WO2024140150A1 - Channel state information feedback enhancement method, apparatus and system, and storage medium - Google Patents

Abstract

The present application relates to a channel state information (CSI) feedback enhancement method, apparatus and system, and a storage medium. The method comprises: acquiring capability information which is reported by a user terminal (S110); acquiring the current service scenario information (S120); according to the capability information and the service scenario information, obtaining a CSI measurement mode (S130); according to the CSI measurement mode, configuring and outputting a CSI measurement parameter (S140); acquiring a CSI measurement result which is reported by the user terminal (S150); according to the CSI measurement result, training an artificial intelligence model to obtain a target model (S160); and obtaining specific CSI information according to the target model, so as to schedule the current service (S170).

Description

Channel state information feedback enhancement method, device, system and storage medium

Related Applications

This application claims priority to the Chinese patent application filed with the China Patent Office on December 27, 2022, with application number 202211680505.8 and invention name “Channel State Information Feedback Enhancement Method, Device, System and Storage Medium”, the entire contents of which are incorporated by reference in this application.

Technical Field

The present application relates to the field of wireless communication technology, and in particular to a channel state information feedback enhancement method, device, system and storage medium.

Background technique

Wireless communication systems have been widely deployed for daily voice, video, data and SMS services. Mobile communications have gone through several stages of development, including 2G (GSM, Global System for Mobile Communications), 3G (TD-SCDMA, UMTS) and 4G (LTE, Long Term Evolution), and have now entered the research and development and deployment stage of 5G (NR, New Radio).

Channel state information CSI (Channel State Information) plays a vital role in the accurate scheduling of services. 3GPP introduced Type I and Type II codebooks in R15. In the design of Type I codebook, the terminal uses oversampled DFT (Discrete Fourier Transformation) vectors as feedback vectors to select the best beam for the entire bandwidth; in the design of Type II codebook, finer-grained feedback is provided for the spatial and frequency domains, resulting in a large feedback overhead. 3GPP optimized the design of Type II codebook in the R16 and R17 protocol versions, compressing the spatial and frequency domains to reduce feedback overhead. However, in the CSI measurement configuration, when the number of antenna ports and subbands configured by the base station for the user terminal is large, the terminal still has to feedback tens or even hundreds of bits of information, resulting in a large channel state information feedback load.

Summary of the invention

According to various embodiments of the present application, a channel state information feedback enhancement method, device, system and storage medium capable of reducing the channel state information feedback load are provided.

In a first aspect, the present application provides a channel state information feedback enhancement method. The method is applied to a base station, and the method includes:

Acquire capability information reported by the user terminal; wherein the capability information includes a model type of an artificial intelligence model supported by the user terminal;

Acquire current service scenario information; wherein the service scenario information includes: channel state information CSI feedback period, user terminal moving speed and channel state information CSI feedback bit number;

Obtaining a channel state information CSI measurement mode according to the capability information and the service scenario information; wherein the channel state information CSI measurement mode includes one of a prediction measurement mode, a compression measurement mode, and a compression and prediction measurement mode;

Configure and output channel state information CSI measurement parameters according to the channel state information CSI measurement method;

Acquire a channel state information CSI measurement result reported by the user terminal; wherein the channel state information CSI measurement result is obtained by the user terminal measuring the channel state information CSI according to the channel state information CSI measurement parameter;

The artificial intelligence model is trained according to the channel state information CSI measurement result to obtain a target model;

The channel state information CSI specific information is obtained according to the target model to schedule the current service.

In a second aspect, the present application also provides a channel state information feedback enhancement method. The method is applied to a user terminal, and the method includes:

Reporting capability information to a base station; wherein the capability information includes a model type of an artificial intelligence model supported by the user terminal;

Acquire a channel state information CSI measurement parameter; wherein the channel state information CSI measurement parameter is obtained by the base station according to a channel state information CSI measurement mode configuration, the channel state information CSI measurement mode is obtained by the base station according to the capability information and the service scenario information, the channel state information CSI measurement mode includes a prediction measurement mode, a compression measurement mode, and a compression and prediction measurement mode, and the service scenario information includes: a channel state information CSI feedback period, a user terminal moving speed, and a channel state information CSI feedback bit number;

Measuring the channel state information CSI according to the channel state information CSI measurement parameter to obtain a channel state information CSI measurement result;

Acquire a target model, and obtain channel state information CSI specific information according to the target model; wherein the target model is obtained by the base station training the artificial intelligence model according to the channel state information CSI measurement result, and the channel state information CSI specific information is used to schedule the current service.

In a third aspect, the present application further provides a channel state information feedback enhancement device. The device is applied to a base station, and the device includes:

A capability information acquisition module, used to acquire capability information reported by a user terminal; wherein the capability information includes a model type of an artificial intelligence model supported by the user terminal;

A service scenario information acquisition module is used to acquire current service scenario information; wherein the service scenario information includes: channel state information CSI feedback period, user terminal moving speed and channel state information CSI feedback bit number;

A measurement mode calculation module, configured to obtain a channel state information CSI measurement mode according to the capability information and the service scenario information; wherein the channel state information CSI measurement mode includes one of a prediction measurement mode, a compression measurement mode, and a compression and prediction measurement mode;

A measurement parameter output module, configured to configure and output a channel state information CSI measurement parameter according to the channel state information CSI measurement mode;

A measurement result acquisition module, used to acquire the channel state information CSI measurement result reported by the user terminal; wherein the channel state information CSI measurement result is obtained by the user terminal measuring the channel state information CSI according to the channel state information CSI measurement parameter;

A model training module, used to train the artificial intelligence model according to the channel state information CSI measurement result to obtain a target model;

The specific information calculation module is used to obtain channel state information CSI specific information according to the target model to schedule the current service.

In a fourth aspect, the present application further provides a channel state information feedback enhancement device. The device is applied to a user terminal, and the device includes:

A capability information reporting module, used to report capability information to a base station; wherein the capability information includes a model type of an artificial intelligence model supported by the user terminal;

A measurement parameter acquisition module, configured to acquire a channel state information CSI measurement parameter; wherein the channel state information CSI measurement parameter is obtained by the base station according to a channel state information CSI measurement mode configuration, the channel state information CSI measurement mode is obtained by the base station according to the capability information and the service scenario information, the channel state information CSI measurement mode includes one of a prediction measurement mode, a compression measurement mode and a compression and prediction measurement mode, and the service scenario information includes: a channel state information CSI feedback period, a user terminal moving speed and a channel state information CSI feedback bit number;

A measurement result calculation module, used to measure the channel state information CSI according to the channel state information CSI measurement parameter to obtain a channel state information CSI measurement result;

A model acquisition module is used to acquire a target model and obtain channel state information CSI specific information according to the target model; wherein the target model is obtained by the base station training the artificial intelligence model according to the channel state information CSI measurement results, and the channel state information CSI specific information is used to schedule the current service.

In the fifth aspect, the present application also provides a channel state information feedback enhancement system, comprising a base station and a user terminal connected to the base station; wherein: the base station is used to execute the steps of the method described in the first aspect above; the user terminal is used to execute the steps of the method described in the second aspect above.

In a sixth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the above method when executed by a processor.

The details of one or more embodiments of the present application are set forth in the following drawings and description. Other features, objects, and advantages of the present application will become apparent from the description, drawings, and claims.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to better describe and illustrate the embodiments and or examples of the inventions disclosed herein, reference may be made to one or more drawings. The additional details or examples used to describe the drawings should not be considered as limiting the scope of the disclosed inventions, the embodiments and or examples currently described, and any of the best modes of these inventions currently understood.

FIG1 is a diagram of an application environment of a channel state information feedback enhancement method according to an embodiment;

FIG2 is a schematic diagram of a flow chart of a channel state information feedback enhancement method implemented from the perspective of a base station in one embodiment;

FIG3 is a schematic diagram of CSI compression based on a CSI compression model in one embodiment;

FIG4 is a schematic diagram of CSI prediction performed on a user terminal side in one embodiment;

FIG5 is a schematic diagram of CSI prediction at a base station side in an embodiment;

FIG6 is a schematic diagram of a flow chart of a method for obtaining channel state information CSI measurement from a base station perspective in one embodiment;

FIG7 is a schematic diagram of a process of obtaining specific information of channel state information CSI from the perspective of a base station in one embodiment;

FIG8 is a schematic diagram of a user terminal performing CSI compression and CSI prediction functions in one embodiment;

FIG9 is a schematic diagram of a flow chart of obtaining specific information of channel state information CSI from the perspective of a base station in another embodiment;

FIG10 is a schematic diagram of a base station performing a CSI prediction function in one embodiment;

FIG11 is a schematic diagram of a flow chart of a channel state information feedback enhancement method implemented from the perspective of a user terminal in one embodiment;

FIG12 is a schematic diagram of a process of obtaining specific information of channel state information CSI from the perspective of a user terminal in one embodiment;

FIG13 is a schematic diagram of a process of obtaining specific information of channel state information CSI from the perspective of a user terminal in another embodiment;

FIG14 is a schematic diagram of signaling interaction between a base station and a user terminal in one embodiment;

FIG15 is a schematic diagram of signaling interaction between a base station and a user terminal in another embodiment;

FIG16 is a schematic diagram of a module of a channel state information feedback enhancement device implemented from the perspective of a base station in one embodiment;

FIG. 17 is a schematic diagram of a module of a channel state information feedback enhancement device implemented from the perspective of a user terminal in one embodiment.

Detailed ways

In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

The channel state information feedback enhancement method provided in the embodiment of the present application can be applied in the application environment shown in FIG1. The user terminal 102 communicates with the base station 104 via a wireless network. The user terminal 102 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. The portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc.

In one embodiment, as shown in FIG. 2 , a channel state information feedback enhancement method is provided, and the method is applied to the base station 104 in FIG. 1 as an example for explanation, including: Follow these steps:

Step S110, obtaining capability information reported by the user terminal; wherein the capability information includes the model type of the artificial intelligence model supported by the user terminal.

Specifically, the base station 104 can send a user terminal 102 capability query message UECapabilityEnquiry to the user terminal 102. After the user terminal 102 receives the user terminal 102 capability query message, it will report the capability information to the base station 104. The capability information is used to characterize whether the current user terminal 102 supports the artificial intelligence model, and the model type of the artificial intelligence model supported by the current user terminal 102 when supporting the artificial intelligence model. It can be understood that the model types include CSI compression model types (such as CNN, RNN, Transformer, ResNet, etc.) and CSI prediction model types (such as FCN, RNN, 3D-CNN, etc.). Among them, the CSI compression model can compress and reconstruct the CSI to reduce the data volume and feedback overhead of the CSI. The CSI prediction model can predict the CSI value at a certain moment in the future based on the historical CSI measurement value to solve the problem of inaccuracy of the channel state information due to delay.

Step S120, obtaining current service scenario information; wherein the service scenario information includes: channel state information CSI feedback period, user terminal moving speed and channel state information CSI feedback bit number.

Specifically, the base station 104 determines the channel state information CSI feedback period and the channel state information CSI feedback bit number according to the type of service scheduled by the user terminal 102. The base station 104 can obtain the current moving speed of the user terminal 102 in the following ways, for example, estimating the moving speed of the user terminal 102 according to the probability of the change of the downlink channel information and the precoding matrix indication; obtaining the moving speed of the user terminal by calculating the autocorrelation function of the time domain received signal when the transmitter continuously sends a special fixed signal; calculating the estimated value of the complex channel autocorrelation function according to the sampling of the transmitted signal, and estimating the maximum Doppler frequency shift in combination with the angle difference of the transmission signal arrival angle, thereby obtaining the moving speed of the user terminal 102. The method of calculating the moving speed of the user terminal can be changed according to the needs of different application scenarios, and is not limited here.

Step S130, obtaining a channel state information CSI measurement mode according to the capability information and the service scenario information; wherein the channel state information CSI measurement mode includes one of a prediction measurement mode, a compression measurement mode, and a compression and prediction measurement mode.

Specifically, the base station 104 can determine that the user terminal 102 supports the corresponding artificial intelligence model based on the capability information, and then obtain the channel state information CSI measurement method based on the business scenario information. It is also possible to first obtain the business scenario information, and then determine the channel state information CSI measurement method in combination with the acquired capability information. Among them, the channel state information CSI measurement method is used to determine whether the base station 104 and the user terminal 102 perform CSI compression and/or CSI prediction functions. For example, when the base station 104 determines that the current user terminal 102 only supports the CSI compression model based on the capability information, and at the same time determines that the CSI compression function can be performed in the current business scenario based on the business scenario information, the channel state information CSI measurement method is determined to be a compression measurement method. Similarly, when the user terminal 102 supports both the CSI compression model and the CSI prediction model, and can perform the CSI compression function and the CSI prediction function in the current business scenario, the channel state information CSI measurement method is determined to be a compression and prediction measurement method.

Step S140, configure and output the channel state information CSI measurement parameters according to the channel state information CSI measurement mode. Specifically, under different channel state information CSI measurement modes, the training data required by the base station 104 for training the artificial intelligence model is also different. Therefore, the base station 104 needs to configure the corresponding channel state information CSI measurement parameters according to the channel state information CSI measurement mode, and send them to the user terminal 102, so that the user terminal 102 can collect the corresponding training data. For example, when the channel state information CSI measurement mode is determined to be a predictive measurement mode, the training data is a plurality of CSI values at historical moments. At this time, the user terminal 102 needs to control itself to obtain a plurality of CSI values at historical moments by receiving the corresponding channel state information CSI measurement parameters; when the channel state information CSI measurement mode is determined to be a compressed measurement mode, the training data can be the CSI value at the current moment. At this time, the user terminal 102 needs to control itself to obtain the CSI value at the current moment by receiving the corresponding channel state information CSI measurement parameters.

Step S150, obtaining the channel state information CSI measurement result reported by the user terminal; wherein the channel state information CSI measurement result is obtained by the user terminal measuring the channel state information CSI according to the channel state information CSI measurement parameter. Specifically, the user terminal 102 measures the CSI according to the channel state information CSI measurement parameter configured by the base station 104, and then sends the measured channel state information CSI measurement result to the base station 104. For example, when the channel state information CSI measurement mode is determined to be a predictive measurement mode, the channel state information CSI measurement parameter is used to configure the user terminal 102 to measure multiple CSI values at historical moments, and report the multiple CSI values to the base station 104 as the channel state information CSI measurement result.

Step S160: Train an artificial intelligence model according to the channel state information CSI measurement result to obtain a target model.

Specifically, the base station 104 trains the corresponding artificial intelligence model according to the channel state information CSI measurement results reported by the user terminal 102, so as to obtain the target model. It is understandable that the target model may include a CSI compression model and/or a CSI prediction model. When the channel state information CSI measurement method is determined to be a prediction measurement method, the target model obtained by training is the CSI prediction model; when the channel state information CSI measurement method is determined to be a compression measurement method, the target model obtained by training is the CSI compression model; when the channel state information CSI measurement method is determined to be a compression and prediction measurement method, the target model obtained by training includes a CSI compression model and a CSI prediction model. It is understandable that when the base station 104 trains the CSI compression model and the CSI prediction model, the specific training method is determined according to the model type of the artificial intelligence model in the capability information reported by the user terminal 102.

Step S170, obtain channel state information CSI specific information according to the target model to schedule the current service. Specifically, after the target model is trained, the channel state information CSI specific information can be calculated according to the target model, and the base station 104 can schedule the current service according to the CSI specific information.

In a specific example, when the channel state information CSI measurement mode is determined to be a compressed measurement mode, the target model obtained by the base station 104 through training is the channel state information CSI compression model. As shown in Figure 3, it is a schematic diagram of CSI compression based on the CSI compression model in one embodiment. The user terminal 102 preprocesses the channel matrix H and calculates the eigenvector V of the channel matrix for each subband. The obtained eigenvector is input into the CSI compression model for AI compression. The size of the output compressed CSI is smaller than the original eigenvector V, and the compressed CSI obtained by the CSI compression model is output as a floating-point vector. Since the CSI compression model cannot directly process complex inputs, the real and imaginary parts of the eigenvector V need to be extracted and merged together during the processing. The user terminal 102 side converts the compressed CSI of the floating-point vector into a quantized bit sequence by quantization to meet the bit width of the CSI feedback. During preprocessing and post-processing, it is necessary to consider adaptive processing of bandwidth, feedback load and number of antenna ports. CSI feedback information is transmitted from the user to the user through the wireless channel. The terminal 102 transmits the information to the base station. The base station performs preprocessing, dequantization, AI decompression and post-processing on the CSI feedback information to obtain a channel matrix H'. The channel matrix H' is used for scheduling the current service of the user terminal 102.

When the channel state information CSI measurement mode is determined to be a predictive measurement mode, the target model obtained by the base station 104 through training is the channel state information CSI prediction model. Among them, the CSI prediction model can predict the CSI on the user terminal 102 side or the base station side to obtain the corresponding channel state information CSI specific information. As shown in Figure 4, it is a schematic diagram of CSI prediction on the user terminal 102 side in an embodiment. The user terminal 102 predicts the value of CSItN at a certain time in the future through the CSI prediction model based on several (such as 16) historical CSIs measurement values, namely CSIt-15, ..., CSIt-1, CSIt0. As shown in Figure 5, it is a schematic diagram of CSI prediction on the base station side in an embodiment. After the user terminal 102 reports the channel state information CSI measurement result to the base station 104, the base station side predicts the CSI value corresponding to the wireless channel at a certain scheduling time tN in the future. It can be understood that when the channel state information CSI measurement method is determined to be a compression and prediction measurement method, the target model obtained by the base station 104 through training is the channel state information CSI compression model and the channel state information CSI prediction model, and its specific processing process is a combination of the above processing processes.

Since in the process of obtaining the specific information of the channel state information CSI, the current scene is identified and the corresponding target model is generated according to different scenes, the feedback load of the channel state information CSI can be reduced, and the CSI feedback accuracy can be improved under the premise of having the same channel state information CSI feedback load.

In one embodiment, as shown in FIG. 6 , in step S130, the step of obtaining a channel state information CSI measurement method according to the capability information and the service scenario information includes:

Step S131: determine whether the user terminal supports the target model type according to the capability information.

Specifically, in this embodiment, after the base station 104 obtains the capability information and business scenario information, it is first necessary to determine whether the user terminal 102 supports the target model type. It can be understood that the target model type is the model type of the artificial intelligence model used by the base station 104 when performing the CSI compression and/or CSI prediction functions. The target model type can be one or more, and the user can set the target model type as needed. Among them, the model type of the CSI compression model can be CNN, RNN, Transformer, ResNet, etc., and the model type of the CSI prediction model can be FCN, RNN, 3D-CNN, etc. The capability information includes the model type of the artificial intelligence model supported by the user terminal 102. When the model type supported by the user terminal 102 is the same as the target model type, the base station 104 will perform subsequent steps to perform CSI compression and/or CSI prediction functions.

Step S132: When the target model type is supported, the channel state information CSI measurement method is determined according to the service scenario information.

Specifically, the judgment is made based on the capability information reported by the user terminal 102. If the current user terminal 102 supports the target model type, it means that the current user terminal 102 and the base station 104 can perform subsequent CSI compression and/or CSI prediction functions together, and the channel state information CSI measurement method can be further determined based on the service scenario information. If the current user terminal 102 does not support the target model type, it means that the user terminal 102 does not have this function. At this time, the base station 104 can notify the user terminal 102 to perform traditional CSI measurement through RRC (Radio Resource Control) signaling.

In one embodiment, in step S132, the step of determining the channel state information CSI measurement mode according to the business scenario information includes: when the channel state information CSI feedback period is greater than a first threshold or the user terminal moving speed is greater than a second threshold, the channel state information CSI measurement mode is determined as a predictive measurement mode; as a specific example, the first threshold can be set to 20ms, and the second threshold can be set to 30km/h. When the base station 104 detects that the channel state information CSI feedback period in the business scenario information is greater than 20ms, or the user terminal moving speed is greater than 30km/h, the channel state information CSI measurement mode can be determined as a predictive measurement mode, and the base station 104 and the user terminal 102 will start to execute the CSI prediction function. When the number of channel state information CSI feedback bits is greater than the third threshold, the channel state information CSI measurement mode is determined as a compression measurement mode; in a specific example, the third threshold can be set to 40 bits. When the base station 104 detects that the number of channel state information CSI feedback bits in the service scenario information is greater than 40 bits, the channel state information CSI measurement mode can be determined as a compression measurement mode, and the base station 104 and the user terminal 102 will start to perform the CSI compression function. When the channel state information CSI feedback period is greater than the first threshold or the user terminal moving speed is greater than the second threshold, and the number of channel state information CSI feedback bits is greater than the third threshold, the channel state information CSI measurement mode is determined as a compression and prediction measurement mode; in a specific example, when the above two conditions are met at the same time, the channel state information CSI measurement mode is determined as a compression and prediction measurement mode, and the base station 104 and the user terminal 102 will start to perform CSI compression and CSI prediction functions. Through the above-mentioned judgment of the business scenario information, the current business scenario can be accurately identified, and the channel state information CSI measurement method suitable for the current business scenario can be determined to perform the corresponding CSI compression and/or CSI prediction functions, so as to reduce the CSI feedback load or improve the CSI feedback accuracy under the premise of the same CSI feedback load, thereby improving the downlink system throughput.

In one embodiment, as shown in FIG. 7 , the target model includes: a channel state information CSI compression model and a channel state information CSI prediction model. In step S170, the step of obtaining specific information of the channel state information CSI according to the target model includes:

Step S171: Send the channel state information CSI compression model and the channel state information CSI prediction model to the user terminal.

Specifically, in the embodiment of the present application, the channel state information CSI measurement mode determined according to the capability information and the service scenario information is a compression and prediction measurement mode, so the target model trained by the base station 104 according to the channel state information CSI measurement result includes a channel state information CSI compression model and a channel state information CSI prediction model. After the base station 104 has trained the target model, it will send the CSI compression model and the CSI prediction model to the user terminal 102. When the base station 104 sends the CSI compression model, it can first notify the user terminal 102 of the model type and compression parameters of the CSI compression model, wherein the model types of the CSI compression model include: CNN, RNN, Transformer, etc., and the compression parameters include: preprocessing mode, quantization mode, etc.; when the base station 104 sends the CSI prediction model, it can first notify the user terminal 102 of the model type and prediction parameters of the CSI prediction model, wherein the model types of the CSI prediction model include: FCN, RNN, 3D-CNN, etc., and the prediction parameters include: predicted delay time (3ms, 4ms, 5ms), etc. The user terminal 102 downloads the corresponding CSI compression model and CSI prediction model from the OTT (Over-The-Top) server based on the acquired model type and compression parameters of the CSI compression model and the model type and prediction parameters of the CSI prediction model, thereby obtaining the channel state information CSI compression model and the channel state information CSI prediction model in the user terminal 102.

Step S172, obtaining first channel state information CSI compression information reported by the user terminal; wherein the first channel state information CSI compression information is obtained by the user terminal compressing the first channel state information CSI prediction information according to the channel state information CSI compression model, and the first channel state information CSI prediction information is obtained by the user terminal according to the channel state information The CSI prediction model predicts the first channel state information CSI measurement information.

Specifically, in this embodiment, the user terminal 102 will perform CSI compression and CSI prediction functions. As shown in Figure 8, it is a schematic diagram of the user terminal 102 performing CSI compression and CSI prediction functions. The user terminal 102 first performs CSI measurement according to the parameters configured by the base station 104 to obtain first channel state information CSI measurement information, and then performs CSI prediction according to the configured channel state information CSI prediction model (including predicted delay time, etc.) to obtain first channel state information CSI prediction information, and finally performs CSI compression on the first channel state information CSI prediction information according to the configured channel state information CSI compression model to obtain first channel state information CSI compression information. After the user terminal 102 obtains the first channel state information CSI compression information, it reports CSI through the air interface, and the base station side can obtain the first channel state information CSI compression information.

Step S173, decompress the first channel state information CSI compressed information according to the channel state information CSI compression model to obtain channel state information CSI specific information. Specifically, after the base station 104 obtains the first channel state information CSI compressed information, it decompresses the first channel state information CSI compressed information according to the channel state information CSI compression model to obtain the channel state information CSI specific information. The base station 104 subsequently schedules the current service according to the channel state information CSI specific information to complete the data transmission with the user terminal 102.

In one embodiment, as shown in FIG. 9 , the target model includes: a channel state information CSI compression model and a channel state information CSI prediction model. In step S170, the step of obtaining specific information of the channel state information CSI according to the target model includes:

Step S174: Send the channel state information CSI compression model to the user terminal.

Specifically, in the embodiment of the present application, the channel state information CSI measurement mode determined according to the capability information and the service scenario information is a compression and prediction measurement mode, so the target model trained by the base station 104 according to the channel state information CSI measurement result includes a channel state information CSI compression model and a channel state information CSI prediction model. After the base station 104 has trained the target model, it only needs to send the CSI compression model to the user terminal 102. When the base station 104 sends the CSI compression model, it can first notify the user terminal 102 of the model type and compression parameters of the CSI compression model, wherein the model types of the CSI compression model include: CNN, RNN, Transformer, etc., and the compression parameters include: preprocessing mode, quantization mode, etc. The user terminal 102 downloads the corresponding CSI compression model from the OTT (Over-The-Top) server based on the model type and compression parameters of the obtained CSI compression model, thereby obtaining the channel state information CSI compression model in the user terminal 102.

Step S175, obtaining second channel state information CSI compression information reported by the user terminal; wherein the second channel state information CSI compression information is obtained by the user terminal compressing the second channel state information CSI measurement information according to a channel state information CSI compression model.

Specifically, in this embodiment, the CSI compression function is performed in the user terminal 102, and the CSI prediction function is performed in the base station 104. As shown in FIG10, it is a schematic diagram of the base station 104 performing the CSI prediction function. The user terminal 102 first performs CSI measurement according to the parameters configured by the base station 104 to obtain the second channel state information CSI measurement information, and then performs CSI compression on the second channel state information CSI measurement information according to the configured channel state information CSI compression model to obtain the second channel state information CSI compression information. After the user terminal 102 obtains the second channel state information CSI compression information, it reports CSI through the air interface, so that the base station 104 obtains the second channel state information CSI compression information.

Step S176: Decompress the second channel state information CSI compressed information according to the channel state information CSI compression model to obtain second channel state information CSI decompressed information. Specifically, after the base station 104 obtains the second channel state information CSI compressed information, it decompresses it through the trained channel state information CSI compression model to obtain the second channel state information CSI decompressed information.

Step S177, predict the second channel state information CSI decompression information according to the channel state information CSI prediction model to obtain channel state information CSI specific information. Specifically, after the base station 104 obtains the second channel state information CSI decompression information, it performs CSI prediction through the channel state information CSI prediction model to obtain channel state information CSI specific information. The base station 104 subsequently schedules the current service according to the channel state information CSI specific information to complete data transmission with the user terminal 102.

In the above embodiment, the CSI prediction function is respectively performed by the user terminal 102 and the base station 104, and the user can flexibly select according to the specific usage scenario. In the case where the computing power of the user terminal 102 is insufficient, the base station 104 can be used to perform the CSI prediction function, reducing the resource occupation of the user terminal 102. In the case where the load of the base station 104 is high, the user terminal 102 can be used to perform the CSI prediction function, reducing the resource occupation of the base station 104.

In one embodiment, as shown in FIG. 11 , the present application further provides a channel state information feedback enhancement method, which is described by taking the method applied to the user terminal 102 in FIG. 1 as an example, including the following steps:

Step S210, reporting capability information to the base station; wherein the capability information includes the model type of the artificial intelligence model supported by the user terminal.

Specifically, the base station 104 can send a user terminal 102 capability query message UECapabilityEnquiry to the user terminal 102. After the user terminal 102 receives the user terminal 102 capability query message, it will report the capability information to the base station 104. The capability information is used to characterize whether the current user terminal 102 supports the artificial intelligence model, and the model type of the artificial intelligence model supported by the current user terminal 102 when supporting the artificial intelligence model. It can be understood that the model types include CSI compression model types (such as CNN, RNN, Transformer, ResNet, etc.) and CSI prediction model types (such as FCN, RNN, 3D-CNN, etc.). Among them, the CSI compression model can compress and reconstruct the CSI to reduce the data volume and feedback overhead of the CSI. The CSI prediction model can predict the CSI value at a certain moment in the future based on the historical CSI measurement value to solve the problem of inaccuracy of the channel state information due to delay.

Step S220, obtaining channel state information CSI measurement parameters; wherein, the channel state information CSI measurement parameters are configured by the base station according to the channel state information CSI measurement method, the channel state information CSI measurement method is obtained by the base station according to the capability information and the service scenario information, the channel state information CSI measurement method includes a prediction measurement method, a compression measurement method and one of the compression and prediction measurement methods, and the service scenario information includes: a channel state information CSI feedback period, a user terminal moving speed and a channel state information CSI feedback bit number.

Specifically, the base station 104 determines the channel state information CSI feedback period and the number of channel state information CSI feedback bits according to the type of service scheduled by the user terminal 102. The station 104 can obtain the current moving speed of the user terminal 102 in the following ways, for example, estimating the moving speed of the user terminal 102 according to the probability of the change of the downlink channel information and the precoding matrix indication; obtaining the moving speed of the user terminal by calculating the autocorrelation function of the time domain received signal when the transmitter continuously sends a special fixed signal; calculating the estimated value of the complex channel autocorrelation function according to the sampling of the transmitted signal, and estimating the maximum Doppler frequency shift in combination with the angle difference of the transmission signal arrival angle, thereby obtaining the moving speed of the user terminal 102. The method of calculating the moving speed of the user terminal can be changed according to the needs of different application scenarios, and is not limited here.

The base station 104 can obtain the channel state information CSI measurement mode according to the business scenario information when it is determined that the user terminal 102 supports the corresponding artificial intelligence model according to the capability information. It is also possible to first obtain the business scenario information, and then determine the channel state information CSI measurement mode in combination with the acquired capability information. Among them, the channel state information CSI measurement mode is used to determine whether the base station 104 and the user terminal 102 perform CSI compression and/or CSI prediction functions. For a specific example, when the base station 104 determines that the current user terminal 102 only supports the CSI compression model according to the capability information, and at the same time determines that the CSI compression function can be performed in the current business scenario according to the business scenario information, the channel state information CSI measurement mode is determined to be a compression measurement mode. Similarly, when the user terminal 102 supports both the CSI compression model and the CSI prediction model, and can perform the CSI compression function and the CSI prediction function in the current business scenario, the channel state information CSI measurement mode is determined to be a compression and prediction measurement mode.

Under different channel state information CSI measurement modes, the training data required by the base station 104 for training the artificial intelligence model is also different. Therefore, the base station 104 needs to configure the corresponding channel state information CSI measurement parameters according to the channel state information CSI measurement mode, and send them to the user terminal 102, so that the user terminal 102 can collect the corresponding training data. For example, when the channel state information CSI measurement mode is determined to be a predictive measurement mode, the training data is a plurality of CSI values at historical moments. At this time, the user terminal 102 needs to control itself to obtain a plurality of CSI values at historical moments by receiving the corresponding channel state information CSI measurement parameters; when the channel state information CSI measurement mode is determined to be a compressed measurement mode, the training data can be the CSI value at the current moment. At this time, the user terminal 102 needs to control itself to obtain the CSI value at the current moment by receiving the corresponding channel state information CSI measurement parameters.

Step S230, the channel state information CSI is measured according to the channel state information CSI measurement parameters to obtain a channel state information CSI measurement result. Specifically, the user terminal 102 measures the CSI according to the channel state information CSI measurement parameters configured by the base station 104, and then sends the measured channel state information CSI measurement result to the base station 104. For example, when the channel state information CSI measurement mode is determined to be a predictive measurement mode, the channel state information CSI measurement parameters are used to configure the user terminal 102 to measure multiple CSI values at historical moments, and report the multiple CSI values to the base station 104 as the channel state information CSI measurement results.

Step S240, obtain the target model, and obtain the specific information of the channel state information CSI according to the target model; wherein, the target model is obtained by training the artificial intelligence model according to the channel state information CSI measurement results of the base station, and the specific information of the channel state information CSI is used to schedule the current service.

Specifically, the base station 104 trains the corresponding artificial intelligence model according to the channel state information CSI measurement results reported by the user terminal 102, so as to obtain the target model. It is understandable that the target model may include a CSI compression model and/or a CSI prediction model. When the channel state information CSI measurement method is determined to be a prediction measurement method, the trained target model is the CSI prediction model; when the channel state information CSI measurement method is determined to be a compression measurement method, the trained target model is the CSI compression model; when the channel state information CSI measurement method is determined to be a compression and prediction measurement method, the trained target model includes a CSI compression model and a CSI prediction model. It is understandable that when the base station 104 trains the CSI compression model and the CSI prediction model, the specific training method is determined according to the model type of the artificial intelligence model in the capability information reported by the user terminal 102. The training of the channel state information CSI compression model and the channel state information CSI prediction model is the same as in the above embodiment, and will not be repeated here one by one.

Since in the process of obtaining the specific information of the channel state information CSI, the current scene is identified and the corresponding target model is generated according to different scenes, the feedback load of the channel state information CSI can be reduced, and the CSI feedback accuracy can be improved under the premise of having the same channel state information CSI feedback load.

In one embodiment, the channel state information CSI measurement method is obtained by the base station based on capability information and business scenario information, including: the capability information is used to instruct the base station to determine whether the user terminal 102 supports the target model type. When the target model type is supported, the channel state information CSI measurement method is determined based on the business scenario information.

Specifically, in this embodiment, after the base station 104 obtains the capability information and business scenario information, it is first necessary to determine whether the user terminal 102 supports the target model type. It can be understood that the target model type is the model type of the artificial intelligence model used by the base station 104 when performing the CSI compression and/or CSI prediction functions. The target model type can be one or more, and the user can set the target model type as needed. Among them, the model type of the CSI compression model can be CNN, RNN, Transformer, ResNet, etc., and the model type of the CSI prediction model can be FCN, RNN, 3D-CNN, etc. The capability information includes the model type of the artificial intelligence model supported by the user terminal 102. When the model type supported by the user terminal 102 is the same as the target model type, the base station 104 will perform subsequent steps to perform CSI compression and/or CSI prediction functions. The base station 104 makes a judgment based on the capability information reported by the user terminal 102. If the current user terminal 102 supports the target model type, it means that the current user terminal 102 and the base station 104 can perform subsequent CSI compression and/or CSI prediction functions together. At this time, the channel state information CSI measurement method can continue to be determined based on the service scenario information. If the current user terminal 102 does not support the target model type, it means that the user terminal 102 does not have this function. At this time, the base station 104 can notify the user terminal 102 to perform traditional CSI measurement through RRC (Radio Resource Control) signaling.

In one embodiment, the step of determining the channel state information CSI measurement mode according to the service scenario information includes: when the channel state information CSI feedback period is greater than the first threshold or the user terminal moving speed is greater than the second threshold, the channel state information CSI measurement mode is determined as a prediction measurement mode; in a specific example, the first threshold can be set to 20ms, and the second threshold can be set to 30km/h. When the base station 104 detects that the channel state information CSI feedback period in the service scenario information is greater than 20ms, or the user terminal moving speed is greater than 30km/h, the channel state information CSI measurement mode can be determined as a prediction measurement mode, and the base station 104 and the user terminal 102 will start to perform the CSI prediction function. When the number of channel state information CSI feedback bits is greater than the third threshold, the channel state information CSI measurement mode is determined as a compression measurement mode; in a specific example, the third threshold can be set to 40bits. When the base station 104 detects that the number of channel state information CSI feedback bits in the service scenario information is greater than 40bits, the channel state information CSI measurement mode can be determined as a compression measurement mode, and the base station 104 and the user terminal 102 will start to perform the CSI compression function. When the channel state information CSI feedback period is greater than the first threshold or the user terminal moving speed is greater than the second threshold, And if the number of channel state information CSI feedback bits is greater than the third threshold, the channel state information CSI measurement mode is determined as a compression and prediction measurement mode; in a specific example, when the above two conditions are met at the same time, the channel state information CSI measurement mode is determined as a compression and prediction measurement mode, and the base station 104 and the user terminal 102 will start to perform CSI compression and CSI prediction functions. Through the above judgment of the business scenario information, the current business scenario can be accurately identified, and the channel state information CSI measurement mode suitable for the current business scenario can be determined to perform the corresponding CSI compression and/or CSI prediction functions, so as to reduce the CSI feedback load or improve the CSI feedback accuracy under the premise of the same CSI feedback load, thereby improving the downlink system throughput.

In one embodiment, as shown in FIG. 12 , the target model includes: a channel state information CSI compression model and a channel state information CSI prediction model. In step S240 , the step of acquiring the target model and obtaining the channel state information CSI specific information according to the target model includes:

Step S241, obtaining a channel state information CSI compression model and a channel state information CSI prediction model sent by a base station.

Specifically, in the embodiment of the present application, the channel state information CSI measurement mode determined according to the capability information and the service scenario information is a compression and prediction measurement mode, so the target model trained by the base station 104 according to the channel state information CSI measurement result includes a channel state information CSI compression model and a channel state information CSI prediction model. After the base station 104 has trained the target model, it will send the CSI compression model and the CSI prediction model to the user terminal 102. When the base station 104 sends the CSI compression model, it can first notify the user terminal 102 of the model type and compression parameters of the CSI compression model, wherein the model types of the CSI compression model include: CNN, RNN, Transformer, etc., and the compression parameters include: preprocessing mode, quantization mode, etc.; when the base station 104 sends the CSI prediction model, it can first notify the user terminal 102 of the model type and prediction parameters of the CSI prediction model, wherein the model types of the CSI prediction model include: FCN, RNN, 3D-CNN, etc., and the prediction parameters include: predicted delay time (3ms, 4ms, 5ms), etc. The user terminal 102 downloads the corresponding CSI compression model and CSI prediction model from the OTT (Over-The-Top) server based on the acquired model type and compression parameters of the CSI compression model and the model type and prediction parameters of the CSI prediction model, thereby obtaining the channel state information CSI compression model and the channel state information CSI prediction model in the user terminal 102.

Step S242, predict the first channel state information CSI measurement information according to the channel state information CSI prediction model to obtain the first channel state information CSI prediction information. Specifically, in this embodiment, the user terminal 102 will perform CSI compression and CSI prediction functions. As shown in Figure 8, it is a schematic diagram of the user terminal 102 performing CSI compression and CSI prediction functions. The user terminal 102 first performs CSI measurement according to the parameters configured by the base station 104 to obtain the first channel state information CSI measurement information, and then performs CSI prediction according to the configured channel state information CSI prediction model (including the predicted delay time, etc.), thereby obtaining the first channel state information CSI prediction information.

Step S243: compress the first channel state information CSI prediction information according to the channel state information CSI compression model to obtain first channel state information CSI compression information. Specifically, after the user terminal 102 obtains the first channel state information CSI prediction information, it performs CSI compression on the first channel state information CSI prediction information according to the configured channel state information CSI compression model to obtain the first channel state information CSI compression information.

Step S244, reporting the first channel state information CSI compression information to the base station, and obtaining the channel state information CSI specific information; wherein, the channel state information CSI specific information is obtained by the base station decompressing the first channel state information CSI compression information according to the channel state information CSI compression model. Specifically, after the user terminal 102 obtains the first channel state information CSI compression information, it reports CSI through the air interface, and the base station side can obtain the first channel state information CSI compression information. After the base station 104 obtains the first channel state information CSI compression information, it decompresses the first channel state information CSI compression information according to the channel state information CSI compression model to obtain the channel state information CSI specific information. The base station 104 subsequently schedules the current service according to the channel state information CSI specific information, and completes the data transmission with the user terminal 102.

In one embodiment, as shown in FIG. 13 , the target model includes: a channel state information CSI compression model and a channel state information CSI prediction model. In step S240 , the step of acquiring the target model and obtaining the channel state information CSI specific information according to the target model includes:

Step S245: Acquire a channel state information CSI compression model sent by the base station.

Specifically, in the embodiment of the present application, the channel state information CSI measurement mode determined according to the capability information and the service scenario information is a compression and prediction measurement mode, so the target model trained by the base station 104 according to the channel state information CSI measurement result includes a channel state information CSI compression model and a channel state information CSI prediction model. After the base station 104 has trained the target model, it only needs to send the CSI compression model to the user terminal 102. When the base station 104 sends the CSI compression model, it can first notify the user terminal 102 of the model type and compression parameters of the CSI compression model, wherein the model types of the CSI compression model include: CNN, RNN, Transformer, etc., and the compression parameters include: preprocessing mode, quantization mode, etc. The user terminal 102 downloads the corresponding CSI compression model from the OTT (Over-The-Top) server based on the model type and compression parameters of the obtained CSI compression model, thereby obtaining the channel state information CSI compression model in the user terminal 102.

Step S246: compress the second channel state information CSI measurement information according to the channel state information CSI compression model to obtain second channel state information CSI compressed information.

Specifically, in this embodiment, the CSI compression function is performed in the user terminal 102, and the CSI prediction function is performed in the base station 104. As shown in FIG10, it is a schematic diagram of the base station 104 performing the CSI prediction function. The user terminal 102 first performs CSI measurement according to the parameters configured by the base station 104 to obtain the second channel state information CSI measurement information, and then performs CSI compression on the second channel state information CSI measurement information according to the configured channel state information CSI compression model to obtain the second channel state information CSI compression information. After the user terminal 102 obtains the second channel state information CSI compression information, it reports CSI through the air interface, so that the base station 104 obtains the second channel state information CSI compression information.

Step S247, report the second channel state information CSI compression information to the base station to obtain channel state information CSI specific information; wherein, the channel state information CSI specific information is obtained by the base station predicting the second channel state information CSI decompression information according to the channel state information CSI prediction model, and the second channel state information CSI decompression information is obtained by the base station decompressing the second channel state information CSI compression information according to the channel state information CSI compression model.

Specifically, after the base station 104 obtains the second channel state information CSI compression information, it decompresses it through the trained channel state information CSI compression model to obtain the second channel state information CSI decompression information. After the base station 104 obtains the second channel state information CSI decompression information, it performs CSI prediction through the channel state information CSI prediction model to obtain the channel state information CSI specific information. The base station 104 subsequently schedules the current service according to the channel state information CSI specific information to complete the communication with the user terminal 102. data transmission.

In the above embodiment, the CSI prediction function is respectively performed by the user terminal 102 and the base station 104, and the user can flexibly select according to the specific usage scenario. In the case where the computing power of the user terminal 102 is insufficient, the base station 104 can be used to perform the CSI prediction function, reducing the resource occupation of the user terminal 102. In the case where the load of the base station 104 is high, the user terminal 102 can be used to perform the CSI prediction function, reducing the resource occupation of the base station 104.

The channel state information feedback enhancement method of the present application is described in detail below with a specific embodiment. As shown in FIG14 , in one embodiment, a schematic diagram of signaling interaction between a base station 104 and a user terminal 102 is shown, and the CSI prediction function is performed by the user terminal 102. The specific execution steps are as follows:

In step S300, the base station 104 sends a capability query message UECapabilityEnquiry to the user terminal 102 to obtain whether the user terminal 102 supports CSI compression based on an artificial intelligence model or CSI prediction capability based on an artificial intelligence model, and obtains the CSI compression model type (CNN, RNN, Transformer, ResNet, etc.) and CSI prediction model type (FCN, RNN, 3D-CNN, etc.) supported by the user terminal 102 from the capability information.

Step S301, the user terminal 102 sends a capability query result, namely, capability information UECapabilityInformation, to the base station 104, and two information elements are added to the capability information of the user terminal 102, as follows:

ai-CsiEncoderDecoder ENUMERATED{supported}OPTIONAL,

ai-CsiPrediction ENUMERATED{supported}OPTIONAL,

ai-CsiEncoderDecoder&PredictionENUMERATED{supported}OPTIONAL,

ai-CsiEncoderDecoder-Type ENUMERATED{CNN,RNN,Transformer,ResNet,…}OPTIONAL,

ai-CsiPrediction-Type ENUMERATED{FCN,RNN,3D-CNN,…}OPTIONAL,

In step S302, the base station 104 decides whether to execute the CSI compression and CSI prediction functions based on the capability information reported by the user terminal 102 and the current business scenario information. The specific judgment conditions are the same as those in the above embodiment. If the current base station 104 determines that the user terminal 102 meets the requirements for enabling the CSI compression and CSI prediction functions, it proceeds to step S303.

In step S303, the base station 104 configures CSI measurement parameters for artificial intelligence model training through an RRCReconfiguration message according to the capability information of the user terminal 102 obtained in step S301.

In step S304, the user terminal 102 measures the CSI according to the CSI measurement parameters configured by the base station 104, and transmits the CSI measurement result as CSI training data to the base station 104 through the user plane for training the CSI compression model and the CSI prediction model.

In step S305 , the base station 104 performs model and parameter training according to the processes in FIG. 3 and FIG. 4 based on the CSI training data obtained from the user terminal 102 and the CSI compression model and CSI prediction model supported by the user terminal 102 obtained in step S301 .

Step S306, after the base station 104 completes the training of the CSI compression model and the CSI prediction model, it notifies the user terminal 102 of the type of CSI compression model used and compression parameters, and the type of CSI prediction model and prediction parameters, wherein: the types of CSI compression models include: CNN, RNN, Transformer, etc., and the CSI compression parameters include: preprocessing method, quantization method, etc.; the types of CSI prediction models include: FCN, RNN, 3D-CNN, etc., and the CSI prediction parameters include: predicted delay time (such as 3ms, 4ms, 5ms, etc.), etc.

Step S307 , the base station 104 configures CSI measurement parameters for service scheduling through an RRCReconfiguration message.

In step S308, the user terminal 102 downloads the CSI prediction model from the OTT (Over-The-Top) server based on the type and prediction parameters of the CSI prediction model obtained in step S306, and performs CSI prediction reasoning with reference to the process in FIG. 4 after CSI measurement.

In step S309, the user terminal 102 downloads the CSI compression model from the OTT server based on the type and compression parameters of the CSI compression model obtained in step S306, and performs CSI compression inference on the compressed CSI generated in step S309 according to the process in reference Figure 8.

Step S310, the user terminal 102 reports compressed CSI information;

In step S311 , the base station 104 performs CSI decompression inference on the compressed CSI information reported by the user terminal 102 to obtain specific CSI information reported by the user terminal 102 .

In step S312 , the base station 104 schedules the current service based on the acquired CSI specific information and transmits data to the user terminal 102 .

The channel state information feedback enhancement method of the present application is described in detail below with another specific embodiment. As shown in FIG15 , it is a schematic diagram of signaling interaction between the base station 104 and the user terminal 102 in another embodiment, and the CSI prediction function is performed by the base station 104. The specific execution steps are as follows:

In step S400, the base station 104 sends a capability query message UECapabilityEnquiry to the user terminal 102 to obtain whether the user terminal 102 supports CSI compression based on an artificial intelligence model or CSI prediction capability based on an artificial intelligence model, and obtains the CSI compression model type (CNN, RNN, Transformer, ResNet, etc.) and CSI prediction model type (FCN, RNN, 3D-CNN, etc.) supported by the user terminal 102 from the capability information.

Step S401, the user terminal 102 sends a capability query result, namely, capability information UECapabilityInformation, to the base station 104, and two information elements are added to the capability information of the user terminal 102, as follows:

ai-CsiEncoderDecoder ENUMERATED{supported}OPTIONAL,

ai-CsiPrediction ENUMERATED{supported}OPTIONAL,

ai-CsiEncoderDecoder&PredictionENUMERATED{supported}OPTIONAL,

ai-CsiEncoderDecoder-Type ENUMERATED{CNN,RNN,Transformer,ResNet,…}OPTIONAL,

ai-CsiPrediction-Type ENUMERATED{FCN,RNN,3D-CNN,…}OPTIONAL,

Step S402: The base station 104 determines whether to perform CSI compression and CSI prediction functions based on the capability information reported by the user terminal 102 and the current service scenario information. The conditions for the interruption are the same as those in the above embodiment. If the current base station 104 determines that the user terminal 102 meets the requirements for enabling the CSI compression and CSI prediction functions, the process goes to step S403.

In step S403, the base station 104 configures CSI measurement parameters for artificial intelligence model training through an RRCReconfiguration message according to the capability information of the user terminal 102 obtained in step S401.

In step S404, the user terminal 102 measures the CSI according to the CSI measurement parameters configured by the base station 104, and uses the CSI measurement results as CSI training data, and transmits them to the base station 104 through the user plane to train the CSI compression model and the CSI prediction model.

In step S405 , the base station 104 performs model and parameter training according to the processes in FIG. 3 and FIG. 4 based on the CSI training data obtained from the user terminal 102 and the CSI compression model and CSI prediction model supported by the user terminal 102 obtained in step S401 .

In step S406, after the base station 104 completes the training of the CSI compression model and the CSI prediction model, it only needs to notify the user terminal 102 of the type of the CSI compression model used and the compression parameters, where: the types of CSI compression models include: CNN, RNN, Transformer, etc., and the CSI compression parameters include: preprocessing method, quantization method, etc.

Step S407, the base station 104 configures CSI measurement parameters for service scheduling through an RRCReconfiguration message.

In step S408, the user terminal 102 downloads the CSI compression model from the OTT (Over-The-Top) server based on the type and compression parameters of the CSI compression model obtained in step S406, and performs CSI compression inference by referring to the process in FIG. 3 after CSI measurement.

Step S409: The user terminal 102 reports compressed CSI information.

In step S410 , the base station 104 performs CSI decompression inference on the compressed CSI information reported by the user terminal 102 , and obtains the CSI compression inference result reported by the user terminal 102 .

In step S411, the base station 104 performs CSI prediction reasoning based on the CSI compression inference result obtained in step S410 and the CSI prediction model, with reference to the process of the left block diagram in FIG. 10 , to obtain CSI specific information.

In step S412 , the base station 104 schedules the current service based on the acquired CSI specific information and transmits data to the user terminal 102 .

It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

Based on the same inventive concept, the embodiment of the present application also provides a channel state information feedback enhancement device for implementing the channel state information feedback enhancement method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in the one or more channel state information feedback enhancement device embodiments provided below can refer to the limitations of the channel state information feedback enhancement method above, and will not be repeated here.

In one embodiment, as shown in FIG. 16 , a channel state information feedback enhancement device is provided, which is applied to a base station 104, and includes: a capability information acquisition module 510, a service scenario information acquisition module 520, a measurement mode calculation module 530, a measurement parameter output module 540, a measurement result acquisition module 550, a model training module 560 and a specific information calculation module 570, wherein:

The capability information acquisition module 510 is used to acquire the capability information reported by the user terminal 102; wherein the capability information includes the model type of the artificial intelligence model supported by the user terminal 102;

The service scenario information acquisition module 520 is used to acquire the current service scenario information; wherein the service scenario information includes: channel state information CSI feedback period, user terminal moving speed and channel state information CSI feedback bit number;

The measurement mode calculation module 530 is used to obtain a channel state information CSI measurement mode according to the capability information and the service scenario information; wherein the channel state information CSI measurement mode includes one of a prediction measurement mode, a compression measurement mode and a compression and prediction measurement mode;

The measurement parameter output module 540 is used to configure and output the channel state information CSI measurement parameters according to the channel state information CSI measurement mode;

The measurement result acquisition module 550 is used to obtain the channel state information CSI measurement result reported by the user terminal 102; wherein the channel state information CSI measurement result is obtained by the user terminal 102 measuring the channel state information CSI according to the channel state information CSI measurement parameter;

A model training module 560 is used to train an artificial intelligence model according to the channel state information CSI measurement result to obtain a target model;

The specific information calculation module 570 is used to obtain the channel state information CSI specific information according to the target model to schedule the current service.

For the specific limitations of the channel state information feedback enhancement device implemented from the perspective of the base station 104, please refer to the limitations of the channel state information feedback enhancement method implemented from the perspective of the base station 104 above, which will not be repeated here. Each module in the above-mentioned channel state information feedback enhancement device implemented from the perspective of the base station 104 can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules. It should be noted that the division of modules in the embodiment of the present application is schematic and is only a logical function division. There may be other division methods in actual implementation.

In one embodiment, the measurement mode calculation module is further used to determine whether the user terminal supports the target model type according to the capability information; if the target model type is supported, the channel state information CSI measurement mode is determined according to the service scenario information.

In one embodiment, the measurement mode calculation module is further used to determine the channel state information CSI measurement mode as the prediction measurement mode when the channel state information CSI feedback period is greater than a first threshold or the user terminal moving speed is greater than a second threshold; when the channel state information CSI feedback bit number is greater than a third threshold, the channel state information CSI measurement mode is determined as the prediction measurement mode. Determined as a compressed measurement mode; when the channel state information CSI feedback period is greater than the first threshold or the user terminal moving speed is greater than the second threshold, and the number of channel state information CSI feedback bits is greater than the third threshold, the channel state information CSI measurement mode is determined to be a compressed and predicted measurement mode.

In one embodiment, the target model includes: a channel state information CSI compression model and a channel state information CSI prediction model, a specific information calculation module, and is also used to send the channel state information CSI compression model and the channel state information CSI prediction model to a user terminal; obtain first channel state information CSI compression information reported by the user terminal; wherein the first channel state information CSI compression information is obtained by the user terminal compressing the first channel state information CSI prediction information according to the channel state information CSI compression model, and the first channel state information CSI prediction information is obtained by the user terminal predicting the first channel state information CSI measurement information according to the channel state information CSI prediction model; the first channel state information CSI compression information is decompressed according to the channel state information CSI compression model to obtain channel state information CSI specific information.

In one embodiment, the target model includes: a channel state information CSI compression model and a channel state information CSI prediction model, a specific information calculation module, and is also used to send the channel state information CSI compression model to a user terminal; obtain second channel state information CSI compression information reported by the user terminal; wherein the second channel state information CSI compression information is obtained by the user terminal compressing second channel state information CSI measurement information according to the channel state information CSI compression model; decompressing the second channel state information CSI compression information according to the channel state information CSI compression model to obtain second channel state information CSI decompression information; predicting the second channel state information CSI decompression information according to the channel state information CSI prediction model to obtain channel state information CSI specific information.

In one embodiment, as shown in FIG. 17 , a channel state information feedback enhancement device is also provided, which is applied to the user terminal 102, and includes: a capability information reporting module 610, a measurement parameter acquisition module 620, a measurement result calculation module 630 and a model acquisition module 640, wherein:

A capability information reporting module 610 is used to report capability information to the base station; wherein the capability information includes the model type of the artificial intelligence model supported by the user terminal 102;

The measurement parameter acquisition module 620 is used to acquire a channel state information CSI measurement parameter; wherein the channel state information CSI measurement parameter is obtained by the base station according to a channel state information CSI measurement mode configuration, the channel state information CSI measurement mode is obtained by the base station according to capability information and service scenario information, the channel state information CSI measurement mode includes a prediction measurement mode, a compression measurement mode, and a compression and prediction measurement mode, and the service scenario information includes: a channel state information CSI feedback period, a user terminal moving speed, and a channel state information CSI feedback bit number;

The measurement result calculation module 630 is used to measure the channel state information CSI according to the channel state information CSI measurement parameter to obtain the channel state information CSI measurement result;

The model acquisition module 640 is used to acquire the target model and obtain the specific information of the channel state information CSI according to the target model; wherein the target model is obtained by training the artificial intelligence model according to the channel state information CSI measurement results of the base station, and the specific information of the channel state information CSI is used to schedule the current service.

For the specific limitations of the channel state information feedback enhancement device implemented from the perspective of the user terminal 102, please refer to the limitations of the channel state information feedback enhancement method implemented from the perspective of the user terminal 102 above, which will not be repeated here. Each module in the above-mentioned channel state information feedback enhancement device implemented from the perspective of the user terminal 102 can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules. It should be noted that the division of modules in the embodiment of the present application is schematic and is only a logical function division. There may be other division methods in actual implementation.

In one embodiment, the channel state information CSI measurement method is obtained by the base station based on capability information and business scenario information, including: the capability information is used to indicate the base station to determine whether the user terminal supports the target model type. When the target model type is supported, the channel state information CSI measurement method is determined based on the business scenario information.

In one embodiment, the step of determining a channel state information CSI measurement mode according to business scenario information includes: when a channel state information CSI feedback period is greater than a first threshold or a user terminal moving speed is greater than a second threshold, the channel state information CSI measurement mode is determined as a predictive measurement mode; when the number of channel state information CSI feedback bits is greater than a third threshold, the channel state information CSI measurement mode is determined as a compressed measurement mode; when the channel state information CSI feedback period is greater than a first threshold or a user terminal moving speed is greater than a second threshold, and the number of channel state information CSI feedback bits is greater than a third threshold, the channel state information CSI measurement mode is determined as a compressed and predictive measurement mode.

In one embodiment, the target model includes: a channel state information CSI compression model and a channel state information CSI prediction model, and a model acquisition module, which is also used to obtain the channel state information CSI compression model and the channel state information CSI prediction model sent by the base station; predict the first channel state information CSI measurement information according to the channel state information CSI prediction model to obtain the first channel state information CSI prediction information; compress the first channel state information CSI prediction information according to the channel state information CSI compression model to obtain the first channel state information CSI compressed information; report the first channel state information CSI compressed information to the base station to obtain the channel state information CSI specific information; wherein the channel state information CSI specific information is obtained by the base station decompressing the first channel state information CSI compressed information according to the channel state information CSI compression model.

In one embodiment, the target model includes: a channel state information CSI compression model and a channel state information CSI prediction model, and a model acquisition module, which is also used to obtain the channel state information CSI compression model sent by the base station; compress the second channel state information CSI measurement information according to the channel state information CSI compression model to obtain second channel state information CSI compression information; report the second channel state information CSI compression information to the base station to obtain channel state information CSI specific information; wherein the channel state information CSI specific information is obtained by the base station predicting the second channel state information CSI decompression information according to the channel state information CSI prediction model, and the second channel state information CSI decompression information is obtained by the base station decompressing the second channel state information CSI compression information according to the channel state information CSI compression model.

In one embodiment, a channel state information feedback enhancement system is provided, including a base station and a user terminal 102 connected to the base station; wherein: the base station is used to execute the steps of the channel state information feedback enhancement method implemented from the perspective of the base station; the user terminal 102 is used to execute the steps of the channel state information feedback enhancement method implemented from the perspective of the user terminal 102.

In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When a processor executes the computer program, the steps in the above-mentioned method embodiments are implemented.

A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, etc., but are not limited to this.

The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims (22)

A channel state information feedback enhancement method, the method is applied to a base station, the method comprising: Acquire capability information reported by the user terminal; wherein the capability information includes a model type of an artificial intelligence model supported by the user terminal; Acquire current service scenario information; wherein the service scenario information includes: channel state information CSI feedback period, user terminal moving speed and channel state information CSI feedback bit number; Obtaining a channel state information CSI measurement mode according to the capability information and the service scenario information; wherein the channel state information CSI measurement mode includes one of a prediction measurement mode, a compression measurement mode, and a compression and prediction measurement mode; Configure and output channel state information CSI measurement parameters according to the channel state information CSI measurement method; Acquire a channel state information CSI measurement result reported by the user terminal; wherein the channel state information CSI measurement result is obtained by the user terminal measuring the channel state information CSI according to the channel state information CSI measurement parameter; The artificial intelligence model is trained according to the channel state information CSI measurement result to obtain a target model; The channel state information CSI specific information is obtained according to the target model to schedule the current service. According to the method of claim 1, the step of obtaining a channel state information CSI measurement method according to the capability information and the service scenario information comprises: Determining whether the user terminal supports the target model type according to the capability information; When the target model type is supported, the channel state information CSI measurement method is determined according to the service scenario information. According to the method of claim 2, the step of determining the channel state information CSI measurement mode according to the service scenario information comprises: When the channel state information CSI feedback period is greater than a first threshold or the moving speed of the user terminal is greater than a second threshold, determining the channel state information CSI measurement mode as a prediction measurement mode; When the number of channel state information CSI feedback bits is greater than a third threshold, determining the channel state information CSI measurement mode as a compressed measurement mode; When the channel state information CSI feedback period is greater than a first threshold or the moving speed of the user terminal is greater than a second threshold, and the number of channel state information CSI feedback bits is greater than a third threshold, the channel state information CSI measurement mode is determined to be a compression and prediction measurement mode. According to the method of claim 1, the target model includes: a channel state information CSI compression model and a channel state information CSI prediction model, and the step of obtaining channel state information CSI specific information according to the target model includes: Sending the channel state information CSI compression model and the channel state information CSI prediction model to the user terminal; Acquire first channel state information CSI compression information reported by the user terminal; wherein the first channel state information CSI compression information is obtained by the user terminal compressing first channel state information CSI prediction information according to the channel state information CSI compression model, and the first channel state information CSI prediction information is obtained by the user terminal predicting first channel state information CSI measurement information according to the channel state information CSI prediction model; The first channel state information CSI compressed information is decompressed according to the channel state information CSI compression model to obtain the channel state information CSI specific information. According to the method of claim 1, the target model includes: a channel state information CSI compression model and a channel state information CSI prediction model, and the step of obtaining channel state information CSI specific information according to the target model includes: Sending the channel state information CSI compression model to the user terminal; Acquire second channel state information CSI compression information reported by the user terminal; wherein the second channel state information CSI compression information is obtained by the user terminal compressing the second channel state information CSI measurement information according to the channel state information CSI compression model; Decompressing the second channel state information CSI compressed information according to the channel state information CSI compression model to obtain second channel state information CSI decompressed information; The second channel state information CSI decompression information is predicted according to the channel state information CSI prediction model to obtain the channel state information CSI specific information. A channel state information feedback enhancement method, the method is applied to a user terminal, the method comprising: Reporting capability information to a base station; wherein the capability information includes a model type of an artificial intelligence model supported by the user terminal; Acquire a channel state information CSI measurement parameter; wherein the channel state information CSI measurement parameter is obtained by the base station according to a channel state information CSI measurement mode configuration, the channel state information CSI measurement mode is obtained by the base station according to the capability information and the service scenario information, the channel state information CSI measurement mode includes a prediction measurement mode, a compression measurement mode, and a compression and prediction measurement mode, and the service scenario information includes: a channel state information CSI feedback period, a user terminal moving speed, and a channel state information CSI feedback bit number; Measuring the channel state information CSI according to the channel state information CSI measurement parameter to obtain a channel state information CSI measurement result; Acquire a target model, and obtain channel state information CSI specific information according to the target model; wherein the target model is obtained by the base station training the artificial intelligence model according to the channel state information CSI measurement result, and the channel state information CSI specific information is used to schedule the current service. According to the method of claim 6, the channel state information CSI measurement method is obtained by the base station according to the capability information and business scenario information, including: the capability information is used to instruct the base station to determine whether the user terminal supports the target model type, and when the target model type is supported, the channel state information CSI measurement method is determined according to the business scenario information. According to the method of claim 7, the step of determining the channel state information CSI measurement mode according to the service scenario information comprises: When the channel state information CSI feedback period is greater than a first threshold or the moving speed of the user terminal is greater than a second threshold, determining the channel state information CSI measurement mode as a prediction measurement mode; When the number of channel state information CSI feedback bits is greater than a third threshold, determining the channel state information CSI measurement mode as a compressed measurement mode; When the channel state information CSI feedback period is greater than a first threshold or the moving speed of the user terminal is greater than a second threshold, and the number of channel state information CSI feedback bits is greater than a third threshold, the channel state information CSI measurement mode is determined to be a compression and prediction measurement mode. According to the method of claim 6, the target model includes: a channel state information CSI compression model and a channel state information CSI prediction model, and the step of acquiring the target model and obtaining the channel state information CSI specific information according to the target model includes: Acquire the channel state information CSI compression model and the channel state information CSI prediction model sent by the base station; Predicting first channel state information CSI measurement information according to the channel state information CSI prediction model to obtain first channel state information CSI prediction information; Compressing the first channel state information CSI prediction information according to the channel state information CSI compression model to obtain first channel state information CSI compressed information; Report the first channel state information CSI compression information to the base station to obtain the channel state information CSI specific information; wherein the channel state information CSI specific information is obtained by the base station decompressing the first channel state information CSI compression information according to the channel state information CSI compression model. According to the method of claim 6, the target model includes: a channel state information CSI compression model and a channel state information CSI prediction model, and the step of acquiring the target model and obtaining the channel state information CSI specific information according to the target model includes: Acquire the channel state information CSI compression model sent by the base station; Compressing the second channel state information CSI measurement information according to the channel state information CSI compression model to obtain second channel state information CSI compressed information; Report the second channel state information CSI compression information to the base station to obtain the channel state information CSI specific information; wherein the channel state information CSI specific information is obtained by the base station predicting the second channel state information CSI decompression information according to the channel state information CSI prediction model, and the second channel state information CSI decompression information is obtained by the base station decompressing the second channel state information CSI compression information according to the channel state information CSI compression model. A channel state information feedback enhancement device, the device is applied to a base station, and the device comprises: A capability information acquisition module, used to acquire capability information reported by a user terminal; wherein the capability information includes a model type of an artificial intelligence model supported by the user terminal; A service scenario information acquisition module is used to acquire current service scenario information; wherein the service scenario information includes: channel state information CSI feedback period, user terminal moving speed and channel state information CSI feedback bit number; A measurement mode calculation module, configured to obtain a channel state information CSI measurement mode according to the capability information and the service scenario information; wherein the channel state information CSI measurement mode includes one of a prediction measurement mode, a compression measurement mode, and a compression and prediction measurement mode; A measurement parameter output module, configured to configure and output a channel state information CSI measurement parameter according to the channel state information CSI measurement mode; A measurement result acquisition module, used to acquire the channel state information CSI measurement result reported by the user terminal; wherein the channel state information CSI measurement result is obtained by the user terminal measuring the channel state information CSI according to the channel state information CSI measurement parameter; A model training module, used to train the artificial intelligence model according to the channel state information CSI measurement result to obtain a target model; The specific information calculation module is used to obtain channel state information CSI specific information according to the target model to schedule the current service. According to the device according to claim 11, the measurement method calculation module is also used to determine whether the user terminal supports the target model type according to the capability information; when the target model type is supported, the channel state information CSI measurement method is determined according to the business scenario information. According to the device according to claim 12, the measurement mode calculation module is also used to determine the channel state information CSI measurement mode as a predictive measurement mode when the channel state information CSI feedback period is greater than a first threshold or the user terminal moving speed is greater than a second threshold; when the channel state information CSI feedback bit number is greater than a third threshold, determine the channel state information CSI measurement mode as a compressed measurement mode; when the channel state information CSI feedback period is greater than the first threshold or the user terminal moving speed is greater than the second threshold, and the channel state information CSI feedback bit number is greater than the third threshold, determine the channel state information CSI measurement mode as a compressed and predictive measurement mode. According to the device of claim 11, the target model includes: a channel state information CSI compression model and a channel state information CSI prediction model, and the specific information calculation module is also used to send the channel state information CSI compression model and the channel state information CSI prediction model to the user terminal; obtain the first channel state information CSI compression information reported by the user terminal; wherein the first channel state information CSI compression information is obtained by the user terminal compressing the first channel state information CSI prediction information according to the channel state information CSI compression model, and the first channel state information CSI prediction information is obtained by the user terminal predicting the first channel state information CSI measurement information according to the channel state information CSI prediction model; the first channel state information CSI compression information is decompressed according to the channel state information CSI compression model to obtain the channel state information CSI specific information. According to the device of claim 11, the target model includes: a channel state information CSI compression model and a channel state information CSI prediction model, and the specific information calculation module is also used to send the channel state information CSI compression model to the user terminal; obtain the second channel state information CSI compression information reported by the user terminal; wherein the second channel state information CSI compression information is obtained by the user terminal compressing the second channel state information CSI measurement information according to the channel state information CSI compression model; decompressing the second channel state information CSI compression information according to the channel state information CSI compression model to obtain second channel state information CSI decompression information; predicting the second channel state information CSI decompression information according to the channel state information CSI prediction model to obtain the channel state information CSI specific information. information. A channel state information feedback enhancement device, the device is applied to a user terminal, and the device comprises: A capability information reporting module, used to report capability information to a base station; wherein the capability information includes a model type of an artificial intelligence model supported by the user terminal; A measurement parameter acquisition module, configured to acquire a channel state information CSI measurement parameter; wherein the channel state information CSI measurement parameter is obtained by the base station according to a channel state information CSI measurement mode configuration, the channel state information CSI measurement mode is obtained by the base station according to the capability information and the service scenario information, the channel state information CSI measurement mode includes one of a prediction measurement mode, a compression measurement mode and a compression and prediction measurement mode, and the service scenario information includes: a channel state information CSI feedback period, a user terminal moving speed and a channel state information CSI feedback bit number; A measurement result calculation module, used to measure the channel state information CSI according to the channel state information CSI measurement parameter to obtain a channel state information CSI measurement result; A model acquisition module is used to acquire a target model and obtain channel state information CSI specific information according to the target model; wherein the target model is obtained by the base station training the artificial intelligence model according to the channel state information CSI measurement results, and the channel state information CSI specific information is used to schedule the current service. According to the device of claim 16, the channel state information CSI measurement method is obtained by the base station according to the capability information and business scenario information, including: the capability information is used to indicate the base station to determine whether the user terminal supports the target model type, and when the target model type is supported, the channel state information CSI measurement method is determined according to the business scenario information. According to the device of claim 17, the step of determining the channel state information CSI measurement mode according to the business scenario information includes: when the channel state information CSI feedback period is greater than a first threshold or the user terminal moving speed is greater than a second threshold, the channel state information CSI measurement mode is determined as a predictive measurement mode; when the channel state information CSI feedback bit number is greater than a third threshold, the channel state information CSI measurement mode is determined as a compressed measurement mode; when the channel state information CSI feedback period is greater than a first threshold or the user terminal moving speed is greater than a second threshold, and the channel state information CSI feedback bit number is greater than a third threshold, the channel state information CSI measurement mode is determined as a compressed and predictive measurement mode. According to the device of claim 16, the target model includes: a channel state information CSI compression model and a channel state information CSI prediction model, and the model acquisition module is also used to obtain the channel state information CSI compression model and the channel state information CSI prediction model sent by the base station; predict the first channel state information CSI measurement information according to the channel state information CSI prediction model to obtain first channel state information CSI prediction information; compress the first channel state information CSI prediction information according to the channel state information CSI compression model to obtain first channel state information CSI compression information; report the first channel state information CSI compression information to the base station to obtain the channel state information CSI specific information; wherein, the channel state information CSI specific information is obtained by the base station decompressing the first channel state information CSI compression information according to the channel state information CSI compression model. According to the device of claim 16, the target model includes: a channel state information CSI compression model and a channel state information CSI prediction model, and the model acquisition module is also used to obtain the channel state information CSI compression model sent by the base station; compress the second channel state information CSI measurement information according to the channel state information CSI compression model to obtain second channel state information CSI compression information; report the second channel state information CSI compression information to the base station to obtain the channel state information CSI specific information; wherein the channel state information CSI specific information is obtained by the base station predicting the second channel state information CSI decompression information according to the channel state information CSI prediction model, and the second channel state information CSI decompression information is obtained by the base station decompressing the second channel state information CSI compression information according to the channel state information CSI compression model. A channel state information feedback enhancement system comprises a base station and a user terminal connected to the base station; wherein: The base station is used to perform the steps of the method according to any one of claims 1 to 5; The user terminal is used to execute the steps of the method according to any one of claims 6 to 10. A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the method according to any one of claims 1 to 10.