Data processing method, apparatus, terminal, network device and storage medium

Abstract

Provided in the present disclosure are a data processing method, an apparatus, a terminal, a network device and a storage medium. The method comprises: receiving first configuration information of a codebook sent by a network device; obtaining a first codebook on the basis of the first configuration information; determining a second codebook on the basis of the first codebook, the second codebook being a codebook which corresponds to the first codebook and which has not been subjected to decomposition calculation; and on the basis of the first codebook and the second codebook, sending first indication information to the network device, the first indication information being used for indicating whether to use a feedback mode of an AI compression model.

Description

Data processing method, device, terminal, network equipment and storage medium

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims priority to Chinese patent application No. 202311757516.6 filed in China on December 20, 2023, the entire contents of which are incorporated herein by reference. Technical Field

[0003] The present disclosure relates to the field of mobile communication technology, and in particular to a data processing method, apparatus, terminal, network equipment, and storage medium. Background Art

[0004] In the New Radio (NR) system, the terminal mainly relies on the codebook to feedback channel state information (CSI). Currently, it supports CSI type (type) I, type II, enhanced type II (type II enhanced, etype II) and other codebook types for feedback of rank indicator (RI), precoding matrix indicator (PMI), channel quality indicator (CQI) and other information.

[0005] The feedback overhead of a Type II codebook is proportional to the proportion of non-zero coefficients and the number of required quantization parameters. Directly extending a low-rank codebook to a high-rank codebook will significantly increase the overhead. To reduce the high-rank codebook overhead, existing solutions extend the rank without increasing the feedback signaling overhead. The sum of non-zero coefficients reported by all layers cannot be greater than a fixed constant configured by a higher layer. This limits the number of bits for non-zero coefficients. While this controls the codebook feedback overhead, it also reduces the accuracy of codebook feedback. Summary of the Invention

[0006] At least one embodiment of the present disclosure provides a data processing method, apparatus, terminal, network device, and storage medium, for solving the problem in related technologies of reducing codebook feedback accuracy in order to reduce codebook overhead for high Rank.

[0007] In order to solve the above technical problems, the present disclosure is implemented as follows:

[0008] In a first aspect, an embodiment of the present disclosure provides a data processing method, applied to a terminal, comprising:

[0009] receiving first configuration information of a codebook sent by a network device;

[0010] Obtaining a first codebook according to the first configuration information;

[0011] Determine a second codebook based on the first codebook; the second codebook is a codebook corresponding to the first codebook that has not been decomposed and calculated;

[0012] Sending first indication information to the network device according to the first codebook and the second codebook;

[0013] Among them, the first indication information is used to indicate whether to adopt the feedback method of the AI ​​compression model.

[0014] Furthermore, the first configuration information includes:

[0015] The number of oversampled DFT beams, the number of DFT basis vectors, and the number of configurable precoding matrices per subband.

[0016] Further, obtaining a first codebook according to the first configuration information includes:

[0017] Decomposing and calculating the first configuration information according to a preset codebook format to obtain decomposed data;

[0018] The decomposed data is encoded to obtain the first codebook.

[0019] Further, sending first indication information to the network device according to the first codebook and the second codebook includes:

[0020] determining a first coefficient matrix of the first codebook according to the first codebook;

[0021] determining a second coefficient matrix of the second codebook according to the second codebook;

[0022] calculating a difference between the first coefficient matrix and the second coefficient matrix;

[0023] Comparing the difference with a preset threshold to obtain a comparison result;

[0024] According to the comparison result, the first indication information is sent to the network device.

[0025] Furthermore, the sending the first indication information to the network device according to the comparison result includes:

[0026] When the difference is greater than the preset threshold, first indication information is sent to the network device to indicate the use of the AI ​​compression model feedback method.

[0027] Further, after sending the first indication information to the network device according to the first codebook and the second codebook, the method further includes:

[0028] When the first indication information indicates adopting an AI compression model feedback mode, receiving second configuration information and second indication information of the codebook, where the second indication information is used to indicate a target compression model for compressing a coefficient matrix of the codebook;

[0029] Determine a third codebook according to the second configuration information;

[0030] Determine a third coefficient matrix of the third codebook according to the third codebook;

[0031] compressing the third coefficient matrix using the target compression model to obtain a fourth coefficient matrix;

[0032] The fourth coefficient matrix is ​​sent to the network device.

[0033] Furthermore, the second indication information includes:

[0034] used to indicate target configuration parameters of the target compression model;

[0035] The target configuration parameter includes at least one of the following: the number of DFT basis vectors, the number of oversampled DFT beams, the number of configurable precoding matrices for each subband, and the model number.

[0036] Furthermore, the method further comprises:

[0037] Sending the first parameter and the second parameter of the third codebook to the network device;

[0038] The first parameter is used to report the beam group; the second parameter includes a DFT vector used for frequency domain compression.

[0039] In a second aspect, an embodiment of the present disclosure provides a data processing method, applied to a network device, comprising:

[0040] Determining first configuration information of a codebook;

[0041] Sending the first configuration information to a terminal, so that the terminal obtains a first codebook according to the first configuration information, and determines a second codebook according to the first codebook; wherein the second codebook is a codebook corresponding to the first codebook that has not been decomposed and calculated;

[0042] receiving first indication information sent by the terminal according to the first codebook and the second codebook;

[0043] Among them, the first indication information is used to indicate whether to adopt the feedback method of the AI ​​compression model.

[0044] Furthermore, after receiving the first indication information sent by the terminal, the method further includes:

[0045] Upon receiving first indication information indicating the use of an AI compression model feedback mode, determining second configuration information of a codebook and second indication information indicating a target compression model;

[0046] Send the second configuration information and the second indication information to the terminal.

[0047] Furthermore, after sending the second configuration information and the second indication information to the terminal, the method further includes:

[0048] receiving a fourth coefficient matrix sent by the terminal; wherein the fourth coefficient matrix is ​​obtained by the terminal compressing a third coefficient matrix of a third codebook determined according to the second configuration information using the target compression model;

[0049] Decompressing the fourth coefficient matrix using a target decompression model to obtain a fifth coefficient matrix;

[0050] A codebook for a channel is determined according to the fifth coefficient matrix.

[0051] Furthermore, after sending the second configuration information and the second indication information to the terminal, the method further includes:

[0052] receiving a first parameter and a second parameter of the third codebook sent by the terminal;

[0053] The determining a codebook of a channel according to the fifth coefficient matrix includes:

[0054] Determining a codebook for a channel according to the fifth coefficient matrix, the first parameter, and the second parameter;

[0055] The first parameter is used to report the beam group; the second parameter includes a DFT vector used for frequency domain compression.

[0056] Furthermore, the method further comprises:

[0057] Collect downlink channel estimation data;

[0058] Determining optional configuration data of the network device based on the downlink channel estimation data; the configuration data includes: the number of DFT basis vectors, the number of oversampled DFT beams, the number of configurable precoding matrices for each subband, and the number of frequency basis vectors;

[0059] Decomposing a codebook corresponding to the downlink channel estimation data according to the configuration data to obtain a sixth coefficient matrix;

[0060] Grouping the sixth coefficient matrix according to the number of oversampled DFT beams and the number of frequency basis vectors in the sixth coefficient matrix to obtain multiple groups of target data;

[0061] Determining a compression model and a decompression model corresponding to each group of target data;

[0062] The compression model and the decompression model corresponding to each group of the target data are trained using the target data to obtain a model database; the model database is used to store the corresponding relationship between the compression model and the model configuration parameters and the corresponding relationship between the decompression model and the compression model;

[0063] The model database is sent to the terminal.

[0064] Furthermore, the method further comprises:

[0065] Determining a target number of elements of samples for training the compression model and the decompression model according to the number of elements of the sixth coefficient matrix;

[0066] If the number of elements of the sixth coefficient matrix is ​​less than the target number of elements, adjusting the target data by zero padding to obtain target sample data;

[0067] Training the compression model and the decompression model using the target sample data;

[0068] The number of the sixth coefficient matrix is ​​the number of elements corresponding to the matrix composed of the number of the oversampled DFT beams and the number of the frequency basis vectors.

[0069] In a third aspect, an embodiment of the present disclosure provides a data processing device, including:

[0070] A first receiving module, configured to receive first configuration information of a codebook sent by a network device;

[0071] A first determining module, configured to obtain a first codebook according to the first configuration information;

[0072] A second determining module is configured to determine a second codebook based on the first codebook; the second codebook is a codebook corresponding to the first codebook that has not been decomposed and calculated;

[0073] A first sending module, configured to send first indication information to the network device according to the first codebook and the second codebook;

[0074] Among them, the first indication information is used to indicate whether to adopt the feedback method of the AI ​​compression model.

[0075] In a fourth aspect, an embodiment of the present disclosure provides a data processing device, including:

[0076] A third determining module, configured to determine first configuration information of a codebook;

[0077] a second sending module, configured to send the first configuration information to a terminal, so that the terminal obtains a first codebook according to the first configuration information, and determines a second codebook according to the first codebook; wherein the second codebook is a codebook corresponding to the first codebook that has not been decomposed and calculated;

[0078] A second receiving module is configured to receive first indication information sent by the terminal according to the first codebook and the second codebook;

[0079] Among them, the first indication information is used to indicate whether to adopt the feedback method of the AI ​​compression model.

[0080] In a fifth aspect, an embodiment of the present disclosure provides a network device, including a transceiver and a processor, wherein:

[0081] The processor is configured to determine first configuration information of a codebook;

[0082] The transceiver is configured to send the first configuration information to a terminal, so that the terminal obtains a first codebook according to the first configuration information, and determines a second codebook according to the first codebook; wherein the second codebook is a codebook corresponding to the first codebook that has not been decomposed and calculated;

[0083] receiving first indication information sent by the terminal according to the first codebook and the second codebook;

[0084] Among them, the first indication information is used to indicate whether to adopt the feedback method of the AI ​​compression model.

[0085] In a sixth aspect, an embodiment of the present disclosure provides a terminal, including a transceiver and a processor, wherein:

[0086] receiving first configuration information of a codebook sent by a network device;

[0087] Obtaining a first codebook according to the first configuration information;

[0088] Determine a second codebook based on the first codebook; the second codebook is a codebook corresponding to the first codebook that has not been decomposed and calculated;

[0089] Sending first indication information to the network device according to the first codebook and the second codebook;

[0090] Among them, the first indication information is used to indicate whether to adopt the feedback method of the AI ​​compression model.

[0091] In a seventh aspect, an embodiment of the present disclosure 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.

[0092] In an eighth aspect, an embodiment of the present disclosure 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.

[0093] In a ninth aspect, an embodiment of the present disclosure provides a computer-readable storage medium having a program stored thereon, and when the program is executed by a processor, the steps of the method described above are implemented.

[0094] In a tenth aspect, an embodiment of the present disclosure provides a computer program product, comprising computer instructions, which, when executed by a processor, implement the steps of the method described above.

[0095] Compared to related technologies, the data processing method, apparatus, terminal, network device, and storage medium provided in the embodiments of the present disclosure can determine a first codebook by receiving first codebook configuration information; recover the first codebook to obtain a second codebook; and thereby send first indication information to the network device, indicating whether to adopt an AI compression model feedback method, based on the first and second codebooks. The solution of the present disclosure, through the first indication information, can instruct the network device to adopt the AI ​​compression model feedback method, thereby ensuring the accuracy of codebook feedback while reducing codebook overhead. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present disclosure. The same reference symbols are used throughout the drawings to represent the same components. In the drawings:

[0097] FIG1 is a schematic diagram of an application scenario of an embodiment of the present disclosure;

[0098] FIG2 is a schematic diagram of a general decomposition form of a precoding matrix according to an embodiment of the present disclosure;

[0099] FIG3 is a flow chart of the data processing method according to an embodiment of the present disclosure when applied to a terminal side;

[0100] FIG4 is a flow chart of a data processing method according to an embodiment of the present disclosure;

[0101] FIG5 is a flow chart of the data processing method according to an embodiment of the present disclosure when applied to a network device side;

[0102] FIG6 is a schematic structural diagram of a data processing device according to an embodiment of the present disclosure;

[0103] FIG7 is a schematic structural diagram of a data processing device according to another embodiment of the present disclosure;

[0104] FIG8 is a schematic structural diagram of a network device according to an embodiment of the present disclosure;

[0105] FIG9 is a schematic structural diagram of a terminal according to an embodiment of the present disclosure;

[0106] FIG10 is a schematic structural diagram of a terminal according to another embodiment of the present disclosure;

[0107] FIG11 is a schematic structural diagram of a network device according to another embodiment of the present disclosure. DETAILED DESCRIPTION

[0108] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0109] The terms "first", "second", etc. in the specification and claims of the present disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily 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 devices. "And / or" in the specification and claims represents at least one of the connected objects.

[0110] 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 in 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 and 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), Institute of Electrical and Electronics Engineers (IEEE) 802.21 Wireless Fidelity (Wi-Fi), IEEE 802.16 World Interoperability for Microwave Access (WiMAX), IEEE 802.20, and Flash-Orthogonal Frequency Division Multiplexing (Flash-OFDM).UTRA and E-UTRA are parts of the Universal Mobile Telecommunications System (UMTS). LTE and more advanced LTE (such as LTE-A) are new versions of UMTS that use E-UTRA. UTRA, E-UTRA, UMTS, LTE, LTE-A, and GSM are described in documents from an organization called the 3rd Generation Partnership Project (3GPP). CDMA2000 and UMB are described in documents from an organization called the 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 NR terminology is used in most of the following description, although these techniques may also be applicable to applications other than NR system applications.

[0111] 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 function and arrangement of the elements discussed without departing from the spirit and scope of this disclosure. The various examples may appropriately omit, substitute, 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.

[0112] Please refer to Figure 1, which shows a block diagram of a wireless communication system applicable to embodiments of the present disclosure. The wireless communication system includes a terminal 11 and a network device 12. Terminal 11 may also be referred to as a user terminal or user equipment (UE). Terminal 11 may be a terminal-side device such as a mobile phone, tablet personal computer, laptop computer, personal digital assistant (PDA), mobile internet device (MID), wearable device, or vehicle-mounted device. It should be noted that the specific type of terminal 11 is not limited in the embodiments of the present disclosure. 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 the fifth generation mobile communication technology (5G) and later versions (for example: the next generation base station (gNB), 5G NR node B (NodeB, NB), etc.), or a base station in other communication systems (for example: eNB, wireless local area network (WLAN) access point, or other access point, etc.), wherein the base station can be called node B, evolved node B, access point, base transceiver station (Base Transceiver Station, BTS), radio base station, radio transceiver, basic service set (BSS), extended service set (ESS), B node, evolved node B (Evolved Node B, eNB), home B node, home evolved B node, WLAN access point, WiFi node or other appropriate terms 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 this disclosure, only the base station in the NR system is taken as an example, but the specific type of the base station is not limited.

[0113] The base station can communicate with the terminal 11 under the control of a base station controller, which in various examples 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 operation on multiple carriers (waveform signals of different frequencies). A multi-carrier transmitter can transmit modulated signals on these multiple carriers simultaneously. 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.

[0114] 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 ​​an access point can be divided into sectors that constitute only a portion of the coverage area. A wireless communication system may include different types of base stations (e.g., macro base stations, micro base stations, or pico base stations). The base stations may also utilize different radio technologies, such as cellular or WLAN radio access technologies. The base stations 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.

[0115] A communication link in a wireless communication system may include an uplink for carrying uplink (UL) transmissions (e.g., from terminal 11 to network device 12), or a downlink for carrying downlink (DL) transmissions (e.g., from network device 12 to terminal 11). UL transmissions may also be referred to as reverse link transmissions, while DL transmissions may also be referred to as forward link transmissions. Downlink transmissions may be performed using a licensed frequency band, an unlicensed frequency band, or both. Similarly, uplink transmissions may be performed using a licensed frequency band, an unlicensed frequency band, or both.

[0116] In the embodiment of the present disclosure, the number of transmitting antennas on the base station side is N, the number of receiving antennas on the user side is N1, and the Rank value RI of the channel matrix H (whose dimension is N1×N) is calculated. The number of subbands is N sb , the number of precoding matrices is N3, N3=N sb×R, R is configured by the high-level parameter number Of PMI Subbands PerCQI Subband-r16 and takes the value {1,2}, which indicates the number of precoding matrices that can be configured for each subband. The channel matrix corresponding to the frequency f of the physical resource block (PRB) selected in the subband is denoted as N3. Perform eigenvalue decomposition (superscript H indicates conjugate transpose), record the first RI eigenvalues ​​and sort them from large to small, and their corresponding eigenvectors are recorded as V f1 ,V f2 ,…,V fRI , concatenate the feature vectors corresponding to the lth layer (stream) of all sub-bands to obtain [V 1l ,V 2l ,…,V N3l ], which is recorded as W l . Codebook W (RI) It can be expressed as follows:

[0117] According to the etypeII codebook, W 1 It can be broken down as follows:

[0118] Among them, W1 is used to report the beam group, and the form of W1 is [b0,L,b L-1 ] corresponds to L oversampled Discrete Fourier Transform (DFT) beams, The matrix consists of the DFT vectors used for frequency domain compression.

[0119] Specifically, the general decomposition form of the R16 eType II precoding matrix is ​​shown in Figure 2, where N is the number of Channel State Information-Reference Signal (CSI-RS) ports (corresponding to N antennas) and M is the number of frequency basis vectors.

[0120] The current eType II codebook transmission method requires quantizing the coefficient matrix by mapping it to a low-precision quantization set and then transmitting it to the base station;

[0121] For different RIs, Recorded as Will W f .

[0122] As shown in FIG3 , an embodiment of the present disclosure provides a data processing method, which is applied to a terminal and includes the following steps:

[0123] Step 301: receiving first configuration information of a codebook sent by a network device;

[0124] Step 302: Obtain a first codebook according to the first configuration information;

[0125] Step 303: Determine a second codebook based on the first codebook; the second codebook is a codebook corresponding to the first codebook that has not been decomposed and calculated;

[0126] Step 304: Send first indication information to the network device according to the first codebook and the second codebook;

[0127] The first indication information is used to indicate whether to adopt an artificial intelligence (AI) compression model feedback method.

[0128] Optionally, the terminal is a terminal that has accessed the network device and has entered a Radio Resource Control (RRC-CONNECTED) state.

[0129] In the embodiment of the present disclosure, the wireless access network AI model training system (deployed in the centralized unit (CU) and / or distributed unit (DU) of the base station, or on a logical entity across CUs) collects downlink channel estimation data, sets the number of oversampled DFT beams and R that the network device can optionally oversample; and decomposes the channel estimation data in the etype II codebook format according to the downlink self-carrying number to obtain unquantized

[0130] Afterwards, a model (auto-encoder) is designed based on the data dimensions and trained offline. The trained compression model, number, and model number table (optional) related to the model features are synchronized to the terminal. The decompression model, number, and model number table (optional) are synchronized to the network equipment (such as a base station).

[0131] Finally, the base station pre-sets the difference threshold δ according to the accuracy requirement of the feedback channel and RI (1≤RI≤RImax) RI ; Among them, the higher the accuracy requirement of the feedback channel, the higher the RI, and the lower the difference threshold.

[0132] In the embodiment of the present disclosure, the network device (such as a base station) broadcasts RI (1≤RI≤RImax) and the corresponding difference threshold δRI Send to the terminal;

[0133] And configure the first configuration information (the number of oversampled DFT beams L, the number of DFT basis vectors P for determining the DFT basis vectors) according to historical information, complexity, performance and other information v and R) sent to the terminal;

[0134] The terminal determines the first codebook according to the first configuration information; the first codebook is obtained by decomposition and calculation, and then recombined in the general form of the precoding matrix as shown in FIG2 ;

[0135] The terminal restores the first codebook into a decomposed and calculated codebook format (the second codebook).

[0136] The data processing method of the disclosed embodiment can determine a first codebook by receiving first codebook configuration information; recover the first codebook to obtain a second codebook; and then send first indication information to a network device based on the first and second codebooks, indicating whether to adopt an AI compression model feedback method. The disclosed solution, through the first indication information, can instruct the network device to adopt the AI ​​compression model feedback method, thereby ensuring the accuracy of codebook feedback while reducing codebook overhead.

[0137] Optionally, the first configuration information includes:

[0138] The number of oversampled DFT beams, the number of DFT basis vectors, and the number of configurable precoding matrices per subband.

[0139] Optionally, obtaining a first codebook according to the first configuration information includes:

[0140] Decomposing and calculating the first configuration information according to a preset codebook format to obtain decomposed data;

[0141] The decomposed data is encoded to obtain the first codebook.

[0142] Optionally, the sending first indication information to the network device according to the first codebook and the second codebook includes:

[0143] determining a first coefficient matrix of the first codebook according to the first codebook;

[0144] determining a second coefficient matrix of the second codebook according to the second codebook;

[0145] calculating a difference between the first coefficient matrix and the second coefficient matrix;

[0146] Comparing the difference with a preset threshold to obtain a comparison result; the preset threshold is configured by the network device and sent to the terminal;

[0147] According to the comparison result, the first indication information is sent to the network device.

[0148] In the embodiment of the present disclosure, the difference between the first coefficient matrix and the second coefficient matrix is ​​calculated using the norm between matrices, specifically:

[0149] Among them, W (RI) represents the first coefficient matrix, W (RI) "" represents the second coefficient matrix, Δ represents the difference between the first coefficient matrix and the second coefficient matrix, and F represents the norm.

[0150] The data processing method of the embodiment of the present disclosure can determine the impact of the decomposition calculation on the integrity of the codebook by comparing the difference between the coefficient matrices of the codebook before and after decomposition calculation, thereby determining the feedback method of the terminal, and can reduce the codebook overhead while ensuring the accuracy of the codebook feedback.

[0151] Optionally, the sending the first indication information to the network device according to the comparison result includes:

[0152] When the difference is greater than the preset threshold, first indication information is sent to the network device to indicate the use of the AI ​​compression model feedback method.

[0153] Optionally, the threshold is sent by the network device to the terminal.

[0154] Optionally, the format of the first indication information is a flag bit.

[0155] In an embodiment of the present disclosure, when the first indication information is used to indicate the use of an AI compression model feedback method, the flag bit is 1; when the first indication information is used to indicate the use of an etype II codebook feedback method, the flag bit is 0.

[0156] In the embodiment of the present disclosure, the difference is greater than the preset threshold (difference threshold δ RI ) indicates that the decomposition calculation has a significant impact on the integrity of the codebook, and the use of the feedback method of the etype II codebook will result in insufficient feedback accuracy; therefore, it is necessary to send the first indication information indicating the use of the AI ​​compression model feedback method to the network device;

[0157] If the difference is smaller than the preset threshold, the feedback mode of the etype II codebook is adopted.

[0158] Optionally, after sending the first indication information to the network device according to the first codebook and the second codebook, the method further includes:

[0159] When the first indication information indicates adopting an AI compression model feedback mode, receiving second configuration information and second indication information of the codebook, where the second indication information is used to indicate a target compression model for compressing a coefficient matrix of the codebook;

[0160] Determine a third codebook according to the second configuration information;

[0161] Determine a third coefficient matrix of the third codebook according to the third codebook;

[0162] compressing the third coefficient matrix using the target compression model to obtain a fourth coefficient matrix;

[0163] The fourth coefficient matrix is ​​sent to the network device.

[0164] Optionally, sending the fourth coefficient matrix to the network device includes:

[0165] Sending the fourth coefficient matrix on a first resource;

[0166] The first resource is determined according to the resource information in the second indication information.

[0167] Optionally, the determining a third codebook according to the second configuration information includes:

[0168] The second configuration information is calculated according to a preset codebook format to obtain the third codebook.

[0169] In the embodiment of the present disclosure, when the first indication information is used to indicate the use of the AI ​​compression model feedback method, the network setting is based on the received first indication information, and reselects the configuration information according to historical experience and other information to determine the second configuration information (L, the number of oversampled DFT beams P v , R, and time-frequency resources (e.g., resource elements (RE)) required for the terminal to feedback CSI compression feedback information; and sending the second configuration information and configuration information of the compression feedback model (the second indication information) to the terminal through a downlink channel;

[0170] The terminal measures the CSI-RS of the downlink channel reference signal, and calculates the unquantized third codebook according to the received second configuration information in the etype II codebook format;

[0171] A channel compression model is selected according to the calculation model number in the second configuration information, and the third codebook is compressed using the channel compression model.

[0172] In the embodiment of the present disclosure, compressing the third codebook includes: compressing the third coefficient matrix of the third codebook Compress and get the fourth coefficient matrix

[0173] The first resource is determined according to resource configuration, and the fourth coefficient matrix is ​​sent to the network device on the first resource.

[0174] The data processing method of the embodiment of the present disclosure can reduce data overhead by compressing the third coefficient matrix and then sending it to the network device.

[0175] Optionally, the second indication information includes:

[0176] used to indicate target configuration parameters of the target compression model;

[0177] The target configuration parameter includes at least one of the following: the number of DFT basis vectors, the number of oversampled DFT beams, the number of configurable precoding matrices for each subband, and the model number.

[0178] Optionally, the second indication information further includes:

[0179] The terminal is configured to send resource information of the fourth coefficient matrix;

[0180] The resource information includes: time-frequency resources.

[0181] The data processing method of the embodiment of the present disclosure determines the third codebook through the received second configuration information, and compresses the third codebook according to the received second indication information; so that the network device parses the third codebook according to the compressed third codebook to obtain all data of the third codebook, thereby ensuring the integrity and accuracy of the third codebook data.

[0182] In the embodiment of the present disclosure, compressing the third codebook includes: compressing the third coefficient matrix of the third codebook Compress and get the fourth coefficient matrix

[0183] The first resource is determined according to resource configuration, and the fourth coefficient matrix is ​​sent to the network device on the first resource.

[0184] The data processing method of the embodiment of the present disclosure can reduce data overhead by compressing the third coefficient matrix and then sending it to the network device.

[0185] Optionally, the method further includes:

[0186] Sending the first parameter and the second parameter of the third codebook to the network device;

[0187] The first parameter is used to report the beam group; the second parameter includes a DFT vector used for frequency domain compression.

[0188] Optionally, the sending the first parameter and the second parameter of the third codebook to the network device includes:

[0189] The first parameter and the second parameter are fed back to the network device through the etype II codebook feedback mode.

[0190] In the embodiment of the present disclosure, the first parameter W1(t) and the second parameter W are fed back to the network device through the etype II codebook feedback mode. f (t).

[0191] The data processing method of the embodiment of the present disclosure can reduce data overhead by compressing the third coefficient matrix and sending it to the network device; and by sending the first parameter and the second parameter to the network device, the network device parses the third codebook based on the fourth coefficient matrix, the first parameter, and the second parameter, thereby obtaining all data information of the third codebook.

[0192] In the embodiment of the present disclosure, the network device receives the fourth coefficient matrix Then, the fourth coefficient matrix is ​​decompressed by the decompression model Decompress and get the fifth coefficient matrix And according to the first parameter W1(t) and the second parameter W f (t) for the fifth coefficient matrix Perform additional calculations to obtain the codebook W (RI) (T);

[0193] Subsequent terminals continue to monitor CSI-RS;

[0194] The network equipment uses the codebook W (RI) (T) (mainly considering the complexity and performance of the codebook) to adjust L, P v and R, and configure the terminal through high-layer signaling (such as Radio Resource Control (RRC)), and adjust the model configuration information of the compression feedback model to the terminal;

[0195] The terminal adjusts the calculation of the codebook and the selection of the model according to the model configuration information adjusted by the network device, thereby completing the compression feedback of the channel.

[0196] As shown in FIG4 , the data processing method of the embodiment of the present disclosure is as follows:

[0197] 1. Collect downlink channel data and calculate the unquantized Based on Offline training of AI-based channel compression / decompression model, synchronization of model and model number table to base station and terminal; at the same time, base station pre-configures difference threshold δ RI ;

[0198] 2. The terminal has connected to the network and entered the RRC-CONNECTED state;

[0199] 3. Base station sending difference threshold δ RI To the terminal, and configure to send CSI-RS, L, P v and R to the terminal;

[0200] 4. The terminal measures CSI-RS, calculates the etype II codebook and recovers it to W (RI) ″, calculate its difference with W (RI) The difference Δ, compare Δ with the difference threshold δ RI ; and feedback the flag bit to the base station, if Δ is different from the difference threshold δ RI , then go to step 5, otherwise go to step 9;

[0201] 5. The base station reselects L and P according to the flag bit fed back by the terminal. v And R, and configure the time-frequency resources and model configuration information fed back by the terminal to the terminal;

[0202] 6. Terminal measurement calculation is quantitative And determine the compression model pair based on the model configuration information Compress to get Feedback to base station W1(t) and W f (t);

[0203] 7. Base station pair Decompress and calculate according to W1(t) and W f (t) Calculate the etype II codebook;

[0204] 8. The terminal continuously monitors CSI-RS and modulates L and P according to the feedback codebook. v and R, thereby adjusting the model configuration information; periodically testing W (RI) ″ and W (RI)The difference is calculated and the flag is fed back to the base station. If the base station determines that the etype II codebook needs to be used instead, the resources are reallocated and the process goes to step 9. Otherwise, the process goes to step 8.

[0205] 9. The base station configures the terminal to feedback resources for the eType II codebook. The terminal feedbacks the eType II codebook according to the configuration of the base station and then proceeds to capture 4.

[0206] As shown in FIG5 , an embodiment of the present disclosure further provides a data processing method, which is applied to a network device and includes the following steps:

[0207] Step 501: Determine first configuration information of a codebook;

[0208] Step 502: Send the first configuration information to a terminal, so that the terminal obtains a first codebook according to the first configuration information, and determines a second codebook based on the first codebook, wherein the second codebook is a codebook corresponding to the first codebook without undergoing decomposition calculation.

[0209] Step 503: Receive first indication information sent by the terminal according to the first codebook and the second codebook;

[0210] Among them, the first indication information is used to indicate whether to adopt the feedback method of the AI ​​compression model.

[0211] The data processing method of the embodiment of the present disclosure determines the first configuration information of the codebook through the network device, so that the terminal can determine the first indication information based on the first configuration information, so that the network device determines whether to adopt the feedback method of the AI ​​compression model based on the first configuration information, thereby ensuring the accuracy of the codebook feedback while reducing the codebook overhead.

[0212] Optionally, after receiving the first indication information sent by the terminal, the method further includes:

[0213] Upon receiving first indication information indicating the use of an AI compression model feedback mode, determining second configuration information of a codebook and second indication information indicating a target compression model;

[0214] Send the second configuration information and the second indication information to the terminal.

[0215] Optionally, the second indication information includes:

[0216] Configuration parameters and resource information of the target compression model;

[0217] The configuration parameters of the target compression model include at least one of the following: the number of DFT basis vectors, the number of oversampled DFT beams, the number of configurable precoding matrices for each subband, and the model number.

[0218] The data processing method of the embodiment of the present disclosure, when determining that a feedback method of an AI compression model needs to be adopted, adjusts the configuration information of the codebook to obtain the second configuration information, and the configuration parameters of the target compression model for compressing the coefficient matrix of the codebook, so that the terminal can determine the third coefficient matrix of the codebook according to the second configuration information, and determine the target model for compressing the third coefficient matrix through the configuration parameters of the target compression model.

[0219] Optionally, after sending the second configuration information and the second indication information to the terminal, the method further includes:

[0220] receiving a fourth coefficient matrix sent by the terminal; wherein the fourth coefficient matrix is ​​obtained by the terminal compressing a third coefficient matrix of a third codebook determined according to the second configuration information using the target compression model;

[0221] Decompressing the fourth coefficient matrix using a target decompression model to obtain a fifth coefficient matrix;

[0222] A codebook for a channel is determined according to the fifth coefficient matrix.

[0223] It should be noted that the target decompression model is a model corresponding to the target compression model.

[0224] Optionally, receiving a fourth coefficient matrix sent by the terminal includes:

[0225] Receiving the fourth coefficient matrix on a first resource; the first resource is determined according to the second indication information;

[0226] The first resources include: time-frequency resources.

[0227] The data processing method of the embodiment of the present disclosure can obtain the fifth coefficient matrix by decompressing the fourth coefficient matrix, and thus determine the codebook based on the fifth coefficient matrix, thereby reducing codebook overhead while ensuring the accuracy of codebook feedback.

[0228] Optionally, after sending the second configuration information and the second indication information to the terminal, the method further includes:

[0229] receiving a first parameter and a second parameter of a third codebook sent by the terminal;

[0230] The determining a channel codebook according to the fifth coefficient matrix includes:

[0231] Determining a codebook for a channel according to the fifth coefficient matrix, the first parameter, and the second parameter;

[0232] The first parameter is used to report the beam group; the second parameter includes a DFT vector used for frequency domain compression.

[0233] The data processing method of the embodiment of the present disclosure can restore all the data of the codebook through the fifth coefficient matrix, the first parameter, and the second parameter, thereby reducing the codebook overhead while ensuring the accuracy of codebook feedback.

[0234] Optionally, the method further includes:

[0235] Collect downlink channel estimation data;

[0236] Determining optional configuration data of the network device based on the downlink channel estimation data; the configuration data includes: the number of DFT basis vectors, the number of oversampled DFT beams, the number of configurable precoding matrices for each subband, and the number of frequency basis vectors;

[0237] Decomposing a codebook corresponding to the downlink channel estimation data according to the configuration data to obtain a sixth coefficient matrix;

[0238] Grouping the sixth coefficient matrix according to the number of oversampled DFT beams and the number of frequency basis vectors in the sixth coefficient matrix to obtain multiple groups of target data;

[0239] Determining a compression model and a decompression model corresponding to each group of target data;

[0240] The compression model and the decompression model corresponding to each group of the target data are trained using the target data to obtain a model database; the model database is used to store the corresponding relationship between the compression model and the model configuration parameters and the corresponding relationship between the decompression model and the compression model;

[0241] The model database is sent to the terminal.

[0242] Optionally, the sixth coefficient matrix is ​​unquantized channel data obtained by calculating the codebook of the uplink channel according to a preset codebook format.

[0243] In one embodiment of the present disclosure, the sixth coefficient matrix is ​​calculated according to the etype II codebook format and is not quantized.

[0244] In the embodiment of the present disclosure, the wireless access network AI model training system (deployed on the CU and / or DU of the base station, or on a logical entity across CUs) collects downlink channel estimation data, sets the number of oversampled DFT beams and R that the network device can optionally oversample; and decomposes the channel estimation data in the etype II codebook format according to the downlink self-carrying number to obtain unquantized

[0245] Afterwards, a model (auto-encoder) is designed based on the data dimension and trained offline. The trained compression model, number, and model number table related to the model features (optional) are synchronized to the terminal, and the decompression model, number, and model number table (optional) are synchronized to the network device (such as a base station).

[0246] Optionally, the method further includes:

[0247] Determining a target number of elements of samples for training the compression model and the decompression model according to the number of elements of the sixth coefficient matrix;

[0248] If the number of elements of the sixth coefficient matrix is ​​less than the target number of elements, adjusting the target data by padding with zeros to obtain target sample data;

[0249] Training the compression model and the decompression model using the target sample data;

[0250] The number of the sixth coefficient matrix is ​​the number of elements corresponding to the matrix composed of the number of the oversampled DFT beams and the number of the frequency basis vectors.

[0251] In the data processing method of the embodiment of the present disclosure, the compression model and the decompression model are both models whose number of input elements can be generalized, and model training can be performed for different configuration parameters, thereby improving the applicability of the model.

[0252] In the embodiment of the present disclosure, the training method of the model may be:

[0253] First, based on the collected downlink channel samples, the base station's optional L, P v And R,M=P v ×N3×R, assuming that the data is grouped and the model is designed based on the difference of L×M, the input of the model is 2×L×M, forming a table as shown in Table 1:

[0254] Table 1 Model number table corresponding to L×M

[0255] During the model training process, it is necessary to Elements in Obtained by compression / decompression model in sequence Compression / decompression process RI times, the obtained Put them together and get That is W (RI) ′.

[0256] In the embodiment of the present disclosure, the training method of the model can also be: according to the collected samples Chinese elements The dimension is 2×L×M. The largest 2×L×M value is selected as [2×L×M]max as the number of input elements to design the AI ​​model. When the model is trained / inferred, when the sample Elements in When the dimension is 2×L×M<[2×L×M]max, After expanding into a vector, it is padded with zeros to extend its length to [2×L×M]max. Then, it passes through the compression / decompression model in sequence. Model training / inference requires RI times in total. Based on the above design, this model is generalizable to the number of input elements, so only one is required.

[0257] The wireless access network AI model training system (deployed on the CU and / or DU of the base station, or on the logical entity across CUs) collects downlink channel estimation data and sets the base station's optional L and P v And R, according to different L, P v , R and the number of downlink subbands to decompose the data in the etype II codebook format to obtain the unquantized Design a model (auto-encoder) based on the data dimension and perform offline training. Synchronize the trained compression model, number, and Table 1 to the terminal. Synchronize the decompression model, number, and Table 1 to the base station. The base station pre-sets the difference threshold δ based on the accuracy requirements of the feedback channel and RI. RI =0.1;

[0258] The terminal has connected to the network and entered the RRC-CONNECTED state;

[0259] The base station sends (can be broadcast or unicast, multicast) RI (1≤RI≤RImax) and its corresponding δ RI To the terminal, configure the number of downlink subbands to 16, and send CSI-RS to the terminal according to the configuration, and configure L=4, P according to historical information, complexity, performance and other information. v =1 / 8, R=1 and sent to the terminal;

[0260] The terminal measures the downlink channel reference signal CSI-RS and calculates its etype II codebook and RI, and restores the decomposed codebook to the format before decomposition, which is recorded as W (RI) ″, calculate its difference with W(RI) The difference Δ between them is compared with the threshold 0.1. When Δ is greater than 0.1, the flag bit 1 of the AI ​​compression feedback is fed back to the base station;

[0261] If the base station receives the flag bit of AI compression feedback, it will reselect L, P according to historical experience and other information. v And R, such as the following configuration: L = 4, P v =1 / 8, R=1, thus calculating M=2, according to M=2, L=4, select the corresponding model number 3 in Table 1, configure the time-frequency resources (such as the number of REs) required for the terminal to feedback the CSI compression feedback information, and send the compression feedback model configuration information (including the reselected L, P v , R and resources) are sent to the terminal through the Physical Downlink Control Channel (PDCCH) / Physical Downlink Shared Channel (PDSCH);

[0262] The terminal measures the downlink channel reference signal CSI-RS and calculates the value of the CSI-RS according to the P in the model configuration information. v , L, R are calculated according to the etypeII codebook to obtain the unquantized The terminal selects the channel compression model number 3 according to the received channel compression feedback model configuration information and completes Compression, the compressed channel information is recorded as And based on resource allocation feedback To the base station, W1(t) and W are fed back simultaneously according to the etype II codebook feedback mode. f (t);

[0263] The base station receives W1(t) fed back by the terminal, and W f (t), will By decompressing the model pass Calculate W (RI) (T).

[0264] The wireless access network AI model training system (deployed on the CU and / or DU of the base station, or on the logical entity across CUs) collects downlink channel estimation data and sets the base station's optional L and P v and R, according to L, P v , R and the number of downlink subbands to decompose the data into the etype II codebook format and calculate the unquantized Record [2×L×M]max, and use [2×L×M]max as the number of model input elements to design the model (auto-encoder), and at the same time, The samples with dimension 2×L×M<[2×L×M]max are padded with zeros to form new samples, and the model is trained offline. The trained compression model is synchronized to the terminal, and the decompression model is synchronized to the base station. The base station pre-sets the difference threshold δ according to the accuracy requirements of the feedback channel and RI. RI =0.1;

[0265] The terminal has connected to the network and entered the RRC-CONNECTED state;

[0266] The base station sends (can be broadcast or unicast, multicast) RI (1≤RI≤RImax) and its corresponding δ RI To the terminal, configure the number of downlink subbands to 16, and send CSI-RS to the terminal according to the configuration, and configure L=4, P according to historical information, complexity, performance and other information. v =1 / 8, R=1 and sent to the terminal;

[0267] The terminal measures the downlink channel reference signal CSI-RS and calculates its etype II codebook and RI, and restores the decomposed codebook to the format before decomposition, which is recorded as W (RI) ″, calculate its difference with W (RI) The difference Δ between them is compared with the threshold 0.1. When Δ is greater than 0.1, the flag bit 1 of the AI ​​compression feedback is fed back to the base station;

[0268] If the base station receives the flag bit of AI compression feedback, it will reselect L, P according to historical experience and other information. v And R, such as the following configuration: L = 4, P v =1 / 8, R=1, configure the time-frequency resources (such as the number of REs) required for the terminal to feedback the CSI compression feedback information, and compress the feedback model configuration information (including the reselected L, P v , R and resources) are sent to the terminal through PDCCH / PDSCH;

[0269] The terminal measures the downlink channel reference signal CSI-RS and calculates the value of the CSI-RS according to the P in the model configuration information. v , L, R are calculated according to the etypeII codebook to obtain the unquantized The elements The samples with dimension 2×L×M<[2×L×M]max are padded with zeros and completed Compression, the compressed channel information is recorded as And based on resource allocation feedback To the base station, W1(t) and W are fed back simultaneously according to the etype II codebook feedback mode.f (t);

[0270] The base station receives W1(t) fed back by the terminal, and W f (t), will By decompressing the model pass Calculate W (RI) (T).

[0271] The above describes various methods of the embodiments of the present disclosure. The following further provides apparatuses for implementing the above methods.

[0272] As shown in FIG6 , the embodiment of the present disclosure further provides a data processing device 600, including:

[0273] A first receiving module 601 is configured to receive first configuration information of a codebook sent by a network device;

[0274] A first determining module 602 is configured to obtain a first codebook according to the first configuration information;

[0275] A second determining module 603 is configured to determine a second codebook based on the first codebook; the second codebook is a codebook corresponding to the first codebook that has not been decomposed and calculated;

[0276] A first sending module 604 is configured to send first indication information to the network device according to the first codebook and the second codebook;

[0277] Among them, the first indication information is used to indicate whether to adopt the feedback method of the AI ​​compression model.

[0278] It should be noted that the device in this embodiment is a device corresponding to the method applied to the terminal described above, and the implementation methods in the above embodiments are all applicable to the embodiments of this device and can achieve the same technical effects. The above-mentioned device provided in the embodiments of the present disclosure can implement all the method steps implemented in the above-mentioned method embodiments and can achieve the same technical effects. The parts and beneficial effects of this embodiment that are the same as those in the method embodiments will not be described in detail here.

[0279] As shown in FIG7 , the embodiment of the present disclosure further provides a data processing device 700, including:

[0280] A third determining module 701 is configured to determine first configuration information of a codebook;

[0281] A second sending module 702 is configured to send the first configuration information to a terminal, so that the terminal obtains a first codebook according to the first configuration information and determines a second codebook based on the first codebook; wherein the second codebook is a codebook corresponding to the first codebook that has not been decomposed and calculated;

[0282] A second receiving module 703 is configured to receive first indication information sent by the terminal according to the first codebook and the second codebook;

[0283] Among them, the first indication information is used to indicate whether to adopt the feedback method of the AI ​​compression model.

[0284] It should be noted that the device in this embodiment is a device corresponding to the method applied to the network side, and the implementation methods in the above embodiments are all applicable to the embodiments of this device and can achieve the same technical effects. The above-mentioned device provided in the embodiments of the present disclosure can implement all the method steps implemented in the above-mentioned method embodiments and can achieve the same technical effects. The parts and beneficial effects of this embodiment that are the same as those in the method embodiments will not be described in detail here.

[0285] Referring to FIG8 , an embodiment of the present disclosure further provides a network device 800 , including a transceiver 810 and a processor 820 , wherein:

[0286] The processor is configured to determine first configuration information of a codebook;

[0287] The transceiver is configured to send the first configuration information to a terminal, so that the terminal obtains a first codebook according to the first configuration information, and determines a second codebook according to the first codebook; wherein the second codebook is a codebook corresponding to the first codebook that has not been decomposed and calculated;

[0288] receiving first indication information sent by the terminal according to the first codebook and the second codebook;

[0289] Among them, the first indication information is used to indicate whether to adopt the feedback method of the AI ​​compression model.

[0290] Referring to FIG9 , an embodiment of the present disclosure further provides a terminal including a transceiver 910 and a processor 920 , wherein:

[0291] receiving first configuration information of a codebook sent by a network device;

[0292] Obtaining a first codebook according to the first configuration information;

[0293] Determine a second codebook based on the first codebook; the second codebook is a codebook corresponding to the first codebook that has not been decomposed and calculated;

[0294] Sending first indication information to the network device according to the first codebook and the second codebook;

[0295] Among them, the first indication information is used to indicate whether to adopt the feedback method of the AI ​​compression model.

[0296] Please refer to Figure 10. The embodiment of the present disclosure also provides a terminal 1000, including a processor 1001, a memory 1002, and a computer program stored in the memory 1002 and capable of running on the processor 1001. When the computer program is executed by the processor 1001, the various processes of the above-mentioned data processing method embodiment executed by the terminal are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0297] Please refer to Figure 11. An embodiment of the present disclosure further provides a network device 1100, including a processor 1101, a memory 1102, and a computer program stored in the memory 1102 and executable on the processor 1101. When the computer program is executed by the processor 1101, the various processes of the above-mentioned data processing method embodiment executed by the network device are implemented, and the same technical effect can be achieved. To avoid repetition, they will not be described here.

[0298] The present disclosure also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the computer program implements the various processes of the above-mentioned data processing method embodiment and can achieve the same technical effect. To avoid repetition, the details are not described here. The computer-readable storage medium is, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0299] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or apparatus comprising the element.

[0300] Through the description of the above embodiments, those skilled in the art can clearly understand that the above embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better embodiment. Based on this understanding, the technical solution of the present disclosure, 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, magnetic disk, optical disk), and includes a number of instructions for enabling a terminal (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present disclosure.

[0301] The embodiments of the present disclosure are described above in conjunction with the accompanying drawings, but the present disclosure 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 disclosure, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present disclosure and the claims, all of which are protected by the present disclosure.

Claims

1. A data processing method, applied to a terminal, comprising: Receiving first configuration information of a codebook sent by a network device; Obtaining a first codebook according to the first configuration information; Determine a second codebook according to the first codebook; The second codebook is a codebook corresponding to the first codebook that has not been decomposed and calculated; Sending first indication information to the network device according to the first codebook and the second codebook; Among them, the first indication information is used to indicate whether to adopt the feedback method of the AI ​​compression model.

2. The data processing method according to claim 1, wherein: The first configuration information includes: The number of oversampled discrete Fourier transform (DFT) beams, the number of DFT basis vectors, and the number of configurable precoding matrices for each subband.

3. The data processing method according to claim 1, wherein: Obtaining a first codebook according to the first configuration information includes: Decomposing and calculating the first configuration information according to a preset codebook format to obtain decomposed data; The decomposed data is encoded to obtain the first codebook.

4. The data processing method according to claim 1, wherein: Sending first indication information to the network device according to the first codebook and the second codebook includes: Determine a first coefficient matrix of the first codebook according to the first codebook; Determine a second coefficient matrix of the second codebook according to the second codebook; Calculating the difference between the first coefficient matrix and the second coefficient matrix; Comparing the difference with a preset threshold to obtain a comparison result; According to the comparison result, the first indication information is sent to the network device.

5. The data processing method according to claim 4, wherein: The sending the first indication information to the network device according to the comparison result includes: When the difference is greater than the preset threshold, first indication information is sent to the network device to indicate the use of an artificial intelligence (AI) compression model feedback method.

6. The data processing method according to claim 1, wherein: After sending first indication information to the network device according to the first codebook and the second codebook, the method further includes: When the first indication information indicates that a feedback mode of an AI compression model is adopted, receiving second configuration information of a codebook and second indication information for indicating a target compression model; Determine a third codebook according to the second configuration information; Determine a third coefficient matrix of the third codebook according to the third codebook; Compressing the third coefficient matrix by using the target compression model to obtain a fourth coefficient matrix; The fourth coefficient matrix is ​​sent to the network device.

7. The data processing method according to claim 6, wherein: The second indication information includes: Used to indicate target configuration parameters of the target compression model; wherein the target configuration parameters include at least one of the following: the number of DFT basis vectors, the number of oversampled DFT beams, the number of configurable precoding matrices for each subband, and the model number.

8. The method according to claim 6, further comprising: Sending the first parameter and the second parameter of the third codebook to the network device; The first parameter is used to report the beam group; the second parameter includes a DFT vector used for frequency domain compression.

9. A data processing method, applied to a network device, the method comprising: Determine first configuration information of a codebook; Sending the first configuration information to a terminal, so that the terminal obtains a first codebook according to the first configuration information, and determines a second codebook according to the first codebook; wherein the second codebook is a codebook corresponding to the first codebook that has not been decomposed and calculated; receiving first indication information sent by the terminal according to the first codebook and the second codebook; Among them, the first indication information is used to indicate whether to adopt the feedback method of the AI ​​compression model.

10. The data processing method according to claim 9, wherein: After the receiving terminal sends the first indication information, the method further includes: Upon receiving first indication information for indicating the use of an AI compression model feedback method, second configuration information of a codebook and second indication information for indicating a target compression model are determined; and the second configuration information and the second indication information are sent to the terminal.

11. The data processing method according to claim 10, wherein: After sending the second configuration information and the second indication information to the terminal, the method further includes: receiving a fourth coefficient matrix sent by the terminal; wherein the fourth coefficient matrix is ​​obtained by the terminal compressing a third coefficient matrix of a third codebook determined according to the second configuration information through the target compression model; Decompressing the fourth coefficient matrix using a target decompression model to obtain a fifth coefficient matrix; A codebook of a channel is determined according to the fifth coefficient matrix.

12. The data processing method according to claim 11, wherein: After sending the second configuration information and the second indication information to the terminal, the method further includes: receiving a first parameter and a second parameter of the third codebook sent by the terminal; The step of determining a codebook of a channel according to the fifth coefficient matrix comprises: Determine a codebook of a channel according to the fifth coefficient matrix, the first parameter, and the second parameter; The first parameter is used to report the beam group; the second parameter includes a DFT vector used for frequency domain compression.

13. The data processing method according to claim 9, further comprising: Collecting downlink channel estimation data; Determining optional configuration data of the network device according to the downlink channel estimation data; The configuration data includes: the number of DFT basis vectors, the number of oversampled DFT beams, the number of configurable precoding matrices for each subband, and the number of frequency basis vectors; Decomposing a codebook corresponding to the downlink channel estimation data according to the configuration data to obtain a sixth coefficient matrix; Grouping the sixth coefficient matrix according to the number of oversampled DFT beams and the number of frequency basis vectors in the sixth coefficient matrix to obtain multiple groups of target data; Determine a compression model and a decompression model corresponding to each group of the target data; The compression model and the decompression model corresponding to each group of the target data are trained through the target data to obtain a model database; the model database is used to store the correspondence between the compression model and the model configuration parameters and the correspondence between the decompression model and the compression model; the model database is sent to the terminal.

14. The method according to claim 13, further comprising: Determining a target number of elements of samples for training the compression model and the decompression model according to the number of elements of the sixth coefficient matrix; If the number of elements of the sixth coefficient matrix is ​​less than the target number of elements, adjusting the target data by padding with zeros to obtain target sample data; Training the compression model and the decompression model using the target sample data; The number of the sixth coefficient matrix is ​​the number of elements corresponding to the matrix composed of the number of the oversampled DFT beams and the number of the frequency basis vectors.

15. A data processing device, comprising: A first receiving module, configured to receive first configuration information of a codebook sent by a network device; A first determining module, configured to obtain a first codebook according to the first configuration information; A second determining module, configured to determine a second codebook according to the first codebook; The second codebook is a codebook corresponding to the first codebook that has not been decomposed and calculated; A first sending module, configured to send first indication information to the network device according to the first codebook and the second codebook; Among them, the first indication information is used to indicate whether to adopt the feedback method of the AI ​​compression model.

16. A data processing device, comprising: A third determining module, used to determine first configuration information of a codebook; A second sending module is configured to send the first configuration information to a terminal, so that the terminal obtains a first codebook according to the first configuration information, and determines a second codebook according to the first codebook; wherein the second codebook is a codebook corresponding to the first codebook that has not been decomposed and calculated; A second receiving module, configured to receive first indication information sent by the terminal according to the first codebook and the second codebook; Among them, the first indication information is used to indicate whether to adopt the feedback method of the AI ​​compression model.

17. A network device comprising a transceiver and a processor, wherein: The processor is configured to determine first configuration information of a codebook; The transceiver is configured to send the first configuration information to a terminal, so that the terminal obtains a first codebook according to the first configuration information, and determines a second codebook according to the first codebook; wherein the second codebook is a codebook corresponding to the first codebook that has not been decomposed and calculated; receiving first indication information sent by the terminal according to the first codebook and the second codebook; Among them, the first indication information is used to indicate whether to adopt the feedback method of the AI ​​compression model.

18. A terminal comprising a transceiver and a processor, wherein: Receiving first configuration information of a codebook sent by a network device; Obtaining a first codebook according to the first configuration information; Determine a second codebook according to the first codebook; the second codebook is a codebook corresponding to the first codebook that has not been decomposed and calculated; Sending first indication information to the network device according to the first codebook and the second codebook; Among them, the first indication information is used to indicate whether to adopt the feedback method of the AI ​​compression model.

19. 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 according to any one of claims 1 to 8.

20. A network device comprising: A processor, a memory, and a program stored in the memory and executable on the processor, wherein the program implements the steps of the method according to any one of claims 9 to 14 when executed by the processor.

21. A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the method according to any one of claims 1 to 14.

22. A computer program product comprising computer instructions, which when executed by a processor implement the steps of the method according to any one of claims 1 to 14.

Images