Data processing method, communication method and device, terminal equipment and network equipment

Abstract

The invention relates to a data processing method, a communication method and device, terminal equipment and network equipment. The data processing method comprises the steps of performing data enhancement on first channel state information (CSI) data based on feature information of the first CSI data to obtain a plurality of pieces of second CSI data; the feature information comprises at least one of spatial domain features and / or frequency domain features; and at least taking the first CSI data and the plurality of pieces of second CSI data as CSI sample data. In this way, the real CSI data can be enhanced.

Description

Data processing method, communication method and device, terminal equipment and network equipment Technical Field

[0001] The present application relates to the field of communications, and more specifically, to a data processing method, a communication method and apparatus, a terminal device, and a network device. Background Art

[0002] Leveraging the nonlinear fitting capabilities of neural networks to compress and feed back channel state information can significantly improve compression efficiency and feedback accuracy. However, this solution requires massive amounts of training data to support model construction. However, collecting such a large amount of channel state information presents challenges in terms of both cost and difficulty.

[0003] Summary of the Invention

[0004] The embodiments of the present application provide a data processing method, a communication method and apparatus, a terminal device, and a network device, which can perform data enhancement on real CSI data, and the CSI data obtained after data enhancement processing retains the spatial domain characteristics and / or frequency domain characteristics of the original CSI data.

[0005] The present invention provides a data processing method, including:

[0006] Based on characteristic information of first channel state information (CSI) data, data enhancement is performed on the first CSI data to obtain multiple second CSI data; the characteristic information includes at least one of the following: spatial domain characteristics and / or frequency domain characteristics;

[0007] At least the first CSI data and the multiple second CSI data are used as CSI sample data.

[0008] An embodiment of the present application provides a communication method, including:

[0009] The terminal device sends the first information, which is obtained by inputting the target channel state information CSI data into the first target model for encoding processing; the first target model is obtained by training the first preset model based on the CSI sample data obtained by the above method.

[0010] The present invention provides a communication method, including:

[0011] The network device receives the first information, inputs the first information into the second target model for decoding processing, and obtains the target CSI data corresponding to the first information; wherein, the second target model is obtained by training the second preset model based on the CSI sample data obtained by the above method.

[0012] An embodiment of the present application provides a data processing device, including:

[0013] a data enhancement processing unit, configured to perform data enhancement on the first channel state information (CSI) data based on characteristic information of the first CSI data to obtain a plurality of second CSI data; the characteristic information comprising at least one of the following: spatial domain characteristics and / or frequency domain characteristics;

[0014] The sample data determining unit is configured to use at least the first CSI data and the plurality of second CSI data as CSI sample data.

[0015] An embodiment of the present application provides a terminal device, including:

[0016] A sending unit is used to send first information, where the first information is obtained by inputting target channel state information CSI data into a first target model for encoding processing; the first target model is obtained by training a first preset model based on the CSI sample data obtained based on the above method.

[0017] An embodiment of the present application provides a network device, including:

[0018] A receiving unit is used to receive first information, input the first information into a second target model for decoding processing, and obtain target CSI data corresponding to the first information; wherein, the second target model is obtained by training a second preset model based on the CSI sample data obtained by the above method.

[0019] An embodiment of the present application provides a terminal device, comprising a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to call and execute the computer program stored in the memory, so that the terminal device executes the above-mentioned communication method applied to the terminal device side.

[0020] An embodiment of the present application provides a network device, comprising a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory, so that the network device executes the above-mentioned communication method applied to the network device side.

[0021] An embodiment of the present application provides a chip for implementing the above-mentioned data processing method or communication method.

[0022] Specifically, the chip includes: a processor for calling and running a computer program from a memory, so that a device equipped with the chip executes the above-mentioned data processing method or communication method.

[0023] An embodiment of the present application provides a computer-readable storage medium for storing a computer program. When the computer program is executed by a device, the device executes the above-mentioned data processing method or communication method.

[0024] An embodiment of the present application provides a computer program product, including computer program instructions, which enable a computer to execute the above-mentioned data processing method or communication method.

[0025] An embodiment of the present application provides a computer program, which, when executed on a computer, enables the computer to execute the above-mentioned data processing method or communication method.

[0026] In an embodiment of the present application, data enhancement can be performed on real CSI data (e.g., using real CSI data as first CSI data), and the CSI data obtained after data enhancement (e.g., second CSI data) retains the spatial domain features and / or frequency domain features of the original CSI data (e.g., first CSI data). This lays the foundation for effectively solving the problem of the existing small sample size that cannot support model training. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] FIG1 is a schematic diagram of an application scenario according to an embodiment of the present application.

[0028] FIG2 is a schematic diagram of a neuron structure according to an embodiment of the present application.

[0029] FIG3 is a schematic diagram of a channel state information feedback system according to another embodiment of the present application.

[0030] FIG4 is a schematic flowchart of a data processing method 400 according to an embodiment of the present application.

[0031] FIG5 is a schematic flowchart of a communication method 500 according to an embodiment of the present application.

[0032] FIG6 is a schematic flowchart of a communication method 600 according to an embodiment of the present application.

[0033] FIG7 is a schematic diagram of a first DFT vector space according to an embodiment of the present application.

[0034] FIG8 is a first schematic block diagram of a data processing device 800 according to an embodiment of the present application.

[0035] FIG9 is a second schematic block diagram of a data processing device 800 according to an embodiment of the present application.

[0036] FIG10 is a schematic block diagram of a terminal device 1000 according to an embodiment of the present application.

[0037] FIG11 is a schematic block diagram of a network device 1100 according to an embodiment of the present application.

[0038] FIG12 is a schematic block diagram of a communication device 1200 according to an embodiment of the present application.

[0039] FIG13 is a schematic block diagram of a chip 1300 according to an embodiment of the present application.

[0040] FIG14 is a schematic block diagram of a communication system 1400 according to an embodiment of the present application. DETAILED DESCRIPTION

[0041] The technical solutions in the embodiments of the present application will be described below in conjunction with the drawings in the embodiments of the present application.

[0042] The technical solutions of the embodiments of the present application can be applied to various communication systems, such as: Global System of Mobile communication (GSM) system, Code Division Multiple Access (CDMA) system, Wideband Code Division Multiple Access (WCDMA) system, General Packet Radio Service (GPRS), Long Term Evolution (LTE) system, Advanced Long Term Evolution (LTE-A) system, New Radio (NR) system, NR system evolution system, LTE on unlicensed spectrum (LTE-U) system, NR on unlicensed spectrum (NR-U) system, Non-Terrestrial Networks (NTN) system, Universal Mobile Telecommunication System (UMTS), Wireless Local Area Networks (WLAN), Wireless Fidelity (Wireless Fidelity) system. Fidelity, WiFi), fifth-generation communication (5th-Generation, 5G) system or other communication systems, etc.

[0043] Generally speaking, traditional communication systems support a limited number of connections and are easy to implement. However, with the development of communication technology, mobile communication systems will not only support traditional communications, but will also support, for example, device-to-device (D2D) communication, machine-to-machine (M2M) communication, machine-type communication (MTC), vehicle-to-vehicle (V2V) communication, or vehicle-to-everything (V2X) communication, etc. The embodiments of the present application can also be applied to these communication systems.

[0044] In one embodiment, the communication system in the embodiment of the present application can be applied to a carrier aggregation (CA) scenario, a dual connectivity (DC) scenario, and a standalone (SA) networking scenario.

[0045] In one embodiment, the communication system in the embodiment of the present application can be applied to an unlicensed spectrum, wherein the unlicensed spectrum can also be considered as a shared spectrum; or, the communication system in the embodiment of the present application can also be applied to an authorized spectrum, wherein the authorized spectrum can also be considered as an unshared spectrum.

[0046] The embodiments of the present application describe various embodiments in conjunction with network devices and terminal devices, wherein the terminal device may also be referred to as user equipment (UE), access terminal, user unit, user station, mobile station, mobile station, remote station, remote terminal, mobile device, user terminal, terminal, wireless communication device, user agent or user device, etc.

[0047] The terminal device can be a station (STAION, ST) in a WLAN, a cellular phone, a cordless phone, a Session Initiation Protocol (SIP) phone, a Wireless Local Loop (WLL) station, a Personal Digital Assistant (PDA) device, a handheld device with wireless communication capabilities, a computing device or other processing device connected to a wireless modem, a vehicle-mounted device, a wearable device, a terminal device in a next-generation communication system such as an NR network, or a terminal device in a future evolved Public Land Mobile Network (PLMN) network, etc.

[0048] In an embodiment of the present application, the terminal device can be deployed on land, including indoors or outdoors, handheld, wearable or vehicle-mounted; it can also be deployed on the water surface (such as ships, etc.); it can also be deployed in the air (such as airplanes, balloons and satellites, etc.).

[0049] In an embodiment of the present application, the terminal device may be a mobile phone, a tablet computer, a computer with wireless transceiver function, a virtual reality (VR) terminal device, an augmented reality (AR) terminal device, a wireless terminal device in industrial control, a wireless terminal device in self-driving, a wireless terminal device in remote medical, a wireless terminal device in a smart grid, a wireless terminal device in transportation safety, a wireless terminal device in a smart city, or a wireless terminal device in a smart home, etc.

[0050] As an example and not a limitation, in the embodiment of the present application, the terminal device may also be a wearable device. Wearable devices may also be called wearable smart devices, which are a general term for wearable devices that are intelligently designed and developed using wearable technology for daily wear, such as glasses, gloves, watches, clothing, and shoes. A wearable device is a portable device that is worn directly on the body or integrated into the user's clothes or accessories. Wearable devices are not only hardware devices, but also achieve powerful functions through software support, data interaction, and cloud interaction. Broadly speaking, wearable smart devices include those that are fully functional, large in size, and can achieve complete or partial functions without relying on smartphones, such as smart watches or smart glasses, as well as those that only focus on a certain type of application function and need to be used in conjunction with other devices such as smartphones, such as various smart bracelets and smart jewelry for vital sign monitoring.

[0051] In an embodiment of the present application, the network device may be a device for communicating with a mobile device. The network device may be an access point (AP) in WLAN, a base station (BTS) in GSM or CDMA, a base station (NodeB, NB) in WCDMA, an evolved base station (eNB or eNodeB) in LTE, or a relay station or access point, or a vehicle-mounted device, a wearable device, and a network device (gNB) in an NR network, or a network device in a future evolved PLMN network or a network device in an NTN network, etc.

[0052] As an example and not a limitation, in an embodiment of the present application, the network device may have a mobile feature, for example, the network device may be a mobile device. Alternatively, the network device may be a satellite or a balloon station. For example, the satellite may be a low earth orbit (LEO) satellite, a medium earth orbit (MEO) satellite, a geostationary earth orbit (GEO) satellite, a high elliptical orbit (HEO) satellite, etc. Optionally, the network device may also be a base station set up in a location such as land or water.

[0053] In an embodiment of the present application, the network device can provide services for a cell, and the terminal device communicates with the network device through the transmission resources used by the cell (for example, frequency domain resources, or spectrum resources). The cell can be a cell corresponding to the network device (for example, a base station). The cell can belong to a macro base station or a base station corresponding to a small cell. The small cells here may include: metro cells, micro cells, pico cells, femto cells, etc. These small cells have the characteristics of small coverage and low transmission power, and are suitable for providing high-speed data transmission services.

[0054] FIG1 exemplarily illustrates a communication system 100. The communication system includes a network device 110 and two terminal devices 120. In one embodiment, the communication system 100 may include multiple network devices 110, and each network device 110 may include a different number of terminal devices 120 within its coverage area, which is not limited in this embodiment of the present application.

[0055] In one embodiment, the communication system 100 may further include other network entities such as a Mobility Management Entity (MME) and an Access and Mobility Management Function (AMF), which is not limited in this embodiment of the present application.

[0056] Among them, the network equipment may include access network equipment and core network equipment. That is, the wireless communication system also includes multiple core networks for communicating with the access network equipment. The access network equipment can be an evolutionary base station (evolutional node B, abbreviated as eNB or e-NodeB) macro base station, micro base station (also called "small base station"), pico base station, access point (AP), transmission point (TP) or new generation base station (new generation Node B, gNodeB), etc. in a long-term evolution (LTE) system, a next-generation (mobile communication system) (next radio, NR) system or an authorized auxiliary access long-term evolution (LAA-LTE) system.

[0057] It should be understood that in the embodiments of the present application, a device having a communication function in a network / system may be referred to as a communication device. Taking the communication system shown in Figure 1 as an example, the communication device may include a network device and a terminal device having a communication function. The network device and the terminal device may be specific devices in the embodiments of the present application and will not be described in detail here. The communication device may also include other devices in the communication system, such as a network controller, a mobility management entity, and other network entities, which are not limited in the embodiments of the present application.

[0058] It should be understood that the terms "system" and "network" are often used interchangeably herein. The term "and / or" is simply a description of an association between related objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A exists alone, A and B exist simultaneously, or B exists alone. Furthermore, the character " / " generally indicates that the related objects are in an "or" relationship.

[0059] It should be understood that the "indication" mentioned in the embodiments of this application can be a direct indication, an indirect indication, or an indication of an association. For example, "A indicates B" can mean that A directly indicates B, for example, B can be obtained through A; it can also mean that A indirectly indicates B, for example, A indicates C, and B can be obtained through C; it can also mean that there is an association between A and B.

[0060] In the description of the embodiments of the present application, the term "corresponding" may indicate a direct or indirect correspondence between the two, or an association relationship between the two, or a relationship between indication and being indicated, configuration and being configured, etc.

[0061] To facilitate understanding of the technical solutions of the embodiments of the present application, the relevant technologies of the embodiments of the present application are described below. The following relevant technologies can be arbitrarily combined with the technical solutions of the embodiments of the present application as optional solutions, and they all fall within the protection scope of the embodiments of the present application.

[0062] First, codebook-based CSI data feedback

[0063] For 5G New Radio (NR) systems, a codebook-based approach can be used to extract and feedback channel features in the design of Channel State Information (CSI) feedback. That is, after performing channel estimation at the transmitter and obtaining the corresponding precoding matrix based on the channel estimation results, the coding matrix that best matches the precoding matrix is ​​selected from a pre-set codebook according to a certain optimization criterion. The index of the coding matrix that best matches the precoding matrix is ​​then fed back to the receiver via the air interface feedback link for precoding. Codebooks can be divided into three schemes: Type I, Type II, and eType II.

[0064] Second, neural network

[0065] A neural network is a computational model consisting of multiple interconnected neuron nodes. The connections between nodes represent weighted values ​​from input signals to output signals, called weights (e.g., w1 to wn). As shown in Figure 2, each node performs a weighted sum (Sum) of different input signals (e.g., a1 to an) and outputs the result using a specific activation function. In practical applications, neural networks can be classified as fully connected neural networks, convolutional neural networks, recurrent neural networks, and so on.

[0066] Third, CSI data feedback based on artificial intelligence (AI)

[0067] Given the tremendous success of AI in areas such as computer vision and natural language processing, the communications sector is beginning to explore new approaches to address technical challenges limited by traditional approaches, such as deep learning. The neural network architecture commonly used in deep learning is nonlinear and data-driven. It can extract features from actual channel state information and restore the compressed channel state information fed back by the UE to the greatest extent possible on the base station side. This ensures the restoration of channel state information while also reducing the CSI data feedback overhead on the UE side.

[0068] As shown in Figure 3, the channel state information feedback system can be divided into two parts: an encoder (such as a CSI encoder) and a decoder (such as a CSI decoder), and the two parts are deployed on the UE side and the base station (BS) side respectively. After the UE side obtains the channel state information through channel estimation, it compresses and encodes the matrix of the channel state information through the neural network of the encoder, and feeds back the compressed bit stream (that is, the compressed CSI data) to the BS side through the air interface feedback link. The BS side recovers or reconstructs the channel state information based on the feedback bit stream through the decoder to obtain the complete channel state information fed back. On the encoder side, the structure shown in Figure 3 can use several fully connected layers for encoding, and correspondingly, a convolutional neural network structure can be used for decoding on the decoder side.

[0069] It is understandable that, when the encoding and decoding framework remains unchanged, the model structure inside the encoder and decoder can be flexibly designed using other models, and the present disclosure does not impose specific restrictions on this.

[0070] Fourth, data enhancement

[0071] Existing data augmentation methods in computer vision primarily perform operations such as flipping, rotating, cropping, and scaling on image data to amplify small amounts of image data into larger amounts. Notably, these methods do not distort the meaning of the image due to these operations; the original content remains in the generated data.

[0072] Here, leveraging the nonlinear fitting capabilities of neural networks to compress and feed back channel state information can significantly improve compression efficiency and feedback accuracy. However, this solution requires massive amounts of training data to support model construction. However, collecting such a large amount of channel state information presents challenges in terms of both cost and difficulty.

[0073] Furthermore, existing data augmentation methods for image data in the computer vision field are not applicable to CSI data. This is because, unlike image matrices, CSI data matrices contain spatial and frequency domain features common in the communications field. Scaling, flipping, and cropping CSI data, similar to image operations, can affect or even erase the physical meaning of the CSI data. While large amounts of data can be generated, they cannot support the training of AI-based CSI data feedback models.

[0074] Based on this, the disclosed solution provides a method for data enhancement based on a small amount of data, thereby solving the problem that the existing sample size is too small to support the training of the AI-based CSI data feedback model.

[0075] Specifically, FIG4 is a schematic flow chart of a data processing method 400 according to an embodiment of the present application. The method can optionally be applied to the system shown in FIG1 , but is not limited thereto. The method includes at least part of the following contents.

[0076] S410. Based on characteristic information of first channel state information CSI data, perform data enhancement on the first CSI data to obtain multiple second CSI data; the characteristic information includes at least one of the following: spatial domain characteristics and / or frequency domain characteristics.

[0077] S420: Use at least the first CSI data and the multiple second CSI data as CSI sample data.

[0078] It can be understood that the data processing method of the disclosed solution can be executed in any entity in the system shown in Figure 1, or in other electronic devices with computing capabilities outside the system shown in Figure 1, such as personal computers, servers, server clusters, etc., and the disclosed solution does not impose any restrictions on this.

[0079] In one embodiment, the first CSI data is CSI data estimated by the terminal device, which may also be referred to as true CSI data. Accordingly, the second CSI data is CSI data obtained by performing data enhancement on the true CSI data.

[0080] Because the second CSI data is obtained by data enhancement based on the spatial and / or frequency domain features of the first CSI data, it effectively preserves the original CSI data, namely, the communication physical characteristics of the first CSI data. This provides training data for the AI-based CSI data feedback model, effectively resolving the issue of insufficient real data to support model training. This also lays the foundation for improving the accuracy of the prediction results of the AI-based CSI data feedback model.

[0081] In a specific example of the disclosed solution, after obtaining the CSI sample data, the AI-based CSI data feedback model may be trained based on the CSI sample data. Specifically, the method further includes training of at least one of the following:

[0082] Training 1: Training the preset model for encoding processing on the transmitting end, such as the terminal device side; specifically, based on the CSI sample data, the first preset model is trained to obtain a first target model, wherein the first target model is used to encode the channel state information CSI data and obtain the target CSI data.

[0083] Training 2: Training the preset model for decoding processing on the receiving end, such as the base station side; specifically, based on the CSI sample data, the second preset model is trained to obtain a second target model, wherein the second target model is used to decode the target CSI data to obtain the CSI corresponding to the target CSI data.

[0084] The above-mentioned training one and training two can be performed selectively, or both can be performed, and the present disclosure scheme does not impose any restrictions on this; moreover, the present disclosure scheme does not impose any restrictions on the execution order of the above-mentioned two trainings, for example, simultaneous training, or separate training, or training of another model after one model training is completed, etc.; as long as the CSI sample data of the present disclosure scheme is used to train the model, it is within the protection scope of the present disclosure scheme.

[0085] In one example, the first target model and the second target model are different templates in the AI-based CSI data feedback model. For example, the first target model is a sub-model for encoding processing in the AI-based CSI data feedback model; accordingly, the first target model is a sub-model for decoding processing in the AI-based CSI data feedback model; further, as shown in FIG3 , the first target model can be specifically located in the encoder on the terminal device side; accordingly, the second target model can be specifically located in the decoder on the base station side. In one example, in this scenario, the above two types of training can be carried out jointly, or, after one model training is completed, the trained model is used to train another model. For example, after the first preset model is trained to obtain the first target model, the first target model, such as the output result of the first target model, is used to train the second preset model, thereby effectively improving the accuracy of model prediction.

[0086] Here, in the scenario where the second target model is trained based on the output result of the first target model, the model training of the second preset model based on the CSI sample data can be understood as: indirectly training the second preset model based on the CSI sample data. It should be noted that, whether the model training is performed directly using the CSI sample data obtained by the disclosed solution (such as directly using the CSI sample data as the input of the model to be trained) or the model training is performed indirectly using the CSI sample data obtained by the disclosed solution (such as using the data obtained after processing the CSI sample data as the input of the model to be trained), both fall within the scope of "model training of the first preset model or the second preset model based on the CSI sample data", that is, both are within the protection scope of the disclosed solution.

[0087] In this way, since corresponding model training can be performed based on the CSI sample data obtained based on the disclosed solution, the problem of insufficient real data to support model training is effectively solved, and the manpower, material and other costs of obtaining real data are greatly reduced; at the same time, it also lays the foundation for improving the accuracy of the prediction results of the AI-based CSI data feedback model.

[0088] It should be noted that, experimentally, there is no overfitting phenomenon when training the model based on the sample data of the disclosed solution, and the performance of the obtained model is good.

[0089] The present disclosure provides two methods for performing data enhancement on the first CSI data. The following describes the two methods in detail. Specifically:

[0090] The first data enhancement method: using a codebook method to enhance the first CSI data, such as real CSI data, is as follows:

[0091] In a specific example of the present disclosure, data enhancement of the first CSI data may be implemented based on the first basis vector group. Specifically, the method further includes:

[0092] Determining, based on a first discrete Fourier transform (DFT) vector space constructed by a preset codebook, a first basis vector group representing spatial domain characteristics of the first CSI data;

[0093] Based on this, step S410 specifically includes: performing data enhancement on the first CSI data based on at least the first basis vector group representing the spatial characteristics of the first CSI data to obtain multiple second CSI data.

[0094] In practical applications, the first CSI data can be represented by a matrix, for example, the matrix W∈C Nt×Nsb, where C represents the complex space, Nt represents the number of transmit antenna ports, and Nsb represents the number of frequency domain subbands corresponding to the transmitter. In this way, the spatial characteristics of the first CSI data are quantitatively expressed based on the matrix W, providing support for effectively extracting the communication physical characteristics of the first CSI data and achieving data enhancement. This approach is simple and highly interpretable. It should be understood that the above is only an example of a matrix expression. In actual applications, the rows and columns of the matrix can be adjusted based on needs. The present disclosure does not limit the specific expression of the matrix.

[0095] Here, since the first basis vector group can represent the spatial characteristics of the first CSI data, the second CSI data obtained after data enhancement based on the first basis vector group can at least represent the original CSI data, that is, the spatial characteristics of the first CSI data. Compared with the existing image data method, the data enhancement method of the disclosed scheme can effectively retain the physical meaning information within the original CSI data, providing data support for subsequent model training while also laying the foundation for improving the accuracy of the model training results.

[0096] In a specific example, the preset codebook can be any of the three schemes of Type I, Type II, and eType II. The present disclosure does not impose any restrictions on this. As long as the DFT space vector can be constructed based on the preset codebook, it is within the protection scope of the present disclosure. Furthermore, the present disclosure does not impose any specific restrictions on the construction method of the first DFT vector space constructed based on the preset codebook. The following specific example provides a method for constructing the first DFT vector space based on eType II. Please refer to the following description and will not be repeated here. It can be understood that this method is only an example and is not used to limit the present disclosure.

[0097] In a specific example of the present disclosure, the first basis vector group may be obtained in the following manner. Specifically, the first DFT vector space constructed based on the preset codebook described above determines the first basis vector group representing the spatial domain characteristics of the first CSI data, specifically including:

[0098] L orthogonal basis vectors are selected from the orthogonal basis vector group included in the first DFT vector space, where the correlation between the L orthogonal basis vectors and the spatial domain characteristics of the first CSI data satisfies a first preset condition (for example, the L orthogonal basis vectors are the L vectors selected from the first DFT vector space that have the highest correlation with the spatial domain characteristics of the first CSI data), and L is a natural number greater than or equal to 1; the first basis vector group is obtained based on the selected L orthogonal basis vectors; that is, the L orthogonal basis vectors constitute the first basis vector group.

[0099] It is understandable that the magnitude of the correlation can be measured by the similarity between vectors. Further, the L is a natural number greater than or equal to 1 and less than the dimension of the spatial domain feature of the first CSI data; for example, in the matrix W∈C Nt×Nsb At this time, L is a natural number greater than or equal to 1 and less than Nt / 2.

[0100] Furthermore, in one example, the average correlation can be used as an indicator to select L orthogonal basis vectors from the first DFT vector space. In this case, the L orthogonal basis vectors can also be specifically L vectors selected from the first DFT vector space and having the highest average correlation with the spatial domain features of the first CSI data.

[0101] It should be noted that, in actual applications, the first preset condition can be set based on actual needs, and the present disclosure does not impose any specific restrictions on this.

[0102] Furthermore, in a specific example of the disclosed solution, a target orthogonal basis vector group may be first selected from a plurality of orthogonal basis vector groups included in the first DFT vector space, and then L orthogonal basis vectors may be selected from the target orthogonal basis vector group. Specifically, the method further includes:

[0103] Selecting, from a plurality of orthogonal basis vector groups included in the first DFT vector space, a target orthogonal basis vector group whose correlation with the spatial domain features of the first CSI data satisfies a second preset condition (for example, selecting a group with the strongest correlation as the target orthogonal basis vector group);

[0104] Based on this, the above-mentioned selecting L orthogonal basis vectors from the orthogonal basis vector group included in the first DFT vector space specifically includes: selecting the L orthogonal basis vectors from the target orthogonal basis vector group.

[0105] It can be understood that the correlation in this example can also be specifically an average correlation. In this case, the target orthogonal basis vector group can be specifically: from the multiple orthogonal basis vector groups contained in the first DFT vector space, the group with the highest average correlation with the spatial domain characteristics of the first CSI data is selected, and the group with the highest average correlation with the spatial domain characteristics of the first CSI data is used as the target orthogonal basis vector group.

[0106] It should be noted that, in actual applications, the second preset condition can be set based on actual needs, and the present disclosure does not impose any specific restrictions on this.

[0107] In a specific example of the present disclosure, a target orthogonal basis vector group may be selected from multiple orthogonal basis vector groups included in the first DFT vector space in the following manner. That is, the aforementioned selection of a target orthogonal basis vector group whose correlation with the spatial domain features of the first CSI data satisfies a second preset condition from the multiple orthogonal basis vector groups included in the first DFT vector space specifically includes:

[0108] Projecting the first CSI data (e.g., a matrix of the first CSI data) into a vector space spanned by a diagonal block matrix to obtain a first projection information matrix corresponding to each of the orthogonal basis vector groups, wherein the diagonal block matrix is ​​obtained based on the orthogonal basis vector group included in the first DFT vector space; elements in the first projection information matrix represent information related to projection coefficients after spatial features of the first CSI data are projected onto the diagonal block matrix;

[0109] Based on the first projection information matrix corresponding to each of the orthogonal basis vector groups, a target orthogonal basis vector group whose correlation with the spatial domain features of the first CSI data meets a second preset condition is selected from the multiple orthogonal basis vector groups.

[0110] It can be understood that for each orthogonal basis vector group, each corresponds to a diagonal block matrix. Accordingly, a first projection information matrix can also be obtained based on the above method. In this way, the first projection information matrix corresponding to each orthogonal basis vector group can be obtained, and then based on the first projection information matrix corresponding to each orthogonal basis vector group, the target orthogonal basis vector group can be selected from the multiple orthogonal basis vector groups.

[0111] Furthermore, in a specific example, the first projection information matrix can be specifically a first projection coefficient absolute value matrix. In this case, the elements in the first projection coefficient absolute value matrix represent the absolute values ​​of the projection coefficients after the spatial domain features of the first CSI data are projected onto the diagonal block matrix.

[0112] In a specific example of the disclosed solution, each first projection information matrix can be processed in the following manner to obtain L first target elements corresponding to each group (for example, L maximum value elements), and then determine a target orthogonal basis vector group based on the L first target elements corresponding to each group. Specifically, the method further includes:

[0113] Data processing is performed on the elements in the first projection information matrix to obtain a first vector corresponding to the orthogonal basis vector group, wherein the first vector can represent the total relevant information of the projection coefficients corresponding to the frequency domain characteristics of the first CSI data, for example, the sum of the projection coefficients corresponding to the frequency domain characteristics of the first CSI data; specifically, for the first CSI data, the matrix W∈C Nt×Nsb For example, at this time, the first vector is obtained by summing the first projection information matrix by column, and represents the sum of the projection coefficients corresponding to the frequency domain characteristics of the first CSI data.

[0114] Furthermore, some elements in the first vector and other elements in the first vector are processed to obtain a second vector corresponding to the orthogonal basis vector group, wherein the second vector can represent the total relevant information of the projection coefficient corresponding to the spatial characteristics of the first CSI data; for example, the matrix of the first CSI data is W∈C Nt×Nsb For example, at this time, the second vector is obtained by adding the first Nt / 2 elements and the last Nt / 2 elements in the first vector.

[0115] Furthermore, L first target elements corresponding to the orthogonal basis vector groups are selected from the second vector. Thus, based on the above method, L first target elements corresponding to each group are obtained. Here, to maximize the extraction of the communication physical characteristics of the first CSI data, the L first target elements can be specifically the L maximum value elements in the second vector.

[0116] Further, based on this, the above-described method of selecting, from the multiple orthogonal basis vector groups, a target orthogonal basis vector group whose correlation with the spatial characteristics of the first CSI data satisfies a second preset condition based on the first projection information matrix corresponding to each of the orthogonal basis vector groups, includes:

[0117] The target orthogonal basis vector group is selected from the multiple orthogonal basis vector groups based on the L first target elements corresponding to each of the orthogonal basis vector groups.

[0118] For example, based on the L maximum value elements corresponding to each of the orthogonal basis vector groups, the target orthogonal basis vector group is selected from the multiple orthogonal basis vector groups.

[0119] Furthermore, in a specific example of the disclosed solution, the target orthogonal basis vector group is selected from the multiple orthogonal basis vector groups based on the L first target elements corresponding to each of the orthogonal basis vector groups, specifically including: obtaining the target sum corresponding to each of the orthogonal basis vector groups based on the L first target elements corresponding to each of the orthogonal basis vector groups; then selecting a target value from the multiple target sums, and using the orthogonal basis vector group corresponding to the selected target value as the target orthogonal basis vector group. For example, based on the L maximum value elements corresponding to each of the orthogonal basis vector groups, the maximum value sum corresponding to each of the orthogonal basis vector groups is obtained; then selecting the group with the largest value from the multiple maximum value sums, and using the orthogonal basis vector group with the largest value selected as the target orthogonal basis vector group.

[0120] In one example, after obtaining the target orthogonal basis vector set, the L first target elements in the target orthogonal basis vector set can be directly used as the L orthogonal basis vectors finally selected. Alternatively, the L orthogonal basis vectors can be selected from the target orthogonal basis vector set based on other rules.

[0121] It should be noted that a specific example is given below to illustrate in detail how to obtain the first basis vector group by using the first DFT vector space constructed based on the preset codebook. For details, please refer to step a in the following example, which will not be repeated here.

[0122] Here, it can be understood that the following example is based on the matrix of the first CSI data as W∈C Nt×Nsb For illustration purposes, in actual applications, the rows and columns of the matrix of the first CSI data may be swapped. In this case, the specific processing method given below only requires making corresponding data changes based on the principle of matrix multiplication.

[0123] In a specific example of the disclosed solution, data enhancement of the first CSI data can be performed based on the first basis vector group and the second basis vector group. In this way, the second CSI data obtained after data enhancement has both spatial and frequency domain characteristics of the first CSI data. Specifically, the method further includes:

[0124] Determining, based on the second DFT vector space constructed by the preset codebook and the first basis vector group representing the spatial domain characteristics of the first CSI data, a second basis vector group representing at least the frequency domain characteristics of the first CSI data;

[0125] Based on this, the above-mentioned data enhancement is performed on the first CSI data based on at least the first basis vector group representing the spatial characteristics of the first CSI data to obtain multiple second CSI data, including:

[0126] Based on the first basis vector group and the second basis vector group, data enhancement is performed on the first CSI data to obtain multiple second CSI data.

[0127] In this way, the second CSI data obtained after the data enhancement processing has both the spatial domain characteristics and the frequency domain characteristics of the original CSI data, that is, the first CSI data.

[0128] It should be noted that the present disclosure does not impose any specific restrictions on the method of constructing the second DFT vector space based on the preset codebook. The following specific example provides a method for constructing a second DFT vector space based on eTypeII. Please refer to the following description and will not be repeated here. It can be understood that this method is only an example and is not intended to limit the present disclosure.

[0129] Furthermore, in a specific example of the present disclosure, the following method may be used to obtain a second basis vector group. Specifically, the second DFT vector space constructed based on the preset codebook and the first basis vector group representing the spatial domain characteristics of the first CSI data described above are used to determine the second basis vector group representing at least the frequency domain characteristics of the first CSI data, specifically including:

[0130] Projecting the first CSI data (e.g., a matrix of the first CSI data) into the space spanned by the first basis vector group to obtain a first projection coefficient matrix; wherein elements in the first projection coefficient matrix represent projection coefficients of the spatial features of the first CSI data in the space spanned by the first basis vector group;

[0131] Based on the first projection coefficient matrix, M orthogonal basis vectors are selected from the orthogonal basis vectors included in the second DFT vector space constructed by the preset codebook, where correlations between the M orthogonal basis vectors and the projection coefficients corresponding to the frequency domain features of the first CSI data in the first projection coefficient matrix satisfy a third preset condition (for example, the M orthogonal basis vectors are: M vectors in the second DFT vector space that have the highest correlations with the projection coefficients corresponding to the frequency domain features of the first CSI data in the first projection coefficient matrix); where M is a natural number greater than or equal to 1;

[0132] Furthermore, the second basis vector group is obtained based on the selected M orthogonal basis vectors, that is, the M orthogonal basis vectors constitute the second basis vector group.

[0133] It is understandable that the magnitude of the correlation can be measured by the similarity between vectors. Further, the M is a natural number greater than or equal to 1 and less than the dimension of the frequency domain feature of the first CSI data; for example, in the matrix W∈C Nt×Nsb, at this time, the M is a natural number greater than or equal to 1 and less than Nsb.

[0134] Furthermore, in one example, average correlation can be used as an indicator to select M orthogonal basis vectors from the second DFT vector space. In this case, the M orthogonal basis vectors can be specifically M vectors selected from the second DFT vector space that have the highest average correlation with the projection coefficients corresponding to the frequency domain characteristics of the first CSI data in the first projection coefficient matrix. In this case, the M orthogonal basis vectors can better represent the original CSI data, that is, the frequency domain characteristics of the first CSI data.

[0135] It should be noted that, in actual applications, the third preset condition can be set based on actual needs, and the present disclosure does not impose any specific restrictions on this.

[0136] Furthermore, in a specific example of the presently disclosed solution, the first CSI data can be projected in the following manner. Specifically, the above-described projecting of the first CSI data into the space spanned by the first basis vector group specifically includes: projecting the first CSI data into the space spanned by the first basis vector matrix, wherein the first basis vector matrix is ​​a diagonal block matrix constructed based on the first basis vector group.

[0137] In a specific example of the disclosed solution, M orthogonal basis vectors may be obtained in the following manner; specifically, the method further includes:

[0138] Projecting the first projection coefficient matrix into the space spanned by the orthogonal basis vectors of the second DFT vector space, obtaining a second projection information matrix; elements in the second projection information matrix represent information related to the projection coefficients of the frequency domain characteristics of the first CSI data corresponding to the first projection coefficient matrix in the space spanned by the orthogonal basis vectors of the second DFT vector space; in other words, the elements in the second projection information matrix represent information related to the projection coefficients of the frequency domain characteristics of the first CSI data in the space spanned by the orthogonal basis vectors of the second DFT vector space;

[0139] Furthermore, the above-described selecting M orthogonal basis vectors from the orthogonal basis vectors contained in the second DFT vector space constructed by the preset codebook based on the first projection coefficient matrix may specifically include: based on the second projection information matrix, selecting M orthogonal basis vectors from the orthogonal basis vectors contained in the second DFT vector space.

[0140] In a specific example, the second projection information matrix can be specifically a second projection coefficient absolute value matrix. In this case, the elements in the second projection coefficient absolute value matrix represent the absolute values ​​of the projection coefficients of the frequency domain characteristics of the first CSI data corresponding to the first projection coefficient matrix in the space spanned by the orthogonal basis vectors of the second DFT vector space, that is, the elements in the second projection coefficient absolute value matrix represent the absolute values ​​of the projection coefficients of the frequency domain characteristics of the first CSI data in the space spanned by the orthogonal basis vectors of the second DFT vector space.

[0141] Furthermore, in a specific example of the disclosed solution, based on the second projection information matrix, M orthogonal basis vectors are selected from the orthogonal basis vectors contained in the second DFT vector space, including:

[0142] Data processing is performed on the elements in the second projection information matrix to obtain a third vector, wherein the third vector represents the total relevant information of the projection coefficients corresponding to the frequency domain characteristics of the first CSI data; for example, the sum of the projection coefficients corresponding to the frequency domain characteristics of the first CSI data; specifically, for the first CSI data, the matrix is ​​W∈C Nt×Nsb For example, at this time, the third vector is obtained by summing the elements in the second projection coefficient absolute value matrix by column;

[0143] Furthermore, M second target elements are selected from the third vector; and orthogonal basis vectors corresponding to the M second target elements are selected from the second DFT vector space to obtain the M orthogonal basis vectors.

[0144] In one example, to maximize the extraction of the communication physical characteristics of the first CSI data, the M second target elements may be specifically the M maximum elements in the third vector. Specifically, the first M maximum elements in the third vector are selected, and the orthogonal basis vectors corresponding to the selected M maximum elements in the second DFT vector space are used as the final M orthogonal basis vectors.

[0145] It should be noted that a specific example is given below to illustrate in detail how to obtain the second basis vector group from the second DFT vector space constructed based on the preset codebook. For details, please refer to step b in the following example, which will not be repeated here.

[0146] Here, it can be understood that the following example is based on the matrix of the first CSI data as W∈C Nt×Nsb For illustration purposes, in actual applications, the rows and columns of the matrix of the first CSI data may be swapped. In this case, the specific processing method given below only requires making corresponding data changes based on the principle of matrix multiplication.

[0147] In a specific example of the disclosed solution, after obtaining the first basis vector group and the second basis vector group, the method further includes:

[0148] Projecting the first projection coefficient matrix into the space spanned by the second basis vector group to obtain a second projection coefficient matrix, wherein elements in the second projection coefficient matrix represent projection coefficients corresponding to the spatial features of the first CSI data in the space spanned by the second basis vector group;

[0149] Furthermore, the above-described performing data enhancement on the first CSI data based on the first basis vector group and the second basis vector group to obtain multiple second CSI data includes:

[0150] Based on the second projection coefficient matrix, the first basis vector matrix and the second basis vector matrix, multiple second CSI data are obtained; wherein the first basis vector matrix is ​​based on the first basis vector group, and the second basis vector matrix is ​​a matrix formed based on the second basis vector group.

[0151] In a specific example, multiple second CSI data are obtained based on the matrix product of the second projection coefficient matrix, the first basis vector matrix, and the second basis vector matrix. In this way, data enhancement of the first CSI data can be achieved through matrix product processing. This method is simple and highly interpretable.

[0152] It should be noted that a specific example is given below to illustrate in detail the method of obtaining the second projection coefficient matrix. For details, please refer to step c in the following example, which will not be repeated here.

[0153] Here, it can be understood that the following example is based on the matrix of the first CSI data as W∈C Nt×Nsb For illustration purposes, in actual applications, the rows and columns of the matrix of the first CSI data may be swapped. In this case, the specific processing method given below only requires making corresponding data changes based on the principle of matrix multiplication.

[0154] In a specific example of the disclosed solution, after obtaining the second projection coefficient matrix, the second projection coefficient matrix may be scrambled. Specifically, the method further includes:

[0155] performing phase adjustment and / or amplitude adjustment on elements in the second projection coefficient matrix to obtain a plurality of third projection coefficient matrices;

[0156] Furthermore, the above-mentioned obtaining a plurality of second CSI data based on the second projection coefficient matrix, the first basis vector matrix, and the second basis vector matrix specifically includes:

[0157] A plurality of second CSI data are obtained based on the plurality of third projection coefficient matrices, the first basis vector matrix, and the second basis vector matrix.

[0158] Further, in one example, the above-described obtaining the plurality of second CSI data based on the plurality of third projection coefficient matrices, the first basis vector matrix, and the second basis vector matrix specifically includes:

[0159] Based on the matrix product of the third projection coefficient matrix, the first basis vector matrix and the second basis vector matrix, a plurality of vectors to be processed are obtained; and the plurality of vectors to be processed are normalized to obtain a plurality of second CSI data.

[0160] In a specific example of the presently disclosed scheme, the second projection coefficient matrix can be scrambled in the following manner. Specifically, the above-mentioned phase or amplitude adjustment of the elements in the second projection coefficient matrix specifically includes: taking the elements in the second projection coefficient matrix as the center, adjusting the center in terms of phase and / or amplitude.

[0161] That is to say, the disclosed solution adopts at least one of the two methods, such as only phase perturbation, or only amplitude adjustment, or phase perturbation and amplitude perturbation, to scramble the second projection coefficient matrix and achieve data enhancement of the first CSI data.

[0162] It should be noted that a specific example is given below to illustrate in detail how to perform scrambling on the second projection coefficient matrix. For details, please refer to step d in the following example, which will not be repeated here.

[0163] Here, it can be understood that the following example is based on the matrix of the first CSI data as W∈C Nt×Nsb For illustration purposes, in actual applications, the rows and columns of the matrix of the first CSI data may be swapped. In this case, the specific processing method given below only requires making corresponding data changes based on the principle of matrix multiplication.

[0164] In this way, the codebook method is used to enhance the real CSI data to obtain multiple second CSI data. Compared with existing enhancement methods for image data in the computer field, the disclosed solution fully considers the communication physical characteristics of real CSI data. Therefore, the disclosed solution is more suitable for data enhancement of CSI data.

[0165] The second data enhancement method: using the singular value decomposition method to enhance the first CSI data, such as the real CSI data

[0166] In a specific example of the disclosed solution, the method further includes:

[0167] Obtaining a target autocorrelation matrix of the first CSI data, where the target autocorrelation matrix can represent spatial domain characteristics and frequency domain characteristics of the first CSI data;

[0168] Performing singular value decomposition on the target autocorrelation matrix to obtain a singular value decomposition result, wherein the singular value decomposition result includes a singular vector matrix representing spatial domain characteristics and frequency domain characteristics of the first CSI data, and a singular value matrix;

[0169] Based on this, the above-mentioned feature information of the first CSI data is used to perform data enhancement on the first CSI data to obtain multiple second CSI data, including:

[0170] The singular value decomposition result is processed using the constructed target random matrix to obtain a plurality of second CSI data after data enhancement.

[0171] It should be noted that a specific example is given below to illustrate in detail the steps of using the constructed target random matrix to process the singular value decomposition results and obtain multiple second CSI data. For details, please refer to steps d and e in the following example, which will not be repeated here.

[0172] Here, it can be understood that the following example is based on the matrix of the first CSI data as W∈C Nt×Nsb For illustration purposes, in actual applications, the rows and columns of the matrix of the first CSI data may be swapped. In this case, the specific processing method given below only requires making corresponding data changes based on the principle of matrix multiplication.

[0173] In a specific example of the disclosed solution, the target autocorrelation matrix may be obtained by the following method; specifically, the method further includes:

[0174] Obtain a first autocorrelation matrix of the first CSI data, wherein the first autocorrelation matrix can characterize the spatial characteristics of the first CSI data; for example, the matrix of the first CSI data is W∈C Nt×Nsb For example, at this time, the first autocorrelation matrix R1 = WW H ∈C Nt×Nt , the first autocorrelation matrix R1 can represent the spatial characteristics of the first CSI data, such as spatial statistical characteristics.

[0175] Obtain a second autocorrelation matrix of the first CSI data, wherein the second autocorrelation matrix can characterize the frequency domain characteristics of the first CSI data; for example, continue to use the matrix of the first CSI data as W∈C Nt×Nsb For example, at this time, the second autocorrelation matrix R2 = W H W∈C Nsb×NsbThe second autocorrelation matrix R2 can represent the frequency domain characteristics of the first CSI data, such as frequency domain statistical characteristics.

[0176] Based on this, the above-mentioned obtaining of the target autocorrelation matrix of the first CSI data specifically includes: obtaining the target autocorrelation matrix based on the first autocorrelation matrix and the second autocorrelation matrix. For example, continuing to take the matrix of the first CSI data as W∈C Nt×Nsb For example, at this time, the target autocorrelation matrix The target autocorrelation matrix R3 can represent the spatial domain characteristics (such as spatial domain statistical characteristics) and frequency domain characteristics (such as frequency domain statistical characteristics) of the first CSI data.

[0177] In a specific example of the disclosed solution, processing the singular value decomposition result using the target random matrix to obtain a plurality of second CSI data after data enhancement specifically includes:

[0178] Using the target random matrix, replacing the first matrix in the singular vector matrix;

[0179] The target random matrix, the second matrix in the singular vector matrix, and the singular value matrix are subjected to matrix product processing to obtain multiple second CSI data; wherein the target autocorrelation matrix is ​​a symmetric matrix, and the first matrix and the second matrix satisfy a symmetric relationship.

[0180] In a specific example, the singular value decomposition result can be expressed as: [V, D, U];

[0181] Here, V∈C NsbNt×NsbNt With U∈C NsbNt×NsbNt They are the left singular vector matrix, the right singular vector matrix, and the diagonal matrix D∈C NsbNt×NsbNt is a singular value matrix, where the diagonal elements are singular values;

[0182] At this time, the first matrix is ​​a right singular vector matrix in the singular vector matrix, such as U∈C NsbNt×NsbNt ; The second matrix is ​​the left singular vector matrix in the singular vector matrix, such as V∈C NsbNt×NsbNt At this time, V=U H .

[0183] Furthermore, performing matrix product processing on the target random matrix, the second matrix in the singular vector matrix, and the singular value matrix to obtain a plurality of second CSI data specifically includes:

[0184] Performing matrix product processing on the target random matrix, the second matrix in the singular vector matrix, and the singular value matrix to obtain an enhanced data matrix; and arranging the enhanced data matrix based on the vector dimension of the first CSI data to obtain multiple second CSI data after data enhancement.

[0185] Furthermore, the enhanced data matrix is ​​arranged based on the vector dimension of the first CSI data to obtain a plurality of second CSI data after data enhancement, specifically including:

[0186] The enhanced data matrix is ​​arranged based on the vector dimension of the first CSI data to obtain an arranged enhanced data matrix; and the arranged enhanced data matrix is ​​normalized to obtain multiple second CSI data.

[0187] In this way, the real CSI data is enhanced using the singular value decomposition method to obtain multiple second CSI data. Compared with existing enhancement methods for image data in the computer field, the disclosed solution fully considers the communication physical characteristics of real CSI data and is therefore more suitable for CSI data enhancement.

[0188] Thus, the disclosed solution designs a data enhancement method for AI-based CSI data feedback models, generating a large amount of data using a small amount of real data, which is insufficient for model training. Furthermore, compared to existing enhancement methods for image data in the computer field, the disclosed solution is more suitable for CSI data enhancement.

[0189] FIG5 is a schematic flow chart of a communication method 500 according to an embodiment of the present application. The method can optionally be applied to the system shown in FIG1 , but is not limited thereto. The method includes at least part of the following contents.

[0190] S510. The terminal device sends first information, which is obtained by inputting channel state information CSI data into a first target model for encoding processing; the first target model is obtained by training a first preset model based on the CSI sample data obtained by the above data processing method.

[0191] The method for obtaining CSI sample data in this embodiment can refer to the relevant description in the above method 400, and for the sake of brevity, it will not be repeated here.

[0192] FIG6 is a schematic flow chart of a communication method 600 according to an embodiment of the present application. The method can optionally be applied to the system shown in FIG1 , but is not limited thereto. The method includes at least part of the following contents.

[0193] S610. The network device receives the first information, inputs the first information into the second target model for decoding processing, and obtains the CSI data corresponding to the first information; wherein, the second target model is obtained by training the second preset model with the CSI sample data obtained by the above data processing method.

[0194] The method for obtaining CSI sample data in this embodiment can refer to the relevant description in the above method 400, and for the sake of brevity, it will not be repeated here.

[0195] The following describes the disclosed solution in detail with reference to specific examples. Specifically, the disclosed solution performs data enhancement on a small amount of real CSI data (i.e., first CSI data, hereinafter collectively referred to as real CSI data). That is, a small amount of real CSI data is used as a seed to generate a large amount of data for training an AI-based CSI data feedback model. This solves the problem of insufficient real data to support model training.

[0196] Specifically, the disclosed solution can achieve data enhancement based on the following methods, including: Method 1, using a codebook method to perform data enhancement on real CSI data; Method 2, using a singular value decomposition method to perform data enhancement on real CSI data. Compared with existing data enhancement methods, such as enhancement methods for image data, the disclosed solution fully considers the communication physical characteristics of real CSI data and is therefore more suitable for data enhancement of real CSI data in communication scenarios. Moreover, the data set constructed based on the disclosed solution can support model training, thereby providing data support for improving the accuracy of model training; at the same time, since the disclosed solution can generate a large amount of data based on a small number of real samples, the corresponding cost of obtaining real data is greatly reduced, thereby reducing the cost of model training.

[0197] The following embodiments involve relatively abstract steps such as high-dimensional vector space projection. For accurate expression, mathematical formulas supplemented by corresponding physical explanations are considered to be used as much as possible for explanation.

[0198] Example 1: Using the codebook method to enhance real CSI data:

[0199] For the 5G NR system, the CSI data feedback design mainly uses a codebook-based method to realize the extraction and feedback of channel features. That is, after the channel is estimated at the transmitting end and the corresponding precoding matrix (that is, the real CSI data) is obtained based on the channel estimation result, the coding matrix that best matches the precoding matrix is ​​selected from the pre-set codebook according to a certain optimization criterion, and the index of the coding matrix that best matches the precoding matrix is ​​fed back to the receiving end through the feedback link of the air interface, so that the receiving end can implement precoding. Specifically, the codebook can be divided into three schemes: Type I, Type II, and eType II. The disclosed scheme takes the DFT vector space constructed by eType II as an example for detailed explanation.

[0200] Here, the real CSI data can be represented by the matrix W, that is, the precoding matrix to be fed back by the transmitter is represented as the matrix W∈C Nt×Nsb , where C represents the complex space and Nt represents the number of ports of the transmitting antenna. For example, taking the transmitting antenna of the transmitting end as a two-dimensional array antenna, Nt = 2N1N2, where N1 and N2 represent the number of ports in the first and second dimensions of the two-dimensional array antenna, such as a two-dimensional planar array dual-polarization antenna. Nsb represents the number of frequency domain subbands corresponding to the transmitting end; further, each column of the matrix W represents a vector used for precoding shared by multiple subcarriers on each frequency domain subband. Each row of the matrix W represents a vector corresponding to a port of the transmitting antenna.

[0201] First, the first DFT vector space can be constructed in the following manner. Specifically, generally speaking, the number of orthogonal two-dimensional DFT vectors of length Nt is at most Nt, which can be expressed as:

[0202]

[0203] where b m,n Pairwise orthogonal, c m and p n are the one-dimensional DFT vectors, represents the Kronecker product; c m and p n The expression is as follows:

[0204] c m =[1,...,exp(j2π(N1-1)m) / N1] T ,

[0205] p n =[1,...,exp(j2π(N2-1)m) / N2] T .

[0206] Here, in order to improve the quantization accuracy of the codebook and increase the number of basis vectors in the DFT vector space, oversampling can be used to increase the number of two-dimensional DFT vectors. For example, the oversampling factors of the two-dimensional array antenna in the first dimension and the second dimension are set to O1 and O2 respectively. In this case, the total number of oversampled two-dimensional DFT vectors can be increased to N1O1N2O2, which can be expressed as:

[0207]

[0208] Among them, v m and u n Represent the oversampled one-dimensional DFT vector, represents the Kronecker product, that is:

[0209] v m =[1,...,exp(j2π(N1-1)m) / N1O1] T ,

[0210] u n =[1,...,exp(j2π(N2-1)m) / N2O2] T .

[0211] N1O1N2O2 DFT vectors, that is, a m,n , which is the first DFT vector space constructed by the disclosed solution; further, the first DFT vector space contains O1O2 groups, where the first dimension corresponds to the O1 group and the second dimension corresponds to the O2 group. Each group has a total of N1N2 DFT vectors, and they are orthogonal to each other. The N1N2 vectors in the o1th group (o1=1,...,O1) in the first dimension and the o2th group (o2=1,...,O2) in the second dimension can be expressed as:

[0212]

[0213]

[0214] in,

[0215] v m =[1,...,exp(j2π(N1-1)m) / N1O1] T ,

[0216] u n =[1,...,exp(j2π(N2-1)m) / N2O2] T .

[0217] Secondly, the second DFT vector space can be constructed in the following manner. Here, the construction method is similar to the above method. The second DFT vector space can be expressed as:

[0218] q m =[1,...,exp(j2π(Nsb-1)m) / Nsb] T ,m=0,1,2,...,Nsb-1;

[0219] The second DFT vector space includes Nsb orthogonal basis vectors with a length of Nsb.

[0220] It is understandable that the above method of constructing the first DFT vector space and the second DFT vector space is only an example and is not intended to limit the present disclosure. In actual applications, there may be other construction methods, which are not limited by the present disclosure.

[0221] Furthermore, based on the first DFT vector space and the second DFT vector space, the following method can be used to obtain the matrix W∈C for each real CSI data (also referred to as a real CSI data sample): Nt×Nsb Data enhancement includes the following steps:

[0222] Step a: Select a target orthogonal basis vector group Q containing N1N2 orthogonal basis vectors from the O1O2 group (i.e., O1O2 orthogonal basis vector groups) of the first DFT vector space, and select L orthogonal basis vectors from the selected target orthogonal basis vector group Q to obtain a first basis vector group Q′.

[0223] Here, step a can be divided into the following sub-steps:

[0224] Sub-step a1: Select a group from the O1O2 group and set each element in the group, i.e., the DFT vector a′ m,n Construct the matrix B′ by column, and then construct the diagonal block matrix W′1=[B′,0;0,B′];

[0225] For example, N1=N2=O1=O2=4, and L=4 are used as an example for explanation. As shown in FIG7 , each dot (including solid dots and hollow dots) represents a vector in the DFT vector space, wherein the solid dot can be represented as a DFT vector b without oversampling. m,n , there are 16 non-oversampled DFT vectors; the hollow dots represent oversampled two-dimensional DFT vectors, a total of 15×16; based on this, all the sampled two-dimensional DFT vectors a m,n , a total of N1O1N2O2=256.

[0226] Furthermore, a group of orthogonal basis vectors is selected from the O1O2=16 groups. The wireframe (including the dotted frame and the solid frame) is the selected group of orthogonal basis vectors; and the solid frame is the L orthogonal basis vectors finally selected.

[0227] Sub-step a2: Project the matrix W of the real CSI data into the space spanned by the column vectors of the diagonal block matrix W′1 to obtain the first projection coefficient absolute value matrix (also known as the first projection information matrix mentioned above) here,(·) H represents conjugate transpose, abs(·) represents the absolute value of matrix elements;

[0228] Sub-step a3: Sum the first projection coefficient absolute value matrix (i.e., the first projection information matrix) W′2 by column to obtain the first vector w′2∈R Nt×1 , where R represents the real number space;

[0229] Sub-step a4: Add the first Nt / 2=N1N2 elements and the last Nt / 2=N1N2 elements of the first vector w′2 to obtain the second vector w″2∈R Nt / 2×1 ;

[0230] Sub-step a5: Sort the elements in the second vector w″2, and record the indexes corresponding to the first L maximum value elements as I={i1,i2,...,i L}, the sum of the first L maximum value elements is recorded as c;

[0231] Sub-step a6: Execute sub-steps a1 to a5 for each of the O1O2 groups to obtain O1O2 index sets, denoted as At the same time, we obtain the sum of the first L maximum elements of each group of O1O2, which is recorded as

[0232] Sub-step a7: Take the index p corresponding to the maximum value in C. The group corresponding to index p is the target orthogonal basis vector group Q. The target orthogonal basis vector group Q contains N1N2 basis vectors;

[0233] Sub-step a8: Get the I corresponding to index p all The set I P , according to set I P The indices included in select L orthogonal basis vectors from the target orthogonal basis vector group Q, and the L orthogonal basis vectors selected from the target orthogonal basis vector group Q are used as the first basis vector group Q′.

[0234] In the present application, each real CSI data can be represented by a matrix, such as the matrix W∈C Nt×Nsb The Nsb columns are vectors corresponding to the Nsb frequency domain subbands, and the Nt rows are vectors corresponding to the Nt transmitting antennas, that is, the rows and columns of the matrix W correspond to the spatial domain features and the frequency domain features, respectively.

[0235] Furthermore, step a can be understood as:

[0236] In the first DFT vector space, find L orthogonal basis vectors with the highest average correlation between the vectors corresponding to the Nsb frequency domain subbands. Specifically, first, find a group with the highest average correlation (i.e., the target orthogonal basis vector group) from the O1O2 orthogonal basis vector groups (corresponding to sub-steps a1 to a7). Second, find L orthogonal basis vectors with the highest average correlation between the vectors corresponding to the Nsb frequency domain subbands from the found group (i.e., the target orthogonal basis vector group) (corresponding to sub-step a8). Because the found L orthogonal basis vectors have the highest correlation, they can better represent the spatial characteristics of the actual CSI data.

[0237] Step b: Select M orthogonal basis vectors from the Nsb orthogonal basis vectors in the second DFT vector space to obtain a second basis vector group P.

[0238] Here, step b is divided into the following sub-steps:

[0239] Sub-step b1: Arrange the elements of the first basis vector group Q′ in columns to form a matrix B, thereby forming a first basis vector matrix W1 = [B, 0; 0, B];

[0240] Sub-step b2: Project the matrix W of the real CSI data into the space spanned by the column vectors of the first basis vector matrix W1 to obtain the first projection coefficient matrix W 2temp =W1 H W∈C 2L×Nsb ;

[0241] Sub-step b3: Sub-projection coefficient matrix W 2temp , projected into the space spanned by the Nsb orthogonal basis vectors of the second DFT vector space, and the second projection coefficient absolute value matrix (also known as the second projection information matrix) is obtained. Among them, W f ' is the matrix composed of Nsb basis vectors of the second DFT vector space arranged in columns, (·) T represents the non-conjugate transpose;

[0242] Sub-step b4: Sum the elements of the second projection coefficient absolute value matrix W″2 by column to obtain the third vector w″′2∈R Nsb×1 ;

[0243] Sub-step b5: Take the index corresponding to the first M maximum value elements in the third vector w″′2 and record it as J={j1,...,j M}, based on the index contained in J, select M orthogonal basis vectors from the second DFT vector space to obtain a second basis vector group P.

[0244] Here, step b can be understood as: the first projection coefficient matrix W of the real CSI data on the first basis vector matrix W1 2temp =W1 H W∈C 2L×Nsb Each column in represents the projection coefficient of the frequency domain subband (i.e., spatial domain feature) corresponding to the real CSI data on the first basis vector matrix W1, and the frequency domain feature of the real CSI data still exists between each column. Further, similar to step a, step b searches for the first projection coefficient matrix W from the second DFT vector space. 2temp Here, since the M orthogonal basis vectors found are the basis vectors with the highest correlation, the M orthogonal basis vectors can better represent the frequency domain characteristics of the real CSI data.

[0245] Step c: The first projection coefficient matrix W 2temp Projected onto the space spanned by the M orthogonal basis vectors (i.e., the second basis vector matrix) of the second basis vector group P, the second projection coefficient matrix W2=W is obtained. f H W 2temp T ∈C Nsb×2L , where W f is the second basis vector matrix, that is, a matrix formed by arranging the M orthogonal basis vectors of the second vector group P in columns.

[0246] Here, step c can be understood as projecting the actual CSI data onto the space spanned by the column vectors of the first basis vector matrix and the column vectors of the second basis vector matrix to obtain projection coefficients. For example, the matrix W of the actual CSI data is left-multiplied by the conjugate transpose of the first basis vector matrix and then right-multiplied by the conjugate transpose of the second basis vector matrix to obtain the second projection coefficient matrix.

[0247] Step d: Scramble the amplitude or phase of the elements in the second projection coefficient matrix W2 to obtain multiple third projection coefficient matrices W 2aug ∈C Nsb×2L ; For example, by performing g scrambling, g third projection coefficient matrices can be obtained;

[0248] Furthermore, the scrambling process can be performed in the following manner:

[0249] Method 1: Amplitude perturbation. Each element in W2 is multiplied by a random number λ1, where λ1 follows a uniform distribution from U1 to [1-ε, 1-ε]. In practice, this method resamples each element to obtain a random number λ1. In this case, each element is multiplied by a different random number λ1.

[0250] Method 2: Phase perturbation. Each element in W2 is multiplied by a random number λ2, where λ2 is sampled from the interval exp(αj), and α follows a uniform distribution from U2 to [-θ, θ]. In practice, this method can resample each element to obtain a random number λ2. In this case, each element is multiplied by a different random number λ2.

[0251] It is understandable that, in actual applications, either of the above two methods can be used to implement scrambling processing, or both methods can be used to implement scrambling processing together, and the present disclosure does not limit this.

[0252] Here, step d can be understood as: scrambling from the amplitude dimension or the phase dimension with each coefficient in the second projection coefficient matrix as the center; further, it can be understood as: the second projection coefficient matrix is ​​a point in the space spanned by the column vectors of the first basis vector matrix or the second basis vector matrix. At this time, after amplitude and phase scrambling, it can be expanded from a point to a sphere.

[0253] Step e: Combine each third projection coefficient matrix with the first basis vector matrix W1 and the second basis vector matrix W f Multiply to obtain the unnormalized enhanced sample W aug,raw =W1W 2aug T W f H ;

[0254] Here, step e can be understood as: multiplying the coefficient corresponding to each point in the sphere obtained in step d by the first basis vector matrix and the second basis vector matrix. Since the first basis vector matrix and the second basis vector matrix can better represent the spatial domain characteristics and frequency domain characteristics of the real CSI data, step e achieves data amplification while retaining the spatial domain characteristics and frequency domain characteristics of the real CSI data.

[0255] Step f: Unnormalized enhanced sample W aug,raw Normalize by frequency domain subband to obtain the final enhanced sample W aug =[w aug,1 / norm(w aug,1 ),...,w aug,Nsb / norm(w aug,Nsb ], where norm(·) represents the second norm, w aug,t , t=1,...,Nsb represents W aug Different column vectors in .

[0256] Therefore, compared with existing image data enhancement methods in the computer field, the disclosed solution fully considers the communication physical characteristics of real CSI data. Therefore, the disclosed solution is more suitable for data enhancement of CSI data.

[0257] Furthermore, the codebook method of the disclosed solution can also be used to construct a sample set to be trained on the model, thereby effectively supporting model training. Here, since a large amount of sample data for model training can be obtained using only a small amount of real data, the manpower, material resources and other costs of obtaining real data can be greatly reduced, thereby reducing the cost of model training. At the same time, it also lays the foundation for improving the efficiency and accuracy of model processing.

[0258] Example 2: Data enhancement using singular value decomposition:

[0259] The embodiment of the present application further considers using the singular value decomposition method to perform data enhancement. In general, the singular vector matrix after the singular value decomposition of the sample can better represent the physical properties corresponding to the matrix of the original CSI data. Based on this, taking the singular vector matrix as the starting point can achieve data enhancement while retaining the corresponding physical properties. Considering that the existing AI-based CSI data feedback data set contains K CSI data (that is, K real CSI data samples), each real CSI data is represented by the matrix W∈C Nt×Nsb , where C represents the complex space and Nt represents the number of ports of the transmitting antenna. For example, taking the transmitting antenna of the transmitting end as a two-dimensional array antenna, Nt = 2N1N2, where N1 and N2 represent the number of ports in the first and second dimensions of the two-dimensional array antenna, such as a two-dimensional planar array dual-polarization antenna. Nsb represents the number of frequency domain subbands corresponding to the transmitting end; further, each column of the matrix W represents a vector used for precoding shared by multiple subcarriers on each frequency domain subband. Each row of the matrix W represents a vector corresponding to a port of the transmitting antenna.

[0260] Furthermore, for each real CSI data, data enhancement using the singular value decomposition method can be specifically divided into the following steps.

[0261] Step a: Obtain the first autocorrelation matrix R1 = WW of the real CSI data H ∈C Nt×Nt ;

[0262] Note 8: R1 retains the original CSI data, that is, the spatial statistical characteristics of the matrix of the real CSI data.

[0263] Step b: Obtain the second autocorrelation matrix R2 = W of the real CSI data H W∈C Nsb×Nsb ;

[0264] Note 9: R2 retains the original CSI data, that is, the frequency domain statistical characteristics of the matrix of the real CSI data.

[0265] Step c: Obtain the target autocorrelation matrix of the real CSI data in represents the Kronecker product;

[0266] Note 10: The target autocorrelation matrix R3 retains the original CSI data, that is, the spatial and frequency domain statistical characteristics of the matrix of the real CSI data.

[0267] Step d: Perform singular value decomposition on the target autocorrelation matrix, [V, D, U] = svd(R3), where V∈C NsbNt×NsbNt With U∈C NsbNt×NsbNt They are the left singular vector matrix, the right singular vector matrix, and the diagonal matrix D∈C NsbNt×NsbNt is a singular value matrix, where the diagonal elements are singular values;

[0268] Note 11: Since R3 is a symmetric matrix, V = U H Moreover, the multiple singular vectors in V retain the original CSI data, that is, the spatial and frequency domain statistical characteristics of the matrix of the real CSI data.

[0269] Step e: Construct the target random matrix U′∈C NsbNt×S , where each element in the target random matrix U′ obeys the complex normal distribution, S represents the number of enhanced samples; further, the enhanced sample set matrix (also known as the enhanced data matrix) W is obtained aug,all :

[0270] W aug,all =Vsqrt(D)U′=[w aug,1 ′,w aug,2 ′,...,w aug,S ′]∈C NsbNt×S ;

[0271] Among them, w aug,d ′∈C NsbNt×1 (d=1, 2, ..., S) represents each enhanced sample (i.e., the second CSI data), and sqrt(·) represents the square root of each element of the matrix;

[0272] Note 12: Since the enhancement is based on the singular vectors in V, the enhanced samples still retain the original CSI data, that is, the spatial and frequency domain statistical characteristics of the real CSI data.

[0273] Step f: For each column w in the enhanced sample set matrix aug,d ′∈C NsbNt×1The elements of (d=1,2,...,S) are rearranged into the original CSI data, that is, the dimension of the real CSI data, to obtain the enhanced sample W aug,d ∈C Nt×Nsb (d=1,2,...,S).

[0274] Step g: Normalize the unnormalized enhanced samples according to the frequency domain subband to obtain the final enhanced samples Where norm(·) represents the second norm, w aug,d,t (t=1,2,...,Nsb) represents W aug,d Different column vectors in .

[0275] Compared to existing computer science methods for image data enhancement, this method leverages the physical properties of CSI data, making it more suitable for CSI data enhancement. Furthermore, the dataset constructed using singular value decomposition can support model training. Furthermore, since only a small number of real samples are used, the human and material costs of acquiring real data can be significantly reduced.

[0276] FIG8 is a schematic block diagram of a data processing device 800 according to an embodiment of the present application. The terminal device 800 may include:

[0277] The data enhancement processing unit 810 is configured to perform data enhancement on the first channel state information (CSI) data based on characteristic information of the first CSI data to obtain multiple second CSI data; the characteristic information includes at least one of the following: spatial domain characteristics and / or frequency domain characteristics;

[0278] The sample data determining unit 820 is configured to use at least the first CSI data and the multiple second CSI data as CSI sample data.

[0279] In a specific example of the disclosed solution, the data enhancement processing unit is further configured to:

[0280] Determining, based on a first discrete Fourier transform (DFT) vector space constructed by a preset codebook, a first basis vector group representing spatial domain characteristics of the first CSI data;

[0281] Based at least on the first basis vector group representing the spatial characteristics of the first CSI data, data enhancement is performed on the first CSI data to obtain multiple second CSI data.

[0282] In a specific example of the disclosed solution, the data enhancement processing unit is specifically configured to:

[0283] Selecting L orthogonal basis vectors from a group of orthogonal basis vectors included in the first DFT vector space, where correlations between the L orthogonal basis vectors and the spatial domain features of the first CSI data meet a first preset condition, and L is a natural number greater than or equal to 1;

[0284] The first basis vector group is obtained based on the selected L orthogonal basis vectors.

[0285] In a specific example of the disclosed solution, the data enhancement processing unit is further configured to:

[0286] Selecting, from a plurality of orthogonal basis vector groups included in the first DFT vector space, a target orthogonal basis vector group whose correlation with the spatial domain feature of the first CSI data satisfies a second preset condition;

[0287] The L orthogonal basis vectors are selected from the target orthogonal basis vector group.

[0288] In a specific example of the disclosed solution, the data enhancement processing unit is specifically configured to:

[0289] Projecting the first CSI data into a vector space spanned by a diagonal block matrix to obtain a first projection information matrix corresponding to each of the orthogonal basis vector groups, wherein the diagonal block matrix is ​​obtained based on the orthogonal basis vector group included in the first DFT vector space; elements in the first projection information matrix represent information related to projection coefficients after spatial features of the first CSI data are projected onto the diagonal block matrix;

[0290] Based on the first projection information matrix corresponding to each of the orthogonal basis vector groups, a target orthogonal basis vector group whose correlation with the spatial domain features of the first CSI data meets a second preset condition is selected from the multiple orthogonal basis vector groups.

[0291] In a specific example of the disclosed solution, the data enhancement processing unit is further configured to:

[0292] Performing data processing on elements in the first projection information matrix to obtain a first vector corresponding to the orthogonal basis vector group, wherein the first vector can represent total relevant information of the projection coefficient corresponding to the frequency domain characteristics of the first CSI data;

[0293] performing data processing on some elements of the first vector and another part of the elements of the first vector to obtain a second vector corresponding to the orthogonal basis vector group, wherein the second vector can represent total relevant information of the projection coefficients corresponding to the spatial characteristics of the first CSI data;

[0294] Selecting L first target elements corresponding to the orthogonal basis vector group from the second vector;

[0295] The target orthogonal basis vector group is selected from the multiple orthogonal basis vector groups based on the L first target elements corresponding to each of the orthogonal basis vector groups.

[0296] In a specific example of the disclosed solution, the data enhancement processing unit is specifically configured to:

[0297] Based on the L first target elements corresponding to each of the orthogonal basis vector groups, obtaining a target sum corresponding to each of the orthogonal basis vector groups;

[0298] A target value is selected from the multiple target sums, and an orthogonal basis vector group corresponding to the selected target value is used as the target orthogonal basis vector group.

[0299] In a specific example of the disclosed solution, the data enhancement processing unit is further configured to:

[0300] Determining, based on the second DFT vector space constructed by the preset codebook and the first basis vector group representing the spatial domain characteristics of the first CSI data, a second basis vector group representing at least the frequency domain characteristics of the first CSI data;

[0301] Based on the first basis vector group and the second basis vector group, data enhancement is performed on the first CSI data to obtain multiple second CSI data.

[0302] In a specific example of the disclosed solution, the data enhancement processing unit is specifically configured to:

[0303] Projecting the first CSI data into the space spanned by the first basis vector group to obtain a first projection coefficient matrix; wherein elements in the first projection coefficient matrix represent projection coefficients of the spatial domain features of the first CSI data in the space spanned by the first basis vector group;

[0304] Based on the first projection coefficient matrix, select M orthogonal basis vectors from the orthogonal basis vectors included in the second DFT vector space constructed by the preset codebook, where correlations between the M orthogonal basis vectors and the projection coefficients corresponding to the frequency domain features of the first CSI data in the first projection coefficient matrix meet a third preset condition; where M is a natural number greater than or equal to 1;

[0305] The second basis vector group is obtained based on the selected M orthogonal basis vectors.

[0306] In a specific example of the disclosed solution, the data enhancement processing unit is specifically configured to:

[0307] The first CSI data is projected into a space spanned by the first basis vector matrix, wherein the first basis vector matrix is ​​a diagonal block matrix formed based on the first basis vector group.

[0308] In a specific example of the disclosed solution, the data enhancement processing unit is further configured to:

[0309] Projecting the first projection coefficient matrix into a space spanned by orthogonal basis vectors of the second DFT vector space, obtaining a second projection information matrix; elements in the second projection information matrix represent information related to projection coefficients of frequency domain features of the first CSI data corresponding to the first projection coefficient matrix in the space spanned by orthogonal basis vectors of the second DFT vector space;

[0310] Based on the second projection information matrix, M orthogonal basis vectors are selected from the orthogonal basis vectors included in the second DFT vector space.

[0311] In a specific example of the disclosed solution, the data enhancement processing unit is specifically configured to:

[0312] Performing data processing on elements in the second projection information matrix to obtain a third vector, wherein the third vector represents total relevant information of projection coefficients corresponding to the frequency domain characteristics of the first CSI data;

[0313] Select M second target elements from the third vector;

[0314] Orthogonal basis vectors corresponding to the M second target elements are selected from the second DFT vector space to obtain the M orthogonal basis vectors.

[0315] In a specific example of the disclosed solution, the data enhancement processing unit is further configured to:

[0316] Projecting the first projection coefficient matrix into the space spanned by the second basis vector group to obtain a second projection coefficient matrix, wherein elements in the second projection coefficient matrix represent projection coefficients corresponding to the spatial features of the first CSI data in the space spanned by the second basis vector group;

[0317] Based on the second projection coefficient matrix, the first basis vector matrix and the second basis vector matrix, multiple second CSI data are obtained; wherein the first basis vector matrix is ​​based on the first basis vector group, and the second basis vector matrix is ​​a matrix formed based on the second basis vector group.

[0318] In a specific example of the disclosed solution, the data enhancement processing unit is further configured to:

[0319] performing phase adjustment and / or amplitude adjustment on elements in the second projection coefficient matrix to obtain a plurality of third projection coefficient matrices;

[0320] A plurality of second CSI data are obtained based on the plurality of third projection coefficient matrices, the first basis vector matrix, and the second basis vector matrix.

[0321] In a specific example of the disclosed solution, the data enhancement processing unit is specifically configured to:

[0322] Obtaining a plurality of vectors to be processed based on a matrix product of the third projection coefficient matrix, the first basis vector matrix, and the second basis vector matrix;

[0323] Normalization is performed on the multiple vectors to be processed to obtain multiple second CSI data.

[0324] In a specific example of the disclosed solution, the data enhancement processing unit is specifically configured to:

[0325] Taking an element in the second projection coefficient matrix as the center, the center is adjusted in terms of phase and / or amplitude.

[0326] In a specific example of the disclosed solution, the data enhancement processing unit is further configured to:

[0327] Obtaining a target autocorrelation matrix of the first CSI data, where the target autocorrelation matrix can represent spatial domain characteristics and frequency domain characteristics of the first CSI data;

[0328] Performing singular value decomposition on the target autocorrelation matrix to obtain a singular value decomposition result, wherein the singular value decomposition result includes a singular vector matrix representing spatial domain characteristics and frequency domain characteristics of the first CSI data, and a singular value matrix;

[0329] The singular value decomposition result is processed using the constructed target random matrix to obtain a plurality of second CSI data after data enhancement.

[0330] In a specific example of the disclosed solution, the data enhancement processing unit is further configured to:

[0331] Obtaining a first autocorrelation matrix of the first CSI data, where the first autocorrelation matrix can represent spatial characteristics of the first CSI data;

[0332] Obtaining a second autocorrelation matrix of the first CSI data, wherein the second autocorrelation matrix can represent frequency domain characteristics of the first CSI data;

[0333] The target autocorrelation matrix is ​​obtained based on the first autocorrelation matrix and the second autocorrelation matrix.

[0334] In a specific example of the disclosed solution, the data enhancement processing unit is specifically configured to:

[0335] Using the target random matrix, replacing the first matrix in the singular vector matrix;

[0336] The target random matrix, the second matrix in the singular vector matrix, and the singular value matrix are subjected to matrix product processing to obtain multiple second CSI data; wherein the target autocorrelation matrix is ​​a symmetric matrix, and the first matrix and the second matrix satisfy a symmetric relationship.

[0337] In a specific example of the disclosed solution, the data enhancement processing unit is specifically configured to:

[0338] Performing matrix product processing on the target random matrix, the second matrix in the singular vector matrix, and the singular value matrix to obtain an enhanced data matrix;

[0339] The enhanced data matrix is ​​arranged based on the vector dimension of the first CSI data to obtain a plurality of second CSI data after data enhancement.

[0340] In a specific example of the disclosed solution, the data enhancement processing unit is specifically configured to:

[0341] Arrange the enhanced data matrix based on the vector dimension of the first CSI data to obtain an arranged enhanced data matrix;

[0342] The arranged enhanced data matrix is ​​normalized to obtain a plurality of second CSI data.

[0343] FIG9 is a second schematic block diagram of a data processing device 800 according to an embodiment of the present application. The data processing device 800 may include:

[0344] A model training unit 910 is configured to perform model training on a first preset model based on the CSI sample data to obtain a first target model, wherein the first target model is used to encode the channel state information CSI data and obtain target CSI data; and / or,

[0345] Based on the CSI sample data, a second preset model is trained to obtain a second target model, wherein the second target model is used to decode the target CSI data to obtain the CSI corresponding to the target CSI data.

[0346] FIG10 is a schematic block diagram of a terminal device 1000 according to an embodiment of the present application. The terminal device 1000 may include:

[0347] The sending unit 1010 is used to send the first information, which is obtained by inputting the channel state information CSI data into the first target model for encoding processing; the first target model is obtained by training the first preset model based on the CSI sample data obtained by the method described above in the embodiment of the present application.

[0348] The terminal device 1000 of the embodiment of the present application can implement the corresponding functions of the terminal device in the aforementioned method embodiment. The processes, functions, implementation methods and beneficial effects corresponding to the various modules (sub-modules, units or components, etc.) in the terminal device 1000 can be found in the corresponding descriptions in the above-mentioned method embodiments, which will not be repeated here. It should be noted that the functions described in the various modules (sub-modules, units or components, etc.) in the terminal device 1000 of the embodiment of the application can be implemented by different modules (sub-modules, units or components, etc.) or by the same module (sub-module, unit or component, etc.).

[0349] FIG11 is a schematic block diagram of a network device 1100 according to an embodiment of the present application. The network device 1100 may include:

[0350] The receiving unit 1110 is used to receive the first information, input the first information into the second target model for decoding processing, and obtain the CSI data corresponding to the first information; wherein, the second target model is obtained by training the second preset model based on the CSI sample data obtained by the method in the embodiment of the present application.

[0351] The network device 1100 of the embodiment of the present application can implement the corresponding functions of the network device in the aforementioned method embodiment. The processes, functions, implementation methods and beneficial effects corresponding to each module (sub-module, unit or component, etc.) in the network device 1100 can be found in the corresponding description in the above method embodiment, and will not be repeated here. It should be noted that the functions described in the various modules (sub-module, unit or component, etc.) in the network device 1100 of the embodiment of the application can be implemented by different modules (sub-module, unit or component, etc.) or by the same module (sub-module, unit or component, etc.).

[0352] Figure 12 is a schematic structural diagram of a communication device 1200 according to an embodiment of the present application. The communication device 600 includes a processor 1210, which can call and execute a computer program from a memory to enable the communication device 1200 to implement the method in the embodiment of the present application.

[0353] In one embodiment, the communication device 1200 may further include a memory 1220. The processor 1210 may call and execute a computer program from the memory 1220 to enable the communication device 1200 to implement the method in the embodiment of the present application.

[0354] The memory 1220 may be a separate device independent of the processor 1210 , or may be integrated into the processor 1210 .

[0355] In one embodiment, the communication device 1200 may further include a transceiver 1230 , and the processor 1210 may control the transceiver 1230 to communicate with other devices. Specifically, the transceiver 1230 may send information or data to other devices, or receive information or data sent by other devices.

[0356] The transceiver 1230 may include a transmitter and a receiver. The transceiver 1230 may further include an antenna, and the number of antennas may be one or more.

[0357] In one embodiment, the communication device 1200 may be a network device of an embodiment of the present application, and the communication device 1200 may implement the corresponding processes implemented by the network device in each method of the embodiment of the present application. For the sake of brevity, they will not be repeated here.

[0358] In one embodiment, the communication device 1200 may be a terminal device of an embodiment of the present application, and the communication device 1200 may implement the corresponding processes implemented by the terminal device in each method of the embodiment of the present application. For the sake of brevity, they will not be repeated here.

[0359] 13 is a schematic structural diagram of a chip 1300 according to an embodiment of the present application. The chip 1300 includes a processor 1310, which can call and execute a computer program from a memory to implement the method according to the embodiment of the present application.

[0360] In one embodiment, the chip 1300 may further include a memory 1320. The processor 1310 may call and execute a computer program from the memory 1320 to implement the method executed by the terminal device or the network device in the embodiment of the present application.

[0361] The memory 1320 may be a separate device independent of the processor 1310 , or may be integrated into the processor 1310 .

[0362] In one embodiment, the chip 1300 may further include an input interface 1330. The processor 1310 may control the input interface 1330 to communicate with other devices or chips, and specifically, may obtain information or data sent by other devices or chips.

[0363] In one embodiment, the chip 1300 may further include an output interface 1340. The processor 1310 may control the output interface 1340 to communicate with other devices or chips, and specifically, may output information or data to other devices or chips.

[0364] In one embodiment, the chip can be applied to the network device in the embodiments of the present application, and the chip can implement the corresponding processes implemented by the network device in each method of the embodiments of the present application. For the sake of brevity, they will not be repeated here.

[0365] In one embodiment, the chip can be applied to the terminal device in the embodiments of the present application, and the chip can implement the corresponding processes implemented by the terminal device in each method of the embodiments of the present application. For the sake of brevity, they will not be repeated here.

[0366] The chips used in the network device and the terminal device may be the same chip or different chips.

[0367] It should be understood that the chip mentioned in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.

[0368] The processor mentioned above may be a general-purpose processor, a digital signal processor (DSP), a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), or other programmable logic devices, transistor logic devices, discrete hardware components, etc. The general-purpose processor mentioned above may be a microprocessor or any conventional processor, etc.

[0369] The memory mentioned above may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. The non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM).

[0370] It should be understood that the above-mentioned memories are exemplary but not restrictive. For example, the memories in the embodiments of the present application may also be static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct RAM RAM (DR RAM), etc. In other words, the memories in the embodiments of the present application are intended to include, but are not limited to, these and any other suitable types of memories.

[0371] FIG14 is a schematic block diagram of a communication system 1400 according to an embodiment of the present application. The communication system 800 includes a terminal device 1410 and a network device 1420.

[0372] The terminal device 1410 sends the first information, which is obtained by inputting the channel state information CSI data into the first target model for encoding processing; the first target model is obtained by training the first preset model based on the CSI sample data obtained by the above method.

[0373] The network device 1420 receives the first information, inputs the first information into the second target model for decoding processing, and obtains the CSI data corresponding to the first information; wherein, the second target model is obtained by training the second preset model based on the CSI sample data obtained by the above method.

[0374] The terminal device 1410 can be used to implement the corresponding functions implemented by the terminal device in the above method, and the network device 1420 can be used to implement the corresponding functions implemented by the network device in the above method. For the sake of brevity, they are not described here in detail.

[0375] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When software is used for implementation, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function in accordance with the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website, computer, server or data center by wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) mode to another website, computer, server or data center. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrations. The available medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a DVD), or a semiconductor medium (eg, a solid state disk (SSD)).

[0376] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0377] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0378] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present application should be included within the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A data processing method, comprising: Based on the characteristic information of the first channel state information CSI data, the first CSI data is enhanced to obtain a plurality of second CSI data; The feature information includes at least one of the following: spatial domain features and / or frequency domain features; At least the first CSI data and the multiple second CSI data are used as CSI sample data.

2. The method according to claim 1, further comprising: Determining, based on a first discrete Fourier transform (DFT) vector space constructed by a preset codebook, a first basis vector group representing spatial domain characteristics of the first CSI data; The performing data enhancement on the first CSI data based on the feature information of the first CSI data to obtain the plurality of second CSI data includes: Based at least on the first basis vector group representing the spatial characteristics of the first CSI data, data enhancement is performed on the first CSI data to obtain multiple second CSI data.

3. The method according to claim 2, wherein: The first DFT vector space constructed based on the preset codebook determines a first basis vector group representing spatial characteristics of the first CSI data, including: Selecting L orthogonal basis vectors from a group of orthogonal basis vectors included in the first DFT vector space, where correlations between the L orthogonal basis vectors and the spatial domain features of the first CSI data meet a first preset condition, and L is a natural number greater than or equal to 1; The first basis vector group is obtained based on the selected L orthogonal basis vectors.

4. The method according to claim 3, further comprising: Selecting, from a plurality of orthogonal basis vector groups included in the first DFT vector space, a target orthogonal basis vector group whose correlation with the spatial domain feature of the first CSI data satisfies a second preset condition; The step of selecting L orthogonal basis vectors from the orthogonal basis vector group contained in the first DFT vector space includes: The L orthogonal basis vectors are selected from the target orthogonal basis vector group.

5. The method according to claim 4, wherein The selecting, from a plurality of orthogonal basis vector groups included in the first DFT vector space, a target orthogonal basis vector group whose correlation with the spatial domain feature of the first CSI data satisfies a second preset condition, includes: Projecting the first CSI data into a vector space spanned by a diagonal block matrix to obtain a first projection information matrix corresponding to each of the orthogonal basis vector groups, wherein the diagonal block matrix is ​​obtained based on the orthogonal basis vector group included in the first DFT vector space; elements in the first projection information matrix represent information related to projection coefficients after spatial features of the first CSI data are projected onto the diagonal block matrix; Based on the first projection information matrix corresponding to each of the orthogonal basis vector groups, a target orthogonal basis vector group whose correlation with the spatial domain features of the first CSI data meets a second preset condition is selected from the multiple orthogonal basis vector groups.

6. The method according to claim 5, further comprising: Performing data processing on elements in the first projection information matrix to obtain a first vector corresponding to the orthogonal basis vector group, wherein the first vector can represent total relevant information of the projection coefficient corresponding to the frequency domain characteristics of the first CSI data; performing data processing on some elements of the first vector and another part of the elements of the first vector to obtain a second vector corresponding to the orthogonal basis vector group, wherein the second vector can represent total relevant information of the projection coefficients corresponding to the spatial characteristics of the first CSI data; Selecting L first target elements corresponding to the orthogonal basis vector group from the second vector; The selecting, based on the first projection information matrix corresponding to each of the orthogonal basis vector groups, a target orthogonal basis vector group whose correlation with the spatial domain features of the first CSI data satisfies a second preset condition from the multiple orthogonal basis vector groups includes: The target orthogonal basis vector group is selected from the multiple orthogonal basis vector groups based on the L first target elements corresponding to each of the orthogonal basis vector groups.

7. The method according to claim 6, wherein: The step of selecting the target orthogonal basis vector group from the plurality of orthogonal basis vector groups based on the L first target elements corresponding to each of the orthogonal basis vector groups includes: Based on the L first target elements corresponding to each of the orthogonal basis vector groups, obtaining a target sum corresponding to each of the orthogonal basis vector groups; A target value is selected from the multiple target sums, and an orthogonal basis vector group corresponding to the selected target value is used as the target orthogonal basis vector group.

8. The method according to any one of claims 2 to 7, further comprising: Determining, based on the second DFT vector space constructed by the preset codebook and the first basis vector group representing the spatial domain characteristics of the first CSI data, a second basis vector group representing at least the frequency domain characteristics of the first CSI data; The performing data enhancement on the first CSI data based at least on the first basis vector group representing the spatial characteristics of the first CSI data to obtain the plurality of second CSI data includes: Based on the first basis vector group and the second basis vector group, data enhancement is performed on the first CSI data to obtain multiple second CSI data.

9. The method according to claim 8, wherein The determining, based on the second DFT vector space constructed by the preset codebook and the first basis vector group representing the spatial domain characteristics of the first CSI data, a second basis vector group representing at least the frequency domain characteristics of the first CSI data includes: Projecting the first CSI data into the space spanned by the first basis vector group to obtain a first projection coefficient matrix; wherein elements in the first projection coefficient matrix represent projection coefficients of the spatial domain features of the first CSI data in the space spanned by the first basis vector group; Based on the first projection coefficient matrix, select M orthogonal basis vectors from the orthogonal basis vectors included in the second DFT vector space constructed by the preset codebook, where correlations between the M orthogonal basis vectors and the projection coefficients corresponding to the frequency domain features of the first CSI data in the first projection coefficient matrix meet a third preset condition; where M is a natural number greater than or equal to 1; The second basis vector group is obtained based on the selected M orthogonal basis vectors.

10. The method according to claim 9, wherein: The projecting the first CSI data into a space spanned by the first basis vector group includes: The first CSI data is projected into a space spanned by the first basis vector matrix, wherein the first basis vector matrix is ​​a diagonal block matrix formed based on the first basis vector group.

11. The method according to claim 9 or 10, further comprising: Projecting the first projection coefficient matrix into the space spanned by the orthogonal basis vectors of the second DFT vector space to obtain a second projection information matrix; The elements in the second projection information matrix represent relevant information of projection coefficients of frequency domain features of the first CSI data corresponding to the first projection coefficient matrix in the space spanned by orthogonal basis vectors of the second DFT vector space; The selecting, based on the first projection coefficient matrix, M orthogonal basis vectors from the orthogonal basis vectors contained in the second DFT vector space constructed by the preset codebook includes: Based on the second projection information matrix, M orthogonal basis vectors are selected from the orthogonal basis vectors included in the second DFT vector space.

12. The method according to claim 11, wherein The selecting, based on the second projection information matrix, M orthogonal basis vectors from the orthogonal basis vectors contained in the second DFT vector space includes: Performing data processing on elements in the second projection information matrix to obtain a third vector, wherein the third vector represents total relevant information of projection coefficients corresponding to the frequency domain characteristics of the first CSI data; Select M second target elements from the third vector; Orthogonal basis vectors corresponding to the M second target elements are selected from the second DFT vector space to obtain the M orthogonal basis vectors.

13. The method according to any one of claims 9 to 12, further comprising: Projecting the first projection coefficient matrix into the space spanned by the second basis vector group to obtain a second projection coefficient matrix, wherein elements in the second projection coefficient matrix represent projection coefficients corresponding to the spatial features of the first CSI data in the space spanned by the second basis vector group; The performing data enhancement on the first CSI data based on the first basis vector group and the second basis vector group to obtain multiple second CSI data includes: Based on the second projection coefficient matrix, the first basis vector matrix and the second basis vector matrix, multiple second CSI data are obtained; wherein the first basis vector matrix is ​​based on the first basis vector group, and the second basis vector matrix is ​​a matrix formed based on the second basis vector group.

14. The method according to claim 13, further comprising: performing phase adjustment and / or amplitude adjustment on elements in the second projection coefficient matrix to obtain a plurality of third projection coefficient matrices; The obtaining of the plurality of second CSI data based on the second projection coefficient matrix, the first basis vector matrix, and the second basis vector matrix includes: A plurality of second CSI data are obtained based on the plurality of third projection coefficient matrices, the first basis vector matrix, and the second basis vector matrix.

15. The method according to claim 14, wherein The obtaining, based on the plurality of third projection coefficient matrices, the first basis vector matrix, and the second basis vector matrix, a plurality of second CSI data, includes: Obtaining a plurality of vectors to be processed based on a matrix product of the third projection coefficient matrix, the first basis vector matrix, and the second basis vector matrix; Normalization is performed on the multiple vectors to be processed to obtain multiple second CSI data.

16. The method according to claim 14 or 15, wherein: The performing phase or amplitude adjustment on the elements in the second projection coefficient matrix includes: Taking an element in the second projection coefficient matrix as the center, the center is adjusted in terms of phase and / or amplitude.

17. The method according to any one of claims 1 to 16, further comprising: Obtaining a target autocorrelation matrix of the first CSI data, where the target autocorrelation matrix can represent spatial domain characteristics and frequency domain characteristics of the first CSI data; Performing singular value decomposition on the target autocorrelation matrix to obtain a singular value decomposition result, wherein the singular value decomposition result includes a singular vector matrix representing spatial domain characteristics and frequency domain characteristics of the first CSI data, and a singular value matrix; The performing data enhancement on the first CSI data based on the feature information of the first CSI data to obtain the plurality of second CSI data includes: The singular value decomposition result is processed using the constructed target random matrix to obtain a plurality of second CSI data after data enhancement.

18. The method according to claim 17, further comprising: Obtaining a first autocorrelation matrix of the first CSI data, where the first autocorrelation matrix can represent spatial characteristics of the first CSI data; Obtaining a second autocorrelation matrix of the first CSI data, wherein the second autocorrelation matrix can represent frequency domain characteristics of the first CSI data; The obtaining of a target autocorrelation matrix of the first CSI data includes: The target autocorrelation matrix is ​​obtained based on the first autocorrelation matrix and the second autocorrelation matrix.

19. The method according to claim 17 or 18, wherein The processing of the singular value decomposition result by using the target random matrix to obtain a plurality of second CSI data after data enhancement includes: Using the target random matrix, replacing the first matrix in the singular vector matrix; The target random matrix, the second matrix in the singular vector matrix, and the singular value matrix are subjected to matrix product processing to obtain multiple second CSI data; wherein the target autocorrelation matrix is ​​a symmetric matrix, and the first matrix and the second matrix satisfy a symmetric relationship.

20. The method according to claim 19, wherein The performing matrix product processing on the target random matrix, the second matrix in the singular vector matrix, and the singular value matrix to obtain a plurality of second CSI data includes: Performing matrix product processing on the target random matrix, the second matrix in the singular vector matrix, and the singular value matrix to obtain an enhanced data matrix; The enhanced data matrix is ​​arranged based on the vector dimension of the first CSI data to obtain a plurality of second CSI data after data enhancement.

21. The method according to claim 20, wherein The arranging the enhanced data matrix based on the vector dimension of the first CSI data to obtain a plurality of second CSI data after data enhancement includes: Arrange the enhanced data matrix based on the vector dimension of the first CSI data to obtain an arranged enhanced data matrix; The arranged enhanced data matrix is ​​normalized to obtain a plurality of second CSI data.

22. The method according to any one of claims 1 to 21, further comprising: Based on the CSI sample data, a first preset model is trained to obtain a first target model, wherein the first target model is used to encode the channel state information (CSI) data and obtain target CSI data; and / or, Based on the CSI sample data, a second preset model is trained to obtain a second target model, wherein the second target model is used to decode the target CSI data to obtain the CSI corresponding to the target CSI data.

23. A communication method, comprising: The terminal device sends first information, where the first information is obtained by inputting channel state information (CSI) data into a first target model for encoding processing; The first target model is obtained by training a first preset model based on the CSI sample data obtained according to any one of claims 1 to 21.

24. A communication method, comprising: The network device receives the first information, inputs the first information into the second target model for decoding processing, and obtains the CSI data corresponding to the first information; wherein, the second target model is obtained by training the second preset model based on the CSI sample data obtained according to any one of claims 1 to 21.

25. A data processing device comprising: a data enhancement processing unit, configured to perform data enhancement on the first channel state information (CSI) data based on feature information of the first CSI data to obtain a plurality of second CSI data; The feature information includes at least one of the following: spatial domain features and / or frequency domain features; The sample data determining unit is configured to use at least the first CSI data and the plurality of second CSI data as CSI sample data.

26. The device according to claim 25, wherein The data enhancement processing unit is further used to: Determining, based on a first discrete Fourier transform (DFT) vector space constructed by a preset codebook, a first basis vector group representing spatial domain characteristics of the first CSI data; Based at least on the first basis vector group representing the spatial characteristics of the first CSI data, data enhancement is performed on the first CSI data to obtain multiple second CSI data.

27. The apparatus according to claim 26, wherein The data enhancement processing unit is specifically used to: Selecting L orthogonal basis vectors from a group of orthogonal basis vectors included in the first DFT vector space, where correlations between the L orthogonal basis vectors and the spatial domain features of the first CSI data meet a first preset condition, and L is a natural number greater than or equal to 1; The first basis vector group is obtained based on the selected L orthogonal basis vectors.

28. The apparatus according to claim 27, wherein The data enhancement processing unit is further used to: Selecting, from a plurality of orthogonal basis vector groups included in the first DFT vector space, a target orthogonal basis vector group whose correlation with the spatial domain feature of the first CSI data satisfies a second preset condition; The L orthogonal basis vectors are selected from the target orthogonal basis vector group.

29. The apparatus according to claim 28, wherein The data enhancement processing unit is specifically used to: Projecting the first CSI data into a vector space spanned by a diagonal block matrix to obtain a first projection information matrix corresponding to each of the orthogonal basis vector groups, wherein the diagonal block matrix is ​​obtained based on the orthogonal basis vector group included in the first DFT vector space; elements in the first projection information matrix represent information related to projection coefficients after spatial features of the first CSI data are projected onto the diagonal block matrix; Based on the first projection information matrix corresponding to each of the orthogonal basis vector groups, a target orthogonal basis vector group whose correlation with the spatial domain features of the first CSI data meets a second preset condition is selected from the multiple orthogonal basis vector groups.

30. The apparatus according to claim 29, wherein The data enhancement processing unit is further used to: Performing data processing on elements in the first projection information matrix to obtain a first vector corresponding to the orthogonal basis vector group, wherein the first vector can represent total relevant information of the projection coefficient corresponding to the frequency domain characteristics of the first CSI data; performing data processing on some elements of the first vector and another part of the elements of the first vector to obtain a second vector corresponding to the orthogonal basis vector group, wherein the second vector can represent total relevant information of the projection coefficients corresponding to the spatial characteristics of the first CSI data; Selecting L first target elements corresponding to the orthogonal basis vector group from the second vector; The target orthogonal basis vector group is selected from the multiple orthogonal basis vector groups based on the L first target elements corresponding to each of the orthogonal basis vector groups.

31. The apparatus according to claim 30, wherein The data enhancement processing unit is specifically used to: Based on the L first target elements corresponding to each of the orthogonal basis vector groups, obtaining a target sum corresponding to each of the orthogonal basis vector groups; A target value is selected from the multiple target sums, and an orthogonal basis vector group corresponding to the selected target value is used as the target orthogonal basis vector group.

32. The device according to any one of claims 26 to 31, wherein The data enhancement processing unit is further used to: Determining, based on the second DFT vector space constructed by the preset codebook and the first basis vector group representing the spatial domain characteristics of the first CSI data, a second basis vector group representing at least the frequency domain characteristics of the first CSI data; Based on the first basis vector group and the second basis vector group, data enhancement is performed on the first CSI data to obtain multiple second CSI data.

33. The apparatus according to claim 32, wherein The data enhancement processing unit is specifically used to: Projecting the first CSI data into the space spanned by the first basis vector group to obtain a first projection coefficient matrix; wherein elements in the first projection coefficient matrix represent projection coefficients of the spatial domain features of the first CSI data in the space spanned by the first basis vector group; Based on the first projection coefficient matrix, select M orthogonal basis vectors from the orthogonal basis vectors included in the second DFT vector space constructed by the preset codebook, where correlations between the M orthogonal basis vectors and the projection coefficients corresponding to the frequency domain features of the first CSI data in the first projection coefficient matrix meet a third preset condition; where M is a natural number greater than or equal to 1; The second basis vector group is obtained based on the selected M orthogonal basis vectors.

34. The apparatus according to claim 33, wherein The data enhancement processing unit is specifically used to: The first CSI data is projected into a space spanned by the first basis vector matrix, wherein the first basis vector matrix is ​​a diagonal block matrix formed based on the first basis vector group.

35. The apparatus according to claim 33 or 34, wherein The data enhancement processing unit is further used to: Projecting the first projection coefficient matrix into a space spanned by orthogonal basis vectors of the second DFT vector space, obtaining a second projection information matrix; elements in the second projection information matrix represent information related to projection coefficients of frequency domain features of the first CSI data corresponding to the first projection coefficient matrix in the space spanned by orthogonal basis vectors of the second DFT vector space; Based on the second projection information matrix, M orthogonal basis vectors are selected from the orthogonal basis vectors included in the second DFT vector space.

36. The apparatus of claim 35, wherein: The data enhancement processing unit is specifically used to: Performing data processing on elements in the second projection information matrix to obtain a third vector, wherein the third vector represents total relevant information of projection coefficients corresponding to the frequency domain characteristics of the first CSI data; Select M second target elements from the third vector; Orthogonal basis vectors corresponding to the M second target elements are selected from the second DFT vector space to obtain the M orthogonal basis vectors.

37. The device according to any one of claims 33 to 36, wherein The data enhancement processing unit is further used to: Projecting the first projection coefficient matrix into the space spanned by the second basis vector group to obtain a second projection coefficient matrix, wherein elements in the second projection coefficient matrix represent projection coefficients corresponding to the spatial features of the first CSI data in the space spanned by the second basis vector group; Based on the second projection coefficient matrix, the first basis vector matrix and the second basis vector matrix, multiple second CSI data are obtained; wherein the first basis vector matrix is ​​based on the first basis vector group, and the second basis vector matrix is ​​a matrix formed based on the second basis vector group.

38. The apparatus according to claim 37, wherein The data enhancement processing unit is further used to: performing phase adjustment and / or amplitude adjustment on elements in the second projection coefficient matrix to obtain a plurality of third projection coefficient matrices; A plurality of second CSI data are obtained based on the plurality of third projection coefficient matrices, the first basis vector matrix, and the second basis vector matrix.

39. The apparatus according to claim 38, wherein The data enhancement processing unit is specifically used to: Obtaining a plurality of vectors to be processed based on a matrix product of the third projection coefficient matrix, the first basis vector matrix, and the second basis vector matrix; Normalization is performed on the multiple vectors to be processed to obtain multiple second CSI data.

40. The apparatus according to claim 38 or 39, wherein The data enhancement processing unit is specifically used to: Taking an element in the second projection coefficient matrix as the center, the center is adjusted in terms of phase and / or amplitude.

41. The apparatus according to any one of claims 25 to 40, wherein The data enhancement processing unit is further used to: Obtaining a target autocorrelation matrix of the first CSI data, where the target autocorrelation matrix can represent spatial domain characteristics and frequency domain characteristics of the first CSI data; Performing singular value decomposition on the target autocorrelation matrix to obtain a singular value decomposition result, wherein the singular value decomposition result includes a singular vector matrix representing spatial domain characteristics and frequency domain characteristics of the first CSI data, and a singular value matrix; The singular value decomposition result is processed using the constructed target random matrix to obtain a plurality of second CSI data after data enhancement.

42. The apparatus according to claim 41, wherein The data enhancement processing unit is further used to: Obtaining a first autocorrelation matrix of the first CSI data, where the first autocorrelation matrix can represent spatial characteristics of the first CSI data; Obtaining a second autocorrelation matrix of the first CSI data, wherein the second autocorrelation matrix can represent frequency domain characteristics of the first CSI data; The target autocorrelation matrix is ​​obtained based on the first autocorrelation matrix and the second autocorrelation matrix.

43. The apparatus according to claim 41 or 42, wherein The data enhancement processing unit is specifically used to: Using the target random matrix, replacing the first matrix in the singular vector matrix; The target random matrix, the second matrix in the singular vector matrix, and the singular value matrix are subjected to matrix product processing to obtain multiple second CSI data; wherein the target autocorrelation matrix is ​​a symmetric matrix, and the first matrix and the second matrix satisfy a symmetric relationship.

44. The apparatus of claim 43, wherein: The data enhancement processing unit is specifically used to: Performing matrix product processing on the target random matrix, the second matrix in the singular vector matrix, and the singular value matrix to obtain an enhanced data matrix; The enhanced data matrix is ​​arranged based on the vector dimension of the first CSI data to obtain a plurality of second CSI data after data enhancement.

45. The apparatus of claim 44, wherein: The data enhancement processing unit is specifically used to: Arrange the enhanced data matrix based on the vector dimension of the first CSI data to obtain an arranged enhanced data matrix; The arranged enhanced data matrix is ​​normalized to obtain a plurality of second CSI data.

46. ​​The apparatus according to any one of claims 25 to 45, further comprising: Model training unit, wherein the model training unit is used to Based on the CSI sample data, a first preset model is trained to obtain a first target model, wherein the first target model is used to encode the channel state information (CSI) data and obtain target CSI data; and / or, Based on the CSI sample data, a second preset model is trained to obtain a second target model, wherein the second target model is used to decode the target CSI data to obtain the CSI corresponding to the target CSI data.

47. A terminal device comprising: A sending unit is used to send first information, where the first information is obtained by inputting channel state information CSI data into a first target model for encoding processing; the first target model is obtained by training a first preset model based on the CSI sample data obtained according to any one of claims 1 to 21.

48. A network device comprising: A receiving unit, configured to receive first information, input the first information into a second target model for decoding processing, and obtain CSI data corresponding to the first information; wherein the second target model is obtained by training a second preset model based on the CSI sample data obtained according to any one of claims 1 to 21.

49. A terminal device comprising: A processor and a memory, the memory being used to store a computer program, the processor being used to call and run the computer program stored in the memory so that the terminal device executes the method according to claim 23.

50. A network device comprising: A processor and a memory, the memory being used to store a computer program, the processor being used to call and run the computer program stored in the memory so as to enable the network device to perform the method according to claim 24.

51. A chip, comprising: A processor, configured to call and run a computer program from a memory, so that a device equipped with the chip executes the method according to any one of claims 1 to 24.

52. A computer-readable storage medium for storing a computer program, which, when executed by a device, causes the device to perform the method according to any one of claims 1 to 24.

53. A computer program product comprising computer program instructions for causing a computer to perform the method of any one of claims 1 to 24.

54. A computer program causing a computer to execute the method according to any one of claims 1 to 24.

55. A communication system comprising: A terminal device, configured to execute the method according to claim 23; A network device, configured to execute the method according to claim 24.

Images