Method and user equipment for determining a precoder for a MIMO system

Abstract

A method and a UE for determining a precoder in a MIMO system are provided. Specifically, the BS sends at least one CSI-RS to the UE. The UE receives the at least one CSI-RS and estimates at least one covariance matrix of at least one downlink channel matrix based on the at least one CSI-RS at different times and frequencies. Then, the UE sends at least one covariance matrix or at least one parameter associated with the at least one covariance matrix to the BS for the BS to reconstruct the at least one covariance matrix and determine the precoder based on the at least one reconstructed covariance matrix. This invention achieves the beneficial effect of improving the accuracy of the precoder.

Description

Technical Field

[0001] The disclosed embodiments generally relate to wireless communication systems, and more specifically to methods and user equipment for determining a precoder for a multi-input multi-output (MIMO) system. Background Technology

[0002] In legacy networks using 3GPP 5G New Radio (NR), under MIMO networks, user equipment (UE) can measure the Channel State Information Reference Signal (CSI-RS) transmitted from the base station (BS) and determine the downlink channel matrix based on the CSI-RS. The UE can then calculate a precoder based on the downlink channel matrix and report the compressed / quantized precoder to the BS using one of the prescribed codebooks via a encoding matrix indicator (PMI). Therefore, the BS can use this precoder to transmit subsequent Physical Downlink Shared Channel (PDSCH). However, due to inaccuracies in the UE's calculated and reported precoder calculations, subsequent transmissions of the PDSCH with the applied precoder may suffer from severe distortion. Summary of the Invention

[0003] A method for determining the precoder of a multi-input multi-output (MIMO) system is proposed. Specifically, the BS can send at least one CSI-RS to the UE. The UE can receive at least one CSI-RS and estimate at least one covariance matrix of at least one downlink channel matrix based on at least one CSI-RS at different times and frequencies. Then, the UE can send at least one covariance matrix or at least one parameter associated with at least one covariance matrix to the BS for the BS to reconstruct at least one covariance matrix and determine the precoder based on at least one reconstructed covariance matrix.

[0004] A method for determining a precoder for a MIMO system includes a UE receiving at least one CSI-RS from a network; the UE estimating at least one covariance matrix of at least one downlink channel matrix based on the at least one CSI-RS at different times and frequencies; and the UE transmitting the at least one first covariance matrix or at least one parameter associated with the at least one first covariance matrix to the network.

[0005] A method for determining a precoder for a MIMO system includes a BS transmitting at least one CSI-RS to a UE for the UE to estimate at least one first covariance matrix of at least one downlink channel matrix based on the at least one CSI-RS at different times and frequencies; the BS receiving the at least one first covariance matrix or at least one parameter associated with the at least one first covariance matrix from the UE; the BS reconstructing at least one second covariance matrix based on the at least one first covariance matrix or the at least one parameter associated with the at least one first covariance matrix; and the BS deriving a precoder based on the at least one second covariance matrix.

[0006] A UE for determining a precoder of a MIMO system includes: a transceiver module for receiving at least one CSI-RS from a network; and a channel state information processing circuit for estimating at least one covariance matrix of at least one downlink channel matrix based on the at least one CSI-RS at different times and frequencies, wherein the transceiver module transmits the at least one first covariance matrix or at least one parameter associated with the at least one first covariance matrix to the network.

[0007] A method and UE for determining the precoder of a MIMO system are proposed, which achieves the beneficial effect of improving the accuracy of the precoder.

[0008] Other embodiments and beneficial effects are described in the detailed description below. This summary is not intended to define the invention. The invention is defined by the claims. Attached Figure Description

[0009] The accompanying drawings illustrate embodiments of the invention, wherein the same numbers indicate the same components.

[0010] Figure 1 An exemplary 5G new radio network supporting the determination of a precoder for a MIMO system is illustrated in an embodiment of the present invention.

[0011] Figure 2 This is a simplified block diagram of the gNB and UE entities according to an embodiment of the present invention.

[0012] Figure 3 An embodiment of message transmission according to an embodiment of the present invention is shown.

[0013] Figure 4A to 4G An embodiment of message transmission according to an embodiment of the present invention is shown.

[0014] Figure 5 This is a flowchart of a method for determining the precoder of a MIMO system according to an embodiment of the present invention. Detailed Implementation

[0015] Reference is now made to some embodiments of the present invention, examples of which are described in the accompanying drawings.

[0016] Figure 1 Exemplary 5G new radio (NR) network 100 supporting the determination of a precoder for a MIMO system is illustrated according to various aspects of the present invention. The 5G NR network 100 includes a UE 110 and a gNB 121 communicatively connected to an access network 120 operating in a licensed frequency band (e.g., 30 GHz to 300 GHz in millimeter wave (mmWave)). The access network 120 provides radio access using Radio Access Technology (RAT) (e.g., 5G NR technology). The access network 120 is connected to a 5G core network 130 via an NG interface, and more specifically, to a User Plane Function (UPF) via an NG user-plane part (NG-u), and to an Access and Mobility Management Function (AMF) via an NG control-plane part (NG-c). For load sharing and redundancy purposes, a gNB may be connected to multiple UPFs / AMFs. UE 110 can be a smartphone, wearable device, Internet of Things (IoT) device, or tablet computer, etc. Alternatively, UE 110 can be a notebook computer (NB) or personal computer (PC) with a data card inserted or installed, wherein the data card includes a modem and radio frequency transceiver to provide wireless communication capabilities.

[0017] gNB 121 can provide communication coverage for a geographic coverage area, where communication with UE 110 is supported via communication link 101. Communication link 101 shown in 5G NR network 100 may include UL transmission from UE 110 to gNB 121 (e.g., on the Physical Uplink Control Channel (PUCCH) or Physical Uplink Shared Channel (PUSCH)) or downlink (DL) transmission from gNB 121 to UE 110 (e.g., on the Physical Downlink Control Channel (PDCCH) or Physical Downlink Shared Channel (PDSCH)).

[0018] Figure 2 This is a simplified block diagram of gNB 121 and UE 110 according to an embodiment of the present invention. For gNB 121, antenna 197 transmits and receives radio signals. An RF transceiver module 196 (including a transmitter and a receiver) coupled to the antenna receives RF signals from the antenna, converts the RF signals into baseband signals, and sends them to processor 193. RF transceiver module 196 also converts the baseband signals received from processor 193 into RF signals and sends them to antenna 197. Processor 193 processes the received baseband signals and invokes different functional modules and circuits to execute functional features in gNB 121. Memory 192 stores program instructions and data 190 to control the operation of gNB 121.

[0019] Similarly, for UE 110, antenna 177 transmits and receives RF signals in a MIMO network. RF transceiver module 176 is coupled to the antenna, receives RF signals from the antenna, converts the RF signals into baseband signals, and sends them to processor 173. RF transceiver module 176 also converts baseband signals received from processor 173 into RF signals and sends them to antenna 177. Processor 173 processes the received baseband signals and invokes different functional modules and circuits to execute functional features in UE 110. Memory 172 stores program instructions and data 170 to control the operation of UE 110.

[0020] gNB 121 and UE 110 also include several functional modules and circuits that can be implemented and configured to perform embodiments of the present invention. Figure 2In the example, gNB 121 includes a set of control function modules and circuitry 180. CSI processing circuitry 182 processes CSI and associated network parameters for UE 110. Configuration and control circuitry 181 provides various parameters to configure and control UE 110. UE 110 includes a set of control function modules and circuitry 160. CSI processing circuitry 162 processes CSI and associated network parameters. Configuration and control circuitry 161 processes configuration and control parameters from gNB 121.

[0021] Please note that different functional modules and circuits can be implemented and configured through software, firmware, hardware, and any combination thereof. When executed by processors 193 and 173 (e.g., via executable program codes 190 and 170), gNB 121 and UE 110 are allowed to execute embodiments of the present invention.

[0022] Figure 3 An embodiment of message transmission according to a novel aspect is illustrated. Specifically, gNB 121 sends at least one channel state information reference signal (CSI-RS) 1210 to UE 110. UE 110 receives at least one CSI-RS 1210 from gNB 121.

[0023] Then, UE 110 estimates at least one covariance matrix of at least one downlink channel matrix based on at least one CSI-RS at different times and frequencies. UE 110 sends at least one covariance matrix or at least one parameter associated with at least one covariance matrix to BS121.

[0024] Upon receiving at least one covariance matrix or at least one parameter associated with at least one covariance matrix, BS121 may: (1) derive a precoder based on at least one covariance matrix; or (2) reconstruct at least one covariance matrix based on at least one parameter associated with at least one covariance matrix and derive a precoder based on at least one covariance matrix.

[0025] In some embodiments, UE 110 estimates the covariance matrix of the downlink channel matrix based on CSI-RS at different times and frequencies. Specifically, after receiving CSI-RS at different times and frequencies, UE 110 estimates n R ×n T The downlink channel matrix H[n,m] of the MIMO channel is as follows:

[0026]

[0027] n RThis refers to the number of receiving antennas (i.e., antenna 177). T This represents the number of transmitting antennas (i.e., antenna 197). n is the time-domain index. m is the frequency-domain index.

[0028] Then, UE 110 estimates the covariance matrix Ψ[n,m] based on the downlink channel matrix H[n,m] as follows:

[0029] Ψ[n,m]=H H [n,m]H[n,m],

[0030] Please note that Ψ[n,m] should be... and And the time-frequency offset Φ[n,m]. Since Φ[n,m] is a diagonal matrix with a unit phase factor, Φ H [n,m]Φ[n,m] is the identity matrix I.

[0031] therefore,

[0032] Ψ[n,m]=H H [n,m]Φ H [n,m]Φ[n,m]H[n,m]

[0033] =H H [n,m]IH[n,m]

[0034] =H H [n,m]H[n,m],

[0035] This means that the time-frequency offset Φ[n,m] can be ignored when estimating the covariance matrix Ψ[n,m] in UE 110.

[0036] In some implementations, such as Figure 4A As shown, UE 110 sends CSI 1100, which includes the covariance matrix Ψ[n,m], to BS121. Therefore, BS121 can use, for example, eigenvalue decomposition or singular value decomposition (SVD), to derive the precoder W from the covariance matrix Ψ[n,m].

[0037] Then, BS121 sends the subsequent PDSCH to UE 110 using the precoder W.

[0038] In some implementations, UE 110 compresses the covariance matrix Ψ[n,m] and sends at least one parameter associated with the compressed covariance matrix to BS 121. Specifically, Ψ ij [n,m] represents the entry in the i-th row and j-th column of Ψ[n,m]. Each entry Ψ in the covariance matrix Ψ[n,m] is... ijIt is considered a function of time n and frequency m. Entries Ψ with different times (total number N) and frequencies (total number M). ij Form Ψ[n,m].

[0039] The covariance matrix Ψ[n,m] is compressed in the following ways: (1) by projecting it onto a two-dimensional (2D) time-frequency basis matrix; or (2) by representing it as a linear combination of 2D sine matrices. The entry in the nth row and mth column of the 2D sine matrix is ​​e. -j2πmτ e j2πnv , where τ represents the time delay and ν represents the Doppler frequency shift. More specifically, the compressed covariance matrix can be expressed as:

[0040]

[0041] B kl These are parameters known to both UE 110 and BS121.

[0042] UE 110 selects a subset S from a 2D time-frequency basis matrix (or a 2D sine matrix). ij and the coefficients corresponding to the subsets (c ijkl :(k,l)∈S ij ), and then, as Figure 4B As shown, a subset S is sent to BS121. ij and the corresponding coefficient (C) ijkl :(k,l)∈S ij CSI 1102.

[0043] Then, BS121 according to subset S ij and the corresponding coefficient (c) ijkl :(k,l)∈S ij )pass Reconstruct the covariance matrix Ψ ij Furthermore, the pre-encoder W is derived from the covariance matrix Ψ[n,m] using, for example, eigenvalue decomposition or SVD.

[0044] Then, BS121 sends the subsequent PDSCH to UE 110 using the precoder W.

[0045] In some implementations, UE 110 compresses the covariance matrix Ψ[n,m] and sends at least one parameter associated with the compressed covariance matrix to BS 121. Specifically, Ψ ij [n,m] represents the entry in the i-th row and j-th column of Ψ[n,m]. Each entry Ψ in the covariance matrix Ψ[n,m] is... ij It is considered a function of time n and frequency m. The entries Ψ of different frequencies (total number M) within time n. ij Forming vector Ψij [n].

[0046] Vector Ψ ij [n] is compressed in the following ways: (1) by projecting onto a one-dimensional (1D) frequency basis vector; or (2) by representing it as a linear combination of 1D sinusoidal vectors. The m-th entry of the 1D sinusoidal vector is e -j2πmτ , where τ represents the time delay. More specifically, after compression, the compressed vector can be represented as:

[0047]

[0048] And b k These are parameters known to both UE 110 and BS121.

[0049] UE 110 selects a subset S from the 1D frequency basis vector (or from the 1D sine vector). ij [n] and the corresponding subset S ij The coefficient of [n] (c ijk [n]:k∈S ij [n]), then as Figure 4C As shown, a subset S is sent to BS 121. ij [n] and corresponding coefficients (c) ijk [n]:k∈S ij [n]) of CSI 1104.

[0050] Then, BS121 according to subset S ij [n] and the corresponding coefficient (c) ijk [n]:k∈S ij [n]) through Reconstructing vector Ψ ij [n]. Next, BS121 is based on the vector ψ ij [n] Reconstruct the covariance matrix Ψ[n,m], and then derive the precoder W from the covariance matrix Ψ[n,m] using, for example, eigenvalue decomposition or SVD.

[0051] Then, BS121 sends the subsequent PDSCH to UE 110 using the precoder W.

[0052] In some implementations, the UE 110 calculates (or approximates) the decomposition of the covariance matrix Ψ[n,m] as follows:

[0053] Ψ[n,m]=U H [n,m]U[n,m],

[0054] Next, UE 110 compresses the decomposition matrix U[n,m] and sends at least one parameter associated with the decomposition matrix of the covariance matrix to BS121. Specifically, u ij[n,m] represents the entry in the i-th row and j-th column of U[n,m]. Entries u with different times (total number N) and frequencies (total number M) ij Forming matrix U ij .

[0055] Matrix U ij Compression can be achieved by: (1) projecting onto a 2D time-frequency basis matrix; or (2) representing it as a linear combination of 2D sine matrices. More specifically, after compression, the compressed decomposition matrix can be represented as:

[0056]

[0057] And B kl These are parameters known to both UE 110 and BS121.

[0058] UE 110 selects a subset S from a 2D time-frequency basis matrix (or from a 2D sine matrix). ij and the coefficients corresponding to the subsets (c ijkl :(k,l)∈S ij In, then, as Figure 4D As shown, a subset S is sent to BS121. ij and corresponding coefficient (c) ijkl :(k,l)∈S ij CSI 1106.

[0059] Then, BS121 according to subset S ij and the corresponding coefficient (c) ijkj :(k,l)∈S ij )pass Reconstructed matrix U ij BS121 from U ij Reconstruct the covariance matrix Ψ[n,m] and derive the precoder W from the covariance matrix Ψ[n,m] using, for example, eigenvalue decomposition or SVD.

[0060] Then, BS121 sends the subsequent PDSCH to UE 110 using the precoder W.

[0061] In some implementations, the UE 110 calculates (or approximates) the decomposition of the covariance matrix Ψ[n,m] as follows:

[0062] Ψ[n,m]=U H [n,m]U[n,m],

[0063] Next, UE 110 compresses the decomposition matrix U[n,m] and sends at least one parameter associated with the decomposition matrix of the covariance matrix to BS121. Specifically, u ij[n,m] represents the entry in the i-th row and j-th column of U[n,m]. The entries u with different frequencies (total number M) in time n. ij Forming vector u ij [n].

[0064] Vector u ij [n] is compressed in the following ways: (1) by projecting onto a 1D frequency basis vector; or (2) by representing it as a linear combination of 1D sine vectors. More specifically, after compression, the compressed decomposition vector can be represented as:

[0065]

[0066] And b k These are parameters known to both UE 110 and BS121.

[0067] UE 110 selects a subset S from the 1D frequency basis vector (or from the 1D sine vector). ij [n] and the corresponding subset S ij The coefficient of [n] (c ijk [n]:k∈S ij [n]), then as Figure 4E As shown, a subset S is sent to BS 121. ij [n] and corresponding coefficients (c) ijk [n]:k∈S ij [n]) of CSI 1108.

[0068] Then, BS121 according to subset S ij [n] and the corresponding coefficient (c) ijk [n]:k∈S ij [n]) through Reconstructed vector u ij [n], and from u ij [n] Reconstruct the covariance matrix Ψ[n,m], and derive the preencoder W from the covariance matrix Ψ[n,m] using, for example, eigenvalue decomposition or SVD.

[0069] Next, BS121 sends the subsequent PDSCH to UE 110 using the precoder W.

[0070] In some implementations, the UE 110 calculates (or approximates) the decomposition of the covariance matrix Ψ[n,m] as follows:

[0071] Ψ[n,m]=U H [n,m]U[n,m],

[0072] U[n,m] can be further decomposed into two terms U v [n,m] and U fU v [n,m] represents time-frequency variation, U f It is time-frequency fixed. For example, U[n,m] is U v [n,m] and U f The matrix product, i.e., U[n,m]=U v [n,m]U f or U f U[n,m]=U v [n,m].

[0073] In some implementations, U v [n,m] includes singular vectors and U f This includes the average of the square root singular values. In some implementations, U v [n,m] includes the square root singular value and U f This includes the average value of singular vectors.

[0074] Next, the UE 110 compression decomposition matrix U v [n,m] and sends at least one parameter associated with the decomposition matrix of the covariance matrix to BS121. Specifically, u v,ij [n,m] represents U v The entries in the i-th row and j-th column of [n, m]. Entries u representing different times (total number N) and frequencies (total number M). v,ij Forming matrix U v,ij .

[0075] Matrix U v,ij Compression can be achieved by: (1) projecting onto a 2D time-frequency basis matrix; or (2) representing it as a linear combination of 2D sine matrices. More specifically, after compression, the compressed decomposition matrix can be represented as:

[0076]

[0077] And B kl These are parameters known to both UE 110 and BS121.

[0078] UE 110 selects a subset S from a 2D time-frequency basis matrix (or a 2D sine matrix). ij and the coefficients corresponding to the subsets (c ijkl :(k,l)∈S ij ), and then, as Figure 4E As shown, a U signal is sent to BS121. f subset S ij and the corresponding coefficient (c) ijkl :(k,l)∈S ij CSI 1110.

[0079] Then, BS121 can be based on subset S ij and the corresponding coefficient (c) ijkl :(k,l)∈S ij )pass Reconstructed matrix U v,ij and from U v,ij and U f Reconstruct the covariance matrix Ψ[n,m]. BS121 derives the precoder W from the covariance matrix Ψ[n,m] using, for example, eigenvalue decomposition or SVD.

[0080] Then, BS121 sends the subsequent PDSCH to UE 110 using the precoder W.

[0081] In some implementations, the UE 110 calculates (or approximates) the decomposition of the covariance matrix Ψ[n,m] as follows:

[0082] Ψ[n,m]=U H [n,m]U[n,m],

[0083] U[n,m] can be further decomposed into two terms U v [n,m] and U f U v [n,m] represents time-frequency variation, U f It is time-frequency fixed. For example, U[n,m] is U v [n,m] and U f The matrix product, i.e., U[n,m]=U v [n,m]U f or U f U[n,m]=U v [n,m]. In some implementations, U v [n,m] includes singular vectors and U f This includes the average of the square root singular values. In some implementations, U v [n,m] includes the square root singular value and U f This includes the average value of singular vectors.

[0084] Next, the UE 110 compression decomposition matrix U v [n,m] and sends at least one parameter associated with the decomposition matrix of the covariance matrix to BS121. Specifically, u v,ij [n,m] represents U v The entries in the i-th row and j-th column of [n, m]. Entries u with different frequencies (total number M). v,ij Forming vector U v,ij [n].

[0085] Vector U v,ij [n] is compressed in the following ways: (1) by projecting onto a 1D frequency basis matrix; or (2) by representing it as a linear combination of 1D sinusoidal vectors. More specifically, after compression, the compressed decomposition vector can be represented as:

[0086]

[0087] And b k These are parameters known to both UE 110 and BS121.

[0088] UE 110 selects a subset S from the 1D frequency basis vector (or from the 1D sine vector). ij [n] and the coefficients (c) corresponding to the subset ijk [n]:k∈S ij [n]), then as Figure 4G As shown, a U signal is sent to BS121. f subset S ij [n] and corresponding coefficients (c) ijk [n]:k∈S ij [n]) of CSI 1112.

[0089] Then, BS121 according to subset S ij [n] and the corresponding coefficient (c) ijk [n]:k∈S ij [n]) through Reconstructed vector U v,ij [n], and from U v,ij [n] and U f Reconstruct the covariance matrix Ψ[n,m], and derive the precoder W from the covariance matrix Ψ[n,m] using, for example, eigenvalue decomposition or SVD.

[0090] Next, BS121 sends the subsequent PDSCH to UE 110 using the precoder W.

[0091] It should be noted that the parameters related to 2D time-frequency basis matrix projection, 1D frequency basis vector projection, linear combination of 2D sine matrices, and linear combination of 1D sine vectors can be calculated using two-dimensional discrete Fourier transform (DFT), 2D estimating signal parameter via rotational invariance technique (ESPRIT), 2D multiple signal classification (MUSIC), 1D-DFT, 1D-ESPRIT, and 1D-MUSIC algorithms.

[0092] Figure 5 This is a flowchart of a method for determining a precoder for a MIMO system based on a novel aspect. In step 501, the BS sends at least one CSI-RS to the UE. In step 502, the UE receives at least one CSI-RS. In step 503, the UE estimates at least one first covariance matrix of at least one downlink channel matrix based on at least one CSI-RS at different times and frequencies.

[0093] In step 504, the UE sends at least one first covariance matrix or at least one parameter associated with the first first covariance matrix to the BS. In step 505, the BS receives at least one first covariance matrix or at least one parameter associated with the first first covariance matrix from the UE.

[0094] In step 506, the BS reconstructs at least one second covariance matrix based on at least one first covariance matrix or at least one parameter associated with at least one first covariance matrix. In step 507, the BS derives the precoder based on at least one second covariance matrix. In step 508, the BS sends a PDSCH to the UE using the precoder.

[0095] While the invention has been described in conjunction with specific embodiments for illustrative purposes, it is not limited thereto. Therefore, various modifications, adaptations, and combinations of the features of the described embodiments can be made without departing from the scope of the invention as set forth in the claims.

Claims

1. A method for determining a precoder for a multiple-input multiple-output system, comprising: The user equipment receives at least one channel state information reference signal from the network; The user equipment estimates at least one covariance matrix of at least one downlink channel matrix based on the at least one channel state information reference signal at different times and frequencies; The user equipment sends at least one parameter associated with the at least one covariance matrix to the network; as well as The user equipment compresses the at least one covariance matrix, wherein the at least one covariance matrix is ​​compressed in the time and frequency dimensions in the following manner: Projected onto at least one two-dimensional time-frequency basis matrix or at least one one-dimensional frequency basis vector; or It is represented as a linear combination of at least one two-dimensional sine matrix or a linear combination of at least one one-dimensional sine vector.

2. The method for determining the pre-encoder of a multiple-input multiple-output system according to claim 1, characterized in that, The at least one parameter associated with the at least one covariance matrix includes: A subset of the following: the at least one two-dimensional time-frequency basis matrix, the at least one one-dimensional frequency basis vector, a linear combination of the at least one two-dimensional sine matrices, or the at least one one-dimensional sine vector; and At least one coefficient corresponding to that subset.

3. A method for determining a precoder for a multiple-input multiple-output system, comprising: The user equipment receives at least one channel state information reference signal from the network; The user equipment estimates at least one covariance matrix of at least one downlink channel matrix based on the at least one channel state information reference signal at different times and frequencies; The user equipment calculates the decomposition matrix of the at least one covariance matrix; The user equipment compresses the decomposition matrix or decomposes the decomposition matrix into a first term and a second term, wherein the first term is time-frequency variable, the second term is time-frequency fixed, and the decomposition matrix is ​​the matrix product of the first term and the second term; and The user equipment sends at least one parameter associated with the at least one covariance matrix to the network.

4. The method for determining the pre-encoder of a multiple-input multiple-output system according to claim 3, characterized in that, The user equipment compresses the decomposition matrix in the following way: Projected onto at least one two-dimensional time-frequency basis matrix or at least one one-dimensional frequency basis vector; or It is represented as a linear combination of at least one two-dimensional sine matrix or a linear combination of at least one one-dimensional sine vector.

5. The method for determining the pre-encoder of a multiple-input multiple-output system according to claim 4, characterized in that, The at least one parameter associated with the decomposition matrix of the at least one covariance matrix includes: A subset of the following: the at least one two-dimensional time-frequency basis matrix, the at least one one-dimensional frequency basis vector, a linear combination of the at least one two-dimensional sine matrices, or the at least one one-dimensional sine vector; and At least one coefficient corresponding to that subset.

6. The method for determining the pre-encoder of a multiple-input multiple-output system according to claim 3, characterized in that, The user equipment decomposes the decomposition matrix, further including: The user equipment compresses the first term and the second term of the decomposition matrix in the following manner: Projected onto at least one two-dimensional time-frequency basis matrix or at least one one-dimensional frequency basis vector; or It is represented as a linear combination of at least one two-dimensional sine matrix or a linear combination of at least one one-dimensional sine vector.

7. The method for determining the pre-encoder of a multiple-input multiple-output system according to claim 6, characterized in that, The at least one parameter is associated with the first term of the decomposition matrix of the at least one covariance matrix, and the at least one parameter includes: A subset of the following: the at least one two-dimensional time-frequency basis matrix, the at least one one-dimensional frequency basis vector, a linear combination of the at least one two-dimensional sine matrices, or the at least one one-dimensional sine vector; and At least one coefficient corresponding to that subset.

8. The method for determining the pre-encoder of a multiple-input multiple-output system according to claim 3, characterized in that, The first term includes at least one singular vector and the second term includes the average of at least one square root singular value; or The first term includes at least one square root singular value and the second term includes the average of at least one singular vector.

9. A method for determining a precoder for a multiple-input multiple-output system, comprising: The base station sends at least one channel state information reference signal to the user equipment, so that the user equipment can estimate at least one covariance matrix of at least one downlink channel matrix based on the at least one channel state information reference signal at different times and frequencies. The base station receives from the user equipment at least one parameter associated with the at least one covariance matrix, wherein the at least one covariance matrix is ​​compressed in the time and frequency dimensions by being projected onto at least one two-dimensional time-frequency basis matrix or at least one one-dimensional frequency basis vector; or represented as a linear combination of at least one two-dimensional sine matrices or at least one one-dimensional sine vector. The base station reconstructs at least one second covariance matrix based on the at least one covariance matrix or the at least one parameter associated with the at least one covariance matrix; and The base station derives the precoder based on the at least one second covariance matrix.

10. A user equipment for determining a pre-encoder of a multiple-input multiple-output system, comprising: A transceiver used to receive at least one channel state information reference signal from a network; as well as A channel state information processing circuit is configured to estimate at least one covariance matrix of at least one downlink channel matrix based on at least one channel state information reference signal at different times and frequencies, and to compress the at least one covariance matrix in the time and frequency dimensions, wherein the at least one covariance matrix is ​​compressed in the time and frequency dimensions by: projecting it onto at least one two-dimensional time-frequency basis matrix or at least one one-dimensional frequency basis vector; or representing it as a linear combination of at least one two-dimensional sine matrices or at least one one-dimensional sine vectors. The transceiver sends at least one parameter associated with the at least one covariance matrix to the network.

11. The user equipment for determining the pre-encoder of a multiple-input multiple-output system according to claim 10, characterized in that, The at least one parameter associated with the at least one covariance matrix includes: A subset of the following: the at least one two-dimensional time-frequency basis matrix, the at least one one-dimensional frequency basis vector, a linear combination of the at least one two-dimensional sine matrices, or the at least one one-dimensional sine vector; and At least one coefficient corresponding to that subset.

12. A user equipment for determining a pre-encoder of a multiple-input multiple-output system, comprising: A transceiver used to receive at least one channel state information reference signal from a network; as well as A channel state information processing circuit is configured to estimate at least one covariance matrix of at least one downlink channel matrix based on at least one channel state information reference signal at different times and frequencies, calculate a decomposition matrix of the at least one covariance matrix, and compress or decompose the decomposition matrix into a first term and a second term, wherein the first term is time-frequency variable, the second term is time-frequency fixed, and the decomposition matrix is ​​the matrix product of the first term and the second term. The transceiver sends at least one parameter associated with the at least one covariance matrix to the network.

13. The user equipment for determining the pre-encoder of a multiple-input multiple-output system according to claim 12, characterized in that, The channel state information processing circuit further includes: The decomposition matrix is ​​compressed in the following way: Projected onto at least one two-dimensional time-frequency basis matrix or at least one one-dimensional frequency basis vector; or It is represented as a linear combination of at least one two-dimensional sine matrix or a linear combination of at least one one-dimensional sine vector.

14. The user equipment for determining the pre-encoder of a multiple-input multiple-output system according to claim 13, characterized in that, The at least one parameter associated with the decomposition matrix of the at least one covariance matrix includes: A subset of the following: the at least one two-dimensional time-frequency basis matrix, the at least one one-dimensional frequency basis vector, a linear combination of the at least one two-dimensional sine matrices, or the at least one one-dimensional sine vector; and At least one coefficient corresponding to that subset.

15. The user equipment for determining the pre-encoder of a multiple-input multiple-output system according to claim 12, characterized in that, The channel state information processing circuit further includes: The first and second terms of the decomposition matrix are compressed in the following way: Projected onto at least one two-dimensional time-frequency basis matrix or at least one one-dimensional frequency basis vector; or It is represented as a linear combination of at least one two-dimensional sine matrix or a linear combination of at least one one-dimensional sine vector.

16. The user equipment for determining the pre-encoder of a multiple-input multiple-output system according to claim 15, characterized in that, The at least one parameter is associated with the first term of the decomposition matrix of the at least one covariance matrix, and the at least one parameter includes: A subset of the following: the at least one two-dimensional time-frequency basis matrix, the at least one one-dimensional frequency basis vector, a linear combination of the at least one two-dimensional sine matrices, or the at least one one-dimensional sine vector; and At least one coefficient corresponding to that subset.

Images