Communication method and device

Abstract

The invention provides a communication method and device. In the method, a first device side performs first processing on first channel information based on at least one model to obtain N modulation symbols; n is a positive integer; determining a performance monitoring result of the at least one model based on the N modulation symbols and the M known modulation symbols; m is a positive integer; wherein the first processing comprises coding and modulation. Based on the method and the device, the performance of at least one model can be effectively monitored.

Description

Technical Field

[0001] This application relates to the field of wireless communication, and more particularly to a communication method and apparatus. Background Technology

[0002] In wireless communication systems, artificial intelligence (AI) models can be used to compress and reconstruct (or restore) channel information. The sender of a CSI report (such as a terminal device) can use one or more AI models to encode and modulate channel information to obtain a Channel State Information (CSI) report, which is then transmitted. The receiver of the CSI report (such as a network device) can use one or more AI models to decode and reconstruct the CSI report based on the received report.

[0003] There is an urgent need to provide a solution for monitoring the performance of the AI ​​model used by the sender of CSI reports in order to ensure the reliability of CSI reports. Summary of the Invention

[0004] This application provides a communication method and apparatus to facilitate the monitoring of the performance of the AI ​​model used for CSI feedback, i.e., the AI ​​model used by the sender of the CSI report.

[0005] Firstly, a communication method is provided, which can be applied to a first device.

[0006] For example, the first device side can be replaced by the terminal side. The terminal side can be a terminal device; or, a component deployed in the terminal device, such as circuits or chips inside the terminal device (such as modem chips, baseband chips, or system-on-chip (SoC) chips or system-in-package (SIP) chips containing modem cores, etc.); or, it can be a device deployed outside the terminal device (such as a server, such as the host of an over-the-top (OTT) system or a cloud server) or a component in the device (such as chips, processors, or circuits inside the device, etc.); or, it can be a logic module or software that can implement all or part of the functions of the terminal device, etc.

[0007] For example, the first device side can be replaced by the network side. The network side can be a network device; or it can be a component deployed in the network device, such as the circuits or chips inside the network device (such as a modem chip, or a baseband chip, or a SoC chip or SIP chip containing a modem core, etc.); or it can be a device outside the network device (such as a smart network element or server on the network side, etc.) or a component in the device (such as a chip, processor or circuit inside the device, etc.); or it can be a logic module or software that can realize all or part of the functions of the network device, etc.

[0008] For example, the method includes: performing a first processing on first channel information based on at least one model to obtain N modulation symbols; N is a positive integer; determining the performance monitoring result of at least one model based on the N modulation symbols and M known modulation symbols; M is a positive integer; wherein the first processing includes coding and modulation.

[0009] Based on the above technical solution, the first device can perform first processing on the first channel information based on at least one model to obtain N modulation symbols. Therefore, the first device can determine the performance monitoring results of at least one model based on M known modulation symbols and N modulation symbols, which is beneficial to determining effective performance monitoring results.

[0010] For example, the above encoding may include source coding and channel coding, or the above encoding may be joint source-channel coding. Source coding can also be referred to as compression.

[0011] For example, the type of the first channel information is a channel matrix, a channel feature matrix, channel response information, or a precoding matrix.

[0012] This application does not limit the use of at least one model. For example, at least one model includes a first model and a second model; the first model is used to encode the first channel information to obtain a first bit sequence; the second model is used to modulate the first bit sequence to obtain N modulation symbols.

[0013] In conjunction with the first aspect, in some implementations of the first aspect, the M known modulation symbols are predefined modulation symbols; or, the method further includes: receiving the M known modulation symbols.

[0014] Based on the above technical solution, if the M known modulation symbols are predefined modulation symbols, the second device does not need to send the M known modulation symbols to the first device, thus reducing the interaction between the first and second devices. If the second device sends the M known modulation symbols to the first device, the processing complexity of the first device in obtaining the M known modulation symbols can be reduced.

[0015] For example, performance monitoring results may be related to one or more of the following: a first mean absolute error (MAE); a first mean square error (MSE); a first square generalized cosine similarity (SGCS); a second MAE; a second MSE; a second SGCS; a weighted sum or weighted average of at least two of the first MAE, first MSE, first SGCS, second MAE, second MSE, or second SGCS. For instance, performance monitoring results may include one or more of the above, or be determined based on one or more of the above.

[0016] The first MAE is determined based on the MAE of the first modulation symbol and each of the K second modulation symbols.

[0017] The first MSE is determined based on the first modulation symbol and the MSE of each of the K second modulation symbols.

[0018] The first SGCS is determined based on the SGCS of the first modulation symbol and each of the K second modulation symbols.

[0019] The second MAE is determined based on the MAE of the third modulation symbol and each of the K fourth modulation symbols.

[0020] The second MSE is determined based on the MSE of the third modulation symbol and each of the K fourth modulation symbols.

[0021] The second SGCS is determined based on the SGCS of the third modulation symbol and each of the K fourth modulation symbols.

[0022] Where K is a positive integer less than or equal to M.

[0023] The first modulation symbol belongs to N modulation symbols, and the K second modulation symbols belong to M known modulation symbols; the geometric distance between the second modulation symbol and the first modulation symbol is less than or equal to the geometric distance between any of the M known modulation symbols other than the K second modulation symbols and the first modulation symbol.

[0024] The third modulation symbol belongs to N modulation symbols, and the K fourth modulation symbols belong to M known modulation symbols; the geometric distance between the fourth modulation symbol and the third modulation symbol is less than or equal to the geometric distance between any of the M known modulation symbols other than the K fourth modulation symbols and the third modulation symbol.

[0025] For example, determining the performance monitoring results of at least one model based on N modulation symbols and M known modulation symbols includes: determining the performance monitoring results of at least one model based on a first modulation symbol and K second modulation symbols.

[0026] For example, determining the performance monitoring results of at least one model based on a first modulation symbol and K second modulation symbols includes: determining the performance monitoring results of at least one model based on one or more of a first MAE, a first MSE, or a first SGCS.

[0027] For example, determining the performance monitoring results of at least one model based on N modulation symbols and M known modulation symbols includes: determining the performance monitoring results of at least one model based on a first modulation symbol, K second modulation symbols, a third modulation symbol, and K fourth modulation symbols.

[0028] For example, determining the performance monitoring result of at least one model based on a first modulation symbol, K second modulation symbols, a third modulation symbol, and K fourth modulation symbols includes: determining the performance monitoring result of at least one model based on one or more of a first MAE, a first MSE, a first SGCS, a second MAE, a second MSE, or a second SGCS.

[0029] In conjunction with the first aspect, in some implementations of the first aspect, before performing the first processing on the first channel information based on at least one model, the method further includes: acquiring the first channel information through channel measurement or simulation.

[0030] Based on the above technical solution, if the first device side obtains the first channel information through channel measurement or simulation, and the M known modulation symbols are predefined modulation symbols, then the first device side can monitor the performance of at least one model without receiving a dataset from the second device side.

[0031] In conjunction with the first aspect, in some implementations of the first aspect, before performing the first processing on the first channel information based on at least one model, the method further includes: receiving the first channel information and a second bit sequence from another device, wherein the second bit sequence is obtained after the other device encodes the first channel information; determining the performance monitoring result of at least one model based on N modulation symbols and M known modulation symbols, including: determining a first monitoring result included in the performance monitoring result based on N modulation symbols and M known modulation symbols; and determining a second monitoring result included in the performance monitoring result based on the first bit sequence and the second bit sequence.

[0032] Based on the above technical solution, if at least one model includes a first model and a second model, the first device can also determine a second monitoring result based on the first bit sequence and the second bit sequence, thereby realizing performance monitoring of the first model.

[0033] For example, the second monitoring result includes one or more of the following: first binary cross-entropy (BCE); third MSE; third MAE; a weighted sum or weighted average of at least two of the first BCE, third MSE and third MAE.

[0034] The first BCE is determined based on the probability of each bit in the first bit sequence being 1 and the second bit sequence.

[0035] The third MSE is the MSE between the first bit sequence and the second bit sequence.

[0036] The third MAE is the MAE between the first bit sequence and the second bit sequence.

[0037] In conjunction with the first aspect, in some implementations of the first aspect, the method also includes: sending performance monitoring results.

[0038] For example, based on the first bit sequence and the second bit sequence, the performance monitoring results are determined to include a second monitoring result, which includes: determining the second monitoring result based on one or more of the first BCE, the third MSE, or the third MAE.

[0039] Based on the above technical solution, the first device sends performance monitoring results to the second device, which helps the second device to promptly detect that the performance of at least one model does not meet the performance requirements, and then promptly perform operations such as updating, switching, or rolling back at least one model.

[0040] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: receiving first information, the first information being used to instruct retraining of at least one model.

[0041] Based on the above technical solution, the second device sends first information to the first device to instruct the first device to perform operations such as retraining, switching, or rollback on at least one model. This can avoid the problem of communication performance degradation caused by the first device using at least one model whose performance does not meet the requirements for CSI feedback.

[0042] In conjunction with the first aspect, in some implementations of the first aspect, determining the performance monitoring results of at least one model includes: determining the performance monitoring results of at least one model during the process of applying at least one model for model inference.

[0043] Based on the above technical solution, the first device side determines the performance monitoring results of at least one model during the model inference process using at least one model. This helps to determine whether the inference results output by at least one model are reliable based on the performance monitoring results of at least one model, thereby helping to avoid the degradation of communication performance caused by the first device side and the second device side communicating based on unreliable inference results.

[0044] In conjunction with the first aspect, in certain implementations of the first aspect, determining the performance monitoring results of at least one model includes: determining the performance monitoring results of at least one model during the training of at least one model.

[0045] Based on the above technical solution, the first device determines the performance monitoring results of at least one model during the training of at least one model, which is conducive to determining whether at least one model has been trained based on the performance monitoring results of at least one model.

[0046] Secondly, a communication method is provided, which can be applied to a first device. A description of the first device can be found in the description of the first aspect above.

[0047] For example, the method includes: receiving a dataset from a second device, the dataset including first channel information and a first modulation symbol sequence, the first modulation symbol sequence corresponding to the first channel information; the dataset is used for performance monitoring of at least one model.

[0048] Based on the above technical solution, the second device can send the first channel information and the first modulation symbol sequence to the first device, which is conducive to the first device performing performance monitoring of at least one model based on the received dataset.

[0049] Here, the first channel information corresponds to the first modulation symbol sequence. This can be understood as the first modulation symbol sequence being the modulation symbol sequence of the first channel information, or as the first modulation symbol sequence being obtained after the second device processes the first channel information. For example, the second device encodes and modulates the first channel information to obtain the first modulation symbol sequence.

[0050] The dataset is used for performance monitoring of at least one model, including: using first channel information as input to at least one model and using a first modulation symbol sequence as a label corresponding to the output of at least one model or a reference for performance monitoring, and performing performance monitoring on at least one model. It should be understood that when the first channel information corresponds to the first modulation symbol sequence, the first modulation symbol sequence can be used as a label corresponding to the output of at least one model or a reference for performance monitoring when the first channel information is used as input to at least one model. Furthermore, when channel information other than the first channel information is used as input to at least one model, the first modulation symbol sequence cannot be used as a label corresponding to the output of at least one model or a reference for performance monitoring.

[0051] For example, the type of the first channel information is a channel matrix, a channel feature matrix, channel response information, or a precoding matrix.

[0052] In conjunction with the second aspect, in some implementations of the second aspect, the method further includes: performing a first processing on the first channel information based on at least one model to obtain a second modulation symbol sequence; determining the performance monitoring result of at least one model based on the first modulation symbol sequence and the second modulation symbol sequence; wherein the first processing includes coding and modulation.

[0053] Based on the above technical solution, at least one model is used to perform first processing on the first channel information and output a second modulation symbol sequence. Then, the first device can effectively monitor the performance of at least one model based on the first modulation symbol sequence and the second modulation symbol sequence.

[0054] For example, the above encoding may include source coding and channel coding, or the above encoding may be joint source-channel coding. Channel coding can also be referred to as compression.

[0055] For example, performance monitoring results may include one or more of the following: the fourth MAE; the fourth MSE; the fourth SGCS; a weighted sum or weighted average of at least two of the fourth MAE, fourth MSE, or fourth SGCS.

[0056] The fourth MAE is the MAE between the first modulation symbol sequence and the second modulation symbol sequence.

[0057] The fourth MSE is the MSE between the first modulation symbol sequence and the second modulation symbol sequence.

[0058] The fourth SGCS is the SGCS between the first modulation symbol sequence and the second modulation symbol sequence.

[0059] For example, determining the performance monitoring results of at least one model based on a first modulation symbol sequence and a second modulation symbol sequence includes: determining the performance monitoring results of at least one model based on one or more of a fourth MAE, a fourth MSE, or a fourth SGCS.

[0060] In conjunction with the second aspect, in some implementations of the second aspect, at least one model includes a first model and a second model; the first model is used to encode the first channel information to obtain a first bit sequence; the second model is used to modulate the first bit sequence to obtain a second modulation symbol sequence; the dataset also includes a second bit sequence, which is obtained by encoding the first channel information of the second device; based on the first modulation symbol sequence and the second modulation symbol sequence, the performance monitoring results of at least one model are determined, including: based on the first modulation symbol sequence and the second modulation symbol sequence, determining a first monitoring result included in the performance monitoring results; and based on the first bit sequence and the second bit sequence, determining a second monitoring result included in the performance monitoring results.

[0061] Based on the above technical solution, the second device can also send a second bit sequence to the first device. Then, the first device encodes the first channel information based on the first model and outputs the first bit sequence. After that, the first device can effectively monitor the performance of the first model based on the first bit sequence and the second bit sequence.

[0062] For example, the second monitoring result is the monitoring result of the first model. That is, the first channel information and the second bit sequence are used to monitor the performance of the first model. The use of the first channel information and the second bit sequence for performance monitoring of the first model includes: using the first channel information as input to the first model, and using the second bit sequence as a tag corresponding to the output of the first model or a reference for performance monitoring, to perform performance monitoring of the first model. It should be understood that when the first channel information and the second bit sequence correspond, when the first channel information is used as input to the first model, the second bit sequence can be used as a tag corresponding to the output of the first model or a reference for performance monitoring. In other words, when channel information other than the first channel information is used as input to the first model, the second bit sequence cannot be used as a tag corresponding to the output of the first model or a reference for performance monitoring.

[0063] For example, the second monitoring result includes one or more of the following: first BCE; third MSE; third MAE; a weighted sum or weighted average of at least two of the first BCE, third MSE and third MAE.

[0064] The first BCE is determined based on the probability of each bit in the first bit sequence being 1 and the second bit sequence.

[0065] The third MSE is the MSE between the first bit sequence and the second bit sequence.

[0066] The third MAE is the MAE between the first bit sequence and the second bit sequence.

[0067] For example, based on the first bit sequence and the second bit sequence, the performance monitoring results are determined to include a second monitoring result, which includes: determining the second monitoring result based on one or more of the first BCE, the third MSE, or the third MAE.

[0068] In conjunction with the second aspect, in some implementations of the second aspect, the method further includes: sending performance monitoring results.

[0069] Based on the above technical solution, the first device sends performance monitoring results to the second device, which helps the second device to promptly detect that the performance of at least one model does not meet the performance requirements, and then promptly perform operations such as updating, switching, or rolling back at least one model.

[0070] In conjunction with the second aspect, in some implementations of the second aspect, the method further includes: receiving first information, the first information being used to instruct retraining of at least one model.

[0071] Based on the above technical solution, the second device sends first information to the first device to instruct the first device to perform operations such as retraining, switching, or rollback on at least one model. This can avoid the problem of communication performance degradation caused by the first device using at least one model whose performance does not meet the requirements for CSI feedback.

[0072] In conjunction with the second aspect, in some implementations of the second aspect, determining the performance monitoring results of at least one model includes: determining the performance monitoring results of at least one model during the process of applying at least one model for model inference.

[0073] Based on the above technical solution, the first device side determines the performance monitoring results of at least one model during the model inference process using at least one model. This helps to determine whether the inference results output by at least one model are reliable based on the performance monitoring results of at least one model, thereby helping to avoid the degradation of communication performance caused by the first device side and the second device side communicating based on unreliable inference results.

[0074] In conjunction with the second aspect, in certain implementations of the second aspect, determining the performance monitoring results of at least one model includes: determining the performance monitoring results of at least one model during the training of at least one model.

[0075] Based on the above technical solution, the first device determines the performance monitoring results of at least one model during the training of at least one model, which is conducive to determining whether at least one model has been trained based on the performance monitoring results of at least one model.

[0076] Thirdly, a communication method is provided that can be applied to the second device side.

[0077] For example, the second device side can be replaced by the terminal side or the network side. For more details on the terminal side and the network side, please refer to the first aspect above.

[0078] For example, the method includes: sending M known modulation symbols; the M known modulation symbols are used for performance monitoring of at least one model, the at least one model is used to perform a first processing on the input first channel information to obtain N modulation symbols; the N modulation symbols and the M known modulation symbols are used to determine the performance monitoring result of at least one model; the first processing includes encoding and modulation; M and N are both positive integers.

[0079] This application does not limit the use of at least one model. For example, at least one model includes a first model and a second model; the first model is used to encode the first channel information to obtain a first bit sequence; the second model is used to modulate the first bit sequence to obtain N modulation symbols.

[0080] For example, performance monitoring results may include one or more of the following: first MAE; first MSE; first SGCS; second MAE; second MSE; second SGCS; a weighted sum or weighted average of at least two of the first MAE, first MSE, first SGCS, second MAE, second MSE, or second SGCS.

[0081] The first MAE is determined based on the MAE of the first modulation symbol and each of the K second modulation symbols.

[0082] The first MSE is determined based on the first modulation symbol and the MSE of each of the K second modulation symbols.

[0083] The first SGCS is determined based on the SGCS of the first modulation symbol and each of the K second modulation symbols.

[0084] The second MAE is determined based on the MAE of the third modulation symbol and each of the K fourth modulation symbols.

[0085] The second MSE is determined based on the MSE of the third modulation symbol and each of the K fourth modulation symbols.

[0086] The second SGCS is determined based on the SGCS of the third modulation symbol and each of the K fourth modulation symbols.

[0087] Where K is a positive integer less than or equal to M.

[0088] The first modulation symbol belongs to N modulation symbols, and the K second modulation symbols belong to M known modulation symbols; the geometric distance between the second modulation symbol and the first modulation symbol is less than or equal to the geometric distance between any of the M known modulation symbols other than the K second modulation symbols and the first modulation symbol.

[0089] The third modulation symbol belongs to N modulation symbols, and the K fourth modulation symbols belong to M known modulation symbols; the geometric distance between the fourth modulation symbol and the third modulation symbol is less than or equal to the geometric distance between any of the M known modulation symbols other than the K fourth modulation symbols and the third modulation symbol.

[0090] In conjunction with the third aspect, in some implementations of the third aspect, at least one model includes a first model and a second model; the first model is used to encode the first channel information to obtain a first bit sequence; the second model is used to modulate the first bit sequence to obtain N modulation symbols; the N modulation symbols and M known modulation symbols are used to determine the first monitoring result included in the performance monitoring result.

[0091] In conjunction with the third aspect, in some implementations of the third aspect, the method further includes: sending first channel information and a second bit sequence, wherein the second bit sequence is obtained by encoding the first channel information; the first bit sequence and the second bit sequence are used to determine the second monitoring result included in the performance monitoring result.

[0092] For example, the second monitoring result includes one or more of the following: first BCE; third MSE; third MAE; a weighted sum or weighted average of at least two of the first BCE, third MSE and third MAE.

[0093] The first BCE is determined based on the probability of each bit in the first bit sequence being 1 and the second bit sequence.

[0094] The third MSE is the MSE between the first bit sequence and the second bit sequence.

[0095] The third MAE is the MAE between the first bit sequence and the second bit sequence.

[0096] In conjunction with the third aspect, in some implementations of the third aspect, the method also includes: receiving performance monitoring results.

[0097] In conjunction with the third aspect, in some implementations of the third aspect, the method further includes: sending a first message when performance monitoring results indicate that the performance of at least one model does not meet performance requirements, the first message being used to instruct retraining of at least one model.

[0098] The third aspect of this application corresponds to the first aspect of this application. The beneficial effects achieved by the third aspect are similar to those of the first aspect and the corresponding feasible implementation methods, and will not be described again.

[0099] Fourthly, a communication method is provided that can be applied to the second device side.

[0100] For example, the second device side can be replaced by the terminal side or the network side. For more details on the terminal side and the network side, please refer to the first aspect above.

[0101] For example, the method includes: transmitting a dataset, the dataset including first channel information and a first modulation symbol sequence, the first modulation symbol sequence being obtained by performing a first processing on the first channel information; the dataset is used for performance monitoring of at least one model.

[0102] In conjunction with the second aspect, in some implementations of the second aspect, at least one model is used to perform a first processing on the input first channel information to obtain a second modulation symbol sequence; the first modulation symbol sequence and the second modulation symbol sequence are used to determine the performance monitoring results of at least one model; the first processing includes coding and modulation.

[0103] For example, performance monitoring results may include one or more of the following: the fourth MAE; the fourth MSE; the fourth SGCS; a weighted sum or weighted average of at least two of the fourth MAE, fourth MSE, or fourth SGCS.

[0104] The fourth MAE is the MAE between the first modulation symbol sequence and the second modulation symbol sequence.

[0105] The fourth MSE is the MSE between the first modulation symbol sequence and the second modulation symbol sequence.

[0106] The fourth SGCS is the SGCS between the first modulation symbol sequence and the second modulation symbol sequence.

[0107] In conjunction with the fourth aspect, in some implementations of the fourth aspect, at least one model includes a first model and a second model; the first model is used to encode the input first channel information to obtain a first bit sequence; the second model is used to modulate the first bit sequence to obtain a first modulation symbol sequence; the first modulation symbol sequence and the second modulation symbol sequence are used to determine the first monitoring result included in the performance monitoring result.

[0108] In conjunction with the fourth aspect, in some implementations of the fourth aspect, the method further includes: transmitting a second bit sequence, the second bit sequence being obtained by encoding the first channel information, the first bit sequence and the second bit sequence being used to determine the second monitoring result included in the performance monitoring result.

[0109] For example, the second monitoring result includes one or more of the following: first BCE; third MSE; third MAE; a weighted sum or weighted average of at least two of the first BCE, third MSE and third MAE.

[0110] The first BCE is determined based on the probability of each bit in the first bit sequence being 1 and the second bit sequence.

[0111] The third MSE is the MSE between the first bit sequence and the second bit sequence.

[0112] The third MAE is the MAE between the first bit sequence and the second bit sequence.

[0113] In conjunction with the fourth aspect, in some implementations of the fourth aspect, the method also includes: receiving performance monitoring results.

[0114] In conjunction with the fourth aspect, in some implementations of the fourth aspect, the method further includes: sending a first message when performance monitoring results indicate that the performance of at least one model does not meet performance requirements, the first message being used to instruct retraining of at least one model.

[0115] The fourth aspect of this application corresponds to the technical solution of the second aspect of this application. The beneficial effects achieved by the fourth aspect are similar to those of the second aspect and the corresponding feasible implementation methods, and will not be described again.

[0116] Fifthly, an apparatus is provided. This apparatus may include functional modules corresponding to each of the methods / operations / steps / actions described in any possible implementation of the first aspect, or may include functional modules corresponding to each of the methods / operations / steps / actions described in any of the second aspect, or may include functional modules corresponding to each of the methods / operations / steps / actions described in any of the third aspect, or may include functional modules corresponding to each of the methods / operations / steps / actions described in any of the fourth aspect. The module may be a hardware circuit, software, or a combination of hardware circuitry and software implementation.

[0117] In one design, the device may include a processing module and a communication module. The communication module is used to perform the sending and receiving actions performed by the first device side in the method described in the first or second aspect above, while the processing module is used to perform processing-related actions performed by the first device side in the method described in the first or second aspect above.

[0118] In one design, the device can be a terminal device, or a device, module, circuit, or chip configured in the terminal device, or a device that can be used in conjunction with the terminal device, such as an OTT host or cloud server.

[0119] In one design, the device may include a processing module and a communication module. The communication module is used to perform the sending and receiving actions performed by the second device side in the method described in the third or fourth aspect above, while the processing module is used to perform processing-related actions performed by the second device side in the method described in the third or fourth aspect above.

[0120] In one design, the device can be a network device, or a device, module, circuit, or chip configured in the network device, or a device that can be used in conjunction with the network device, such as an intelligent network element with a radio access network (RAN) intelligent controller (RIC) deployed thereon.

[0121] A sixth aspect provides an apparatus comprising a processor and a storage medium storing instructions which, when executed by the processor, cause a method as described in the first aspect or any possible implementation thereof to be implemented, or cause a method as described in the second aspect or any possible implementation thereof to be implemented, or cause a method as described in the third aspect or any possible implementation thereof to be implemented, or cause a method as described in the fourth aspect or any possible implementation thereof to be implemented.

[0122] A seventh aspect provides an apparatus comprising a processing circuit for processing data and / or information such that a method as in the first aspect or any possible implementation thereof is implemented, or a method as in the second aspect or any possible implementation thereof is implemented, or a method as in the third aspect or any possible implementation thereof is implemented, or a method as in the fourth aspect or any possible implementation thereof is implemented.

[0123] The processing circuit may include one or more processors, or all or part of the circuitry in one or more processors used for control or processing functions.

[0124] Optionally, the apparatus may further include a memory for storing programs or instructions, and the processor for executing the programs or instructions to implement the methods as in the first aspect or any possible implementation thereof, or to implement the methods as in the second aspect or any possible implementation thereof, or to implement the methods as in the third aspect or any possible implementation thereof, or to implement the methods as in the fourth aspect or any possible implementation thereof.

[0125] Optionally, the device may also include the transceiver circuit, or an input / output interface.

[0126] Eighthly, a chip is provided, including processing circuitry for running a program or instructions to cause the method as described in the first aspect or any possible implementation thereof to be implemented, or to cause the method as described in the second aspect or any possible implementation thereof to be implemented, or to cause the method as described in the third aspect or any possible implementation thereof to be implemented, or to cause the method as described in the fourth aspect or any possible implementation thereof to be implemented.

[0127] Optionally, the chip may further include a memory for storing programs or instructions.

[0128] Optionally, the chip may also include transceiver circuitry, or input / output interfaces.

[0129] A ninth aspect provides a computer-readable storage medium comprising instructions that, when executed by a processor, cause the method of the first aspect or any possible implementation thereof to be implemented, or cause the method of the second aspect or any possible implementation thereof to be implemented, or cause the method of the third aspect or any possible implementation thereof to be implemented, or cause the method of the fourth aspect or any possible implementation thereof to be implemented.

[0130] In a tenth aspect, a computer program product is provided, the computer program product comprising computer program code or instructions, which, when executed, cause the method of the first aspect and any possible implementation thereof to be implemented, or cause the method of the second aspect and any possible implementation thereof to be implemented, or cause the method of the third aspect and any possible implementation thereof to be implemented, or cause the method of the fourth aspect and any possible implementation thereof to be implemented.

[0131] Eleventhly, a communication system is provided, the communication system comprising one or more of the following: means for performing the first aspect and any possible implementation thereof, means for performing the second aspect and any possible implementation thereof, means for performing the third aspect and any possible implementation thereof, and means for performing the fourth aspect and any possible implementation thereof. Attached Figure Description

[0132] Figure 1 This is a schematic diagram of a communication system applicable to the communication method in the embodiments of this application;

[0133] Figure 2 This is a schematic diagram of another communication system applicable to the communication method of the embodiments of this application;

[0134] Figure 3 This is a schematic diagram of a possible application framework in a communication system;

[0135] Figure 4 This is a schematic diagram of another possible application framework in a communication system;

[0136] Figure 5 This is a schematic diagram of CSI feedback using an auto-encoder (AE) model provided in an embodiment of this application;

[0137] Figure 6 This is another schematic diagram of CSI feedback using the AE model provided in this application;

[0138] Figure 7 This is a schematic diagram of CSI feedback using joint source-channel coding (JSCC) provided in this application;

[0139] Figure 8 This is an example diagram of a neuron structure;

[0140] Figure 9 An example diagram of a deep neural network (DNN) is shown.

[0141] Figure 10 This is a schematic diagram illustrating the data set connection between the network side and the terminal side;

[0142] Figure 11 This is a schematic diagram illustrating the model interoperability between the network side and the terminal side;

[0143] Figure 12 This is a schematic diagram of the communication method provided in an embodiment of this application;

[0144] Figure 13 This is another schematic diagram of the communication method provided in the embodiments of this application;

[0145] Figure 14 This is another schematic diagram of the communication method provided in the embodiments of this application;

[0146] Figure 15 This is a schematic block diagram of the communication device provided in the embodiments of this application;

[0147] Figure 16 This is a schematic block diagram of another communication device provided in the embodiments of this application;

[0148] Figure 17 This is a schematic block diagram of the AI ​​processor provided in the embodiments of this application. Detailed Implementation

[0149] The technical solution provided in this application will now be described with reference to the accompanying drawings.

[0150] To facilitate understanding of the embodiments of this application, the following points will be explained first:

[0151] First, in this application, the terminal side can also be referred to as the user equipment (UE) side, terminal-side equipment, etc., including: terminal equipment (or user equipment, terminal, etc.), components deployed in the terminal equipment (such as circuits or chips inside the terminal equipment), equipment deployed outside the terminal equipment (such as the host or cloud server of an OTT system, hereinafter referred to as the OTT system server), or components deployed in equipment outside the terminal equipment (such as circuits or chips inside the equipment). The network side (NW side) can also be referred to as network-side equipment, including: network equipment communicating with the terminal equipment, components deployed in the network equipment (such as circuits or chips inside the network equipment with near real-time radio access network (RAN) intelligent control functions), equipment deployed outside the network equipment (such as intelligent network elements, for example, intelligent network elements with near real-time RAN intelligent control functions), or components deployed in the intelligent network element (such as circuits or chips inside the intelligent network element). Among them, network equipment can include: access network equipment, core network equipment, or operation administration and maintenance (OAM).

[0152] Second, in this application, the indication includes direct indication (also known as explicit indication) and indirect indication (also known as implicit indication). Directly indicating information A means including information A; indirectly indicating information A can mean indicating information A through the correspondence between information A and information B and by directly indicating information B; or by indicating information A through a preset rule that can be used to determine A based on B and by directly indicating information B. The correspondence between information A and information B, and the preset rule, can be predefined, pre-stored, pre-burned, or pre-configured.

[0153] Third, in this application, "at least one" means one or more, and "more than one" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A alone, A and B simultaneously, or B alone, where A and B can be singular or plural. The character " / " generally indicates an "or" relationship between the preceding and following related objects, but it does not exclude the possibility of indicating an "and" relationship; the specific meaning can be understood in context. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can mean: a, b, c; a and b; a and c; b and c; or a and b and c. Here, a, b, and c can be single or multiple.

[0154] Fourth, the use of prefixes such as "first" and "second" in this application is merely for the purpose of distinguishing and describing different things belonging to the same name category, and does not constrain the order, size, or quantity of things. For example, "first model" and "second model" are simply different models, and do not limit the number of models, their deployment location, size relationship, or priority relationship.

[0155] Fifth, in this application, "send" and "receive" indicate the direction of signal transmission. For example, "send information to a terminal device" can be understood as the destination of the information being the terminal device, which may include direct transmission via the air interface or indirect transmission by other units or modules via the air interface. "Receive information from a network device" can be understood as the source of the information being the network device, which may include direct reception from the network device via the air interface or indirect reception from the network device by other units or modules via the air interface. "Send" can also be understood as the "output" of the chip interface, and "receive" can also be understood as the "input" of the chip interface. In other words, sending and receiving can occur between devices, such as between a terminal device and a computing node, or within a device, such as between components, modules, chips, software modules, or hardware modules within the device via a bus, wiring, or interface.

[0156] Sixth, in the embodiments of this application, "when," "if," and "if" all refer to the device making corresponding processing under certain objective circumstances, and are not limited to a time, nor do they require the device to make a judgment action when it is implemented, nor do they mean that there are other limitations.

[0157] Seventh, in this application, the words "example," "exemplarily," "for example," or "such as" are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "example," "exemplarily," "for example," or "such as" in this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the words "example," "exemplarily," "for example," or "such as" is intended to present the relevant concepts in a specific manner.

[0158] Eighth, for ease of distinction and explanation, this paper introduces the terms encoder and decoder. These names are given only to distinguish different functions and do not limit the structure of the device. For example, if the terminal side (such as a terminal device or OTT system server) has the inference function of the encoder, it can be said that the terminal side contains the encoder, which can be understood as a functional module of the terminal side; if the network side has the inference function of the decoder, it can be said that the network side contains the decoder, which can be understood as a functional module of the network side. The encoder has the quantization function, so it can be said that the encoder contains the quantizer. The decoder has the dequantization function, so it can be said that the decoder contains the dequantizer. In specific implementations, the encoder and decoder can be implemented through a neural network model. More specifically, the functions of the encoder and decoder, quantizer and dequantizer can be implemented separately in hardware, or in software, or in a combination of hardware and software; this application does not limit this.

[0159] The technical solutions provided in this application can be applied to various communication systems, such as 5th generation (5G) or new radio (NR) systems, frequency division duplex (FDD) systems, time division duplex (TDD) systems, wireless local area network (WLAN) systems, satellite communication systems, future communication systems, or integrated systems of multiple systems. The technical solutions provided in this application can also be applied to device-to-device (D2D) communication, vehicle-to-everything (V2X) communication, machine-to-machine (M2M) communication, machine-type communication (MTC), and Internet of Things (IoT) communication systems or other communication systems.

[0160] In a communication system, a device can send signals to or receive signals from another device. These signals can include information, signaling, or data. The device can also be replaced by an entity, network entity, network element, communication equipment, communication module, node, communication node, etc. This disclosure uses a device as an example. For instance, a communication system can include at least one terminal device and at least one network device. The network device can send downlink signals to the terminal device, and / or the terminal device can send uplink signals to the network device. It is understood that the terminal device in this disclosure can be replaced by a first device, and the network device can be replaced by a second device, both performing the corresponding methods described in this disclosure.

[0161] Figure 1 This is a schematic diagram of a communication system applicable to the communication method in the embodiments of this application. For example... Figure 1 As shown, the communication system 100A may include at least one access network device, such as Figure 1 The access network device 110 shown; the communication system 100A may also include at least one terminal device, such as Figure 1 Terminal devices 120 and 130 are shown. Access network device 110 can communicate with terminal devices (such as terminal devices 120 and 130) via a wireless link. Communication devices in this communication system, for example, between access network device 110 and terminal device 120, can communicate via multi-antenna technology.

[0162] In wireless communication networks, such as mobile communication networks, the services supported by the networks are becoming increasingly diverse, thus requiring increasingly diverse demands. For example, networks need to support ultra-high speeds, ultra-low latency, and / or massive connectivity. This characteristic makes network planning, network configuration, and / or resource scheduling increasingly complex. Furthermore, as network functions become more powerful, such as supporting higher spectrum levels, supporting higher-order multiple-input multiple-output (MIMO) technologies, supporting beamforming (BF), and supporting beam management, network energy efficiency has become a hot research topic. These new demands, new scenarios, and new characteristics bring unprecedented challenges to network planning, operation, and efficient operation. To meet these challenges, artificial intelligence (AI) technology can be introduced into wireless communication networks to achieve network intelligence. To support AI technology in wireless networks, AI nodes may also be introduced. AI nodes can be AI network elements or AI modules. Figure 2 This is a schematic diagram of another communication system applicable to the communication method in the embodiments of this application. Compared to Figure 1 Regarding the communication system 100A shown, Figure 2 The communication system 100B shown also includes an AI network element 140. The AI ​​network element 140 is used to perform AI-related operations, such as building datasets or AI models. The AI ​​network element can also be simply referred to as an intelligent network element. In this disclosure, the AI ​​model can be simply referred to as a model.

[0163] In one possible implementation, access network device 110 can send data related to the training of the AI ​​model to AI network element 140, whereby AI network element 140 constructs a dataset and trains the AI ​​model. For example, the data related to the training of the AI ​​model may include data reported by terminal devices. AI network element 140 can send the results of operations related to the AI ​​model to access network device 110, and then forward them to terminal devices via access network device 110. For example, the results of operations related to the AI ​​model may include at least one of the following: a trained AI model, model evaluation results, or test results, etc. Exemplarily, a portion of the trained AI model may be deployed on access network device 110, and another portion on terminal devices 120 and / or 130. Alternatively, the trained AI model may be deployed on access network device 110. Or, the trained AI model may be deployed on terminal devices 120 and / or 130.

[0164] It should be understood that Figure 2This explanation only uses the direct connection between AI network element 140 and access network device 110 as an example. In other scenarios, AI network element 140 can also be connected to terminal devices. Alternatively, AI network element 140 can be connected to both access network device 110 and terminal devices simultaneously. Alternatively, AI network element 140 can also be connected to access network device 110 through a third-party network element. This application embodiment does not limit the connection relationship between AI network element and other network elements. For example, AI network element 140 can also be configured as a module in access network device and / or terminal device, for example, configured in... Figure 1 In the access network device 110 or terminal device shown.

[0165] It should be noted that, Figure 1 and Figure 2 This is a simplified illustration for ease of understanding only. For example, the communication system may also include other devices, such as wireless relay devices and / or wireless backhaul devices. Figure 1 and Figure 2 The figures are not shown. In practical applications, this communication system may include multiple access network devices or multiple terminal devices. This application does not limit the number of access network devices and terminal devices included in the communication system.

[0166] In the embodiments of this application, the terminal device may also be referred to as UE, access terminal, user unit, user station, mobile station, mobile station, remote station, remote terminal, mobile device, user terminal, terminal, wireless communication device, user agent, or user equipment.

[0167] Terminal devices can be devices that provide voice / data, such as handheld devices with wireless connectivity, in-vehicle devices, etc. Currently, examples of terminals include: mobile phones, tablets, laptops, PDAs, mobile internet devices (MIDs), wearable devices, virtual reality (VR) devices, augmented reality (AR) devices, wireless terminals in industrial control, wireless terminals in self-driving, wireless terminals in remote medical surgery, wireless terminals in smart grids, wireless terminals in transportation safety, wireless terminals in smart cities, wireless terminals in smart homes, cellular phones, cordless phones, session initiation protocol (SIP) phones, wireless local loop (WLL) stations, personal digital assistants (PDAs), handheld devices with wireless communication capabilities, computing devices or other processing devices connected to a wireless modem, wearable devices, terminal devices in 5G networks, or future public land mobile communication networks. Terminal devices in a network (PLMN), etc., are not limited to this in the embodiments of this application.

[0168] By way of example and not limitation, in this embodiment, the terminal device can also be a wearable device. Wearable devices, also known as wearable smart devices, are a general term for devices that utilize wearable technology to intelligently design and develop everyday wearables, such as glasses, gloves, watches, clothing, and shoes. Wearable devices are portable devices that are worn directly on the body or integrated into the user's clothing or accessories. Wearable devices are not merely hardware devices, but also achieve powerful functions through software support, data interaction, and cloud interaction. Broadly speaking, wearable smart devices include those that are feature-rich, large in size, and can achieve complete or partial functions without relying on a smartphone, such as smartwatches or smart glasses, as well as those that focus on a specific type of application function and require the use of other devices such as smartphones, such as various smart bracelets and smart jewelry for vital sign monitoring.

[0169] In this embodiment, the device for implementing the functions of the terminal device can be the terminal device itself, or it can be any device capable of supporting the terminal device in implementing those functions, such as a chip system. This device can be installed in or used in conjunction with the terminal device. In this embodiment, the chip system can be composed of chips or may include chips and other discrete components. This embodiment only uses the terminal device as an example to illustrate the device for implementing the functions of the terminal device, and does not constitute a limitation on the solution of this embodiment.

[0170] The network device in this application embodiment can be a device for communicating with a terminal device. This network device may include access network equipment, core network equipment, or other equipment in the communication system. The access network equipment may be, for example, a base station. In this application embodiment, the access network equipment may refer to a RAN node (or device) that connects the terminal device to the wireless network. A base station can broadly encompass, or be replaced by, various names such as: NodeB, evolved NodeB (eNB), next-generation NodeB (gNB), relay station, access point, transmit / receive point (TRP), transmitting point (TP), master station, auxiliary station, motor slide retainer (MSR) node, home base station, network controller, access node, wireless node, access point (AP), transmission node, transceiver node, baseband unit (BBU), remote radio unit (RRU), active antenna unit (AAU), remote radio head (RRH), central unit (CU), distributed unit (DU), radio unit (RU), positioning node, etc. A base station can be a macro base station, micro base station, relay node, donor node, or similar entities, or combinations thereof. A base station can also refer to a communication module, modem, or chip installed within the aforementioned equipment or apparatus. A base station can also be a mobile switching center, a device that performs base station functions in D2D, V2X, and M2M communications, or a device that performs base station functions in future communication systems. A base station can support networks using the same or different access technologies. Optionally, a RAN node can also be a server, wearable device, vehicle, or in-vehicle equipment. For example, the access network equipment in V2X technology can be a roadside unit (RSU). The embodiments of this application do not limit the specific technology or equipment form used in the access network equipment.

[0171] Base stations can be fixed or mobile. For example, a helicopter or drone can be configured to act as a mobile base station, and one or more cells can move depending on the location of the mobile base station. In other examples, a helicopter or drone can be configured as a device to communicate with another base station.

[0172] In some deployments, the access network equipment mentioned in the embodiments of this application may be a device including a CU, or a DU, or a device including both CU and DU, or a device with a control plane CU node (central unit-control plane (CU-CP)) and a user plane CU node (central unit-user plane (CU-UP)) and a DU node. For example, the access network equipment may include gNB-CU-CP, gNB-CU-UP, and gNB-DU.

[0173] In some deployments, multiple RAN nodes collaborate to assist terminals in achieving wireless access, with different RAN nodes each implementing some of the base station's functions. For example, RAN nodes can be CUs, DUs, CU-CPs, CU-UPs, or RUs. CUs and DUs can be configured separately or included in the same network element, such as a BBU. RUs can be included in radio frequency equipment or radio frequency units, such as RRUs, AAUs, or RRHs.

[0174] RAN nodes can support one or more types of fronthaul interfaces, each corresponding to a DU and RU with different functions. If the fronthaul interface between the DU and RU is a Common Public Radio Interface (CPRI), the DU is configured to implement one or more baseband functions, and the RU is configured to implement one or more radio frequency functions. If the fronthaul interface between the DU and RU is another interface, relative to CPRI, it moves some downlink and / or uplink baseband functions—for example, for downlink, precoding, digital beamforming, or one or more of inverse fast Fourier transform (IFFT) / adding a cyclic prefix (CP)—from the DU to the RU; and for uplink, digital beamforming, or one or more of fast Fourier transform (FFT) / removing CP—from the DU to the RU. In one possible implementation, this interface can be an enhanced common public radio interface (eCPRI). Under the eCPRI architecture, the splitting methods between DU and RU are different, corresponding to different types (category, Cat) of eCPRI, such as eCPRI Cat A, B, C, D, E, F.

[0175] Taking eCPRI Cat A as an example, for downlink transmission, layer mapping is used as the dividing line. The DU is configured to implement one or more functions preceding layer mapping (i.e., coding, rate matching, scrambling, modulation, and layer mapping itself), while other functions following layer mapping (e.g., resource element (RE) mapping, digital beamforming, or one or more of IFFT / CP addition) are implemented in the RU. For uplink transmission, de-RE mapping is used as the dividing line. The DU is configured to implement one or more functions preceding de-mapping (i.e., decoding, rate matching de-matching, descrambling, demodulation, inverse discrete Fourier transform (IDFT), channel equalization, and de-RE mapping itself), while other functions following de-mapping (e.g., digital BF or FFT / CP removal) are implemented in the RU. It is understood that descriptions of the functions of the DU and RU corresponding to various types of eCPRI can be found in the eCPRI protocol and will not be elaborated upon here.

[0176] In one possible design, the processing unit in the BBU used to implement baseband functions is called the baseband high (BBH) unit, and the processing unit in the RRU / AAU / RRH used to implement baseband functions is called the baseband low (BBL) unit.

[0177] In different systems, CU (or CU-CP and CU-UP), DU, or RU may have different names, but those skilled in the art will understand their meaning. For example, in an open RAN (ORAN) architecture, CU can also be called open CU (open-CU, O-CU), DU can also be called open DU (open-DU, O-DU), CU-CP can also be called open CU-CP (open-CU-CP) O-CU-CP, CU-UP can also be called open CU-UP (open-CU-UP, O-CU-UP), and RU can also be called open RU (open-RU, O-RU). Any of the units among CU (or CU-CP, CU-UP), DU, and RU in this application can be implemented through software modules, hardware modules, or a combination of software modules and hardware modules.

[0178] In this embodiment, the apparatus for implementing the functions of a network device can be a network device itself; it can also be an apparatus capable of supporting the network device in implementing those functions, such as a chip system, hardware circuit, software module, or a hardware circuit plus a software module. This apparatus can be installed in the network device or used in conjunction with the network device. In this embodiment, the example of a network device being used to implement the functions of a network device is provided only and does not constitute a limitation on the solutions described in this embodiment.

[0179] Network devices and / or terminal devices can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; they can also be deployed on water; and they can also be deployed in the air on airplanes, balloons, and satellites. This application does not limit the scenario in which the network devices and terminal devices are located. Furthermore, terminal devices and network devices can be hardware devices, or software functions running on dedicated hardware or general-purpose hardware, such as virtualization functions instantiated on a platform (e.g., a cloud platform), or entities that include dedicated or general-purpose hardware devices and software functions. This application does not limit the specific form of the terminal devices and network devices.

[0180] Optionally, the AI ​​node can be deployed in one or more of the following locations within the communication system: access network equipment, terminal equipment, or core network elements. Alternatively, the AI ​​node can also be deployed independently, for example, in a location other than any of the aforementioned devices, such as a server, or in a host or cloud server of an over-the-top (OTT) system. The AI ​​node can communicate with other devices in the communication system, which can be, for example, one or more of the following: access network equipment, terminal equipment, or core network elements.

[0181] It is understood that this application does not limit the number of AI nodes. For example, when there are multiple AI nodes, these nodes can be divided based on function, such as different AI nodes being responsible for different functions.

[0182] It can also be understood that AI nodes can be independent devices, or they can be integrated into the same device to achieve different functions. Alternatively, they can be network elements in hardware devices, software functions running on dedicated hardware, or virtualization functions instantiated on a platform (e.g., a cloud platform). This application does not limit the specific form of the aforementioned AI nodes.

[0183] Figure 3 This is a schematic diagram of a possible application framework in a communication system. For example... Figure 3As shown, network elements in a communication system are connected via interfaces (such as next-generation (NG) interfaces or Xn interfaces) or air interfaces. The NG interface is the interface between the radio access network and the 5G core network. The Xn interface is the interface between access network devices, and the air interface is the interface between access network devices and terminal devices. These network element nodes, such as core network devices, RAN nodes, terminal devices, or one or more devices in the OAM, are equipped with one or more AI modules (for clarity, ...). Figure 3 (Only one is shown in the image). The access network node can be a single RAN node or can include multiple RAN nodes, such as RU, CU, and DU. One or more of the RU, CU, or DU can also be equipped with one or more AI modules. Optionally, the CU can also be split into CU-CP and CU-UP. One or more AI modules are provided in CU-CP and / or CU-UP.

[0184] The AI ​​module is used to implement corresponding AI functions. AI modules deployed in different network elements can be the same or different. Depending on the parameter configuration, the AI ​​module can implement different functions. The AI ​​module model can be configured based on one or more of the following parameters: structural parameters (e.g., at least one of the following: number of neural network layers, neural network width, inter-layer connections, neuron weights, neuron activation function, or bias in the activation function), input parameters (e.g., type and / or dimension of input parameters), or output parameters (e.g., type and / or dimension of output parameters). The bias in the activation function can also be referred to as the neural network bias.

[0185] An AI module can have one or more models. A model can infer an output, which includes one or more parameters. The learning, training, or inference processes of different models can be deployed on different nodes or devices, or they can be deployed on the same node or device.

[0186] For example, an AI model can be based on one or more AI processors, or on an AI cluster.

[0187] For example, an AI module may include all or part of one or more AI processors, or all or part of an AI cluster.

[0188] The network device can be a network device equipped with one or more AI modules. The network device may include... Figure 3 This refers to one or more devices within the core network, RAN, or OAM. For example, the AI ​​module could be... Figure 4The RAN intelligent controller (RIC) shown can be a near-real-time (near-RT) RIC or a non-real-time (non-RT) RIC. For example, a near-real-time RIC is located in a RAN node (e.g., in a CU or DU), while a non-real-time RIC is located in an OAM, a cloud server, a core network device, or other access network devices. The RIC can obtain subsets from multiple terminal devices from RAN nodes (e.g., CU, CU-CP, CU-UP, DU, and / or RU), reassemble them into a dataset, and train based on the dataset. Exemplarily, near-real-time and non-real-time RICs can also be set up as separate network elements, and access network devices can be either near-real-time or non-real-time RICs.

[0189] Figure 4 This is a schematic diagram of another possible application framework in a communication system. Figure 4 The communication system shown includes access network nodes (CU, DU, and RU shown in the figure) and terminals, as well as a RIC. For example, the RIC could be... Figure 3 The AI ​​module shown can be used to implement AI-related functions. The RIC includes near real-time RIC and non-real-time RIC. Non-real-time RIC primarily processes non-real-time information, such as data that is not sensitive to latency, with latency on the order of seconds. Real-time RIC primarily processes near real-time information, such as data that is relatively sensitive to latency, with latency on the order of tens of milliseconds.

[0190] The near real-time RIC is used for model training and inference. For example, it can be used to train an AI model and then use that AI model for inference. The near real-time RIC can obtain network-side and / or terminal-side information from RAN nodes (e.g., CU, CU-CP, CU-UP, DU, and / or RU) and / or terminals. This information can be used as training data or inference data. Optionally, the near real-time RIC can deliver inference results to RAN nodes and / or terminals. Optionally, inference results can be exchanged between CU and DU, and / or between DU and RU. For example, the near real-time RIC delivers the inference result to the DU, and the DU sends it to the RU.

[0191] The non-real-time RIC is also used for model training and inference. For example, it can be used to train an AI model and then use that model for inference. The non-real-time RIC can obtain network-side and / or terminal-side information from RAN nodes (e.g., one or more of CU, CU-CP, CU-UP, DU, or RU) and / or terminals. This information can be used as training data or inference data, and the inference results can be delivered to the RAN nodes and / or terminals. Optionally, inference results can be exchanged between CU and DU, and / or between DU and RU; for example, the non-real-time RIC delivers the inference results to the DU, which then forwards them to the RU.

[0192] The near real-time RIC and non-real-time RIC can also be set up as separate network elements. Optionally, the near real-time RIC and non-real-time RIC can also be part of other devices. For example, the near real-time RIC can be set in the RAN node (e.g., in CU, DU), while the non-real-time RIC can be set in the OAM, cloud server, core network device, or other network device.

[0193] With the development of wireless communication technology and the increasing number of supported services, higher demands are being placed on communication systems in terms of system capacity, communication latency, and other indicators. Among these advancements, massive MIMO (Multi-User MIMO) systems, by configuring large-scale antenna arrays at the transceiver end, can achieve spatial diversity gain, thereby significantly increasing system capacity. For example, access network equipment can simultaneously transmit data to multiple terminal devices using the same time-frequency resources, i.e., multi-user MIMO (MU-MIMO); or, access network equipment can simultaneously transmit multiple data streams to the same terminal device, i.e., single-user MIMO (SU-MIMO). The data between these multiple terminal devices or the multiple data streams within the same terminal device are spatially multiplexed, thus becoming a key direction in the evolution of communication systems.

[0194] Access network equipment can obtain channel information of the downlink channel to determine one or more of the following configurations for scheduling terminal equipment: downlink data channel resources, modulation and coding scheme (MCS), precoding, etc.

[0195] Taking precoding as an example, in massive MIMO, access network equipment uses a precoding matrix to precode downlink data. By using precoding technology, access network equipment can achieve spatial multiplexing between terminal devices or data streams. That is, it spatially isolates data between different terminal devices or between different data streams of the same terminal device, thereby reducing interference between different terminal devices or different data streams and improving the received signal-to-interference-plus-noise ratio (SINR) of the terminal devices. To calculate the precoding matrix, the access network equipment obtains the channel information of the downlink channel and determines the precoding matrix based on the channel information.

[0196] In TDD systems, due to the reciprocity of uplink and downlink channels, access network devices can obtain uplink channel information by measuring uplink reference signals, and then infer relatively accurate downlink channel information, for example, using uplink channel information as downlink channel information. However, in FDD systems, uplink and downlink reciprocity cannot be guaranteed. Downlink channel information is obtained by terminal devices measuring downlink reference signals, such as channel state information reference signals (CSI-RS) or synchronizing signal blocks (SSBs). Therefore, terminal devices need to generate CSI reports according to predefined protocols or configured access network devices, and feed the generated CSI reports back to the access network devices so that they can obtain downlink channel information.

[0197] In FDD systems, a crucial part of CSI report feedback is the precoding matrix indicator (PMI), which uses 0-1 bits in the CSI to quantize the channel matrix or precoding matrix. PMI design (also known as codebook design) is a fundamental issue in mobile communication systems. Traditional codebook design methods predefine (or agree upon) a series of precoding matrices and their corresponding numbers in the protocol; these precoding matrices are called codewords. The channel matrix or precoding matrix can be approximated using predefined codewords or linear combinations of multiple predefined codewords. Therefore, terminal equipment can use the PMI to feed back the corresponding codeword numbers and one or more weighting coefficients to the access network equipment, which is then used by the access network equipment to reconstruct the channel matrix or precoding matrix.

[0198] As the antenna array size of MIMO systems continues to increase, the number of supported antenna ports also increases, leading to a growth in the dimensionality of the corresponding channel matrix and precoding matrix. To enable terminal devices to estimate (or measure) the downlink channel, the overhead of the access network equipment transmitting reference signals increases. Simultaneously, the error of approximating large-scale channel matrices and precoding matrices with a finite number of predefined codewords increases. One method to improve channel reconstruction accuracy is to increase the number of codewords in the codebook, but this simultaneously increases the overhead of CSI feedback (including the corresponding codeword number and one or more weighting coefficients), thereby reducing the available resources for data transmission and causing system capacity loss. Therefore, it is necessary to investigate how to more effectively compress and represent channel information without increasing the overhead of transmitting reference signals and CSI feedback, and how to more effectively reconstruct the channel based on feedback information. Correlation exists between different elements in the downlink channel matrix between access network equipment and terminal equipment; furthermore, correlation exists between downlink channel matrices in different time slots. For example, the correlation between different elements in the channel matrix implies the existence of a basis (which can be represented by matrices U1 and U2). Projecting the channel matrix H onto this basis yields a sparse equivalent channel, i.e., H' = U1H. H U2 is a sparse matrix, where the superscript H denotes the conjugate transpose operation. Theoretically, the channel matrix H can be reconstructed simply by estimating the non-zero elements in H' using the reference signal and feeding them back. Therefore, the overhead of transmitting the reference signal and CSI feedback has room for compression. However, traditional CSI feedback schemes, such as the codebook-based feedback method mentioned above, do not fully utilize the channel compression space, and the channel compression process may cause significant information loss. Machine learning methods (such as deep learning (DL)) have stronger nonlinear feature extraction capabilities, thus enabling more effective extraction of correlations between channel matrices. Consequently, compared to traditional schemes, they can more effectively compress and represent channel information, and more effectively reconstruct channel information based on the feedback information.

[0199] To facilitate understanding of the embodiments of this application, the terms involved in this application will be briefly explained below.

[0200] 1. Channel Information: Channel information, also known as channel state information (CSI) or channel environment information, is a type of information that reflects channel characteristics and channel quality.

[0201] Channel information measurement refers to the process by which the receiver deciphers the channel information based on a reference signal transmitted by the transmitter; that is, it involves estimating the channel information using channel estimation methods. For example, the reference signal may include one or more of the following: channel state information reference signal (CSI-RS), synchronizing signal / physical broadcast channel block (SSB), sounding reference signal (SRS), or demodulation reference signal (DMRS). One or more of CSI-RS, SSB, and DMRS can be used to measure downlink channel information. SRS and / or DMRS can be used to measure uplink channel information.

[0202] Channel information can be determined based on channel measurements of a reference signal. Alternatively, channel information can be derived from channel measurements of a reference signal.

[0203] In this embodiment of the application, the channel information can be one or more of the following: channel response information (such as channel response matrix, frequency domain channel response information, time domain channel response information), weight information corresponding to the channel response, precoding matrix information corresponding to the channel response, channel matrix, channel feature matrix, or precoding matrix.

[0204] Optionally, if the terminal device also performs joint source channel coding on other forms of channel information before reporting it to the network side, then these forms of channel information are also included within the scope of this application's embodiments. For example, other channel information may include one or more of the following: precoding matrix indicator (PMI), CSI-RS resource indicator (CRI), reference signal receiving power (RSRP), signal-to-interference plus noise ratio (SINR), identifier of the best beam (ID), identifier of the optimal beam (ID), or ID of the top K beams. The signal-to-interference plus noise ratio can also be called the signal-to-interference-plus-noise ratio. A best beam may refer to a beam that achieves received or transmitted energy greater than or equal to a threshold. An optimal beam may refer to a beam that maximizes received or transmitted energy. For example, if the receiver uses different receiving beams to receive signals, the optimal beam may include the beam with the largest RSRP among multiple different receiving beams. For example, if the transmitter uses different transmit beams to send signals, the optimal beam can include the beam with the highest RSRP (Receiving Power Ratio) of the signal when it arrives at the receiver. Similarly, the optimal top K beams refer to the top K beams that maximize the received or transmitted energy.

[0205] 2. AI Model: A function model that maps an input of a certain dimension to an output of a certain dimension. Its parameters can be obtained through machine learning (ML). For example, f(x) = ax 2 +b is a quadratic function model, which can be viewed as an AI model. a and b correspond to the parameters of the model and can be obtained through machine learning training.

[0206] 3. Autoencoder (AE) model: This generally refers to a network structure consisting of two AI models, such as an encoder and a decoder. Each model can be an independent AI model. AE models are also called bilateral models, two-end models, collaborative models, etc. The encoder and decoder of an AE are usually trained together and can be used in a coordinated manner.

[0207] In this application, CSI feedback can be implemented based on AE's AI model. Figure 5This is a schematic diagram of CSI feedback using an AE model provided in an embodiment of this application.

[0208] like Figure 5 As shown, the terminal compresses the channel information using an encoder, and the network reconstructs the compressed channel information using a decoder. For example, the terminal can use the channel information as input to the encoder, which can compress the channel information to obtain compressed channel information. The terminal can then quantize the compressed channel information to obtain quantized channel information. The terminal can then send this quantized channel information to the network as CSI feedback information, for example, through a CSI report. Here, quantization can be replaced by normalization. Compression can also be referred to as source coding.

[0209] The network side can first dequantize the CSI feedback information to obtain compressed channel information with quantization loss. The network side can then use this compressed channel information as input to the decoder, which can decompress it to obtain the reconstructed channel information. Dequantization can be replaced by normalized recovery. Decompression can also be called source decoding.

[0210] The quantizer used to perform quantization can be predefined, such as protocol predefined, or it can be indicated by the network side; this application does not limit this.

[0211] In another implementation, the encoder on the terminal side can also compress and quantize the channel information, outputting CSI feedback information. The decoder on the network side can also dequantize and decompress the CSI feedback information to obtain the reconstructed channel information. In this case, Figure 5 The encoder output is CSI feedback information, and the decoder input is also CSI feedback information. Here, quantization can be replaced by normalization, and correspondingly, dequantization can be replaced by normalized recovery.

[0212] Figure 6This is another schematic diagram of CSI feedback using the AE model provided in this application. As shown in the figure, the model on the terminal side is an encoder used to compress CSI-RS measurement data (such as channel information). The terminal side sends the compressed CSI-RS measurement data (such as CSI feedback information) to the network side. After receiving the compressed CSI-RS measurement data, the network side uses its decoder to recover the compressed CSI-RS measurement data, obtaining the reconstructed CSI-RS measurement data. In this way, the terminal side and the network side can achieve large-port, full-band channel measurement with limited CSI-RS measurement overhead based on the dual-end model. During CSI-RS measurement, multi-port channel measurement can be achieved by indicating multiple resource sets. Currently, one resource set can support channel measurement for up to 32 antenna ports. By adding multiple resource sets, channel measurement with larger arrays can be supported. In related technologies, to support high-bandwidth channel measurements, the number of resource blocks (RBs) that the terminal needs to measure can be indicated. Density indicates the number of CSI-RS resources measured within one RB, with values ​​of 0.5, 1, and 3. Specifically, a density of 0.5 indicates that one CSI-RS resource is measured within every two RBs, a density of 1 indicates that one CSI-RS resource is measured within each RB, and a density of 3 indicates that three CSI-RS resources are measured within each RB.

[0213] like Figure 6 In the CSI-RS measurement scenario shown, the terminal can measure only a portion of the CSI-RS bandwidth and compress the corresponding CSI-RS measurement data. Correspondingly, the network side recovers the compressed CSI-RS measurement data for that portion of the bandwidth to obtain the full-bandwidth CSI-RS measurement data. Alternatively, the terminal can measure only a portion of the CSI-RS bandwidth and compress the corresponding CSI-RS measurement data; however, the network side not only recovers the compressed CSI-RS measurement data for that portion of the bandwidth but can also predict the CSI-RS measurement data for the unmeasured portion of the bandwidth to obtain the full-bandwidth CSI-RS measurement data. Alternatively, the terminal can measure only a portion of the CSI-RS bandwidth and compress the corresponding CSI-RS measurement data; after the terminal sends the compressed CSI-RS measurement data for that portion of the bandwidth multiple times, the network side predicts the full-bandwidth CSI-RS measurement data based on these multiple transmissions of the compressed CSI-RS measurement data for that portion of the bandwidth.

[0214] For example, the CSI-RS measurement data mentioned above can refer to channel information obtained by measuring CSI-RS, or measurement data obtained directly by measuring CSI-RS, such as RSRP, SINR, etc. The definition of channel information can be found above.

[0215] It should be noted that the encoder on the terminal side can be deployed inside the terminal device or in other devices outside the terminal device, such as the aforementioned server; the decoder on the network side can be deployed inside the network device or in other devices outside the network device, such as the aforementioned intelligent network element.

[0216] It should be understood that although the figure shows an encoder and a decoder, this is only a model division from a functional perspective. The encoder can also be called the first model, and the decoder can also be called the third model, etc. This application does not limit this. In addition, this application does not limit the number of models included in the AE model.

[0217] It should also be understood that the AE model is only one possible model for achieving the above functions and should not constitute any limitation on this application. The AE model can also be replaced by other AI models that can achieve the same or similar functions.

[0218] 4. Joint Source-Channel Coding (JSCC): This refers to the joint processing of source coding and channel coding using an AI model. Current research on JSCC mainly focuses on CSI compression feedback, where the terminal side can implement joint source and channel coding through an encoder model.

[0219] The terminal side can implement joint channels and channel coding through the encoder model in the following two ways.

[0220] Method 1: The encoder model on the terminal side can be used for encoding; that is, the encoder model has encoding capabilities. For example... Figure 7 As shown in (a) of the diagram, in Method 1, the input to the encoder model is the channel information to be processed, and the output is the encoded bit sequence. The channel information to be processed can be denoted as V. Furthermore, the output of the encoder model can be used as the input to the AI ​​modulation model, and the output of the AI ​​modulation model is the modulated constellation symbol.

[0221] Accordingly, the network side uses the received symbols as input to the AI ​​demodulation model, and the output of the AI ​​demodulation model is a bit sequence. Then, the bit sequence output by the AI ​​demodulation model can be used as input to the decoder model on the network side to obtain the reconstructed channel information output by the decoder model. The reconstructed channel information is denoted as... The decoder model on the network side can be used for decoding, meaning the decoder model has decoding capabilities.

[0222] Method 2: The encoder model on the terminal side can be used for both encoding and modulation; that is, the encoder model has both encoding and modulation capabilities. For example... Figure 7 As shown in (b) of the diagram, in mode 2, the input to the encoder model is the channel information to be processed, and the output is complex modulation symbols. The channel information to be processed can be denoted as V.

[0223] Accordingly, the network side uses the received symbols as input to the decoder model to obtain the reconstructed channel information output by the decoder model. The reconstructed channel information is denoted as... The decoder model on the network side can be used for demodulation and decoding, meaning the decoder model has the functions of demodulation and decoding.

[0224] The encoding in method 1 or method 2 described above may include source coding and channel coding, or the encoding in method 1 or method 2 may be joint source-channel coding. Source coding can also be referred to as compression. Decoding may include source decoding and channel decoding, or the decoding described above may be joint source-channel decoding. Source decoding can also be referred to as decompression.

[0225] 5. Model training: By selecting an appropriate function (such as a loss function), the model parameters are trained using optimization algorithms to minimize the difference between the model's predicted values ​​and the ground truth (or, the true value, the target value, and the label).

[0226] For example, model training methods include, but are not limited to, supervised learning, self-supervised learning, and knowledge distillation. These methods are briefly explained below.

[0227] Supervised learning, also known as supervised instruction, involves using machine learning algorithms to learn the mapping relationship between sample values ​​and labels based on collected sample values ​​and labels. This learned mapping relationship is then expressed using a machine learning model. The process of training the machine learning model is the process of learning this mapping relationship. For example, in signal detection, the noisy received signal is the sample, and the corresponding real constellation point is the label. Machine learning aims to learn the mapping relationship between samples and labels through training, that is, to enable the machine learning model to learn a signal detector. During training, the model parameters are optimized by calculating the error between the model's predicted values ​​and the real labels. Once the mapping relationship is learned, it can be used to predict the sample label of each new sample. The mapping relationship learned in supervised learning can include linear and nonlinear mappings. Based on the type of label, the learning task can be divided into classification tasks and regression tasks.

[0228] For example, the model training for the joint source-channel coding model for CSI feedback mentioned above can be achieved through supervised learning. For instance, during model training, the input to the model can be the original channel information, and the output can be the CSI feedback information. Then, the model parameters are optimized by calculating the error between the model's output CSI feedback information and the label corresponding to the model's output. For example, the error between the model's output and the label can be characterized by mean absolute error (MAE) or mean square error (MSE). Furthermore, reducing the MAE / MSE between the model's output and the label can be used as the training objective for training the joint source-channel coding model for CSI feedback. For example, if the MAE / MSE between the model's output and the label is less than or equal to a predefined threshold, the model training is considered complete; if the MAE / MSE is greater than the predefined threshold, the model parameters are optimized to achieve the training objective. Here, MAE / MSE can also be called the loss function for model training. The training objective described above can be replaced by increasing the square generalized cosine similarity (SGCS) between the model's output and the label. That is, the completion of model training can be determined by whether the SGCS between the model's output and the label is greater than or equal to a predefined threshold.

[0229] The model training method described above for the joint source-channel coding model used for CSI feedback can also be applied to the modulation model, or to the training of the joint source-channel coding and modulation model used for CSI feedback. In other words, the modulation model can be trained through supervised learning of the labels corresponding to the model's output.

[0230] For example, the joint source-channel coding model used for CSI feedback and the model used for modulation can be the same model; in other words, the model is used for both joint source-channel coding and modulation for CSI feedback. Alternatively, the joint source-channel coding model used for CSI feedback and the model used for modulation can be different models.

[0231] Self-supervised learning is a type of unsupervised learning. Unsupervised learning relies on collected sample values ​​to allow algorithms to discover inherent patterns within the samples themselves. Self-supervised learning uses the samples themselves as supervisory signals; that is, the model learns the mapping relationship from sample to sample. During training, model parameters are optimized by calculating the error between the model's predicted values ​​and the actual samples. Self-supervised learning can be used in signal compression and decompression recovery applications; common algorithms include autoencoders and generative adversarial networks.

[0232] For example, model training for the joint source-channel coding model mentioned above for CSI feedback can be achieved through self-supervised learning. For instance, the model set #a can be trained to achieve the goal of training the joint source-channel coding model for CSI feedback. Here, model set #a contains the joint source-channel coding model for CSI feedback and the joint source-channel decoding model for reconstructing channel information. During model training of model set #a, the input to model set #a is the original channel information, and the output of model set #a is the reconstructed channel information. Then, the model parameters of the models included in model set #a are optimized by calculating the error between the reconstructed channel information and the original channel information. For example, the error between the reconstructed channel information and the original channel information can be characterized by MAE, MSE, or normalized mean squared error (NMSE). Furthermore, reducing the MAE / MSE / NMSE between the reconstructed channel information and the original channel information can be used as the training objective when training model set #a. For example, if the MAE / MSE / NMSE between the reconstructed channel information and the original channel information is less than or equal to a predefined threshold, then the training of model set #a is considered complete, and consequently, the training of the joint source-channel coding model used for CSI feedback is considered complete. If the MAE / MSE / NMSE between the reconstructed channel information and the original channel information is greater than the predefined threshold, then the model parameters are optimized to achieve the training objective. Here, MAE / MSE / NMSE can also be called the loss function for model training. The above training objective can be replaced by increasing the square generalized cosine similarity (SGCS) between the reconstructed channel information and the original channel information; that is, by judging whether the SGCS between the reconstructed channel information and the original channel information is greater than or equal to a predefined threshold, it is determined whether the model training is complete.

[0233] The training method described above for model set #a can also be applied to training models for model set #b. Model set #b includes models for modulation and models for demodulation. The input to the demodulation model is the output of the modulation model. In other words, self-supervised learning on the input to model set #b can be used to train models for model set #b, thereby enabling the training of all models included in model set #b.

[0234] For example, the joint source-channel coding model used for CSI feedback and the model used for modulation can be the same model; in other words, the model is used for both joint source-channel coding and modulation for CSI feedback. Alternatively, the joint source-channel coding model used for CSI feedback and the model used for modulation can be different models.

[0235] Knowledge distillation: Generally, large models are often single complex networks or collections of networks, possessing excellent performance and generalization ability, while small models, due to their smaller network size, have limited expressive power. Therefore, the knowledge learned by the large model can be used to guide the training of the small model; this process is called knowledge distillation. Knowledge distillation can enable small models to achieve performance comparable to large models, but with fewer parameters and shorter inference latency, thus achieving model compression and acceleration. Furthermore, directly training small models with massive amounts of data often does not yield good performance, while training large models with massive amounts of data and then using the large model to perform knowledge distillation on the small model can achieve better continuation results. In addition, knowledge distillation can also be used to integrate and transfer datasets from different domains.

[0236] Knowledge distillation employs a teacher-student model, where a teacher model assists in training a student model. The teacher model is a complex, large model, while the student model is a simple, small model. Because the teacher model has strong learning capabilities, it can transfer the knowledge it learns to the relatively weaker student model, thereby enhancing the student model's generalization ability. The complex, cumbersome, but effective teacher model remains offline, simply acting as a mentor; the flexible and lightweight student model is the one actually deployed for prediction tasks.

[0237] The model training methods listed above are merely examples, and this application does not limit the methods used for model training.

[0238] Model training in this application embodiment may include initial training (or original training) or retraining of the model. Model training enables model enhancement. Model enhancement can refer to improving model functionality through model structure optimization and / or model training, such as adding other functions or neural networks to the existing model functionality, so that the enhanced model can be applied to more scenarios or richer requirements.

[0239] 6. Model Files and Model Parameters: Model files and / or model parameters can be used to determine the model. Therefore, the model in this application can refer to the model itself, or it can refer to the model files and / or model parameters used to determine the model.

[0240] The model file can be used to indicate the model structure, which includes, but is not limited to, deep neural networks (DNNs), feedforward neural networks (FNNs), convolutional neural networks (CNNs), and recurrent neural networks (RNNs). The model file can have a fixed format, such as a standard predefined format or a format pre-negotiated by both ends of the connection. Model parameters can refer to parameters in the neural network model, such as, but not limited to, the number of layers in the neural network, the type and weights of neurons in each layer, etc. This application does not limit the method of distributing model parameters.

[0241] Take DNN as an example. The idea behind DNN comes from the neuronal structure of the brain. Each neuron can perform a weighted summation operation on its inputs and then use the result of the weighted summation operation to generate the output through a nonlinear function. Figure 8 This is an example of a neuron structure. Figure 8 The input to the neuron shown is x = [x0x1…x] Z-1 The weights corresponding to the inputs are w = [w0w1…w] Z-1 The bias of the weighted summation is b. The nonlinear function f() can take many forms; for example, the nonlinear function f() can be the maximum value function max{0, x}. Then the effect of a neuron's execution is... Where Z is a positive integer; n is a positive integer greater than or equal to 0 and less than or equal to (Z-1).

[0242] A typical DNN has multiple neural network layers, including an input layer, one or more hidden layers, and an output layer. Generally, the first layer is the input layer, the last layer is the output layer, and the layers in between are hidden layers. Each layer contains multiple neurons. Layers are fully connected; that is, any neuron in the i-th layer is connected to any neuron in the (i+1)-th layer. The input layer processes the received values ​​(i.e., the DNN's input) through neurons and then passes them to the hidden layers. Similarly, the hidden layers pass the computation results to the final output layer, producing the DNN's output. Figure 9 An example of a DNN is shown. Figure 9 The DNN model shown has three neural network layers: an input layer, a hidden layer, and an output layer.

[0243] It should be understood that the above text, in combination with... Figure 8 and Figure 9The examples provided are for illustrative purposes only and should not be construed as limiting the scope of this application. This application does not limit the structure and parameters used in the AI ​​model.

[0244] One of the model structure or model parameters can be predefined, while the other can be sent by the sender (e.g., the network side). Alternatively, both the model structure and model parameters can be sent by the sender (e.g., the network side). This application does not impose any restrictions on this.

[0245] 7. Two-way connection: This refers to the connection between the sender (e.g., the network side) and the receiver (e.g., the terminal side). It can be a connection between datasets or between models. A dataset is used in machine learning for model training, validation, and testing; the quantity and quality of the dataset affect the effectiveness of machine learning. In this embodiment, the dataset can be used for model training.

[0246] The following section uses CSI compressed inference tasks as an example to explain the connection between the dataset and the model.

[0247] Dataset docking mainly refers to the sending end sending a dataset to the receiving end for model training on the receiving end. Figure 10 This is a schematic diagram illustrating the data set connection between the network side and the terminal side.

[0248] like Figure 10 As shown, the network side can first obtain a dual-end model through joint training, and then obtain the dataset using the obtained dual-end model. For example, the network side can use a CSI compression model to compress the channel information, and then use a CSI reconstruction model to decompress the channel information. That is, the channel information is the input of the CSI compression model, and the output of the CSI compression model can be the compressed CSI. The compressed CSI can be the input of the CSI reconstruction model, and the CSI reconstruction model can output the reconstructed channel information.

[0249] Optionally, the compressed CSI can be further quantized to obtain quantized CSI. During the model training phase, both the compressed CSI and the quantized CSI can be used as CSI feedback information.

[0250] For example, the dataset sent from the network side to the terminal side may include one or more of the following: channel information, CSI feedback information, or reconstructed channel information. Optionally, the dataset may also include a quantization method. In other words, the quantization method is also indicated by the network side. Optionally, the quantization method is predefined, such as protocol predefined. The quantization method is used to determine how to quantize the compressed CSI and / or how to dequantize the quantized CSI. The quantization method can also be replaced by a dequantization method, a quantizer, or a dequantizer.

[0251] The network side can distribute this dataset to the terminal side. The terminal side can then train a model based on the received dataset to obtain a CSI compressed model for the terminal side.

[0252] Model integration mainly refers to the sending end sending model files and / or model parameters to the receiving end for model deployment on the receiving end, or for model development (e.g., model training) and model deployment on the receiving end. Figure 11 This is a schematic diagram illustrating the model interoperability between the network side and the terminal side.

[0253] like Figure 11 As shown, the network side can first obtain the CSI compression model through joint training. The specific process of network-side joint training can be found in the above text. Figure 10 The explanation will not be repeated here. The network side can distribute the CSI compressed model or CSI reconstructed model obtained from joint training to the terminal side. For example, the network side can send the model file and / or model parameters of the CSI compressed model to the terminal side, or send the model file and / or model parameters of the CSI reconstructed model to the terminal side. The terminal side can then train the model based on the received model file and / or model parameters.

[0254] It should be understood that the above text, in combination with... Figure 10 and Figure 11 The CSI compression model shown can be an encoder, and the CSI reconstruction model can be a decoder.

[0255] As mentioned earlier, the terminal can process the acquired CSI data based on an AI model to obtain a CSI report and send it to the network. Therefore, there is an urgent need for a solution to monitor the performance of the model used for CSI feedback on the terminal side to ensure the reliability of the CSI reports submitted by the terminal.

[0256] In view of this, this application provides a method to facilitate the monitoring of the performance of the AI ​​model used for CSI feedback, i.e., the AI ​​model used by the sender of the CSI report.

[0257] The method provided in this application will now be described in detail with reference to the accompanying drawings.

[0258] It should be understood that the communication method provided in this application embodiment can be applied to communication systems including terminal devices and network devices, for example, applied to Figure 1 or Figure 2 The communication architecture shown can also be applied to other communication systems. This application does not limit the application scenario in any way, and any communication system including the devices in the following embodiments is applicable.

[0259] It should also be understood that the embodiments shown below do not particularly limit the specific structure of the execution subject of the method provided in the embodiments of this application, as long as it is possible to communicate according to the method provided in the embodiments of this application by running a program that records the code of the method provided in the embodiments of this application. For example, the execution subject of the method provided in the embodiments of this application may be a device or apparatus; or, it may be a functional module in a device or apparatus that can call and execute a program.

[0260] The following describes in detail the communication method provided in this application embodiment, taking the interaction between the first device side and the second device side as an example. For example, the first device side can be replaced by the terminal side, and the second device side can be replaced by the network side. For another example, the first device side can be replaced by the network side, and the second device side can be replaced by the terminal side. More detailed descriptions of the terminal side and the network side can be found above.

[0261] Figure 12 This is a schematic flowchart of the communication method 1200 provided in the embodiments of this application. Figure 12 The method 1200 shown may include steps S1210 and S1220. Optionally, the method may further include one or more steps of S1201, S1230, and S1240. The various steps in method 1200 are described in detail below.

[0262] S1210, the first device performs first processing on the first channel information based on at least one model to obtain N modulation symbols.

[0263] The first process includes encoding and modulation. N is a positive integer. Encoding may include source coding and channel coding, or the above encoding may be joint source-channel coding. Source coding can also be referred to as compression.

[0264] The first channel information can be of the following types: channel matrix, channel feature matrix, channel response information, or precoding matrix. For a more detailed description of the first channel information, please refer to the explanation of channel information in the terminology section above.

[0265] For example, if the first device side is replaced by the terminal side, then S1210 may specifically include: the terminal device performs a first processing on the first channel information based on at least one model to obtain N modulation symbols; or, S1210 may specifically include: the server performs a first processing on the first channel information based on at least one model to obtain N modulation symbols.

[0266] For example, the terminal-side server mentioned in the embodiments of this application may include an OTT system server, a cloud server, a server provided by an operator or a third party, etc.

[0267] At least one model will be described below.

[0268] For example, at least one model includes a model, for instance, denoted as model #A. Model #A is used to perform a first process on the first channel information to obtain N modulation symbols. In other words, the inference task of model #A is to perform a first process on the first channel information to obtain N modulation symbols.

[0269] For example, at least one model may include two models, such as a first model and a second model. The first model encodes the first channel information to obtain a first bit sequence. The second model modulates the first bit sequence to obtain N modulation symbols. In other words, the inference task of the first model is to encode the first channel information to obtain the first bit sequence, and the inference task of the second model is to modulate the first bit sequence to obtain N modulation symbols.

[0270] This application does not limit the method by which the first device side obtains at least one model.

[0271] As an example, at least one model can be obtained by model training on the first device side. For instance, the first device side can train a pre-configured initial model to obtain at least one model. As another example, the first device side can train an initial model configured by the second device side on the first device side to obtain at least one model.

[0272] In another example, at least one model is configured by the second device side to the first device side. For example, configuring at least one model by the second device side to the first device side includes: the second device side sending the model file and / or model parameters of at least one model to the first device side.

[0273] The method by which the first device obtains the first channel information is described below.

[0274] In one possible implementation, the first device can acquire the first channel information through measurement or simulation. For example, the first device receives a reference signal from the second device and then acquires the first channel information by measuring the reference signal.

[0275] In one possible implementation, the first device can receive first channel information from the second device.

[0276] For example, if the first device side can be replaced by the terminal side and the second device side can be replaced by the network side, then the first device side receiving the first channel information from the second device side can specifically include: the terminal device receiving the first channel information from the network device, such as receiving it through the air interface; or, the terminal device obtaining the first channel information from the terminal-side server, which is received by the terminal-side server from the intelligent network element or server on the network side, such as receiving it through a wired network; or, the chip inside the terminal device, that is, the chip used in the terminal device, receiving the first channel information from the network device.

[0277] In another example, if the first device side is replaced by the terminal side and the second device side is replaced by the network side, then the first device side receives the first channel information from the second device side. Specifically, this may include: the terminal side server receiving the first channel information from the intelligent network element or server on the network side, such as receiving it through a wired network; or, the terminal side server obtaining the first channel information from the terminal device, which is received by the terminal device from the network side, such as a network device, such as receiving it through the air interface.

[0278] The network-side servers mentioned in this application may include cloud servers, servers provided by operators, etc.

[0279] S1220, the first device side determines the performance monitoring results based on N modulation symbols and M known modulation symbols.

[0280] The performance monitoring results determined by the first device side are the performance monitoring results of at least one model. The performance monitoring results of at least one model are used to determine whether the performance of at least one model meets the performance requirements. M is a positive integer.

[0281] For example, if the first device side is replaced by the terminal side, then S1220 may specifically include: the terminal device determining the performance monitoring result based on N modulation symbols and M known modulation symbols; or, S1220 may specifically include: the server on the terminal side determining the performance monitoring result based on N modulation symbols and M known modulation symbols.

[0282] As an example, the first device side can determine the performance monitoring results of at least one model during the process of applying at least one model for model inference.

[0283] In another example, the first device can determine the performance monitoring results of at least one model during the training of at least one model. In other words, the first channel information and M known modulation symbols are used for model training of at least one model. Based on this, the performance monitoring results determined by the first device are used to determine whether the training of at least one model is complete. For example, if the performance monitoring results of at least one model indicate that the performance of at least one model meets the performance requirements, it means that the training of at least one model is complete; if the performance monitoring results of at least one model indicate that the performance of at least one model does not meet the performance requirements, it means that the training of at least one model is not complete.

[0284] The following describes how the first device acquires M known modulation symbols.

[0285] In one possible implementation, the M known modulation symbols are predefined modulation symbols. For example, the M known modulation symbols are predefined modulation symbols in a protocol or standard.

[0286] In another possible implementation, the M known modulation symbols are determined by the first device side based on a predefined method. For example, the first device side can determine the M known modulation symbols by predefining the length of the bit sequence to be modulated and the number of bits mapped to a modulation symbol.

[0287] In another possible implementation, the M known modulation symbols are obtained by the first terminal side from the second device side. Based on this implementation, method 1200 further includes S1201: the first device side receives the M known modulation symbols from the second device side. Accordingly, in S1201, the second device side transmits the M known modulation symbols to the first device side.

[0288] For example, if the first device side is replaced by the terminal side and the second device side is replaced by the network side, then S1201 may specifically include: the terminal device receiving M known modulation symbols from the network device, for example, receiving them through the air interface; or, the terminal device obtaining M known modulation symbols from the terminal-side server, which are received by the terminal-side server from the intelligent network element or server on the network side, for example, receiving them through a wired network; or, the chip inside the terminal device, that is, the chip used in the terminal device, receiving M known modulation symbols from the network device.

[0289] In another example, if the first device side is replaced by the terminal side and the second device side is replaced by the network side, then S1201 may specifically include: the terminal side server receiving M known modulation symbols from the intelligent network element or server on the network side, for example, received through a wired network; or, the terminal side server obtaining M known modulation symbols from the terminal device, which are received by the terminal device from the network side, such as a network device, for example, received through an air interface.

[0290] The following describes how the first device determines the performance monitoring results based on N modulation symbols and M known modulation symbols.

[0291] In one possible implementation, the first device determines the performance monitoring result based on a first modulation symbol and K second modulation symbols. The performance monitoring result may include one or more of the following: a first MAE; a first MSE; a first SGCS; a weighted sum or weighted average of at least two of the first MAE, first MSE, and first SGCS. In other words, the first device determines the performance monitoring result based on one or more of the first MAE, first MSE, or first SGCS.

[0292] The first modulation symbol will be denoted as modulation symbol #1, and the second modulation symbol will be denoted as modulation symbol #2.

[0293] K is a positive integer less than or equal to M. The value of K is a predefined value, or the value of K is indicated by the second device to the first device and is not limited.

[0294] In this context, modulation symbol #1 belongs to N modulation symbols, and K modulation symbols #2 belong to M known modulation symbols. The geometric distance between modulation symbol #2 and modulation symbol #1 is less than or equal to the geometric distance between any of the M known modulation symbols (excluding the K modulation symbols #2) and modulation symbol #1. The geometric distance between two modulation symbols can also be referred to as the MAE between the two modulation symbols.

[0295] The first MAE will be introduced below.

[0296] The first MAE is determined based on the MAE of modulation symbol #1 and each of the K modulation symbols #2. For example, the MAE between modulation symbol #1 and the kth modulation symbol #2 is denoted as MAE. k If k = 1, 2, ..., K, then the first MAE can be MAE1 to MAE2. K The maximum value in, or MAE1 to MAE K The average value, or MAE1 to MAE K The sum. For example, in this application, B is determined based on A, which can also be described as B being related to A, or B being a function of multiple parameters, including A, etc.

[0297] For example, modulation symbol #1 is represented as: s = s real +j·s imag The k-th modulation symbol #2 is represented as: MAE between modulation symbol #1 and the kth modulation symbol #2 kIt can be expressed as or satisfy the following formula (1).

[0298]

[0299] Among them, s real s represents the real part of modulation symbol #1. imag This represents the imaginary part of modulation symbol #1. This represents the real part of the k-th modulation symbol #2. This represents the imaginary part of the k-th modulation symbol #2. j is the imaginary unit, j 2 =-1.

[0300] It should be noted that the MAE between modulation symbol #1 and the kth modulation symbol #2 k It can also be called the geometric distance between modulation symbol #1 and the kth modulation symbol #2, or the L1 norm between modulation symbol #1 and the kth modulation symbol #2, or the Manhattan distance between modulation symbol #1 and the kth modulation symbol #2.

[0301] The first MSE will be introduced below.

[0302] The first MSE is determined based on the MSE of modulation symbol #1 and each of the K modulation symbols #2. For example, the MSE between modulation symbol #1 and the kth modulation symbol #2 is denoted as MSE. k If k = 1, 2, ..., K, then the first MSE can be MSE1 to MSE2. K The maximum value in, or MSE1 to MSE K The average value, or MSE1 to MSE K The sum of.

[0303] For example, modulation symbol #1 is represented as: s = s real +j·s imag The k-th modulation symbol #2 is represented as: MSE between modulation symbol #1 and the kth modulation symbol #2 k It can be expressed as or satisfy the following formula (2).

[0304]

[0305] It should be noted that the MSE between modulation symbol #1 and the kth modulation symbol #2 k It can also be called the L2 norm between modulation symbol #1 and the kth modulation symbol #2.

[0306] The first SGCS will be introduced below.

[0307] The first SGCS is determined based on the SGCS of modulation symbol #1 and each of the K modulation symbols #2. For example, the SGCS between modulation symbol #1 and the kth modulation symbol #2 is denoted as SGCS. k If k = 1, 2, ..., K, then the first SGCS can be SGCS1 to SGCS2. K The maximum value in, or SGCS1 to SGCS K The average value, or SGCS1 to SGCS K The sum of.

[0308] For example, modulation symbol #1 is represented as: s = s real +j·s imag The k-th modulation symbol #2 is represented as: SGCS between modulation symbol #1 and the kth modulation symbol #2 k It can be expressed as or satisfy the following formula (3).

[0309]

[0310] in,

[0311] In one possible implementation, the first device determines the performance monitoring result based on multiple modulation symbols #x and K modulation symbols #y corresponding to each modulation symbol #x. Here, the multiple modulation symbols #x belong to N modulation symbols, and the K modulation symbols #y corresponding to each modulation symbol #x belong to M known modulation symbols. Taking modulation symbol #x1 among the multiple modulation symbols #x as an example, the K modulation symbols #y corresponding to modulation symbol #x1 are denoted as K modulation symbols #y1. Then, the geometric distance between modulation symbol #x1 and modulation symbol #y1 is less than or equal to the geometric distance between any modulation symbol other than the K modulation symbols #y1 and modulation symbol #x1 among the M known modulation symbols.

[0312] For example, if multiple modulation symbols #x include a first modulation symbol and a third modulation symbol, then the first device determines the performance monitoring based on the first modulation symbol, K corresponding second modulation symbols, a third modulation symbol, and K corresponding fourth modulation symbols. The performance monitoring result determined by the first device is a weighted sum or weighted average of at least two of the following: first MAE; first MSE; first SGCS; second MAE; second MSE; second SGCS; first MAE, first MSE, first SGCS, second MAE, second MSE, or second SGCS. In other words, the first device determines the performance monitoring result based on one or more of the first MAE, first MSE, first SGCS, second MAE, second MSE, or second SGCS.

[0313] For example, the performance monitoring results determined by the first device side include the first MAE; the first MSE; the first SGCS; the second MAE; the second MSE; the second SGCS; the weighted sum or weighted average of at least two of the first MAE, first MSE, first SGCS, second MAE, second MSE, or second SGCS.

[0314] The third modulation symbol will be denoted as modulation symbol #3, and the fourth modulation symbol will be denoted as modulation symbol #4.

[0315] For an introduction to the first MSE, the first MAE, and the first SGCS, please refer to the above text.

[0316] The second MAE will be introduced below.

[0317] The second MAE is determined based on the MAE of modulation symbol #3 and each of the K modulation symbols #4. For example, the MAE between modulation symbol #3 and the kth modulation symbol #4 out of the K modulation symbols #4 is denoted as MAE. k If k' = 1, 2, ..., K, then the second MAE can be MAE1' to MAE'. K The maximum value in ', or MAE1' to MAE K The average value of ', or MAE1' to MAE K 'and'.

[0318] For example, the MAE between modulation symbol #3 and the kth modulation symbol #4 k The calculation method for ' can be found in formula (1) above.

[0319] The second MSE will be introduced below.

[0320] The second MSE is determined based on the MSE of modulation symbol #3 and each of the K modulation symbols #4. For example, the MSE between modulation symbol #3 and the kth modulation symbol #4 out of the K modulation symbols #4 is denoted as MSE. k If k' = 1, 2, ..., K, then the second MSE can be MSE1' to MSE... K The maximum value in ', or MSE1' to MSE K The average value of ', or MSE1' to MSE K 'and'.

[0321] For example, the MSE between modulation symbol #3 and the kth modulation symbol #4 k The calculation method for ' can be found in formula (2) above.

[0322] The second SGCS will be introduced below.

[0323] The second SGCS is determined based on the SGCS of modulation symbol #3 and each of the K modulation symbols #4. For example, the SGCS between modulation symbol #3 and the kth modulation symbol #4 among the K modulation symbols #4 is denoted as SGCS. k If k' = 1, 2, ..., K, then the second SGCS can be SGCS1' to SGCS2'. K The maximum value in ', or SGCS1' to SGCS K The average value of ', or SGCS1' to SGCS K 'and'.

[0324] For example, the SGCS between modulation symbol #3 and the kth modulation symbol #4 k The calculation method for ' can be found in formula (3) above.

[0325] The relationship between the performance monitoring results of at least one model and the performance of at least one model is described below.

[0326] For example, if the performance monitoring results include the value of a first performance metric, and the first performance metric is used to measure the similarity between the output of at least one model and the label, then if the value of the first performance metric is greater than or equal to a similarity threshold, the performance monitoring results are used to indicate that the performance of at least one model meets the performance requirements; or, if the value of the first performance metric is less than a similarity threshold, the performance monitoring results are used to indicate that the performance of at least one model does not meet the performance requirements.

[0327] For example, the aforementioned first performance indicator can be the SGCS, meaning that the performance monitoring results determined by the first device side can include the first SGCS and / or the second SGCS. Taking the performance monitoring results including the first SGCS as an example, if the first SGCS is greater than or equal to the SGCS threshold, the performance monitoring results are used to indicate that the performance of at least one model meets the performance requirements; or, if the first SGCS is less than the SGCS threshold, the performance monitoring results are used to indicate that the performance of at least one model does not meet the performance requirements.

[0328] For another example, if the performance monitoring results include the value of a second performance metric, and the second performance metric is used to measure the difference between the output of at least one model and the label, then if the value of the second performance metric is greater than or equal to the difference threshold, the performance monitoring results are used to indicate that the performance of at least one model does not meet the performance requirements; or, if the value of the second performance metric is less than the difference threshold, the performance monitoring results are used to indicate that the performance of at least one model meets the performance requirements.

[0329] For example, the second performance indicator mentioned above can be MSE, meaning that the performance monitoring results determined by the first device side can include the first MSE and / or the second MSE. Taking the performance monitoring results including the first MSE as an example, if the first MSE is greater than or equal to the MSE threshold, the performance monitoring results are used to indicate that the performance of at least one model does not meet the performance requirements; or, if the first MSE is less than the MSE threshold, the performance monitoring results are used to indicate that the performance of at least one model meets the performance requirements.

[0330] For example, if the performance monitoring results include the values ​​of a first performance indicator and a second performance indicator, then if the value of the first performance indicator is greater than or equal to the similarity threshold and the value of the second performance indicator is less than the difference threshold, the performance monitoring results are used to indicate that the performance of at least one model meets the performance requirements; or, if the value of the first performance indicator is less than the similarity threshold and the value of the second performance indicator is greater than or equal to the difference threshold, the performance monitoring results are used to indicate that the performance of at least one model does not meet the performance requirements.

[0331] In this embodiment, the first device performs first processing on the first channel information based on at least one model and outputs N modulation symbols. In this case, the first device acquires M known modulation symbols, which allows the first device to effectively monitor the performance of at least one model based on the M known modulation symbols and the N modulation symbols.

[0332] In one possible implementation, method 1200 may also include S1230.

[0333] S1230, the first device sends the performance monitoring results.

[0334] Correspondingly, the second device receives the performance monitoring results.

[0335] For example, if the first device side is replaced by the terminal side and the second device side is replaced by the network side, then S1230 may specifically include: the terminal device sending the performance monitoring result to the network device; or, the terminal device sending the performance monitoring result to the server on the terminal side, so that the server on the terminal side can send the performance monitoring result to the intelligent network element or server on the network side; or, the chip inside the terminal device, that is, the chip used in the terminal device, sending the performance monitoring result to the network device.

[0336] For example, if the first device side is replaced by the terminal side and the second device side is replaced by the network side, then S1230 may specifically include: the server on the terminal side sending performance monitoring results to the intelligent network element or server on the network side; or, the server on the terminal side may send performance monitoring results to the terminal device, so that the terminal device can send performance monitoring results to the network device.

[0337] It should be understood that after the first device side determines the performance monitoring results of at least one model, it can directly send the performance monitoring results to the second device side, or it can quantify, normalize, or encode the performance monitoring results determined by the first device side before sending the processed performance monitoring results to the second device side.

[0338] It should also be understood that if the first device side obtains a similarity threshold and / or difference threshold for determining whether at least one model meets the performance requirements, the performance monitoring results sent by the first device side to the second device side may include: the performance of at least one model meets the performance requirements, or the performance of at least one model does not meet the requirements.

[0339] In this embodiment, the first device sends performance monitoring results to the second device, which helps the second device to promptly detect that the performance of at least one model does not meet the performance requirements, and then promptly update, switch, or roll back the at least one model.

[0340] In one possible implementation, method 1200 further includes S1240.

[0341] S1240, the second device sends the first information.

[0342] Accordingly, the first device receives the first information.

[0343] The first information is used to instruct at least one model to be retrained, or to instruct at least one model to be shut down, or to instruct at least one model to be switched to another model that has the same function as at least one model.

[0344] It should be understood that after receiving the performance monitoring results from the first device, if the second device determines that the performance of at least one model does not meet the performance requirements based on the performance monitoring results, the second device may send the first information to the first device.

[0345] For example, if the first device side is replaced by the terminal side and the second device side is replaced by the network side, then S1240 may specifically include: the terminal device receiving first information from the network device, such as receiving it through the air interface; or, the terminal device obtaining the first information from the server on the terminal side, which is received by the server on the terminal side from the intelligent network element or server on the network side, such as receiving it through a wired network; or, the chip inside the terminal device, that is, the chip used in the terminal device, receiving the first information from the network device.

[0346] In another example, if the first device side is replaced by the terminal side and the second device side is replaced by the network side, then S1240 may specifically include the terminal side server receiving first information from the intelligent network element or server on the network side, such as receiving it through a wired network; or, the terminal side server obtaining first information from the terminal device, which is received by the terminal device from the network side, such as a network device, such as receiving it through an air interface.

[0347] In this embodiment, the second device sends first information to the first device to instruct the first device to perform operations such as retraining, switching, or rollback on at least one model. This can avoid the problem of communication performance degradation caused by the first device using at least one model whose performance does not meet the requirements for channel information feedback.

[0348] As mentioned above, if at least one model includes a first model and a second model, then the first device can monitor the performance of the first model and the second model respectively. The following section combines... Figure 13 The method of monitoring the performance of the first model and the second model on the first device side is described.

[0349] Figure 13 This is a schematic flowchart of the communication method 1300 provided in the embodiments of this application. Figure 13 The method 1300 shown may include steps S1310 to S1340. Optionally, the method may further include one or more steps of S1301, S1350, and S1360. The various steps in method 1300 are described in detail below.

[0350] S1301, the second device sends M known modulation symbols.

[0351] Accordingly, the first device receives M known modulation symbols.

[0352] For a more detailed description of S1301, please refer to S1201 in Method 1200 above.

[0353] S1310, the second device sends the first channel information and the second bit sequence.

[0354] Accordingly, the first device receives the first channel information and the second bit sequence.

[0355] For example, the second bit sequence is obtained by encoding the first channel information based on the third model on the second device side. The third model can be obtained by the second device side through joint training of model set #1 and model set #2. Model set #1 includes the third model and the fourth model. The fourth model is used to modulate the encoded bit sequence output by the third model to obtain a modulated symbol sequence. Model set #2 is used to demodulate and decode the modulated symbol sequence output by the fourth model to obtain the reconstructed channel information. For example, if the second device side performs joint training of model set #1 and model set #2 based on the first channel information, the second device side can determine whether the training of model set #1 and model set #2 is complete by comparing the reconstructed channel information output by model set #2 with the first channel information. For example, if the SGCS between the first channel information and the reconstructed channel information is greater than or equal to a predefined threshold, it indicates that model set #1 and model set #2 are trained successfully; if the SGCS between the first channel information and the reconstructed channel information is less than the predefined threshold, it indicates that model set #1 and model set #2 are not trained successfully.

[0356] The encoding process can include source coding and channel coding, or the above coding can be combined source-channel coding. Source coding can also be referred to as compression. Decoding can include source decoding and channel decoding, or the above decoding can be combined source-channel decoding. Source decoding can also be referred to as decompression.

[0357] For example, if the first device side can be replaced by the terminal side and the second device side can be replaced by the network side, then S1310 may specifically include: the terminal device receiving first channel information and a second bit sequence from the network device, for example, receiving it through the air interface; or, the terminal device obtaining the first channel information and the second bit sequence from the server on the terminal side, wherein the first channel information and the second bit sequence are received by the server on the terminal side from the intelligent network element or server on the network side, for example, receiving it through a wired network; or, the chip inside the terminal device receiving the first channel information and the second bit sequence from the chip inside the network device.

[0358] In another example, if the first device side is replaced by the terminal side and the second device side is replaced by the network side, then S1310 may specifically include: the server on the terminal side receiving the first channel information and the second bit sequence from the intelligent network element or server on the network side, for example, received through a wired network; or, the server on the terminal side obtaining the first channel information and the second bit sequence from the terminal device, wherein the first channel information and the second bit sequence are received by the terminal device from the network side, such as a network device, for example, received through an air interface.

[0359] The terminal-side servers mentioned in the embodiments of this application may include OTT system servers, cloud servers, servers provided by operators or third parties, etc. The network-side servers mentioned in the embodiments of this application may include cloud servers, servers provided by operators, etc.

[0360] S1320, the first device encodes the first channel information based on the first model to obtain the first bit sequence.

[0361] For example, if the first device side is replaced by the terminal side, then S1320 may specifically include: the terminal device encodes the first channel information based on the first model to obtain the first bit sequence; or, S1320 may specifically include: the server on the terminal side encodes the first channel information based on the first model to obtain the first bit sequence.

[0362] S1330, the first device modulates the first bit sequence based on the second model to obtain N modulation symbols.

[0363] Where N is a positive integer.

[0364] For example, if the first device side is replaced by the terminal side, then S1330 may specifically include: the terminal device modulates the first bit sequence based on the second model to obtain N modulation symbols; or, S1330 may specifically include: the server on the terminal side modulates the first bit sequence based on the second model to obtain N modulation symbols.

[0365] Regarding the method by which the first device side obtains the first model and the second model, you can refer to the method described in Method 1200 above for the method by which the first device side obtains at least one model.

[0366] S1340, the first device determines the performance monitoring result based on the first bit sequence, the second bit sequence, N modulation symbols, and M known modulation symbols.

[0367] The performance monitoring results determined by the first device side are the performance monitoring results of at least one model. The performance monitoring results of at least one model are used to determine whether the performance of at least one model meets the performance requirements. M is a positive integer.

[0368] The method for obtaining M known modulation symbols on the first device side can refer to method 1200 above.

[0369] For example, if the first device side is replaced by the terminal side, then S1340 may specifically include: the terminal device determining the performance monitoring result based on the first bit sequence, the second bit sequence, N modulation symbols and M known modulation symbols; or, S1340 may specifically include: the server on the terminal side determining the performance monitoring result based on the first bit sequence, the second bit sequence, N modulation symbols and M known modulation symbols.

[0370] The following describes how the performance monitoring results are determined on the first equipment side.

[0371] For example, the first device side determines the first monitoring result, which includes the performance monitoring result, based on N modulation symbols and M known modulation symbols.

[0372] The method by which the first device determines the first monitoring result can be referred to the description in S1220 above. Referring to S1220 above, the first monitoring result determined by the first device may include one or more of the following: first MAE; first MSE; first SGCS; first MAE; second MSE; second SGCS; a weighted sum or weighted average of at least two of the first MAE, first MSE, first SGCS, second MAE, second MSE, or second SGCS. In other words, the first device may determine the first monitoring result based on one or more of the first MAE, first MSE, first SGCS, second MAE, second MSE, or second SGCS.

[0373] For further description of the first MAE, first MSE, first SGCS, second MAE, second MSE, or second SGCS, please refer to Method 1200 above.

[0374] It should be understood that the first monitoring result can be used to determine whether the performance of the second model meets the performance requirements. The relationship between the first monitoring result and the performance of the second model can be referred to the relationship between the performance monitoring results and the performance of at least one model described in Method 1200 above.

[0375] It should also be understood that the first monitoring results can also be used to determine whether the performance of the first model and the second model as a whole meets the performance requirements.

[0376] For example, the first device determines the second monitoring result, which is included in the performance monitoring result, based on the first bit sequence and the second bit sequence.

[0377] For example, the second monitoring result may include one or more of the following: a first binary cross-entropy (BCE); a third MSE; a third MAE; a weighted sum or weighted average of at least two of the first BCE, the third MSE, or the third MAE. In other words, the first device can determine the first monitoring result based on one or more of the first BCE, the third MSE, or the third MAE.

[0378] The first BCE is determined based on the probability of each bit in the first bit sequence being 1 and the second bit sequence.

[0379] For example, the first bit sequence consists of L bits, represented as: y = [y1, y2, ..., y... L The second bit sequence consists of L bits, represented as: The probability sequence corresponding to the first bit sequence is represented as: p = [p1, p2, ..., p...]. L ], where p i Indicates y i The probability that the value of is 1. Where L is a positive integer, i = 1, 2, ..., L. The first BCE can be expressed as or satisfies the following formula (4).

[0380]

[0381] The third MSE is the MSE between the first bit sequence and the second bit sequence.

[0382] For example, the first bit sequence is represented as: y = [y1, y2, ..., y L The second bit sequence is represented as: The third MSE can be expressed as or satisfy the following formula (5).

[0383]

[0384] The third MAE is the MAE between the first bit sequence and the second bit sequence.

[0385] For example, the first bit sequence is represented as: y = [y1, y2, ..., y L The second bit sequence is represented as: The third MSE can be expressed as or satisfy the following formula (6).

[0386]

[0387] It should be understood that the second monitoring result can be used to determine whether the performance of the first model meets the performance requirements. The relationship between the second monitoring result and the performance of the first model can be referred to the relationship between the performance monitoring result and the performance of at least one model described in Method 1200 above. For example, if the second monitoring result includes a third MAE, then if the third MAE is less than or equal to the MAE threshold, it indicates that the performance of the first model meets the performance requirements; or, if the third MAE is greater than the MAE threshold, it indicates that the performance of the first model does not meet the performance requirements.

[0388] In this embodiment, the first device performs a first processing on the first channel information based on a first model and a second model, and outputs N modulation symbols. In this case, the first device acquires M known modulation symbols, enabling it to effectively monitor the performance of at least one model based on the M known modulation symbols and the N modulation symbols. Furthermore, the first device can also acquire a second bit sequence, allowing it to effectively monitor the performance of the first model based on the second bit sequence and the first bit sequence output by the first model.

[0389] In one possible implementation, method 1300 may also include S1350.

[0390] S1350, the first device sends performance monitoring results.

[0391] Correspondingly, the second device receives the performance monitoring results.

[0392] For more details on S1350, please refer to the section on S1230 above.

[0393] In one possible implementation, method 1300 further includes S1360.

[0394] S1360, the second device sends the first information.

[0395] Accordingly, the first device receives the first information.

[0396] For more details on S1360, please refer to S1240 above.

[0397] The following is combined with Figure 14 Another method for performance monitoring provided in this application is described.

[0398] Figure 14 This is a schematic flowchart of the communication method 1400 provided in the embodiments of this application. Figure 14 The method 1400 shown may include steps S1410 to S1430. Optionally, the method may further include one or more steps in S1440 and S1450. The various steps in method 1400 are described in detail below.

[0399] S1410, the second device sends the dataset.

[0400] Accordingly, the first device receives the dataset.

[0401] The dataset includes first channel information and first modulation symbol sequence.

[0402] In this context, the first channel information corresponds to the first modulation symbol sequence. This correspondence can be understood as the first modulation symbol sequence being the modulation symbol sequence of the first channel information, or as the first modulation symbol sequence being obtained after the second device processes the first channel information. For example, the second device encodes and modulates the first channel information to obtain the first modulation symbol sequence.

[0403] The dataset is used for performance monitoring of at least one model. For example, using the dataset for performance monitoring of at least one model includes: taking first channel information as input to at least one model and taking a first modulation symbol sequence as a label or performance monitoring reference corresponding to the output of at least one model, and performing performance monitoring on at least one model. It should be understood that when the first channel information corresponds to the first modulation symbol sequence, the first modulation symbol sequence can be used as a label or performance monitoring reference corresponding to the output of at least one model when the first channel information is used as input. Further, when channel information other than the first channel information is used as input to at least one model, the first modulation symbol sequence cannot be used as a label or performance monitoring reference corresponding to the output of at least one model. At least one model can be referred to the description in method 1200 above.

[0404] For example, the first modulation symbol sequence is obtained by the second device side after encoding and modulating the first channel information based on model set #1. Model set #1 can be obtained by the second device side through joint training of model set #1 and model set #2. Model set #2 is used to demodulate and decode the modulation symbol sequence output by model set #1 to obtain the reconstructed channel information. For example, if the second device side performs joint training of model set #1 and model set #2 based on the first channel information, the second device side can determine whether the training of model set #1 and model set #2 is complete by comparing the reconstructed channel information output by model set #2 with the first channel information. For example, if the SGCS between the first channel information and the reconstructed channel information is greater than or equal to a predefined threshold, it means that model set #1 and model set #2 are trained successfully; if the SGCS between the first channel information and the reconstructed channel information is less than the predefined threshold, it means that model set #1 and model set #2 are not trained successfully.

[0405] The first channel information can be of the type of channel matrix, channel feature matrix, channel response information, or precoding matrix. For a more detailed description of the first channel information, please refer to the explanation of channel information in the terminology section above. Encoding can include source coding and channel coding, or, the above coding can be joint source-channel coding. Source coding can also be referred to as compression. Decoding can include source decoding and channel decoding, or, the above decoding can be joint source-channel decoding. Source decoding can also be referred to as decompression.

[0406] Optionally, the dataset may also include a second bit sequence. For example, if model set #1 includes a third model and a fourth model, where the third model is used to encode the first channel information and the fourth model is used to modulate the bit sequence output by the third model, then the dataset may include a second bit sequence obtained by processing the first channel information using the third model.

[0407] Where at least one model includes a first model and a second model, the second bit sequence and the first channel information are used for performance monitoring of the first model. The use of the first channel information and the second bit sequence for performance monitoring of the first model includes: using the first channel information as input to the first model, and using the second bit sequence as a tag corresponding to the output of the first model or a reference for performance monitoring, to perform performance monitoring of the first model. It should be understood that when the first channel information corresponds to the second bit sequence, when the first channel information is used as input to the first model, the second bit sequence can be used as a tag corresponding to the output of the first model or a reference for performance monitoring. In other words, when channel information other than the first channel information is used as input to the first model, the second bit sequence cannot be used as a tag corresponding to the output of the first model or a reference for performance monitoring.

[0408] Optionally, the dataset includes at least one sample, each sample including corresponding channel information and modulation symbol sequences. The correspondence between the channel information and modulation symbol sequences included in each sample can be referenced to the correspondence between the first channel information and the first modulation symbol sequence.

[0409] Optionally, each sample also includes a bit sequence corresponding to the channel information. The correspondence between the channel information and the bit sequence included in each sample can be referenced to the correspondence between the first channel information and the second bit sequence.

[0410] For example, if the first device side can be replaced by the terminal side and the second device side can be replaced by the network side, then S1410 may specifically include: the terminal device receiving a dataset from the network device, for example, receiving it via an air interface; or, the terminal device obtaining a dataset from a server on the terminal side, which is received by the terminal side server from an intelligent network element or server on the network side, for example, receiving it via a wired network; or, the chip inside the terminal device, that is, the chip used in the terminal device, receiving the dataset from the network device.

[0411] In another example, if the first device side is replaced by the terminal side and the second device side is replaced by the network side, then S1410 may specifically include: the terminal side server receiving a dataset from the intelligent network element or server on the network side, for example, received through a wired network; or, the terminal side server obtaining a dataset from the terminal device, which is received by the terminal device from the network side, such as a network device, for example, received through an air interface.

[0412] The terminal-side servers mentioned in the embodiments of this application may include OTT system servers, cloud servers, servers provided by operators or third parties, etc. The network-side servers mentioned in the embodiments of this application may include cloud servers, servers provided by operators, etc.

[0413] S1420, the first device performs a first processing on the first channel information based on at least one model to obtain a second modulation symbol sequence.

[0414] The first process includes encoding and modulation.

[0415] For more details on S1420, please refer to S1210 in Method 1200 above.

[0416] S1430, the first device determines the performance monitoring result based on the first modulation symbol sequence and the second modulation symbol sequence.

[0417] The performance monitoring results determined by the first device side are the performance monitoring results of at least one model. The performance monitoring results of at least one model are used to determine whether the performance of at least one model meets the performance requirements.

[0418] For example, if the first device side is replaced by the terminal side, then S1430 may specifically include: the terminal device determining the performance monitoring result based on the first modulation symbol sequence and the second modulation symbol sequence; or, S1430 may specifically include: the server on the terminal side determining the performance monitoring result based on the first modulation symbol sequence and the second modulation symbol sequence.

[0419] As an example, the first device side can determine the performance monitoring results of at least one model during the process of applying at least one model for model inference.

[0420] In another example, the first device side can determine the performance monitoring results of at least one model during the training of at least one model.

[0421] Optionally, if the dataset also includes a second bit sequence, and during the process of the first device processing the first channel information based on at least one model, the first channel information is first encoded based on the first model to obtain the first bit sequence, and then the first bit sequence is modulated based on the second model to obtain the second modulation symbol sequence, then S1430 specifically includes: the first device determining the first monitoring result included in the performance monitoring result based on the first modulation symbol sequence and the second modulation symbol sequence; the first device determining the second monitoring result included in the performance monitoring result based on the first bit sequence and the second bit sequence.

[0422] The performance monitoring results determined by the first equipment side are described below.

[0423] For example, if the first device determines the performance monitoring result based on the first modulation symbol sequence and the second modulation symbol sequence, the performance monitoring result determined by the first device may include one or more of the following: the fourth MAE; the fourth MSE; the fourth SGCS; a weighted sum or a weighted average of at least two of the fourth MAE, the fourth MSE, or the fourth SGCS. In other words, the first device determines the performance monitoring result based on one or more of the fourth MAE, the fourth MSE, or the fourth SGCS.

[0424] The fourth MAE is the MAE between the first modulation symbol sequence and the second modulation symbol sequence.

[0425] For example, the first modulation symbol sequence includes T modulation symbols, represented as: g = [g1, g2, ..., g T The second modulation symbol sequence consists of T modulation symbols, represented as follows: Where T is a positive integer. The fourth MAE can be expressed as or satisfies the following formula (7).

[0426]

[0427] Among them, g real,t Let g represent the real part of the t-th modulation symbol in the first modulation symbol sequence. imag,t It represents the imaginary part of the t-th modulation symbol in the first modulation symbol sequence. Let represent the real part of the t-th modulation symbol in the second modulation symbol sequence. It represents the imaginary part of the t-th modulation symbol in the second modulation symbol sequence.

[0428] The fourth MSE is the MSE between the first modulation symbol sequence and the second modulation symbol sequence.

[0429] For example, the first modulation symbol sequence includes T modulation symbols, represented as: g = [g1, g2, ..., g T The second modulation symbol sequence consists of T modulation symbols, represented as follows: Where T is a positive integer. The fourth MSE can be expressed as or satisfies the following formula (8).

[0430]

[0431] The fourth SGCS is the SGCS between the first modulation symbol sequence and the second modulation symbol sequence.

[0432] For example, the first modulation symbol sequence includes T modulation symbols, represented as: g = [g1, g2, ..., g T The second modulation symbol sequence consists of T modulation symbols, represented as follows: Where T is a positive integer. The fourth SGCS can be expressed as or satisfies the following formula (9).

[0433]

[0434] Among them, g t =g real,t +j·g imag,t , represents the t-th modulation symbol in the first modulation symbol sequence. The k-th modulation symbol #2 is represented as: This represents the t-th modulation symbol in the second modulation symbol sequence.

[0435] For example, if the first device also determines the performance monitoring result based on the first bit sequence and the second bit sequence, the performance monitoring result determined by the first device may further include one or more of the following: the first BCE, the third MAE; the third MSE; a weighted sum or a weighted average of at least two of the first BCE, the third MAE, or the third MSE. In other words, the first device also determines the performance monitoring result based on one or more of the first BCE, the third MAE, or the third MSE.

[0436] For more details on the first BCE, the third MAE, and the third MSE, please refer to Method 1300 above.

[0437] In this embodiment, the second device can send first channel information and a first modulation symbol sequence to the first device. Then, the first device performs first processing on the first channel information based on at least one model and outputs a second modulation symbol sequence. The first device can then effectively monitor the performance of at least one model based on the first and second modulation symbol sequences.

[0438] Furthermore, the second device can also send a second bit sequence to the first device, thereby enabling the first device to effectively monitor the performance of the first model based on the second bit sequence and the first bit sequence output by the first model when at least one model includes the first model and the second model.

[0439] In one possible implementation, method 1400 may also include S1440.

[0440] S1440, the first device sends the performance monitoring results.

[0441] Correspondingly, the second device receives the performance monitoring results.

[0442] For more details on S1440, please refer to S1230 above.

[0443] In one possible implementation, method 1400 also includes S1450.

[0444] S1450, the second device sends the first information.

[0445] Accordingly, the first device receives the first information.

[0446] For more details on S1450, please refer to the section on S1240 above.

[0447] It should be understood that in the various embodiments shown above in conjunction with the accompanying drawings, the sequence number of each step does not imply the order of execution. The execution order of each step should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0448] In the various embodiments of this application, unless otherwise specified or in case of logical conflict, the terminology and / or descriptions between different embodiments are consistent and can be referenced by each other. Technical features in different embodiments can be combined to form new embodiments based on their inherent logical relationships.

[0449] In the above embodiments, exemplary descriptions are mainly based on devices in the current network architecture (such as terminal devices, network devices, OTT system servers, intelligent network elements, etc.). The specific form of the devices is not limited in the embodiments of this application. For example, devices that can achieve the same function in the future can also be applied to the methods provided in the embodiments of this application.

[0450] It is understood that the methods and operations implemented by devices (such as terminal devices, network devices, OTT system servers, and smart network elements) in the above-described method embodiments can also be implemented by components of the devices (such as chips or circuits).

[0451] The methods provided in the embodiments of this application have been described in detail above with reference to several accompanying drawings. The apparatus provided in the embodiments of this application will now be described with reference to the accompanying drawings.

[0452] Figure 15 and Figure 16 This is a schematic block diagram of possible apparatuses provided in the embodiments of this application. These apparatuses can be used to implement the functions on the terminal side or network side in the above method embodiments, and thus can also achieve the beneficial effects of the above method embodiments.

[0453] Figure 15 A schematic block diagram of the apparatus provided in the embodiments of this application. Figure 15 The device 1900 shown may include a processing module 1910 and a communication module 1920.

[0454] In one possible design, device 1900 can be used to achieve... Figures 12 to 14 The communication method shown in the embodiment is implemented by the first device side. For example, the processing module 1910 is used to implement the processing-related steps performed by the first device side in each method embodiment; the communication module 1920 is used to implement the sending and / or receiving steps performed by the first side in each method embodiment.

[0455] For example, the processing module 1910 can be used to perform a first processing on the first channel information based on at least one model to obtain N modulation symbols; N is a positive integer; the processing module 1910 can also be used to determine the performance monitoring result of at least one model based on the N modulation symbols and M known modulation symbols; M is a positive integer; wherein, the first processing includes coding and modulation. For example, coding may include source coding and channel coding, or joint source-channel coding.

[0456] Optionally, the communication module 1920 can be used to receive M known modulation symbols.

[0457] Optionally, the processing module 1910 can also be used to acquire first channel information through channel measurement or simulation.

[0458] Optionally, the communication module 1920 can also be used to receive first channel information and a second bit sequence.

[0459] Optionally, the processing module 1910 can be used to: determine a first monitoring result included in the performance monitoring result based on N modulation symbols and M known modulation symbols; and determine a second monitoring result included in the performance monitoring result based on a first bit sequence and a second bit sequence.

[0460] Optionally, the communication module 1920 can also be used to send performance monitoring results.

[0461] Optionally, the communication module 1920 can also be used to receive first information, which indicates that at least one model should be retrained.

[0462] For example, the communication module 1920 can be used to receive a dataset, which includes first channel information and a first modulation symbol sequence; the first modulation symbol sequence corresponds to the first channel information, and the dataset is used for performance monitoring of at least one model.

[0463] Optionally, the processing module 1910 can be used to: perform a first processing on the first channel information based on at least one model to obtain a second modulation symbol sequence; and determine the performance monitoring results of at least one model based on the first modulation symbol sequence and the second modulation symbol sequence; wherein the first processing includes coding and modulation.

[0464] Optionally, the dataset also includes a second bit sequence. Specifically, the processing module 1910 can be used to: determine a first monitoring result included in the performance monitoring results based on the first modulation symbol sequence and the second modulation symbol sequence; and determine a second monitoring result included in the performance monitoring results based on the first bit sequence and the second bit sequence.

[0465] Optionally, the communication module 1920 can also be used to send performance monitoring results.

[0466] Optionally, the communication module 1920 can also be used to receive first information, which indicates that at least one model should be retrained.

[0467] For a more detailed description of the processing module 1910 and communication module 1920 mentioned above, please refer to [link / reference]. Figures 12 to 14 The relevant descriptions in the method embodiments shown are directly obtained and will not be repeated here.

[0468] In another possible design, device 1900 can be used to achieve Figures 12 to 14 The communication method shown in the embodiment is implemented by the second device side. For example, the processing module 1910 is used to implement the processing-related steps performed by the second device side in each method embodiment; the communication module 1920 is used to implement the sending and / or receiving steps performed by the second device side in each method embodiment.

[0469] For example, the communication module 1920 can be used to send M known modulation symbols; the M known modulation symbols are used for performance monitoring of at least one model, and the at least one model is used to perform a first processing on the input first channel information to obtain N modulation symbols; the N modulation symbols and the M known modulation symbols are used to determine the performance monitoring result of at least one model; the first processing includes encoding and modulation; M and N are both positive integers.

[0470] Optionally, the communication module 1920 can also be used to transmit first channel information and a second bit sequence, the first bit sequence and the second bit sequence being used to determine the second monitoring result included in the performance monitoring result; N modulation symbols and M known modulation symbols are used to determine the first monitoring result included in the performance monitoring result.

[0471] Optionally, the communication module 1920 can also be used to: receive performance monitoring results; and, if the performance monitoring results indicate that the performance of at least one model does not meet the performance requirements, send first information, the first information being used to instruct the at least one model to be retrained.

[0472] For example, the communication module 1920 can be used to transmit a dataset, which includes first channel information and a first modulation symbol sequence; the first modulation symbol sequence corresponds to the first channel information, and the dataset is used for performance monitoring of at least one model.

[0473] Optionally, the first channel information and the first modulation symbol sequence are used for performance monitoring of at least one model, and the at least one model is used to perform a first process on the input first channel information to obtain a second modulation symbol sequence; the first modulation symbol sequence and the second modulation symbol sequence are used to determine the performance monitoring result of at least one model; the first process includes coding and modulation. For example, coding may include source coding and channel coding, or joint source-channel coding.

[0474] Optionally, the communication module 1920 can also be used to send a second bit sequence, the first bit sequence and the second bit sequence being used to determine the second monitoring result included in the performance monitoring result; the first modulation symbol sequence and the second modulation symbol sequence being used to determine the first monitoring result included in the performance monitoring result.

[0475] Optionally, the communication module 1920 can also be used to: receive performance monitoring results; and, if the performance monitoring results indicate that the performance of at least one model does not meet the performance requirements, send first information, the first information being used to instruct the at least one model to be retrained.

[0476] For a more detailed description of the processing module 1910 and communication module 1920 mentioned above, please refer to [link / reference]. Figures 12 to 14 The relevant descriptions in the method embodiments shown are directly obtained and will not be repeated here.

[0477] It should be noted that the communication module can also be called a transceiver module, transceiver unit, transceiver, transceiver device, or transceiver apparatus, etc. The processing module can also be called a processor, processing board, processing unit, or processing apparatus, etc. Optionally, the communication module is used to perform the sending and receiving operations on the terminal side or network side in the above method. The device in the communication module that implements the receiving function can be considered as the receiving module, and the device in the communication module that implements the sending function can be considered as the sending module; that is, the communication module can include both a receiving module and a sending module.

[0478] It should also be noted that, in one possible design, the aforementioned processing module and / or communication module can be implemented through virtual modules. For example, the processing module can be implemented through software functional units or virtual devices, and the communication module can be implemented through software functions or virtual devices. In another possible design, the processing module or communication module can also be implemented through physical devices. For example, if the device is implemented using a chip / chip circuit, the communication module can be an input / output circuit and / or a communication interface, performing input operations (corresponding to the aforementioned receiving operation) and output operations (corresponding to the aforementioned sending operation); the processing module can be an integrated processor, a microprocessor, or an integrated circuit.

[0479] The module division in this embodiment is illustrative and represents only one logical functional division; in actual implementation, other division methods may be used. Furthermore, the functional modules in the various examples of this embodiment can be integrated into a single processor, exist as separate physical entities, or be integrated into a single module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0480] Figure 16 This is a schematic diagram of the structure of a communication device provided in yet another embodiment of this application. (See attached diagram.) Figure 16 As shown, the device 2000 includes a processing circuit 2010 and a communication circuit 2020. The processing circuit 2010 and the communication circuit 2020 are coupled to each other.

[0481] It can be understood that the processing circuit 2010 can be one or more processors, or it can be all or part of the processing functions of one or more processors.

[0482] It is understandable that the communication circuit 2020 can be a transceiver or an input / output interface.

[0483] Optionally, the device 2000 may further include a memory 2030 for storing instructions executed by the processing circuit 2010, or storing input data required for the running instructions of the processing circuit 2010, or storing data generated after the running instructions of the processing circuit 2010.

[0484] It is understood that the memory 2030 may be located outside the processing circuit 2010 or inside the processing circuit 2010.

[0485] As an example, the processing circuit 2010 is used to implement the functions of the processing module 1910, and the communication circuit 2020 is used to implement the functions of the communication module 1920.

[0486] As an example, device 2000 can be a communication device or a chip used in a communication device.

[0487] When device 2000 is a communication device, the communication circuit can be a transceiver; when device 2000 is a chip, the communication circuit can be an input / output circuit, a bus, pins, or other types of communication interfaces. The input circuit in the input / output circuit can be used for receiving, and the output interface can be used for transmitting.

[0488] In one possible implementation, the apparatus 2000 is used to implement the various processes and steps corresponding to the first device side in the above method embodiments. In another possible implementation, the apparatus 2000 is used to implement the various processes and steps corresponding to the second device side in the above method embodiments.

[0489] It is understood that device 2000 can specifically be the first device or the second device in the above embodiments, or it can be a chip or a chip system. Correspondingly, the communication circuit can be the interface circuit of the chip, or an input / output circuit, which is not limited here.

[0490] When the aforementioned communication device is a chip or OTT device applied to the terminal side, the terminal-side chip or OTT device implements the terminal-side functions described in the above method embodiments, such as implementing terminal-side processing functions. The terminal-side chip or OTT device receiving information from the network side can be understood as the information being first received by other modules on the terminal side (such as a radio frequency module or antenna), and then sent by these modules to the terminal-side chip or OTT device. The terminal-side chip or OTT device sending information to the network side can be understood as the information being first sent by the terminal-side chip or OTT device to other modules on the terminal side (such as a radio frequency module or antenna), and then sent by these modules to the network side.

[0491] When the aforementioned communication device is a chip or OTT device applied to the network side, the network-side chip or OTT device implements the network-side functions described in the above method embodiments, for example, implementing network-side processing functions. The network-side chip or OTT device receiving information from the terminal side can be understood as the information being first received by other modules on the network side (such as radio frequency modules or antennas), and then sent by these modules to the network-side chip or OTT device. The network-side chip or OTT device sending information to the first device can be understood as the information being first sent by the network-side chip or OTT device to other modules on the network side (such as radio frequency modules or antennas), and then sent by these modules to the terminal side.

[0492] This application also provides a computer program product that, when run on a processor, can implement the communication method executed by the first device side or the communication method executed by the second device side in the above method embodiments.

[0493] This application also provides a computer-readable storage medium containing computer instructions that, when executed on a processor, can implement the communication method executed by the first device side or the communication method executed by the second device side in the above method embodiments.

[0494] This application also provides a communication system, including the aforementioned first device side and second device side.

[0495] It is understood that the processor in the embodiments of this application may be any of the following devices or all or part of the circuitry used for processing functions: a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), processors for AI, field-programmable gate arrays (FPGAs), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. A general-purpose processor may be a microprocessor or any conventional processor.

[0496] The processor used for AI can be one or more of the following: graphics processing unit (GPU), neural processing unit (NPU), tensor processing unit (TPU), and data processing unit (DPU).

[0497] For example, one possible implementation of a processor for AI could be... Figure 17 The AI ​​processor 2100 shown.

[0498] like Figure 17 As shown, the AI ​​processor 2100 may include one or more of the following: AI core, digital vision pre-processing (DVPP) module, task scheduler (TS), L3 cache, AI CPU, control CPU, L2 cache, universal serial bus (USB) interface, network card, peripheral component interconnect express (PCIe) interface (PCIe is a high-speed serial computer expansion bus standard), double data rate (DDR) / high bandwidth memory (HBM) interface, generation purpose input / output (GPIO) / inter-integrated circuit (I2C) bus, etc.

[0499] The terms “unit”, “module”, etc., used in this specification may be used to refer to computer-related entities, hardware, firmware, combinations of hardware and software, software, or software in execution.

[0500] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0501] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0502] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0503] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0504] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0505] In the above embodiments, the functions of each functional unit can be implemented entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. This computer program product includes one or more computer instructions (programs). When the computer program instructions (programs) are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., digital video discs (DVDs)), or semiconductor media (e.g., solid-state disks (SSDs)).

[0506] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0507] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A communication method, characterized in that, include: Based on at least one model, the first channel information is processed to obtain N modulation symbols; N is a positive integer; Based on the N modulation symbols and M known modulation symbols, determine the performance monitoring results of at least one model; M is a positive integer; The first process includes encoding and modulation.

2. The method according to claim 1, characterized in that, The at least one model includes a first model and a second model; the first model is used to encode the first channel information to obtain a first bit sequence; the second model is used to modulate the first bit sequence to obtain the N modulation symbols.

3. The method according to claim 1 or 2, characterized in that, The M known modulation symbols are predefined modulation symbols; or, The method further includes: Receive the M known modulation symbols.

4. The method according to any one of claims 1 to 3, characterized in that, The performance monitoring results are related to one or more of the following: first mean absolute error (MAE); first mean square error (MSE); first squared generalized cosine similarity (SGCS); second MAE; second MSE; second SGCS; a weighted sum or weighted average of at least two of the first MAE, first MSE, first SGCS, second MAE, second MSE, or second SGCS. The first MAE is determined based on the MAE of the first modulation symbol and each of the K second modulation symbols; The first MSE is determined based on the first modulation symbol and the MSE of each of the K second modulation symbols; The first SGCS is determined based on the SGCS of the first modulation symbol and each of the K second modulation symbols; The second MAE is determined based on the MAE of the third modulation symbol and each of the K fourth modulation symbols; The second MSE is determined based on the MSE of the third modulation symbol and each of the K fourth modulation symbols; The second SGCS is determined based on the SGCS of the third modulation symbol and each of the K fourth modulation symbols; Where K is a positive integer less than or equal to M; The first modulation symbol belongs to the N modulation symbols, and the K second modulation symbols belong to the M known modulation symbols; the geometric distance between the second modulation symbol and the first modulation symbol is less than or equal to the geometric distance between any of the M known modulation symbols other than the K second modulation symbols and the first modulation symbol; The third modulation symbol belongs to the N modulation symbols, and the K fourth modulation symbols belong to the M known modulation symbols; the geometric distance between the fourth modulation symbol and the third modulation symbol is less than or equal to the geometric distance between any of the M known modulation symbols other than the K fourth modulation symbols and the third modulation symbol.

5. The method according to any one of claims 1 to 4, characterized in that, Before performing the first processing on the first channel information based on the at least one model, the method further includes: The first channel information is obtained through channel measurement or simulation.

6. The method according to claim 2, characterized in that, Before performing the first processing on the first channel information based on the at least one model, the method further includes: Receive the first channel information and the second bit sequence from another device, wherein the second bit sequence is obtained by the other device encoding the first channel information; The determination of the performance monitoring results of at least one model based on the N modulation symbols and M known modulation symbols includes: Based on the N modulation symbols and the M known modulation symbols, the first monitoring result is determined to be included in the performance monitoring result; Based on the first bit sequence and the second bit sequence, the second monitoring result is determined to be included in the performance monitoring result.

7. The method according to claim 6, characterized in that, The second monitoring result includes one or more of the following: first binary cross-entropy (BCE); third MSE; third MAE; a weighted sum or weighted average of at least two of the first BCE, third MSE, and third MAE. The first BCE is determined based on the probability of each bit in the first bit sequence being 1 and the second bit sequence; The third MSE is the MSE between the first bit sequence and the second bit sequence; The third MAE is the MAE between the first bit sequence and the second bit sequence.

8. The method according to any one of claims 1 to 7, characterized in that, The method further includes: Send the performance monitoring results.

9. The method according to claim 8, characterized in that, The method further includes: Receive first information, the first information being used to instruct the at least one model to be retrained.

10. The method according to any one of claims 1 to 9, characterized in that, Determining the performance monitoring results of the at least one model includes: During the process of applying the at least one model to perform model inference, the performance monitoring results of the at least one model are determined.

11. The method according to any one of claims 1 to 9, characterized in that, Determining the performance monitoring results of the at least one model includes: During the training of the at least one model, the performance monitoring results of the at least one model are determined.

12. A communication method, characterized in that, include: Send M known modulation symbols; The M known modulation symbols are used for performance monitoring of at least one model, and the at least one model is used to perform a first process on the input first channel information to obtain N modulation symbols; the N modulation symbols and the M known modulation symbols are used to determine the performance monitoring result of the at least one model; the first process includes encoding and modulation; M and N are both positive integers.

13. The method according to claim 12, characterized in that, The at least one model includes a first model and a second model; the first model is used to encode the first channel information to obtain a first bit sequence; the second model is used to modulate the first bit sequence to obtain the N modulation symbols; The N modulation symbols and the M known modulation symbols are used to determine the first monitoring result included in the performance monitoring results.

14. The method according to claim 13, characterized in that, The method further includes: The first channel information and the second bit sequence are transmitted, wherein the second bit sequence is obtained by encoding the first channel information, and the first bit sequence and the second bit sequence are used to determine the second monitoring result included in the performance monitoring result.

15. The method according to any one of claims 12 to 14, characterized in that, The method further includes: Receive performance monitoring results; If the performance monitoring results indicate that the performance of at least one model does not meet the performance requirements, a first message is sent, which is used to instruct the at least one model to be retrained.

16. A communication device, characterized in that, Includes modules or units for performing the method according to any one of claims 1 to 15.

Images