Channel state information feedback method, terminal, network equipment and storage medium

Abstract

The embodiment of the invention provides a channel state information feedback method, a terminal, network equipment and a storage medium. When the method is applied to the terminal, the method comprises the following steps: obtaining a first downlink channel correlation coefficient matrix of a downlink channel; generating a first amplitude matrix used for representing a downlink channel amplitude correlation coefficient and / or a first phase matrix used for representing a downlink channel phase correlation coefficient according to the first downlink channel correlation coefficient matrix; compressing the first amplitude matrix by using a first amplitude encoder to obtain a first amplitude bit stream, and / or compressing the first phase matrix by using a first phase encoder to obtain a first phase bit stream; and sending the first amplitude bit stream and / or the first phase bit stream to network equipment. According to the embodiment of the invention, the channel feedback overhead can be reduced on the basis of guaranteeing the channel recovery precision by utilizing the reciprocity of the amplitude and the phase.

Description

Technical Field

[0001] The present application relates to the field of mobile communication technology, and in particular to a channel state information feedback method, terminal, network equipment and storage medium. Background Art

[0002] The channel state information compression feedback scheme based on machine learning learns the implicit low-dimensional features in the channel in a data-driven way. In most cases, it can achieve better compression performance than the codebook (such as eType II codebook) used in the existing 3GPP New Radio (NR) standard, and can reduce the feedback overhead of the downlink channel state information of the frequency division duplex (FDD) system. Therefore, the channel state information compression feedback based on artificial intelligence (AI) / machine learning has received widespread attention.

[0003] At present, the input of the AI ​​compression model based on the channel characteristics in the space-frequency domain of the related technology only considers the channel information in the space-frequency domain of the downlink channel, and its channel feedback overhead is still relatively large. How to further reduce the channel feedback overhead while ensuring the accuracy of channel information recovery is an urgent problem to be solved. Summary of the invention

[0004] At least one embodiment of the present application provides a channel state information feedback method, terminal, network device and storage medium, which utilizes the reciprocity of amplitude and phase to significantly reduce channel feedback overhead while ensuring channel recovery accuracy.

[0005] In order to solve the above technical problems, this application is implemented as follows:

[0006] In a first aspect, an embodiment of the present application provides a channel state information feedback method, which is applied to a terminal, including:

[0007] Obtaining a first downlink channel correlation coefficient matrix of a downlink channel;

[0008] generating, according to the first downlink channel correlation coefficient matrix, a first amplitude matrix for representing downlink channel amplitude correlation coefficients and / or a first phase matrix for representing downlink channel phase correlation coefficients;

[0009] Compressing the first amplitude matrix using a first amplitude encoder to obtain a first amplitude bit stream, and / or compressing the first phase matrix using a first phase encoder to obtain a first phase bit stream;

[0010] The first amplitude bit stream and / or the first phase bit stream is sent to a network device.

[0011] Optionally, obtaining a first downlink channel correlation coefficient matrix includes:

[0012] A first downlink channel matrix of downlink channel state information is acquired, and the first downlink channel matrix is ​​transformed into an angle domain and a delay domain to obtain the first downlink channel correlation coefficient matrix.

[0013] Optionally, before using a first amplitude encoder to compress the first amplitude matrix to obtain a first amplitude bit stream, and / or using a first phase encoder to compress the first phase matrix to obtain a first phase bit stream, the method further includes:

[0014] Receive first configuration information sent by a network device, where the first configuration information is used to configure the first amplitude encoder and / or the first phase encoder.

[0015] Optionally, after sending the first amplitude bit stream and / or the first phase bit stream to the network device, the method further includes:

[0016] Receiving second configuration information sent by the network device, where the second configuration information is used to configure a second amplitude encoder and / or a second phase encoder;

[0017] Obtaining a second downlink channel correlation coefficient matrix of the downlink channel;

[0018] generating, according to the second downlink channel correlation coefficient matrix, a second amplitude matrix for representing downlink channel amplitude correlation coefficients and / or a second phase matrix for representing downlink channel phase correlation coefficients;

[0019] Compressing the second amplitude matrix using the second amplitude encoder to obtain a second amplitude bit stream, and / or compressing the second phase matrix using a second phase encoder to obtain a second phase bit stream;

[0020] The second amplitude bit stream and / or the second phase bit stream is sent to a network device.

[0021] Optionally, obtaining a second downlink channel correlation coefficient matrix of the downlink channel includes:

[0022] A second downlink channel matrix of the downlink channel state information is acquired, and the second downlink channel matrix is ​​transformed into an angle domain and a delay domain to obtain the second downlink channel correlation coefficient matrix.

[0023] In a second aspect, an embodiment of the present application provides a channel state information feedback method, which is applied to a network device, including:

[0024] A first amplitude bit stream and / or a first phase bit stream sent by a receiving terminal, wherein the first amplitude bit stream is obtained by compressing a first amplitude matrix used to represent a downlink channel amplitude correlation coefficient using a first amplitude encoder, and the first phase bit stream is obtained by compressing a first phase matrix used to represent a downlink channel phase correlation coefficient using a first phase encoder by the terminal, and the first amplitude matrix and the first phase matrix are generated based on a first downlink channel correlation coefficient matrix obtained by converting the first downlink channel matrix;

[0025] Decoding the first amplitude bit stream using a first amplitude decoder to obtain a first amplitude reasoning matrix; and / or decoding the first phase bit stream using a first phase decoder to obtain a first phase reasoning matrix; generating first downlink channel feedback information based on the first amplitude reasoning matrix and / or the first phase reasoning matrix;

[0026] Inputting the first uplink channel information and the first downlink channel feedback information into a downlink channel recovery model to obtain the first downlink channel information generated by the downlink channel recovery model;

[0027] The first amplitude encoder and the first amplitude decoder both belong to the autoencoder corresponding to the first amplitude correlation interval, and the first phase encoder and the first phase decoder both belong to the autoencoder corresponding to the first phase correlation interval.

[0028] Optionally, before receiving the first amplitude bit stream and / or the first phase bit stream sent by the terminal, the method further includes:

[0029] Sending first configuration information to the terminal, where the first configuration information is used to configure the first amplitude encoder and / or the first phase encoder.

[0030] Optionally, the above method further includes:

[0031] Calculating an amplitude correlation parameter between the first downlink channel information and the first uplink channel information, and determining a second amplitude correlation interval to which the amplitude correlation parameter belongs; and determining a second amplitude encoder and a second amplitude decoder corresponding to the second amplitude correlation interval;

[0032] Calculating a phase correlation parameter between the first downlink channel information and the first uplink channel information, and determining a second phase correlation interval to which the phase correlation parameter belongs; determining a second phase encoder and a second phase decoder corresponding to the second phase correlation interval;

[0033] Sending second configuration information to the terminal, where the second configuration information is used to configure the second amplitude encoder and the second phase encoder.

[0034] Optionally, the above method further includes:

[0035] A second amplitude bit stream and / or a second phase bit stream sent by a receiving terminal, wherein the second amplitude bit stream is obtained by compressing a second amplitude matrix used to represent a downlink channel amplitude correlation coefficient using the second amplitude encoder, and the second phase bit stream is obtained by compressing a second phase matrix used to represent a downlink channel phase correlation coefficient using the second phase encoder by the terminal, and the second amplitude matrix and the second phase matrix are generated based on a second downlink channel correlation coefficient matrix obtained by converting the second downlink channel matrix;

[0036] Decoding the second amplitude bit stream using the second amplitude decoder to obtain a second amplitude reasoning matrix; and / or decoding the second phase bit stream using a second phase decoder to obtain a second phase reasoning matrix; generating second downlink channel feedback information based on the second amplitude reasoning matrix and / or the second phase reasoning matrix;

[0037] The second uplink channel information and the second downlink channel feedback information are input into the downlink channel recovery model to obtain the second downlink channel information generated by the downlink channel recovery model.

[0038] Optionally, the above method further includes:

[0039] Dividing the value range of the amplitude correlation parameter of the uplink and downlink channels into a plurality of continuous and non-overlapping amplitude correlation intervals; dividing the value range of the phase correlation parameter of the uplink and downlink channels into a plurality of continuous and non-overlapping phase correlation intervals;

[0040] Based on a plurality of downlink channel training matrices in the same amplitude correlation interval, a first training data set is generated, wherein the first training data set includes: an amplitude training matrix for representing a downlink channel amplitude correlation coefficient generated based on the downlink channel training matrix;

[0041] Based on a plurality of downlink channel training matrices in the same phase correlation interval, a second training data set is generated, wherein the second training data set includes: a phase training matrix for representing a downlink channel phase correlation coefficient generated based on the downlink channel training matrix;

[0042] The first training data set is used to train the amplitude autoencoder corresponding to the amplitude correlation interval; and / or the second training data set is used to train the phase autoencoder corresponding to the phase correlation interval, wherein the amplitude autoencoder includes an amplitude encoder and an amplitude decoder, and the phase autoencoder includes a phase encoder and a phase decoder.

[0043] Optionally, using the first training data set to train the amplitude autoencoder corresponding to the amplitude correlation interval; and / or using the second training data set to train the phase autoencoder corresponding to the phase correlation interval, includes:

[0044] The amplitude training matrix is ​​input into an amplitude encoder to generate a first training bit stream; the first training bit stream is input into an amplitude decoder to generate an amplitude reference matrix; a first loss value between the amplitude reference matrix and the amplitude training matrix is ​​calculated, and the amplitude encoder and the amplitude decoder are updated according to the first loss value until a preset first training end condition is met, thereby obtaining an amplitude autoencoder corresponding to the amplitude correlation interval;

[0045] and / or,

[0046] The phase training matrix is ​​input into a phase encoder to generate a second training bit stream; the second training bit stream is input into a phase decoder to generate a phase reference matrix; a second loss value between the phase reference matrix and the phase training matrix is ​​calculated, and the phase encoder and phase decoder are updated according to the second loss value until a preset second training end condition is met, thereby obtaining a phase autoencoder corresponding to the amplitude correlation interval.

[0047] Optionally, the first loss value is represented by a mean square error between the amplitude reference matrix and the amplitude training matrix, and the second loss value is represented by a cosine similarity between the phase reference matrix and the phase training matrix.

[0048] Optionally, the above method further includes:

[0049] The uplink channel information corresponding to the amplitude reference matrix, the phase reference matrix and the downlink channel training matrix is ​​input into a downlink channel recovery model, and the downlink channel recovery model corresponding to the amplitude correlation interval is trained with the goal of minimizing the downlink channel training matrix error output by the downlink channel recovery model.

[0050] In a third aspect, an embodiment of the present application provides a terminal, including a transceiver and a processor, wherein:

[0051] The processor is used to obtain a first downlink channel correlation coefficient matrix of a downlink channel; generate a first amplitude matrix for representing downlink channel amplitude correlation coefficients and / or a first phase matrix for representing downlink channel phase correlation coefficients according to the first downlink channel correlation coefficient matrix; compress the first amplitude matrix using a first amplitude encoder to obtain a first amplitude bit stream, and / or compress the first phase matrix using a first phase encoder to obtain a first phase bit stream;

[0052] The transceiver is used to send the first amplitude bit stream and / or the first phase bit stream to a network device.

[0053] Optionally, the processor is further configured to:

[0054] A first downlink channel matrix of downlink channel state information is acquired, and the first downlink channel matrix is ​​transformed into an angle domain and a delay domain to obtain the first downlink channel correlation coefficient matrix.

[0055] Optionally, the transceiver is further used for:

[0056] Receive first configuration information sent by a network device, where the first configuration information is used to configure the first amplitude encoder and / or the first phase encoder.

[0057] Optionally, the transceiver is further used to receive second configuration information sent by the network device after sending the first amplitude bit stream and / or the first phase bit stream to the network device, where the second configuration information is used to configure a second amplitude encoder and / or a second phase encoder;

[0058] The processor is further configured to obtain a second downlink channel correlation coefficient matrix of the downlink channel; generate a second amplitude matrix for representing downlink channel amplitude correlation coefficients and / or a second phase matrix for representing downlink channel phase correlation coefficients according to the second downlink channel correlation coefficient matrix; compress the second amplitude matrix using the second amplitude encoder to obtain a second amplitude bit stream, and / or compress the second phase matrix using the second phase encoder to obtain a second phase bit stream;

[0059] The transceiver is further configured to send the second amplitude bit stream and / or the second phase bit stream to a network device.

[0060] Optionally, the processor is further configured to:

[0061] A second downlink channel matrix of the downlink channel state information is acquired, and the second downlink channel matrix is ​​transformed into an angle domain and a delay domain to obtain the second downlink channel correlation coefficient matrix.

[0062] In a fourth aspect, an embodiment of the present application provides a terminal, comprising: a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the method described in the first aspect.

[0063] In a fifth aspect, an embodiment of the present application provides a network device, including a transceiver and a processor, wherein:

[0064] The transceiver is used to receive a first amplitude bit stream and / or a first phase bit stream sent by a terminal, wherein the first amplitude bit stream is obtained by compressing a first amplitude matrix used to represent a downlink channel amplitude correlation coefficient using a first amplitude encoder, and the first phase bit stream is obtained by compressing a first phase matrix used to represent a downlink channel phase correlation coefficient using a first phase encoder by the terminal, and the first amplitude matrix and the first phase matrix are generated based on a first downlink channel correlation coefficient matrix obtained by converting a first downlink channel matrix;

[0065] The processor is configured to decode the first amplitude bit stream using a first amplitude decoder to obtain a first amplitude reasoning matrix; and / or decode the first phase bit stream using a first phase decoder to obtain a first phase reasoning matrix; generate first downlink channel feedback information based on the first amplitude reasoning matrix and / or the first phase reasoning matrix; input the first uplink channel information and the first downlink channel feedback information into a downlink channel recovery model to obtain the first downlink channel information generated by the downlink channel recovery model;

[0066] The first amplitude encoder and the first amplitude decoder both belong to the autoencoder corresponding to the first amplitude correlation interval, and the first phase encoder and the first phase decoder both belong to the autoencoder corresponding to the first phase correlation interval.

[0067] Optionally, the transceiver is further used for:

[0068] Before receiving the first amplitude bit stream and / or the first phase bit stream sent by the terminal, first configuration information is sent to the terminal, where the first configuration information is used to configure the first amplitude encoder and / or the first phase encoder.

[0069] Optionally, the processor is further configured to:

[0070] Calculating an amplitude correlation parameter between the first downlink channel information and the first uplink channel information, and determining a second amplitude correlation interval to which the amplitude correlation parameter belongs; and determining a second amplitude encoder and a second amplitude decoder corresponding to the second amplitude correlation interval;

[0071] Calculating a phase correlation parameter between the first downlink channel information and the first uplink channel information, and determining a second phase correlation interval to which the phase correlation parameter belongs; determining a second phase encoder and a second phase decoder corresponding to the second phase correlation interval;

[0072] Sending second configuration information to the terminal, where the second configuration information is used to configure the second amplitude encoder and the second phase encoder.

[0073] Optionally, the transceiver is further used to receive a second amplitude bit stream and / or a second phase bit stream sent by the terminal, wherein the second amplitude bit stream is obtained by compressing a second amplitude matrix used to represent a downlink channel amplitude correlation coefficient using the second amplitude encoder, and the second phase bit stream is obtained by compressing a second phase matrix used to represent a downlink channel phase correlation coefficient by the terminal using a second phase encoder, and the second amplitude matrix and the second phase matrix are generated based on a second downlink channel correlation coefficient matrix obtained by converting the second downlink channel matrix;

[0074] The processor is further used to decode the second amplitude bit stream using the second amplitude decoder to obtain a second amplitude reasoning matrix; and / or, decode the second phase bit stream using the second phase decoder to obtain a second phase reasoning matrix; generate second downlink channel feedback information based on the second amplitude reasoning matrix and / or the second phase reasoning matrix; input the second uplink channel information and the second downlink channel feedback information into the downlink channel recovery model to obtain the second downlink channel information generated by the downlink channel recovery model.

[0075] Optionally, the processor is further configured to:

[0076] Dividing the value range of the amplitude correlation parameter of the uplink and downlink channels into a plurality of continuous and non-overlapping amplitude correlation intervals; dividing the value range of the phase correlation parameter of the uplink and downlink channels into a plurality of continuous and non-overlapping phase correlation intervals;

[0077] Based on a plurality of downlink channel training matrices in the same amplitude correlation interval, a first training data set is generated, wherein the first training data set includes: an amplitude training matrix for representing a downlink channel amplitude correlation coefficient generated based on the downlink channel training matrix;

[0078] Based on a plurality of downlink channel training matrices in the same phase correlation interval, a second training data set is generated, wherein the second training data set includes: a phase training matrix for representing a downlink channel phase correlation coefficient generated based on the downlink channel training matrix;

[0079] The first training data set is used to train the amplitude autoencoder corresponding to the amplitude correlation interval; and / or the second training data set is used to train the phase autoencoder corresponding to the phase correlation interval, wherein the amplitude autoencoder includes an amplitude encoder and an amplitude decoder, and the phase autoencoder includes a phase encoder and a phase decoder.

[0080] Optionally, the processor is further configured to:

[0081] The amplitude training matrix is ​​input into an amplitude encoder to generate a first training bit stream; the first training bit stream is input into an amplitude decoder to generate an amplitude reference matrix; a first loss value between the amplitude reference matrix and the amplitude training matrix is ​​calculated, and the amplitude encoder and the amplitude decoder are updated according to the first loss value until a preset first training end condition is met, thereby obtaining an amplitude autoencoder corresponding to the amplitude correlation interval;

[0082] and / or,

[0083] The phase training matrix is ​​input into a phase encoder to generate a second training bit stream; the second training bit stream is input into a phase decoder to generate a phase reference matrix; a second loss value between the phase reference matrix and the phase training matrix is ​​calculated, and the phase encoder and phase decoder are updated according to the second loss value until a preset second training end condition is met, thereby obtaining a phase autoencoder corresponding to the amplitude correlation interval.

[0084] Optionally, the first loss value is represented by a mean square error between the amplitude reference matrix and the amplitude training matrix, and the second loss value is represented by a cosine similarity between the phase reference matrix and the phase training matrix.

[0085] Optionally, the processor is further configured to:

[0086] The uplink channel information corresponding to the amplitude reference matrix, the phase reference matrix and the downlink channel training matrix is ​​input into a downlink channel recovery model, and the downlink channel recovery model corresponding to the amplitude correlation interval is trained with the goal of minimizing the downlink channel training matrix error output by the downlink channel recovery model.

[0087] In a sixth aspect, an embodiment of the present application provides a network device, comprising: a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the method described in the second aspect.

[0088] In a seventh aspect, an embodiment of the present application provides a computer-readable storage medium, on which a program is stored. When the program is executed by a processor, the steps of the method described above are implemented.

[0089] Compared with the prior art, the channel state information feedback method, terminal, network device and storage medium provided in the embodiments of the present application decompose the downlink correlation coefficient matrix into amplitude and phase correlation coefficient matrices, and perform compressed feedback through different compression models (encoders). It can utilize the reciprocity of amplitude and phase, and greatly reduce the channel feedback overhead while ensuring the accuracy of channel information recovery. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present application. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0091] Figure 1 A schematic diagram of channel state information compression feedback based on machine learning in the related art;

[0092] Figure 2 A schematic diagram of a self-encoding network deployed on a terminal and a base station side of the related technology;

[0093] Figure 3 A schematic diagram of an application scenario of an embodiment of the present application;

[0094] Figure 4 This is a flow chart of the channel state information feedback method according to an embodiment of the present application applied to a terminal;

[0095] Figure 5 A schematic diagram of the mean values ​​of the amplitude correlation and the phase correlation according to an embodiment of the present application;

[0096] Figure 6 This is a flow chart of the channel state information feedback method according to an embodiment of the present application applied to a network device;

[0097] Figure 7 An interactive example diagram of the channel state information feedback method according to an embodiment of the present application;

[0098] Figure 8 This is a schematic diagram of the structure of a terminal according to an embodiment of the present application;

[0099] Fig. 9 This is a schematic diagram of the structure of a terminal according to another embodiment of the present application;

[0100] Fig.10 A schematic diagram of the structure of a network device according to an embodiment of the present application;

[0101] Fig.11 A schematic diagram of the structure of a network device according to another embodiment of the present application;

[0102] Fig.12 This is a schematic diagram of the structure of a terminal according to another embodiment of the present application;

[0103] Fig.13 This is a schematic diagram of the structure of a network device according to another embodiment of the present application. DETAILED DESCRIPTION

[0104] The exemplary embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present application and to fully convey the scope of the present application to those skilled in the art.

[0105] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and need not be used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable in appropriate circumstances, so that the embodiments of the present application described herein can be implemented in a sequence other than those illustrated or described herein, for example. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, the process, method, system, product or equipment comprising a series of steps or units need not be limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or equipment. "And / or" in the specification and claims represents at least one of the connected objects.

[0106] The technology described herein is not limited to NR systems and Long Time Evolution (LTE) / LTE-Advanced (LTE-A) systems, and can also be used for various wireless communication systems, such as Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Frequency Division Multiple Access (FDMA), Orthogonal Frequency Division Multiple Access (OFDMA), Single-carrier Frequency Division Multiple Access (SC-FDMA) and other systems. The terms "system" and "network" are often used interchangeably. A CDMA system can implement radio technologies such as CDMA2000, Universal Terrestrial Radio Access (UTRA). UTRA includes Wideband Code Division Multiple Access (WCDMA) and other CDMA variants. A TDMA system can implement radio technologies such as Global System for Mobile Communication (GSM). OFDMA systems can implement radio technologies such as Ultra Mobile Broadband (UMB), Evolution-UTRA (E-UTRA), IEEE 802.21 (Wi-Fi), IEEE 802.16 (WiMAX), IEEE 802.20, Flash-OFDM, etc. UTRA and E-UTRA are parts of Universal Mobile Telecommunications System (UMTS). LTE and more advanced LTE (such as LTE-A) are new UMTS versions that use E-UTRA. UTRA, E-UTRA, UMTS, LTE, LTE-A, and GSM are described in documents from an organization named "3rd Generation Partnership Project" (3GPP). CDMA2000 and UMB are described in documents from an organization named "3rd Generation Partnership Project 2" (3GPP2).The techniques described herein may be used for the systems and radio technologies mentioned above as well as for other systems and radio technologies. However, the following description describes an NR system for example purposes and uses NR terminology in much of the following description, although the techniques may also be applicable to applications other than NR system applications.

[0107] The following description provides examples and does not limit the scope, applicability, or configuration set forth in the claims. Changes may be made to the functions and arrangements of the elements discussed without departing from the spirit and scope of the present disclosure. Various examples may appropriately omit, replace, or add various procedures or components. For example, the described methods may be performed in an order different from that described, and various steps may be added, omitted, or combined. In addition, features described with reference to certain examples may be combined in other examples.

[0108] See also Figure 3 , Figure 3 A block diagram of a wireless communication system applicable to an embodiment of the present application is shown. The wireless communication system includes a terminal 11 and a network device 12. The terminal 11 may also be referred to as a user terminal or user equipment (UE, User Equipment), and the terminal 11 may be a terminal side device such as a mobile phone, a tablet computer (Tablet Personal Computer), a laptop computer (Laptop Computer), a personal digital assistant (Personal Digital Assistant, PDA), a mobile Internet device (Mobile Internet Device, MID), a wearable device (Wearable Device) or a vehicle-mounted device. It should be noted that the specific type of the terminal 11 is not limited in the embodiment of the present application. The network device 12 can be a base station and / or a core network element, wherein the above-mentioned base station can be a base station of 5G and later versions (for example: gNB, 5G NR NB, etc.), or a base station in other communication systems (for example: eNB, WLAN access point, or other access point, etc.), wherein the base station can be called a node B, an evolved node B, an access point, a base transceiver station (Base Transceiver Station, BTS), a radio base station, a radio transceiver, a basic service set (Basic Service Set, BSS), an extended service set (Extended Service Set, ESS), a B node, an evolved B node (eNB), a home B node, a home evolved B node, a WLAN access point, a WiFi node or some other suitable term in the field. As long as the same technical effect is achieved, the base station is not limited to specific technical vocabulary. It should be noted that in the embodiment of the present application, only the base station in the NR system is taken as an example, but the specific type of the base station is not limited.

[0109] The base station can communicate with the terminal 11 under the control of a base station controller, and in various examples, the base station controller can be part of the core network or certain base stations. Some base stations can communicate control information or user data with the core network via a backhaul. In some examples, some of these base stations can communicate with each other directly or indirectly via a backhaul link, which can be a wired or wireless communication link. The wireless communication system can support operations on multiple carriers (waveform signals of different frequencies). A multi-carrier transmitter can transmit modulated signals on these multiple carriers at the same time. For example, each communication link can be a multi-carrier signal modulated according to various radio technologies. Each modulated signal can be sent on a different carrier and can carry control information (e.g., reference signals, control channels, etc.), overhead information, data, etc.

[0110] The base station can communicate wirelessly with the terminal 11 via one or more access point antennas. Each base station can provide communication coverage for its respective coverage area. The coverage area of ​​the access point can be divided into sectors that constitute only a portion of the coverage area. The wireless communication system may include different types of base stations (e.g., macro base stations, micro base stations, or micro-micro base stations). The base station may also utilize different radio technologies, such as cellular or WLAN radio access technologies. The base station may be associated with the same or different access networks or operator deployments. The coverage areas of different base stations (including coverage areas of the same or different types of base stations, coverage areas utilizing the same or different radio technologies, or coverage areas belonging to the same or different access networks) may overlap.

[0111] The communication link in the wireless communication system may include an uplink for carrying an uplink (UL) transmission (e.g., from the terminal 11 to the network device 12), or a downlink for carrying a downlink (DL) transmission (e.g., from the network device 12 to the terminal 11). UL transmission may also be referred to as reverse link transmission, and DL transmission may also be referred to as forward link transmission. Downlink transmission may be performed using a licensed frequency band, an unlicensed frequency band, or both. Similarly, uplink transmission may be performed using a licensed frequency band, an unlicensed frequency band, or both.

[0112] For channel state information compression feedback based on artificial intelligence (AI) / machine learning, its model is generally divided into two parts: encoder and decoder. Figure 1 In actual deployment, an encoder needs to be deployed on the terminal side and a decoder needs to be deployed on the network side. The terminal uses the encoder to compress the estimated channel state information (CSI) into a bit stream, which is sent to the network side via an uplink feedback channel. The network side finally restores the original CSI based on the bit stream.

[0113] Taking the compression recovery of channel feature vector as an example, assume that the number of base station transmitting antennas is Nt=32, the number of terminal receiving antennas is Nr=4, and the number of subbands is Nsb=16. dl The channel matrix (32*4*16) is processed by singular value decomposition (SVD) to obtain the V matrix (32*16*2), where 2 represents the real part and the imaginary part.

[0114] A self-coding network deployed on the terminal and base station side is as follows Figure 2 As shown in the figure, the autoencoder network includes an encoder and a decoder. Considering the weak processing power of the terminal, a lightweight model is usually deployed on the terminal, using a two-layer fully connected network for compression processing of the input original eigenvalue matrix. A more complex model is deployed on the base station side, based on the residual network (Residual Network, Resnet), a one-layer fully connected plus a two-layer convolutional network and residual processing.

[0115] Since the purpose of model construction is to restore the correlation information of the channel feature vector to the greatest extent, the square of the correlation (here the square of the cosine similarity ρ is used) is selected as the basic evaluation index. Where W is the actual channel feature vector, and W' is the estimated channel feature vector restored by the base station.

[0116]

[0117] After the terminal negotiates the AI ​​coding model and AI decoding model required for downlink channel feedback with the base station, the base station periodically sends a channel state information reference signal (CSI-RS). After receiving the CSI-RS, the terminal performs channel estimation, compresses and quantizes it using the coding model, and then sends the compressed channel information to the terminal. After receiving it, the base station decompresses the channel information according to the negotiated AI decoding model to obtain complete downlink channel state information.

[0118] The AI-based CSI channel feedback compression in related technologies mainly treats the channel information matrix or feature vector matrix as a picture, and uses the AI ​​model to mine the channel characteristics of the spatial-frequency domain of the channel information for compression processing. The input of the AI ​​compression model based on the channel characteristics of the spatial-frequency domain only considers the channel information of the spatial-frequency domain of the downlink channel, resulting in a large channel feedback overhead.

[0119] As mentioned above, the input of the AI ​​compression model based on the channel characteristics of the space-frequency domain only considers the channel information of the space-frequency domain of the downlink channel, resulting in a large channel feedback overhead. In addition, the accuracy improvement of the AI ​​compression feedback method based on the space-frequency domain is relatively limited. Furthermore, the performance of channel compression feedback depends largely on the design of the AI ​​model. For example, the performance of Transformer is better, but its AI model is more complex, the training and reasoning overhead is large, and the requirements for data samples are also relatively high.

[0120] To solve at least one of the above problems, an embodiment of the present application provides a channel state information feedback method, which decomposes the downlink correlation coefficient matrix into amplitude and phase correlation coefficient matrices, and performs compressed feedback through different encoders. It can utilize the reciprocity of amplitude and phase, and greatly reduce the channel feedback overhead while ensuring the accuracy of channel information recovery.

[0121] Please refer to Figure 4 , a channel state information feedback method provided in an embodiment of the present application, when applied to a terminal side, includes:

[0122] Step 41: Obtain a first downlink channel correlation coefficient matrix of a downlink channel.

[0123] Here, the first downlink channel matrix of the downlink channel state information can be obtained by measuring the downlink reference signal (such as the downlink CSI-RS), and then the first downlink channel matrix is ​​transformed in the antenna domain and the frequency domain respectively, and the first downlink channel matrix is ​​transformed into the angle domain and the delay domain to obtain the first downlink channel correlation coefficient matrix. Specifically, the above transformation process can be performed by DFT and / or IDFT.

[0124] For example, assuming that the first downlink channel matrix between the terminal and the network device (base station) is H, M represents the number of transmitting antennas of the network device (base station), and N represents the number of subcarriers between the terminal and the network device (base station), the following relationship exists:

[0125]

[0126]

[0127]

[0128] Among them, h(i,j) represents the channel state information corresponding to antenna j on subcarrier i, P(τ l ) represents the basis vector in the frequency domain, a H (θ l ) represents the basis vector of the antenna domain, and L represents the number of subpaths. l Indicates that the lth subpath is in the frequency domain basis vector P(τ l) and antenna domain a H (θ l ), that is, the path fading factor on the lth subpath. The channel correlation coefficient matrix g l is a matrix composed of the path fading factors on the L sub-paths.

[0129] For example, after transforming the M*N downlink channel matrix in the antenna domain and the frequency domain respectively, the channel correlation coefficient matrix g can be obtained. l After transformation to the angle domain and delay domain, the correlation coefficient of each sampling point represents the path fading factor of each subpath in the multipath environment, expressed as a complex number, and the path fading factor g on the lth subpath is l It can be expressed as:

[0130] g l =a+bj=P*exp(jφ), where the amplitude is P and the phase is φ.

[0131] Step 42: Generate a first amplitude matrix for representing downlink channel amplitude correlation coefficients and / or a first phase matrix for representing downlink channel phase correlation coefficients according to the first downlink channel correlation coefficient matrix.

[0132] Considering the partial reciprocity of the uplink and downlink channels, the amplitude of path fading mainly depends on energy attenuation, such as path loss, shadow fading and other factors, and the correlation between the uplink and downlink channels is relatively large. The phase of path fading mainly depends on the frequency difference between the uplink and downlink, the phase change of the frequency in the multipath channel transmission, and considering that the initial phase of the uplink and downlink may be relatively random, a slight difference in frequency may lead to a large difference in the phase of path fading, so the correlation between the uplink and downlink channels is small. The closer the correlation is to 1, the greater the correlation.

[0133] like Figure 5 As shown, the mean values ​​of the amplitude correlation and the phase correlation (the position corresponding to the cross symbol in the figure) are quite different. Therefore, the embodiment of the present application divides the complete path fading factor matrix into amplitude and phase correlation coefficient matrices, and performs compression feedback through two AI models.

[0134] Step 43: compress the first amplitude matrix using a first amplitude encoder to obtain a first amplitude bit stream, and / or compress the first phase matrix using a first phase encoder to obtain a first phase bit stream.

[0135] In an embodiment of the present application, the terminal may pre-receive first configuration information sent by a network device, and the first configuration information is used to configure the first amplitude encoder and / or the first phase encoder. Specifically, the first configuration information may include model parameters of the first amplitude encoder and / or the first phase encoder to configure the relevant encoder; or, the first configuration information may include an ID corresponding to the first amplitude encoder and / or the first phase encoder, so that the terminal can determine the corresponding encoder according to the ID. At this time, the model parameters of the encoders corresponding to all IDs need to be pre-configured to the terminal in advance.

[0136] In this way, in step 43, the terminal can use the first amplitude encoder to compress the first amplitude matrix, or use the first phase encoder to compress the first phase matrix, or use the first amplitude encoder to compress the first amplitude matrix and use the first phase encoder to compress the first phase matrix.

[0137] Step 44: Send the first amplitude bit stream and / or the first phase bit stream to a network device.

[0138] Here, in an embodiment of the present application, the terminal sends the first amplitude bit stream to the network device, or sends the first phase bit stream to the network device, or sends the first amplitude bit stream and the first phase bit stream to the network device.

[0139] Through the above steps, the embodiment of the present application decomposes the downlink correlation coefficient matrix into amplitude and phase correlation coefficient matrices, and performs compression feedback respectively through different encoders (AI compression models). It can utilize the reciprocity of amplitude and phase to greatly reduce the channel feedback overhead while ensuring the accuracy of channel recovery.

[0140] Generally, the higher the accuracy of the encoder, the greater its channel feedback overhead. In an embodiment of the present application, the network device can adjust the amplitude and phase feedback accuracy according to the downlink channel information to select an encoder with appropriate accuracy to reduce the channel feedback overhead. For example, an amplitude encoder and / or phase encoder with high amplitude and phase feedback accuracy can be configured for the terminal first. After the terminal uses the above encoder to perform the above steps 41 to 44, the network device can obtain downlink channel information with high accuracy. In this way, the network device compares the received downlink channel amplitude and phase information with the uplink channel amplitude and phase information, calculates the correlation, and reselects the amplitude encoder and / or phase encoder with suitable feedback accuracy according to the threshold of the correlation, and configures the selected encoder to the terminal. In this way, the embodiment of the present application can use an encoder with low model complexity and low feedback overhead to compress and feedback the amplitude and phase correlation coefficient matrix when the correlation is large, so as to reduce the channel feedback overhead; when the correlation is small, use an encoder with high model complexity and high feedback overhead to compress and feedback the amplitude and phase correlation coefficient matrix to ensure the channel recovery accuracy.

[0141] At this time, after the above step 44, the terminal may also receive the second configuration information sent by the network device, and the second configuration information is used to configure the second amplitude encoder and / or the second phase encoder. Subsequently, the terminal may obtain the second downlink channel matrix of the downlink channel state information by measuring the downlink reference signal (such as the downlink CSI-RS), and then transform the second downlink channel matrix into the angle domain and the delay domain to obtain the second downlink channel correlation coefficient matrix. After obtaining the second downlink channel correlation coefficient matrix, the terminal generates a second amplitude matrix for representing the downlink channel amplitude correlation coefficient and / or a second phase matrix for representing the downlink channel phase correlation coefficient according to the second downlink channel correlation coefficient matrix; then, the second amplitude matrix is ​​compressed using the second amplitude encoder to obtain a second amplitude bit stream, and / or, the second phase matrix is ​​compressed using the second phase encoder to obtain a second phase bit stream; and the second amplitude bit stream and / or the second phase bit stream are sent to the network device.

[0142] Please refer to Figure 6 The channel state information feedback method of the embodiment of the present application, when applied to a network device (such as a base station), includes:

[0143] Step 61, receiving a first amplitude bit stream and / or a first phase bit stream sent by a terminal, wherein the first amplitude bit stream is obtained by compressing a first amplitude matrix for representing a downlink channel amplitude correlation coefficient using a first amplitude encoder, and the first phase bit stream is obtained by compressing a first phase matrix for representing a downlink channel phase correlation coefficient using a first phase encoder by the terminal, and the first amplitude matrix and the first phase matrix are generated based on a first downlink channel correlation coefficient matrix obtained by converting the first downlink channel matrix.

[0144] Here, before step 61, the network device may also send first configuration information to the terminal, and the first configuration information is used to configure the first amplitude encoder and / or the first phase encoder, so that the terminal can use the above encoder to encode and compress the correlation matrix. Specifically, the first configuration information may include model parameters of the first amplitude encoder and / or the first phase encoder to configure the relevant encoder; or the first configuration information may include the ID corresponding to the first amplitude encoder and / or the first phase encoder, so that the terminal can determine the corresponding encoder according to the ID. At this time, the network device needs to pre-configure the model parameters of the encoders corresponding to all IDs to the terminal.

[0145] Step 62: use a first amplitude decoder to decode the first amplitude bit stream to obtain a first amplitude reasoning matrix; and / or use a first phase decoder to decode the first phase bit stream to obtain a first phase reasoning matrix; generate first downlink channel feedback information based on the first amplitude reasoning matrix and / or the first phase reasoning matrix.

[0146] Step 63: input the first uplink channel information and the first downlink channel feedback information into a downlink channel recovery model to obtain the first downlink channel information generated by the downlink channel recovery model.

[0147] The first amplitude encoder and the first amplitude decoder both belong to the autoencoder corresponding to the first amplitude correlation interval, and the first phase encoder and the first phase decoder both belong to the autoencoder corresponding to the first phase correlation interval.

[0148] Through the above steps, in the embodiment of the present application, the network device can recover downlink channel information, such as downlink channel state information, based on the amplitude bit stream and / or phase bit stream sent by the terminal. In the above scheme, the amplitude bit stream and / or phase bit stream sent by the terminal is compressed using different encoders, thereby utilizing the reciprocity of amplitude and phase to reduce channel feedback overhead.

[0149] In an embodiment of the present application, the network device may first configure an amplitude encoder and / or phase encoder with high amplitude and phase feedback accuracy for the terminal. After the terminal uses the above encoder to perform the above steps 41 to 44, the base station can obtain downlink channel information with high accuracy. In this way, the base station compares the received downlink channel amplitude and phase information with the uplink channel amplitude and phase information, calculates the correlation, and reselects an amplitude encoder and / or phase encoder with suitable feedback accuracy based on the correlation threshold, and configures the selected encoder to the terminal. In this way, when the correlation is large, an encoder with low model complexity and low feedback overhead can be used to compress and feedback the amplitude and phase correlation coefficient matrix to reduce the channel feedback overhead; when the correlation is small, an encoder with high model complexity and high feedback overhead can be used to compress and feedback the amplitude and phase correlation coefficient matrix to ensure channel recovery accuracy.

[0150] In an embodiment of the present application, the network device can dynamically adjust the amplitude and phase feedback accuracy according to the downlink channel information to select an encoder with appropriate accuracy to reduce the channel feedback overhead. For example, an amplitude encoder and / or phase encoder with high amplitude and phase feedback accuracy can be configured for the terminal first. After the network device executes the above steps 61 to 63, the network device can obtain downlink channel information with high accuracy. In this way, the network device compares the received downlink channel amplitude and phase information with the uplink channel amplitude and phase information, calculates the correlation, and reselects an amplitude encoder and / or phase encoder with appropriate feedback accuracy according to the threshold of the correlation, and configures the selected encoder to the terminal. In this way, the terminal can use the new encoder to compress and feedback the amplitude and phase correlation coefficient matrix to reduce the channel feedback overhead.

[0151] At this time, after the above step 63, the network device may also calculate the amplitude correlation parameter between the first downlink channel information and the first uplink channel information, determine the second amplitude correlation interval to which the amplitude correlation parameter belongs; determine the second amplitude encoder and the second amplitude decoder corresponding to the second amplitude correlation interval; and calculate the phase correlation parameter between the first downlink channel information and the first uplink channel information, determine the second phase correlation interval to which the phase correlation parameter belongs; determine the second phase encoder and the second phase decoder corresponding to the second phase correlation interval. Then, second configuration information is sent to the terminal, and the second configuration information is used to configure the second amplitude encoder and the second phase encoder.

[0152] Subsequently, the network device receives the second amplitude bit stream and / or the second phase bit stream sent by the terminal, wherein the second amplitude bit stream is obtained by compressing the second amplitude matrix used to represent the downlink channel amplitude correlation coefficient using the second amplitude encoder, and the second phase bit stream is obtained by compressing the second phase matrix used to represent the downlink channel phase correlation coefficient using the second phase encoder by the terminal, and the second amplitude matrix and the second phase matrix are generated based on the second downlink channel correlation coefficient matrix obtained by converting the second downlink channel matrix. Then, the network device decodes the second amplitude bit stream using the second amplitude decoder to obtain a second amplitude reasoning matrix; and / or, decodes the second phase bit stream using the second phase decoder to obtain a second phase reasoning matrix; generates second downlink channel feedback information based on the second amplitude reasoning matrix and / or the second phase reasoning matrix; and inputs the second uplink channel information and the second downlink channel feedback information into the downlink channel recovery model to obtain the second downlink channel information generated by the downlink channel recovery model.

[0153] In the embodiment of the present application, the network device may pre-divide the value range of the amplitude correlation parameter of the uplink and downlink channels into multiple continuous and non-overlapping amplitude correlation intervals. For example, the amplitude correlation parameter value range is from 0 to 1, which is divided into the following three amplitude correlation intervals:

[0154] The first interval is [0, 0.5); the second interval is [0.5, 0.7); the third interval is [0.7, 1].

[0155] Then, a model of an amplitude autoencoder is trained for each amplitude correlation interval, and the amplitude autoencoder includes an amplitude encoder and an amplitude decoder. Generally, when the amplitude correlation parameter in the amplitude correlation interval is large, the model parameters corresponding to the interval are more complex, and accordingly, the channel feedback overhead is larger. Table 1 provides an example of the above relationship:

[0156] Table 1

[0157] Amplitude correlation interval Model complexity Feedback overhead (number of output bits) [0.7,1] Simplified Model 10 [0.5,0.7) Moderately complex models 50 [0,0.5) Complex Model 200

[0158] Specifically, the first training data set may be generated based on a plurality of pre-acquired downlink channel training matrices in the same amplitude correlation interval. The first training data set includes: an amplitude training matrix generated based on the downlink channel training matrix and used to represent the downlink channel amplitude correlation coefficient.

[0159] Then, the first training data set is used to train the amplitude autoencoder corresponding to the amplitude correlation interval.

[0160] The specific training process is as follows:

[0161] The amplitude training matrix is ​​input into an amplitude encoder to generate a first training bit stream; the first training bit stream is input into an amplitude decoder to generate an amplitude reference matrix; a first loss value between the amplitude reference matrix and the amplitude training matrix is ​​calculated, and the amplitude encoder and the amplitude decoder are updated according to the first loss value until a preset first training end condition is met, thereby obtaining an amplitude autoencoder corresponding to the amplitude correlation interval;

[0162] Here, since the amplitude correlation coefficient is sensitive to the absolute size, the first loss value can be represented by the Normalized Mean Squared Error (NMSE) between the amplitude reference matrix and the amplitude training matrix.

[0163] Similarly, the embodiment of the present application can also divide the value range of the phase correlation parameter of the uplink and downlink channels into multiple continuous and non-overlapping phase correlation intervals; then, a phase autoencoder model is trained for each phase correlation interval, and the phase autoencoder includes a phase encoder and a phase decoder. Generally, when the phase correlation parameter in the phase correlation interval has a large value, the model parameters corresponding to the interval are more complex, and accordingly, the channel feedback overhead is larger.

[0164] Then, based on multiple downlink channel training matrices in the same phase correlation interval, a second training data set is generated, and the second training data set includes: a phase training matrix for representing the downlink channel phase correlation coefficient generated based on the downlink channel training matrix; then, the phase training matrix is ​​input into a phase encoder to generate a second training bit stream; the second training bit stream is input into a phase decoder to generate a phase reference matrix; a second loss value between the phase reference matrix and the phase training matrix is ​​calculated, and the phase encoder and phase decoder are updated according to the second loss value until a preset second training end condition is met, and a phase autoencoder corresponding to the amplitude correlation interval is obtained.

[0165] Here, the phase correlation coefficient is sensitive to the relative angle, so the second loss value can be represented by the cosine similarity between the phase reference matrix and the phase training matrix.

[0166] In addition, the embodiment of the present application also needs to pre-train the downlink channel recovery model, and the specific training is as follows: the uplink channel information corresponding to the amplitude reference matrix, the phase reference matrix and the downlink channel training matrix is ​​input into the downlink channel recovery model, and the downlink channel training matrix output by the downlink channel recovery model is minimized as the goal (that is, the downlink channel training matrix is ​​used as the output label of the model training), and the downlink channel recovery model corresponding to the amplitude correlation interval is obtained by training. Here, the loss function can select NMSE or cosine similarity.

[0167] In the embodiment of the present application, the amplitude correlation parameters of the uplink and downlink channels can be calculated in a variety of ways. A specific calculation method is provided below: First, the covariance Cov(X, Y) of the uplink and downlink channel amplitude matrices is calculated, and then the standard deviation S of the uplink channel amplitude matrix is ​​calculated respectively. x and the standard deviation S of the downlink channel amplitude matrix y Then, according to the covariance of the uplink and downlink channel amplitude matrix, the standard deviation of the uplink channel amplitude matrix and the standard deviation of the downlink channel amplitude matrix, the amplitude correlation parameter r of the uplink and downlink channels is calculated. xy , the specific calculation formula is as follows:

[0168]

[0169] The closer the amplitude correlation parameter is to 1, the greater the correlation is; the closer the amplitude correlation parameter is to 0, the smaller the correlation is.

[0170] The calculation of the phase correlation parameter is similar and will not be described again here.

[0171] Figure 7 An example of the interaction process between the terminal (UE) and the base station (gNB) of the above-mentioned method of an embodiment of the present application is further provided, including steps 701 to 708.

[0172] The above describes various methods of the embodiments of the present application. The following further provides devices for implementing the above methods.

[0173] Please refer to Figure 8 , the embodiment of the present application further provides a terminal, including:

[0174] A first obtaining module 801, configured to obtain a first downlink channel correlation coefficient matrix of a downlink channel;

[0175] A first generating module 802 is used to generate a first amplitude matrix for representing downlink channel amplitude correlation coefficients and / or a first phase matrix for representing downlink channel phase correlation coefficients according to the first downlink channel correlation coefficient matrix;

[0176] A first compression module 803 is configured to compress the first amplitude matrix using a first amplitude encoder to obtain a first amplitude bit stream, and / or to compress the first phase matrix using a first phase encoder to obtain a first phase bit stream;

[0177] The first sending module 804 is configured to send the first amplitude bit stream and / or the first phase bit stream to a network device.

[0178] Through the above modules, the embodiment of the present application can utilize the reciprocity of amplitude and phase to significantly reduce the channel feedback overhead while ensuring the accuracy of channel recovery.

[0179] Optionally, the first acquisition module is further used to acquire a first downlink channel matrix of downlink channel state information, transform the first downlink channel matrix into an angle domain and a delay domain, and obtain the first downlink channel correlation coefficient matrix.

[0180] Optionally, the terminal further includes:

[0181] The first receiving module is used to receive first configuration information sent by a network device, where the first configuration information is used to configure the first amplitude encoder and / or the first phase encoder.

[0182] Optionally, the terminal further includes:

[0183] A second receiving module, used to receive second configuration information sent by the network device, where the second configuration information is used to configure a second amplitude encoder and / or a second phase encoder;

[0184] A second obtaining module, used to obtain a second downlink channel correlation coefficient matrix of the downlink channel;

[0185] A second generating module, used for generating a second amplitude matrix for representing downlink channel amplitude correlation coefficients and / or a second phase matrix for representing downlink channel phase correlation coefficients according to the second downlink channel correlation coefficient matrix;

[0186] A second compression module, configured to compress the second amplitude matrix using the second amplitude encoder to obtain a second amplitude bit stream, and / or to compress the second phase matrix using a second phase encoder to obtain a second phase bit stream;

[0187] The second sending module is used to send the second amplitude bit stream and / or the second phase bit stream to the network device.

[0188] Optionally, the second acquisition module is further used to obtain a second downlink channel matrix of the downlink channel state information, transform the second downlink channel matrix into the angle domain and the delay domain, and obtain the second downlink channel correlation coefficient matrix.

[0189] It should be noted that the device in this embodiment is a device corresponding to the method applied to the terminal side, and the implementation methods in the above embodiments are applicable to the embodiments of the device, and can also achieve the same technical effect. The above device provided in the embodiment of the present application can implement all the method steps implemented in the above method embodiment, and can achieve the same technical effect. The parts and beneficial effects that are the same as those in the method embodiment in this embodiment will not be specifically described here.

[0190] Please refer to Fig. 9 , the embodiment of the present application further provides a terminal 900, including: a transceiver 901 and a processor 902;

[0191] The processor 902 is used to obtain a first downlink channel correlation coefficient matrix of a downlink channel; generate a first amplitude matrix for representing downlink channel amplitude correlation coefficients and / or a first phase matrix for representing downlink channel phase correlation coefficients according to the first downlink channel correlation coefficient matrix; compress the first amplitude matrix using a first amplitude encoder to obtain a first amplitude bit stream, and / or compress the first phase matrix using a first phase encoder to obtain a first phase bit stream;

[0192] The transceiver 901 is used to send the first amplitude bit stream and / or the first phase bit stream to a network device.

[0193] Optionally, the processor is further configured to:

[0194] A first downlink channel matrix of downlink channel state information is acquired, and the first downlink channel matrix is ​​transformed into an angle domain and a delay domain to obtain the first downlink channel correlation coefficient matrix.

[0195] Optionally, the transceiver is further used for:

[0196] Receive first configuration information sent by a network device, where the first configuration information is used to configure the first amplitude encoder and / or the first phase encoder.

[0197] Optionally, the transceiver is further used to receive second configuration information sent by the network device after sending the first amplitude bit stream and / or the first phase bit stream to the network device, where the second configuration information is used to configure a second amplitude encoder and / or a second phase encoder;

[0198] The processor is further configured to obtain a second downlink channel correlation coefficient matrix of the downlink channel; generate a second amplitude matrix for representing downlink channel amplitude correlation coefficients and / or a second phase matrix for representing downlink channel phase correlation coefficients according to the second downlink channel correlation coefficient matrix; compress the second amplitude matrix using the second amplitude encoder to obtain a second amplitude bit stream, and / or compress the second phase matrix using the second phase encoder to obtain a second phase bit stream;

[0199] The transceiver is further configured to send the second amplitude bit stream and / or the second phase bit stream to a network device.

[0200] Optionally, the processor is further configured to:

[0201] A second downlink channel matrix of the downlink channel state information is acquired, and the second downlink channel matrix is ​​transformed into an angle domain and a delay domain to obtain the second downlink channel correlation coefficient matrix.

[0202] It should be noted that the device in this embodiment is a device corresponding to the method applied to the terminal side, and the implementation methods in the above embodiments are applicable to the embodiments of the device, and can also achieve the same technical effect. The above device provided in the embodiment of the present application can implement all the method steps implemented in the above method embodiment, and can achieve the same technical effect. The parts and beneficial effects that are the same as those in the method embodiment in this embodiment will not be specifically described here.

[0203] Please refer to Fig.10 , the embodiment of the present application further provides a network device, including:

[0204] The first receiving module 1001 is used to receive a first amplitude bit stream and / or a first phase bit stream sent by a terminal, wherein the first amplitude bit stream is obtained by compressing a first amplitude matrix used to represent a downlink channel amplitude correlation coefficient using a first amplitude encoder, and the first phase bit stream is obtained by compressing a first phase matrix used to represent a downlink channel phase correlation coefficient using a first phase encoder by the terminal, and the first amplitude matrix and the first phase matrix are generated based on a first downlink channel correlation coefficient matrix obtained by converting a first downlink channel matrix;

[0205] A first decoding module 1002 is configured to decode the first amplitude bit stream using a first amplitude decoder to obtain a first amplitude reasoning matrix; and / or decode the first phase bit stream using a first phase decoder to obtain a first phase reasoning matrix; and generate first downlink channel feedback information based on the first amplitude reasoning matrix and / or the first phase reasoning matrix;

[0206] A first obtaining module 1003 is used to input the first uplink channel information and the first downlink channel feedback information into a downlink channel recovery model to obtain the first downlink channel information generated by the downlink channel recovery model;

[0207] The first amplitude encoder and the first amplitude decoder both belong to the autoencoder corresponding to the first amplitude correlation interval, and the first phase encoder and the first phase decoder both belong to the autoencoder corresponding to the first phase correlation interval.

[0208] Optionally, the above network device further includes:

[0209] The first sending module is used to send first configuration information to the terminal before receiving the first amplitude bit stream and / or the first phase bit stream sent by the terminal, where the first configuration information is used to configure the first amplitude encoder and / or the first phase encoder.

[0210] Optionally, the above network device further includes:

[0211] A first determination module is configured to calculate an amplitude correlation parameter between the first downlink channel information and the first uplink channel information, determine a second amplitude correlation interval to which the amplitude correlation parameter belongs; and determine a second amplitude encoder and a second amplitude decoder corresponding to the second amplitude correlation interval;

[0212] A second determination module is used to calculate a phase correlation parameter between the first downlink channel information and the first uplink channel information, determine a second phase correlation interval to which the phase correlation parameter belongs; and determine a second phase encoder and a second phase decoder corresponding to the second phase correlation interval;

[0213] The second sending module is used to send second configuration information to the terminal, where the second configuration information is used to configure the second amplitude encoder and the second phase encoder.

[0214] Optionally, the above network device further includes:

[0215] A second receiving module is used to receive a second amplitude bit stream and / or a second phase bit stream sent by the terminal, wherein the second amplitude bit stream is obtained by compressing a second amplitude matrix used to represent a downlink channel amplitude correlation coefficient using the second amplitude encoder, and the second phase bit stream is obtained by compressing a second phase matrix used to represent a downlink channel phase correlation coefficient using the second phase encoder by the terminal, and the second amplitude matrix and the second phase matrix are generated based on a second downlink channel correlation coefficient matrix obtained by converting the second downlink channel matrix;

[0216] A second decoding module, used to decode the second amplitude bit stream using the second amplitude decoder to obtain a second amplitude reasoning matrix; and / or, to decode the second phase bit stream using a second phase decoder to obtain a second phase reasoning matrix; and generate second downlink channel feedback information based on the second amplitude reasoning matrix and / or the second phase reasoning matrix;

[0217] The second obtaining module is used to input the second uplink channel information and the second downlink channel feedback information into the downlink channel recovery model to obtain the second downlink channel information generated by the downlink channel recovery model.

[0218] Optionally, the above network device further includes:

[0219] Training modules for:

[0220] Dividing the value range of the amplitude correlation parameter of the uplink and downlink channels into a plurality of continuous and non-overlapping amplitude correlation intervals; dividing the value range of the phase correlation parameter of the uplink and downlink channels into a plurality of continuous and non-overlapping phase correlation intervals;

[0221] Based on a plurality of downlink channel training matrices in the same amplitude correlation interval, a first training data set is generated, wherein the first training data set includes: an amplitude training matrix for representing a downlink channel amplitude correlation coefficient generated based on the downlink channel training matrix;

[0222] Based on a plurality of downlink channel training matrices in the same phase correlation interval, a second training data set is generated, wherein the second training data set includes: a phase training matrix for representing a downlink channel phase correlation coefficient generated based on the downlink channel training matrix;

[0223] The first training data set is used to train the amplitude autoencoder corresponding to the amplitude correlation interval; and / or the second training data set is used to train the phase autoencoder corresponding to the phase correlation interval, wherein the amplitude autoencoder includes an amplitude encoder and an amplitude decoder, and the phase autoencoder includes a phase encoder and a phase decoder.

[0224] Optionally, the training module is further used to:

[0225] The amplitude training matrix is ​​input into an amplitude encoder to generate a first training bit stream; the first training bit stream is input into an amplitude decoder to generate an amplitude reference matrix; a first loss value between the amplitude reference matrix and the amplitude training matrix is ​​calculated, and the amplitude encoder and the amplitude decoder are updated according to the first loss value until a preset first training end condition is met, thereby obtaining an amplitude autoencoder corresponding to the amplitude correlation interval;

[0226] and / or,

[0227] The phase training matrix is ​​input into a phase encoder to generate a second training bit stream; the second training bit stream is input into a phase decoder to generate a phase reference matrix; a second loss value between the phase reference matrix and the phase training matrix is ​​calculated, and the phase encoder and phase decoder are updated according to the second loss value until a preset second training end condition is met, thereby obtaining a phase autoencoder corresponding to the amplitude correlation interval.

[0228] Optionally, the first loss value is represented by a mean square error between the amplitude reference matrix and the amplitude training matrix, and the second loss value is represented by a cosine similarity between the phase reference matrix and the phase training matrix.

[0229] Optionally, the training module is further used to:

[0230] The uplink channel information corresponding to the amplitude reference matrix, the phase reference matrix and the downlink channel training matrix is ​​input into a downlink channel recovery model, and the downlink channel recovery model corresponding to the amplitude correlation interval is trained with the goal of minimizing the downlink channel training matrix error output by the downlink channel recovery model.

[0231] It should be noted that the device in this embodiment is a device corresponding to the method applied to the network device side, and the implementation methods in the above embodiments are all applicable to the embodiments of the device, and can also achieve the same technical effect. The above device provided in the embodiment of the present application can implement all the method steps implemented in the above method embodiment, and can achieve the same technical effect, and the parts and beneficial effects that are the same as those in the method embodiment in this embodiment will not be specifically repeated here.

[0232] Please refer to Fig.11 , the embodiment of the present application further provides a network device 1100, including: a transceiver 1101 and a processor 1102;

[0233] The transceiver 1101 is used to receive a first amplitude bit stream and / or a first phase bit stream sent by a terminal, wherein the first amplitude bit stream is obtained by compressing a first amplitude matrix used to represent a downlink channel amplitude correlation coefficient using a first amplitude encoder, and the first phase bit stream is obtained by compressing a first phase matrix used to represent a downlink channel phase correlation coefficient using a first phase encoder by the terminal, and the first amplitude matrix and the first phase matrix are generated based on a first downlink channel correlation coefficient matrix obtained by converting a first downlink channel matrix;

[0234] The processor 1102 is configured to decode the first amplitude bit stream using a first amplitude decoder to obtain a first amplitude reasoning matrix; and / or decode the first phase bit stream using a first phase decoder to obtain a first phase reasoning matrix; generate first downlink channel feedback information based on the first amplitude reasoning matrix and / or the first phase reasoning matrix; input the first uplink channel information and the first downlink channel feedback information into a downlink channel recovery model to obtain the first downlink channel information generated by the downlink channel recovery model;

[0235] The first amplitude encoder and the first amplitude decoder both belong to the autoencoder corresponding to the first amplitude correlation interval, and the first phase encoder and the first phase decoder both belong to the autoencoder corresponding to the first phase correlation interval.

[0236] Optionally, the transceiver is further used for:

[0237] Before receiving the first amplitude bit stream and / or the first phase bit stream sent by the terminal, first configuration information is sent to the terminal, where the first configuration information is used to configure the first amplitude encoder and / or the first phase encoder.

[0238] Optionally, the processor is further configured to:

[0239] Calculating an amplitude correlation parameter between the first downlink channel information and the first uplink channel information, and determining a second amplitude correlation interval to which the amplitude correlation parameter belongs; and determining a second amplitude encoder and a second amplitude decoder corresponding to the second amplitude correlation interval;

[0240] Calculating a phase correlation parameter between the first downlink channel information and the first uplink channel information, and determining a second phase correlation interval to which the phase correlation parameter belongs; determining a second phase encoder and a second phase decoder corresponding to the second phase correlation interval;

[0241] Sending second configuration information to the terminal, where the second configuration information is used to configure the second amplitude encoder and the second phase encoder.

[0242] Optionally, the transceiver is further used to receive a second amplitude bit stream and / or a second phase bit stream sent by the terminal, wherein the second amplitude bit stream is obtained by compressing a second amplitude matrix used to represent a downlink channel amplitude correlation coefficient using the second amplitude encoder, and the second phase bit stream is obtained by compressing a second phase matrix used to represent a downlink channel phase correlation coefficient by the terminal using a second phase encoder, and the second amplitude matrix and the second phase matrix are generated based on a second downlink channel correlation coefficient matrix obtained by converting the second downlink channel matrix;

[0243] The processor is further used to decode the second amplitude bit stream using the second amplitude decoder to obtain a second amplitude reasoning matrix; and / or, decode the second phase bit stream using the second phase decoder to obtain a second phase reasoning matrix; generate second downlink channel feedback information based on the second amplitude reasoning matrix and / or the second phase reasoning matrix; input the second uplink channel information and the second downlink channel feedback information into the downlink channel recovery model to obtain the second downlink channel information generated by the downlink channel recovery model.

[0244] Optionally, the processor is further configured to:

[0245] Dividing the value range of the amplitude correlation parameter of the uplink and downlink channels into a plurality of continuous and non-overlapping amplitude correlation intervals; dividing the value range of the phase correlation parameter of the uplink and downlink channels into a plurality of continuous and non-overlapping phase correlation intervals;

[0246] Based on a plurality of downlink channel training matrices in the same amplitude correlation interval, a first training data set is generated, wherein the first training data set includes: an amplitude training matrix for representing a downlink channel amplitude correlation coefficient generated based on the downlink channel training matrix;

[0247] Based on a plurality of downlink channel training matrices in the same phase correlation interval, a second training data set is generated, wherein the second training data set includes: a phase training matrix for representing a downlink channel phase correlation coefficient generated based on the downlink channel training matrix;

[0248] The first training data set is used to train the amplitude autoencoder corresponding to the amplitude correlation interval; and / or the second training data set is used to train the phase autoencoder corresponding to the phase correlation interval, wherein the amplitude autoencoder includes an amplitude encoder and an amplitude decoder, and the phase autoencoder includes a phase encoder and a phase decoder.

[0249] Optionally, the processor is further configured to:

[0250] The amplitude training matrix is ​​input into an amplitude encoder to generate a first training bit stream; the first training bit stream is input into an amplitude decoder to generate an amplitude reference matrix; a first loss value between the amplitude reference matrix and the amplitude training matrix is ​​calculated, and the amplitude encoder and the amplitude decoder are updated according to the first loss value until a preset first training end condition is met, thereby obtaining an amplitude autoencoder corresponding to the amplitude correlation interval;

[0251] and / or,

[0252] The phase training matrix is ​​input into a phase encoder to generate a second training bit stream; the second training bit stream is input into a phase decoder to generate a phase reference matrix; a second loss value between the phase reference matrix and the phase training matrix is ​​calculated, and the phase encoder and phase decoder are updated according to the second loss value until a preset second training end condition is met, thereby obtaining a phase autoencoder corresponding to the amplitude correlation interval.

[0253] Optionally, the first loss value is represented by a mean square error between the amplitude reference matrix and the amplitude training matrix, and the second loss value is represented by a cosine similarity between the phase reference matrix and the phase training matrix.

[0254] Optionally, the processor is further configured to:

[0255] The uplink channel information corresponding to the amplitude reference matrix, the phase reference matrix and the downlink channel training matrix is ​​input into a downlink channel recovery model, and the downlink channel recovery model corresponding to the amplitude correlation interval is trained with the goal of minimizing the downlink channel training matrix error output by the downlink channel recovery model.

[0256] It should be noted that the device in this embodiment is a device corresponding to the method applied to the network device side, and the implementation methods in the above embodiments are all applicable to the embodiments of the device, and can also achieve the same technical effect. The above device provided in the embodiment of the present application can implement all the method steps implemented in the above method embodiment, and can achieve the same technical effect, and the parts and beneficial effects that are the same as those in the method embodiment in this embodiment will not be specifically repeated here.

[0257] Please refer to Fig.12 The embodiment of the present application also provides a terminal 1200, including a processor 1201, a memory 1202, and a computer program stored in the memory 1202 and executable on the processor 1201. When the computer program is executed by the processor 1201, each process of the channel state information feedback method embodiment executed by the terminal is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0258] Please refer to Fig.13 The embodiment of the present application also provides a network device 1300, including a processor 1301, a memory 1302, and a computer program stored in the memory 1302 and executable on the processor 1301. When the computer program is executed by the processor 1301, each process of the channel state information feedback method embodiment executed by the network device is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0259] The embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, each process of the channel state information feedback method embodiment described above is implemented, and the same technical effect can be achieved. To avoid repetition, it is not repeated here. Among them, the computer-readable storage medium, such as a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0260] It should be noted that, in this article, the terms "include", "include" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0261] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0262] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose of the present application and the scope of protection of the claims, all of which are within the protection of the present application.

Claims

1. A channel state information feedback method, applied to a terminal, characterized in that: include: Obtaining a first downlink channel correlation coefficient matrix of a downlink channel; generating, according to the first downlink channel correlation coefficient matrix, a first amplitude matrix for representing downlink channel amplitude correlation coefficients and / or a first phase matrix for representing downlink channel phase correlation coefficients; Compressing the first amplitude matrix using a first amplitude encoder to obtain a first amplitude bit stream, and / or compressing the first phase matrix using a first phase encoder to obtain a first phase bit stream; The first amplitude bit stream and / or the first phase bit stream is sent to a network device.

2. The method according to claim 1, characterized in that The obtaining of the first downlink channel correlation coefficient matrix comprises: A first downlink channel matrix of downlink channel state information is acquired, and the first downlink channel matrix is ​​transformed into an angle domain and a delay domain to obtain the first downlink channel correlation coefficient matrix.

3. The method according to claim 1, characterized in that Before compressing the first amplitude matrix using a first amplitude encoder to obtain a first amplitude bit stream, and / or compressing the first phase matrix using a first phase encoder to obtain a first phase bit stream, the method further includes: Receive first configuration information sent by a network device, where the first configuration information is used to configure the first amplitude encoder and / or the first phase encoder.

4. The method according to claim 1, characterized in that After sending the first amplitude bit stream and / or the first phase bit stream to the network device, the method further includes: Receiving second configuration information sent by the network device, where the second configuration information is used to configure a second amplitude encoder and / or a second phase encoder; Obtaining a second downlink channel correlation coefficient matrix of the downlink channel; generating, according to the second downlink channel correlation coefficient matrix, a second amplitude matrix for representing downlink channel amplitude correlation coefficients and / or a second phase matrix for representing downlink channel phase correlation coefficients; Compressing the second amplitude matrix using the second amplitude encoder to obtain a second amplitude bit stream, and / or compressing the second phase matrix using a second phase encoder to obtain a second phase bit stream; The second amplitude bit stream and / or the second phase bit stream is sent to a network device.

5. The method according to claim 4, characterized in that The obtaining of a second downlink channel correlation coefficient matrix of the downlink channel comprises: A second downlink channel matrix of the downlink channel state information is acquired, and the second downlink channel matrix is ​​transformed into an angle domain and a delay domain to obtain the second downlink channel correlation coefficient matrix.

6. A channel state information feedback method, applied to a network device, characterized in that: include: A first amplitude bit stream and / or a first phase bit stream sent by a receiving terminal, wherein the first amplitude bit stream is obtained by compressing a first amplitude matrix used to represent a downlink channel amplitude correlation coefficient using a first amplitude encoder, and the first phase bit stream is obtained by compressing a first phase matrix used to represent a downlink channel phase correlation coefficient using a first phase encoder by the terminal, and the first amplitude matrix and the first phase matrix are generated based on a first downlink channel correlation coefficient matrix obtained by converting the first downlink channel matrix; Decoding the first amplitude bit stream using a first amplitude decoder to obtain a first amplitude reasoning matrix; and / or decoding the first phase bit stream using a first phase decoder to obtain a first phase reasoning matrix; generating first downlink channel feedback information based on the first amplitude reasoning matrix and / or the first phase reasoning matrix; Inputting the first uplink channel information and the first downlink channel feedback information into a downlink channel recovery model to obtain the first downlink channel information generated by the downlink channel recovery model; The first amplitude encoder and the first amplitude decoder both belong to the autoencoder corresponding to the first amplitude correlation interval, and the first phase encoder and the first phase decoder both belong to the autoencoder corresponding to the first phase correlation interval.

7. The method according to claim 6, characterized in that Before receiving the first amplitude bit stream and / or the first phase bit stream sent by the terminal, the method further includes: Sending first configuration information to the terminal, where the first configuration information is used to configure the first amplitude encoder and / or the first phase encoder.

8. The method according to claim 6, characterized in that Also includes: calculating an amplitude correlation parameter between the first downlink channel information and the first uplink channel information, and determining a second amplitude correlation interval to which the amplitude correlation parameter belongs; Determine a second amplitude encoder and a second amplitude decoder corresponding to the second amplitude correlation interval; calculating a phase correlation parameter between the first downlink channel information and the first uplink channel information, and determining a second phase correlation interval to which the phase correlation parameter belongs; Determine a second phase encoder and a second phase decoder corresponding to the second phase correlation interval; Sending second configuration information to the terminal, where the second configuration information is used to configure the second amplitude encoder and the second phase encoder.

9. The method according to claim 8, characterized in that Also includes: A second amplitude bit stream and / or a second phase bit stream sent by a receiving terminal, wherein the second amplitude bit stream is obtained by compressing a second amplitude matrix used to represent a downlink channel amplitude correlation coefficient using the second amplitude encoder, and the second phase bit stream is obtained by compressing a second phase matrix used to represent a downlink channel phase correlation coefficient using the second phase encoder by the terminal, and the second amplitude matrix and the second phase matrix are generated based on a second downlink channel correlation coefficient matrix obtained by converting the second downlink channel matrix; Decoding the second amplitude bit stream using the second amplitude decoder to obtain a second amplitude reasoning matrix; and / or decoding the second phase bit stream using a second phase decoder to obtain a second phase reasoning matrix; generating second downlink channel feedback information based on the second amplitude reasoning matrix and / or the second phase reasoning matrix; The second uplink channel information and the second downlink channel feedback information are input into the downlink channel recovery model to obtain the second downlink channel information generated by the downlink channel recovery model.

10. The method according to claim 6, characterized in that Also includes: Dividing the value range of the amplitude correlation parameter of the uplink and downlink channels into a plurality of continuous and non-overlapping amplitude correlation intervals; Dividing the value range of the phase correlation parameter of the uplink and downlink channels into a plurality of continuous and non-overlapping phase correlation intervals; Based on a plurality of downlink channel training matrices in the same amplitude correlation interval, a first training data set is generated, wherein the first training data set includes: an amplitude training matrix for representing a downlink channel amplitude correlation coefficient generated based on the downlink channel training matrix; Based on a plurality of downlink channel training matrices in the same phase correlation interval, a second training data set is generated, wherein the second training data set includes: a phase training matrix for representing a downlink channel phase correlation coefficient generated based on the downlink channel training matrix; The first training data set is used to train the amplitude autoencoder corresponding to the amplitude correlation interval; and / or the second training data set is used to train the phase autoencoder corresponding to the phase correlation interval, wherein the amplitude autoencoder includes an amplitude encoder and an amplitude decoder, and the phase autoencoder includes a phase encoder and a phase decoder.

11. The method according to claim 10, characterized in that The using the first training data set to train the amplitude autoencoder corresponding to the amplitude correlation interval; And / or, using the second training data set to train a phase autoencoder corresponding to the phase correlation interval, comprising: The amplitude training matrix is ​​input into an amplitude encoder to generate a first training bit stream; the first training bit stream is input into an amplitude decoder to generate an amplitude reference matrix; a first loss value between the amplitude reference matrix and the amplitude training matrix is ​​calculated, and the amplitude encoder and the amplitude decoder are updated according to the first loss value until a preset first training end condition is met, thereby obtaining an amplitude autoencoder corresponding to the amplitude correlation interval; and / or, The phase training matrix is ​​input into a phase encoder to generate a second training bit stream; the second training bit stream is input into a phase decoder to generate a phase reference matrix; a second loss value between the phase reference matrix and the phase training matrix is ​​calculated, and the phase encoder and phase decoder are updated according to the second loss value until a preset second training end condition is met, thereby obtaining a phase autoencoder corresponding to the amplitude correlation interval.

12. The method according to claim 11, characterized in that The first loss value is represented by a mean square error between the amplitude reference matrix and the amplitude training matrix, and the second loss value is represented by a cosine similarity between the phase reference matrix and the phase training matrix.

13. The method according to claim 11, characterized in that Also includes: The uplink channel information corresponding to the amplitude reference matrix, the phase reference matrix and the downlink channel training matrix is ​​input into a downlink channel recovery model, and the downlink channel recovery model corresponding to the amplitude correlation interval is obtained by training with the goal of minimizing the downlink channel training matrix error output by the downlink channel recovery model.

14. A terminal, characterized in that: comprising a transceiver and a processor, wherein: The processor is used to obtain a first downlink channel correlation coefficient matrix of a downlink channel; generate a first amplitude matrix for representing downlink channel amplitude correlation coefficients and / or a first phase matrix for representing downlink channel phase correlation coefficients according to the first downlink channel correlation coefficient matrix; compress the first amplitude matrix using a first amplitude encoder to obtain a first amplitude bit stream, and / or compress the first phase matrix using a first phase encoder to obtain a first phase bit stream; The transceiver is used to send the first amplitude bit stream and / or the first phase bit stream to a network device.

15. The terminal according to claim 14, characterized in that The processor is further configured to: A first downlink channel matrix of downlink channel state information is acquired, and the first downlink channel matrix is ​​transformed into an angle domain and a delay domain to obtain the first downlink channel correlation coefficient matrix.

16. The terminal according to claim 14, characterized in that The transceiver is further used for: Receive first configuration information sent by a network device, where the first configuration information is used to configure the first amplitude encoder and / or the first phase encoder.

17. The terminal according to claim 14, characterized in that: The transceiver is further configured to receive second configuration information sent by the network device after sending the first amplitude bit stream and / or the first phase bit stream to the network device, wherein the second configuration information is used to configure a second amplitude encoder and / or a second phase encoder; The processor is further configured to obtain a second downlink channel correlation coefficient matrix of the downlink channel; generate a second amplitude matrix for representing downlink channel amplitude correlation coefficients and / or a second phase matrix for representing downlink channel phase correlation coefficients according to the second downlink channel correlation coefficient matrix; compress the second amplitude matrix using the second amplitude encoder to obtain a second amplitude bit stream, and / or compress the second phase matrix using the second phase encoder to obtain a second phase bit stream; The transceiver is further configured to send the second amplitude bit stream and / or the second phase bit stream to a network device.

18. The terminal according to claim 17, characterized in that The processor is further configured to: A second downlink channel matrix of the downlink channel state information is acquired, and the second downlink channel matrix is ​​transformed into an angle domain and a delay domain to obtain the second downlink channel correlation coefficient matrix.

19. A terminal, characterized in that: include: A processor, a memory, and a program stored in the memory and executable on the processor, wherein when the program is executed by the processor, the steps of the terminal according to any one of claims 1 to 5 are implemented.

20. A network device, characterized in that: comprising a transceiver and a processor, wherein: The transceiver is used to receive a first amplitude bit stream and / or a first phase bit stream sent by a terminal, wherein the first amplitude bit stream is obtained by compressing a first amplitude matrix used to represent a downlink channel amplitude correlation coefficient using a first amplitude encoder, and the first phase bit stream is obtained by compressing a first phase matrix used to represent a downlink channel phase correlation coefficient using a first phase encoder by the terminal, and the first amplitude matrix and the first phase matrix are generated based on a first downlink channel correlation coefficient matrix obtained by converting a first downlink channel matrix; The processor is configured to decode the first amplitude bit stream using a first amplitude decoder to obtain a first amplitude reasoning matrix; and / or decode the first phase bit stream using a first phase decoder to obtain a first phase reasoning matrix; generate first downlink channel feedback information based on the first amplitude reasoning matrix and / or the first phase reasoning matrix; input the first uplink channel information and the first downlink channel feedback information into a downlink channel recovery model to obtain the first downlink channel information generated by the downlink channel recovery model; The first amplitude encoder and the first amplitude decoder both belong to the autoencoder corresponding to the first amplitude correlation interval, and the first phase encoder and the first phase decoder both belong to the autoencoder corresponding to the first phase correlation interval.

21. The network device according to claim 20, characterized in that The transceiver is further used for: Before receiving the first amplitude bit stream and / or the first phase bit stream sent by the terminal, first configuration information is sent to the terminal, where the first configuration information is used to configure the first amplitude encoder and / or the first phase encoder.

22. The network device according to claim 20, characterized in that The processor is further configured to: calculating an amplitude correlation parameter between the first downlink channel information and the first uplink channel information, and determining a second amplitude correlation interval to which the amplitude correlation parameter belongs; Determine a second amplitude encoder and a second amplitude decoder corresponding to the second amplitude correlation interval; calculating a phase correlation parameter between the first downlink channel information and the first uplink channel information, and determining a second phase correlation interval to which the phase correlation parameter belongs; Determine a second phase encoder and a second phase decoder corresponding to the second phase correlation interval; Sending second configuration information to the terminal, where the second configuration information is used to configure the second amplitude encoder and the second phase encoder.

23. The network device according to claim 22, characterized in that Also includes: The transceiver is further used to receive a second amplitude bit stream and / or a second phase bit stream sent by the terminal, wherein the second amplitude bit stream is obtained by compressing a second amplitude matrix used to represent a downlink channel amplitude correlation coefficient using the second amplitude encoder, and the second phase bit stream is obtained by compressing a second phase matrix used to represent a downlink channel phase correlation coefficient using the second phase encoder by the terminal, and the second amplitude matrix and the second phase matrix are generated based on a second downlink channel correlation coefficient matrix obtained by converting the second downlink channel matrix; The processor is further used to decode the second amplitude bit stream using the second amplitude decoder to obtain a second amplitude reasoning matrix; and / or, decode the second phase bit stream using the second phase decoder to obtain a second phase reasoning matrix; generate second downlink channel feedback information based on the second amplitude reasoning matrix and / or the second phase reasoning matrix; input the second uplink channel information and the second downlink channel feedback information into the downlink channel recovery model to obtain the second downlink channel information generated by the downlink channel recovery model.

24. The network device according to claim 20, characterized in that The processor is further configured to: Dividing the value range of the amplitude correlation parameter of the uplink and downlink channels into a plurality of continuous and non-overlapping amplitude correlation intervals; dividing the value range of the phase correlation parameter of the uplink and downlink channels into a plurality of continuous and non-overlapping phase correlation intervals; Based on a plurality of downlink channel training matrices in the same amplitude correlation interval, a first training data set is generated, wherein the first training data set includes: an amplitude training matrix for representing a downlink channel amplitude correlation coefficient generated based on the downlink channel training matrix; Based on a plurality of downlink channel training matrices in the same phase correlation interval, a second training data set is generated, wherein the second training data set includes: a phase training matrix for representing a downlink channel phase correlation coefficient generated based on the downlink channel training matrix; The first training data set is used to train the amplitude autoencoder corresponding to the amplitude correlation interval; and / or the second training data set is used to train the phase autoencoder corresponding to the phase correlation interval, wherein the amplitude autoencoder includes an amplitude encoder and an amplitude decoder, and the phase autoencoder includes a phase encoder and a phase decoder.

25. The network device according to claim 24, characterized in that The processor is further configured to: Inputting the amplitude training matrix into an amplitude encoder to generate a first training bit stream; inputting the first training bit stream into an amplitude decoder to generate an amplitude reference matrix; Calculating a first loss value between the amplitude reference matrix and the amplitude training matrix, and updating the amplitude encoder and the amplitude decoder according to the first loss value until a preset first training end condition is met, thereby obtaining an amplitude autoencoder corresponding to the amplitude correlation interval; and / or, Inputting the phase training matrix into a phase encoder to generate a second training bit stream; Inputting the second training bit stream into a phase decoder to generate a phase reference matrix; A second loss value between the phase reference matrix and the phase training matrix is ​​calculated, and the phase encoder and the phase decoder are updated according to the second loss value until a preset second training end condition is met, thereby obtaining a phase autoencoder corresponding to the amplitude correlation interval.

26. The network device according to claim 25, characterized in that The first loss value is represented by a mean square error between the amplitude reference matrix and the amplitude training matrix, and the second loss value is represented by a cosine similarity between the phase reference matrix and the phase training matrix.

27. The network device according to claim 25, characterized in that The processor is further configured to: The uplink channel information corresponding to the amplitude reference matrix, the phase reference matrix and the downlink channel training matrix is ​​input into a downlink channel recovery model, and the downlink channel recovery model corresponding to the amplitude correlation interval is obtained by training with the goal of minimizing the downlink channel training matrix error output by the downlink channel recovery model.

28. A network device, characterized in that: include: A processor, a memory, and a program stored in the memory and executable on the processor, wherein when the program is executed by the processor, the steps of the method according to any one of claims 6 to 13 are implemented.

29. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 13 are implemented.

Images