.3970Ϊ 0.4579 + 0.4234i]。 对干扰值空间进行多维度量化后得到的量化结果分别为: 最大维度: 码字 14; 次大维度: 码字 13; 第 3大维度: 码字 16; 所在正交向量基组内向量编号分别为: 2,1,4; 第 3大维度向量的遗留干扰估计 SINR为: SINR = 3.0417; 由于 SINR小于门限 因此所以遗留干扰标志为: 4。 假设 UE3主信号信道为- [0.6988 - 0.1689Ϊ, -0.3181 + 0.1407Ϊ, 1.6536 - 0.2026Ϊ, 1.2404 + 0.3565Ϊ 0.1272 - 0.6056Ϊ, -1.3569 - 0.5173Ϊ, 1.1567 + 0.5585Ϊ, -1.3527 - 0.8050i]。 主信号信道码本化后, 得到的 PMI=14。 UE1将多维度码本化结果以及遗留干扰标志反馈给服务基站 BBU1 ; UE2把多维 度码本化结果以及遗留干扰标志反馈给服务基站 BBU2; BBU1 把码本化信息和遗留 干扰标志转发给协作基站 3, BBU2把码本化信息和遗留干扰标志转发给协作基站 3。 协作基站 3在收到协作请求后, 对 BBU1和 BBU2的多维度码字按照表 3和表 4 进行恢复。 αΐ = [0.3905 - 0.3515Ϊ
-0.6424 - 0.2718Ϊ
0.2730 - 0.0746Ϊ
-0.3654 - 0.1543Ϊ]
α2 = [0.3905 - 0.3515Ϊ
-0.6424 - 0.2718Ϊ
0.2730 - 0.0746Ϊ
-0.3654 - 0.1543Ϊ] 对 [αΐ, α2]进行 SVD分解, 结果为: U = [ -0.7071 + 0.0000Ϊ -0.7071 - 0.0000Ϊ
0.5315 + 0.4664Ϊ -0.5315 - 0.4664Ϊ]
D = [1.3229 0 0 0
0 0.5001 0 0]
V =
[-0.2224 + 0.4266Ϊ, -0.5161 - 0.1345Ϊ, 0.2700 - 0.1558Ϊ, -0.3078 - 0.5405Ϊ
0.4733 + 0.3117Ϊ, 0.5645 - 0.0558Ϊ, 0.3452 + 0.3408Ϊ, -0.1429 - 0.3183Ϊ
-0.3517 - 0.0883Ϊ, 0.1583 + 0.4447Ϊ, 0.7557 - 0.1686Ϊ, 0.0558 + 0.2075Ϊ
0.5447 + 0.1384Ϊ, -0.4075 + 0.0703Ϊ, 0.2293 - 0.1239Ϊ, 0.6667 + 0.0273i]。 取其特征值小于门限 a的奇异值对应的右奇异值矩阵的向量组成干扰零空间, 该 干扰零空间为:
[0.2700 - 0.1558Ϊ -0.3078 - 0.5405Ϊ
0.3452 + 0.3408Ϊ -0.1429 - 0.3183Ϊ 0.7557 - 0.1686Ϊ 0.0558 + 0.2075Ϊ
0.2293 - 0.1239Ϊ 0.6667 + 0.0273i]。
BBUl将转发来的标志对应的码字向量, 向干扰零空间进行投影得到 B1 :
[0.2854 - 0.1782Ϊ
0.1070 - 0.2926Ϊ
-0.5321 + 0.073H
-0.1442 + 0.563 li];
BBU2将转发来的标志对应的码字向量, 向干扰零空间进行投影得到 B2:
[0.1329 + 0.2797Ϊ
-0.0817 - 0.0369Ϊ
-0.4556 + 0.0872Ϊ
-0.4110 + 0.2294Ϊ];
根据门限 f进行判断, 得到的遗留干扰空间为:
[0.2854 - 0.1782Ϊ 0.1329 + 0.2797Ϊ
0.1070 - 0.2926Ϊ -0.0817 - 0.0369Ϊ
-0.5321 + 0.073H -0.4556 + 0.0872Ϊ
-0.1442 + 0.563 li -0.4110 + 0.2294i]。
PMI向干扰零空间的投影向量 PI为:
[ 0.5000 + 0.0000Ϊ
-0.0000 - 0.0000Ϊ
-0.5000 + 0.0000Ϊ
-0.0000 + 0.0000Ϊ];
Β1*Ρ1/( |B1|*|P1|)=0.7229; Β2*Ρ1/( |B2|*|P1 |)= 0.3674。 根据门限 c进行判断, 因为超过了门限(;, 所以保留 B1与 B2。 因此, 遗留干扰 空间 B为:
[0.3596 - 0.0866Ϊ 0.0498 + 0.2129Ϊ
0.0848 - 0.5649Ϊ -0.0825 - 0.2129Ϊ
-0.3704 + 0.0854Ϊ -0.2672 + 0.2232Ϊ
-0.0872 + 0.5578Ϊ -0.2577 + 0.2025]。 由于 B的维度 K=2等于干扰零空间的维度(J=2),所以对遗留干扰空间进行 SVD 分解, 分解得到的结果为- U = [ -0.9008 0.4342
-0.4236 + 0.0953Ϊ -0.8788 + 0.1978Ϊ]
D = [1.0523 0 0 0
0 0.3842 0 0]
V =
[-0.3472 - 0.0070Ϊ, 0.1830 - 0.559H, -0.1005 + 0.1755Ϊ, 0.0596 + 0.6992Ϊ
-0.0201 + 0.5618Ϊ, 0.3940 - 0.1940Ϊ, -0.1236 - 0.6714Ϊ, 0.1024 - 0.1205Ϊ
0.4044 - 0.1872Ϊ, 0.0777 - 0.5517Ϊ, 0.6658 - 0.0857Ϊ, -0.1513 - 0.1325Ϊ
0.1600 - 0.5823Ϊ, 0.3866 + 0.0345Ϊ, -0.1532 - 0.1379Ϊ, 0.6635 - 0.0439i]。 选择特征值平方 /所有特征值平方和小于门限 e的特征值对应的右奇异值矩阵中的 向量作为预编码向量空间, 得到的预编码向量空间为:
[ 0.1830 - 0.559H
0.3940 - 0.1940Ϊ
0.0777 - 0.5517Ϊ
0.3866 + 0.0345i]。 将 PMI对应的码字向量向预编码向量空间进行投影得到预编码矩阵, 该预编码矩 阵为-
[0.1277 - 0.2324Ϊ
0.1907 - 0.0530Ϊ 0.0805 - 0.2378Ϊ
0.1683 + 0.0475i]。 在另外一个实施例中, 还提供了一种软件, 该软件用于执行上述实施例及优选实 施方式中描述的技术方案。 在另外一个实施例中, 还提供了一种存储介质, 该存储介质中存储有上述软件, 该存储介质包括但不限于: 光盘、 软盘、 硬盘、 可擦写存储器等。 从以上的描述中, 可以看出, 本发明实现了如下技术效果: UE在向主基站反馈的 信息中携带多维码本化信息和自身估计的遗留干扰标志, 协作基站在接收到反馈的这 些信息后, 可以根据其中的多维码本化信息和遗留干扰标志计算预编码矩阵, 因为其 中携带有遗留干扰标志使得协作基站进一步控制预编码矩阵和遗留干扰空间的相关 性, 使得最终得到的预编码矩阵进一步降低了对邻站的干扰。 通过上述方式解决了现 有技术中难以确定遗留干扰的大小而导致的难以有效降低对邻站 UE的干扰的技术问 题, 达到了有效降低对邻站 UE的干扰, 提高 COMP的增益的技术效果。 显然, 本领域的技术人员应该明白, 上述的本发明的各模块或各步骤可以用通用 的计算装置来实现, 它们可以集中在单个的计算装置上, 或者分布在多个计算装置所 组成的网络上, 可选地, 它们可以用计算装置可执行的程序代码来实现, 从而, 可以 将它们存储在存储装置中由计算装置来执行, 并且在某些情况下, 可以以不同于此处 的顺序执行所示出或描述的步骤, 或者将它们分别制作成各个集成电路模块, 或者将 它们中的多个模块或步骤制作成单个集成电路模块来实现。 这样, 本发明不限制于任 何特定的硬件和软件结合。 以上所述仅为本发明的优选实施例而已, 并不用于限制本发明, 对于本领域的技 术人员来说, 本发明可以有各种更改和变化。 凡在本发明的精神和原则之内, 所作的 任何修改、 等同替换、 改进等, 均应包含在本发明的保护范围之内。

Claims

权 利 要 求 书
1. 一种协作消除干扰的方法, 包括:
用户终端 UE根据信道信息生成多维码本化信息;
所述 UE对遗留干扰进行估计得到遗留干扰标志;
所述 UE通过主基站将所述多维码本化信息和所述遗留干扰标志发送给协 作基站, 用于指示所述协作基站根据所述多维码本化信息和所述遗留干扰标志 得到预编码矩阵。
2. 根据权利要求 1所述的方法, 其中, 所述多维码本化信息包括: 所述 UE主信 号信道的预编码矩阵指示 PMI、 和 /或干扰信道值空间中的向量的量化结果。
3. 根据权利要求 2所述的方法, 其中, 所述干扰值空间中向量的量化结果包括以 下至少之一: 码字维度向量、分量系数幅度量化索引、分量系数相位量化索引。
4. 根据权利要求 2或 3所述的方法, 其中,
所述 UE根据所述信道信息生成所述多维码本化信息包括:
对检测因子与干扰信道相乘后得到的等效干扰信道进行奇异值分解 SVD; 取进行所述 SVD 分解后的大于门限的奇异值对应的右奇异值矩阵的向量 组成的空间作为所述干扰值空间;
对所述干扰值空间中的每个向量分别执行以下操作:
将当前向量向码本中的各个正交组进行投影;
在反馈维度为 Q维的情况下, 计算在每个正交组中投影分量大小位于前 Q 的 Q个分量的平方和, 选择平方和最大的正交组, 其中, Q为正整数;
对所述当前向量在所述选择的正交组中的所述 Q个分量进行量化得到所述 量化结果; 以及
所述 UE对所述遗留干扰进行估计得到遗留干扰标志包括:
对所述干扰值空间中的每个分量执行以下操作:
将当前分量在所述选择的正交组中除所述 Q维之外的其他维度的向量的分 量作为所述遗留干扰; 对所述其他维度的向量的分量进行遗留干扰估计得到所述遗留干扰标志。
5. 根据权利要求 4所述的方法, 其中, 在所述遗留干扰为一个维度的向量的分量的情况下, 进行遗留干扰估计得 到所述遗留干扰标志包括: 如果所述一个维度的向量的分量的遗留干扰估计小 于预定的遗留干扰标志门限, 则将该分量在所述正交组内的向量编号作为所述 遗留干扰标志; 和 /或
在所述遗留干扰为多个维度的向量的分量的情况下, 进行遗留干扰估计得 到所述遗留干扰标志包括: 如果所述多个维度的向量的分量的遗留干扰估计中 最小的遗留干扰估计小于所述预定的遗留干扰标志门限, 则将所述最小的遗留 干扰估计对应的分量在所述正交组内的向量编号作为所述遗留干扰标志。
6. 根据权利要求 5所述的方法, 其中, 按照以下公式计算所述遗留干扰估计:
SINR=主信号能量 / ( (GHW) (GHW) H);
其中, SINR表示所述遗留干扰估计, G表示所述检测因子, H表示干扰信 号信道, W表示当前计算的维度向量的分量, ( · ) H表示转置。
7. 一种协作消除干扰的方法, 包括:
协作基站接收主基站发送的多维码本化信息和遗留干扰标志; 所述协作基站根据所述多维码本化信息计算得到干扰零空间; 所述协作基站根据所述干扰零空间和所述遗留干扰标志得到遗留干扰空 间;
所述协作基站根据所述干扰零空间和遗留干扰空间得到预编码矩阵。
8. 根据权利要求 7所述的方法, 其中, 所述协作基站根据所述多维码本化信息和 所述遗留干扰标志计算所述干扰零空间包括:
所述协作基站根据所述多维码本化信息中携带的所述干扰值空间中的多个 维度的向量的量化结果恢复出干扰空间;
对所述干扰空间联合进行奇异值分解 SVD;
将所述 SVD 分解后得到的特征值中小于零空间向量选取门限的特征值对 应的右奇异值矩阵的向量作为所述干扰零空间。
9. 根据权利要求 7所述的方法, 其中, 所述协作基站根据所述干扰零空间和所述 遗留干扰标志得到遗留干扰空间包括:
对所述遗留干扰标志对应的码字向量向所述干扰零空间进行投影得到遗留 干扰向量;
将所述遗留干扰向量中模值大于或等于遗留干扰空间向量门限的向量作为 遗留干扰空间。
10. 根据权利要求 7所述的方法, 其中, 所述协作基站根据所述干扰零空间和遗留 干扰空间得到预编码矩阵包括:
计算所述遗留干扰空间中的每个向量与所述协作基站的主信号信道的预编 码矩阵指示 PMI在所述干扰零空间的归一化投影向量的相关范数;
如果计算得到的相关范数均小于或等于遗留干扰向量和主信号相关性门 限, 则将所述协作基站的主信号信道的 PMI作为确定的 PMI, 否则, 重新确定 PMI;
将所述确定的 PMI 对应的码本向量向所述干扰零空间进行投影得到所述 预编码矩阵。
11. 根据权利要求 10所述的方法, 其中, 所述重新确定 PMI包括:
将码字作为所述确定的 PMI, 其中, 所述码字是码本中使得所述协作基站 的主信号信道的 PMI与该码字的相关范数大于主信号相关性门限,且使得所述 遗留干扰空间与所述协作基站的主信号信道的 PMI 在所述干扰零空间的归一 化投影向量的相关范数最小的码字。
12. 根据权利要求 7所述的方法, 其中, 所述协作基站根据所述干扰零空间和遗留 干扰空间得到预编码矩阵包括:
计算所述遗留干扰空间中的每个向量与所述协作基站的主信号信道的 PMI 在所述干扰零空间的归一化投影向量的相关范数;
如果计算得到的相关范数均小于或等于遗留干扰向量和主信号相关性门 限, 则将干扰零空间作为预编码向量空间, 否则, 重新确定所述预编码向量空 间;
将所述协作基站的主信号信道的 PMI 对应的码本向量向所述确定的预编 码向量空间进行投影得到所述预编码矩阵。
13. 根据权利要求 12所述的方法, 其中, 重新确定所述预编码向量空间包括: 确定所述遗留干扰空间的维度 N是否小于所述干扰零空间的维度 M,其中, M和 N为自然数;
如果小于, 则根据遗留干扰空间构造 (M-N) 个向量;
将所述维度为 N的遗留干扰空间的向量用所述干扰零空间的向量作为基向 量进行表示;
将用所述干扰零空间的向量作为基向量进行表示的遗留干扰空间的向量按 照模值从大到小的顺序进行排序后, 与构造的(M-N)个向量构成 M个线性无 关向量;
从模值最大的遗留干扰空间向量开始对所述 M个线性无关向量进行正交 化处理;
将正交化处理后得到的正交化结果中从后向前的(M-N)个或多于(M-N) 个向量作为所述重新确定的预编码向量空间。
14. 根据权利要求 12所述的方法, 其中, 重新确定所述预编码向量空间包括: 确定所述遗留干扰空间的维度 N 是否大于或等于所述干扰零空间的维度 M, 其中, M和 N为自然数;
如果大于或等于, 则对所述遗留干扰空间进行奇异值分解 SVD; 将所述 SVD 分解后得到的特征值平方与所有特征值平方和之间的比值小 于遗留干扰空间向量门限的特征值对应的右奇异值矩阵中的向量作为所述重新 确定的预编码向量空间。
15. 一种协作消除干扰的装置, 位于用户终端 UE中, 包括:
多维码本化模块, 设置为根据信道信息生成多维码本化信息; 遗留干扰估计模块, 设置为对遗留干扰进行估计得到遗留干扰标志; 反馈模块, 设置为通过主基站将所述多维码本化信息和所述遗留干扰标志 发送给协作基站, 用于指示所述协作基站根据所述多维码本化信息和所述遗留 干扰标志得到预编码矩阵。
16. 根据权利要求 15所述的装置, 其中, 所述多维码本化信息包括: 所述 UE主信 号信道的预编码矩阵指示 PMI、和 /或干扰信道值空间中的向量的量化结果, 其 中, 所述干扰值空间中向量的量化结果包括以下至少之一: 码字维度向量、 分 量系数幅度量化索引、 分量系数相位量化索引。
17. 根据权利要求 16所述的装置, 其中,
所述多维码本化模块包括:
分解单元, 设置为对检测因子与干扰信道相乘后得到的等效干扰信道进行 奇异值分解 SVD;
干扰值空间确定单元,设置为取进行所述 SVD分解后的大于门限的奇异值 对应的右奇异值矩阵的向量组成的空间作为所述干扰值空间;
量化单元, 设置为对所述干扰值空间中的每个向量分别执行以下操作: 将 当前向量向码本中的各个正交组进行投影; 在反馈维度为 Q维的情况下, 计算 在每个正交组中投影分量大小位于前 Q的 Q个分量的平方和,选择平方和最大 的正交组, 其中, Q为正整数; 对所述当前向量在所述选择的正交组中的所述 Q个分量进行量化得到所述量化结果;
所述遗留干扰估计模块设置为对所述干扰值空间中的每个分量执行以下操 作以确定所述遗留干扰标志: 将当前分量在所述选择的正交组中除所述 Q维之 外的其他维度的向量的分量作为所述遗留干扰; 对所述其他维度的向量的分量 进行遗留干扰估计得到所述遗留干扰标志。
18. 根据权利要求 17所述的装置, 其中,
所述遗留干扰估计模块还设置为在所述遗留干扰为一个维度的向量的分量 的情况下, 如果该一个维度的向量的分量的遗留干扰估计小于预定的遗留干扰 标志门限, 则将该分量在所述正交组内的向量编号作为所述遗留干扰标志; 和 / 或
所述遗留干扰估计模块还设置为在所述遗留干扰为多个维度的向量的分量 的情况下, 如果所述多个维度的向量的分量的遗留干扰估计中最小的遗留干扰 估计小于所述预定的遗留干扰标志门限, 则将所述最小的遗留干扰估计对应的 分量在所述正交基向量组内的向量编号作为所述遗留干扰标志。
19. 一种协作消除干扰的装置, 位于协作基站中, 包括- 接收模块, 设置为接收主基站发送的多维码本化信息和遗留干扰标志; 干扰零空间计算模块, 设置为根据所述多维码本化信息计算得到干扰零空 间; 遗留干扰空间计算模块, 设置为根据所述干扰零空间和所述遗留干扰标志 得到遗留干扰空间;
预编码计算模块, 设置为根据所述干扰零空间和遗留干扰空间得到预编码 矩阵。
20. 根据权利要求 19所述的装置, 其中, 所述干扰零空间计算模块包括:
恢复单元, 设置为根据所述多维码本化信息中携带的所述干扰值空间中的 多个维度的向量的量化结果恢复出干扰空间;
联合分解单元, 设置为对所述干扰空间联合进行奇异值分解 SVD; 干扰零空间确定单元,设置为将所述 SVD分解后得到的特征值中小于零空 间向量选取门限的特征值对应的右奇异值矩阵的向量作为所述干扰零空间。
21. 根据权利要求 19所述的装置, 其中, 所述遗留干扰空间计算模块包括:
遗留干扰向量确定单元, 设置为对所述遗留干扰标志对应的码字向量向所 述干扰零空间进行投影得到遗留干扰向量;
遗留干扰空间确定单元, 设置为将所述遗留干扰向量中模值大于或等于遗 留干扰空间向量门限的向量作为遗留干扰空间。
22. 根据权利要求 19所述的装置, 其中, 所述预编码计算模块包括:
第一相关范数计算单元, 设置为计算所述遗留干扰空间中的每个向量与所 述协作基站的主信号信道的预编码矩阵指示 PMI 在所述干扰零空间的归一化 投影向量的相关范数;
PMI重新确定单元, 设置为在计算得到的相关范数均小于或等于遗留干扰 向量和主信号相关性门限的情况下,将所述协作基站的主信号信道的 PMI作为 确定的 PMI, 否则, 重新确定 PMI;
第一预编码计算单元,设置为将所述确定的 PMI对应的码本向量向所述干 扰零空间进行投影得到所述预编码矩阵。
23. 根据权利要求 19所述的装置, 其中, 所述预编码计算模块包括:
第二相关范数计算单元, 设置为计算所述遗留干扰空间中的每个向量与所 述协作基站的主信号信道的 PMI 在所述干扰零空间的归一化投影向量的相关 范数; 预编码向量空间重新确定单元, 设置为在计算得到的相关范数均小于或等 于遗留干扰向量和主信号相关性门限的情况下, 将干扰零空间作为预编码向量 空间, 否则, 重新确定所述预编码向量空间;
第二预编码计算单元,设置为将所述协作基站的主信号信道的 PMI对应的 码本向量向所述确定的预编码向量空间进行投影得到所述预编码矩阵。
24. 一种协作消除干扰的系统, 包括设置有上述权利要求 15至 18中任一项所述的 协作消除干扰的装置的用户终端 UE, 和设置有上述权利要求 19至 23中任一 项所述的协作消除干扰的装置的协作基站。
PCT/CN2013/082942 2012-12-10 2013-09-04 协作消除干扰的方法、装置及系统 WO2014089992A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201210528022.6A CN103873184B (zh) 2012-12-10 2012-12-10 协作消除干扰的方法、装置及系统
CN201210528022.6 2012-12-10

Publications (1)

Publication Number Publication Date
WO2014089992A1 true WO2014089992A1 (zh) 2014-06-19

Family

ID=50911345

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2013/082942 WO2014089992A1 (zh) 2012-12-10 2013-09-04 协作消除干扰的方法、装置及系统

Country Status (2)

Country Link
CN (1) CN103873184B (zh)
WO (1) WO2014089992A1 (zh)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106559116B (zh) * 2015-09-28 2020-07-28 联芯科技有限公司 K用户系统及其预编码矩阵的确定方法
WO2018027813A1 (zh) * 2016-08-11 2018-02-15 华为技术有限公司 一种反馈参数上报方法和装置
CN109905155B (zh) * 2019-01-21 2021-06-04 郑州轻工业学院 一种基于内外级联预编码的干扰管理方法、无线通信系统
CN111884692B (zh) * 2020-07-28 2021-06-25 西安邮电大学 一种射频反射镜使能的收发端联合空间调制传输方法
CN113095354B (zh) * 2021-03-03 2022-04-29 电子科技大学 基于辐射源特征子空间知识的未知雷达目标识别方法

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102231659A (zh) * 2011-06-22 2011-11-02 中兴通讯股份有限公司 多点协作传输的预编码方法、系统和设备
CN102237974A (zh) * 2010-05-07 2011-11-09 华为技术有限公司 预编码矩阵获取方法及装置
CN102457319A (zh) * 2010-10-28 2012-05-16 中兴通讯股份有限公司 业务数据下发方法和装置
US20120287799A1 (en) * 2011-05-09 2012-11-15 Texas Instruments Incorporated Channel feedback for coordinated multi-point transmissions

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102237974A (zh) * 2010-05-07 2011-11-09 华为技术有限公司 预编码矩阵获取方法及装置
CN102457319A (zh) * 2010-10-28 2012-05-16 中兴通讯股份有限公司 业务数据下发方法和装置
US20120287799A1 (en) * 2011-05-09 2012-11-15 Texas Instruments Incorporated Channel feedback for coordinated multi-point transmissions
CN102231659A (zh) * 2011-06-22 2011-11-02 中兴通讯股份有限公司 多点协作传输的预编码方法、系统和设备

Also Published As

Publication number Publication date
CN103873184B (zh) 2018-04-27
CN103873184A (zh) 2014-06-18

Similar Documents

Publication Publication Date Title
EP2618500A1 (en) Wireless communication control method, wireless communication system, wireless base station and mobile terminal
JP2016500942A (ja) WiFiビームフォーミング、フィードバックおよびサウンディング(WiBEAM)のための方法
CN103312390B (zh) 基于干扰对齐的预编码方法,发射机和设备
WO2009030102A1 (fr) Procédé de codage à l'avance et de planification pour utilisateurs multiples et station de base pour mettre en œuvre le procédé
EP2469724B1 (en) Precoding method and device
WO2014101242A1 (zh) 报告信道状态信息csi的方法、用户设备和基站
CN104871437A (zh) Fdd系统中信道互易性补偿方法和装置
WO2014089992A1 (zh) 协作消除干扰的方法、装置及系统
WO2012055276A1 (zh) 业务数据下发方法和装置
WO2017148429A1 (zh) 传输数据的方法和装置
CN103580789B (zh) 多点协作传输预编码处理方法、装置及系统
WO2012094948A1 (zh) 一种协作预编码方法、协作信息交互方法及系统
EP4092919A1 (en) Method and apparatus for transmitting information
EP3017572B1 (en) System and method for trellis coded quantization with selective feedback and interpolation
CN103259582B (zh) 一种多点协同传输预编码方法、终端及基站
WO2021007810A1 (zh) 预编码的方法和通信设备
US11159346B2 (en) Co-polarized feedback for frequency domain compression
CN104243374B (zh) 一种信号传输方法、装置及系统
CN103607260B (zh) 基于mimo系统总干扰泄漏最小的预编码矩阵组的选择算法
WO2013078743A1 (zh) 协作多点多用户mimo系统的预编码方法及矩阵生成装置
WO2006114030A1 (fr) Procede permettant d’augmenter la capacite dans le sens montant d’un systeme a entrees multiples et a sorties multiples du type wcdma et son appareil
CN105763240B (zh) Mimo干扰广播信道中基于多点协作的干扰对齐方法
CN109314563B (zh) 用于确定mimo无线通信系统中的csi的方法和装置
WO2014049416A2 (en) Method and apparatus for coordinated multipoint downlink transmission between two cells
WO2012113185A1 (zh) 一种发射信号预处理发送方法及装置

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: 13863619

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 13863619

Country of ref document: EP

Kind code of ref document: A1