Communication method and apparatus

Abstract

The present disclosure provides a communication method and apparatus. The method comprises: receiving first indicating information, the first indicating information indicating a first model and / or a first function, and the first model and / or the first function being configured for channel state information processing; and sending second indicating information, the second indicating information indicating a first data processing solution, the first data processing solution being determined according to the first indicating information, and the first data processing solution being used for determining enhanced data. The described solution allows enhanced data to be determined by means of a first data processing solution, thereby reducing data acquisition costs and data transmission overhead.

Description

Communication methods and devices

[0001] This application claims priority to Chinese Patent Application No. 202411926175.5, filed on December 23, 2024, entitled "Communication Method and Apparatus", the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of communication technology, and more specifically, to communication methods and apparatus. Background Technology

[0003] In the process of terminal devices feeding back channel state information (CSI) to network devices, to reduce the overhead of CSI feedback, the terminal devices can compress the information to be fed back using models, such as artificial intelligence (AI) models. Correspondingly, the network devices can decompress / restore the compressed information fed back by the terminal devices using models. Alternatively, the terminal devices can predict the CSI at future times based on current or historical CSI using models.

[0004] Models are obtained through training, which requires acquiring a training dataset. However, the cost / overhead of collecting and / or transmitting data in the training dataset is significant. Therefore, reducing the cost / overhead of data collection and transmission is a pressing issue that needs to be addressed. Summary of the Invention

[0005] This application provides a communication method and apparatus that can reduce the cost / overhead of data collection and transmission.

[0006] Firstly, a communication method is provided. This method can be executed by a first communication device, which can refer to a device on the terminal device side or network device side (e.g., a terminal device or a network device), or a component within the device (e.g., a communication module, processor, circuit, chip, or chip system), or a logic module or software capable of implementing all or part of the functions of the communication device. The terminal device side can include at least one of a terminal device or an artificial intelligence (AI) entity on the terminal device side. The AI ​​entity on the terminal device side can be the terminal device itself or an AI entity serving the terminal device, such as a server, such as an over-the-top (OTT) server or a cloud server. The network device side can include at least one of a network device or an AI entity on the network device side. The AI ​​entity on the network device side can be the network device itself or an AI entity serving the network device, such as a radio access network (RAN) intelligent controller (RIC), operation administration and maintenance (OAM), or a server, such as an OTT server or a cloud server. Communication between servers can be achieved through communication links between terminal devices and network devices, through forwarding via other communication devices outside the server, or through wired links. For ease of description, the following explanation uses the first communication device executing this method as an example.

[0007] The method includes: receiving first indication information, the first indication information indicating a first model and / or a first function, the first model and / or the first function being used for channel state information (CSI) processing; and sending second indication information, the second indication information indicating a first data processing scheme, the first data processing scheme being determined based on the first indication information, the first data processing scheme being used to determine augmented data, the augmented data being used to train at least one model corresponding to the first model and / or the first function.

[0008] Based on the solution provided in the embodiments of this application, by sending instruction information indicating a first data processing scheme, which is determined according to a first model and / or a first function, on the one hand, it is possible to enable the device receiving the instruction information to determine augmented data through the first data processing scheme, thereby reducing the cost of data collection and the overhead of data transmission; on the other hand, it is possible to determine a data processing scheme suitable for the first model and / or the first function, thereby improving the gain brought by using the data augmentation method and ensuring the performance of the model obtained after training at least one model corresponding to the first model and / or the first function.

[0009] In some possible implementations, the method further includes: determining a first data processing scheme based on the first instruction information.

[0010] In some possible implementations, the first data processing scheme is determined based on first indication information, including: the first data processing scheme is determined based on a first model and / or a first function and a first correspondence, the first correspondence indicating the relationship between at least one model and / or at least one function and at least one data processing scheme for determining augmented data, the at least one model and / or at least one function including the first model and / or the first function, and the at least one data processing scheme including the first data processing scheme corresponding to the first model and / or the first function.

[0011] In some possible implementations, the first data processing scheme is determined based on the first indication information, including: the first data processing scheme is determined based on M and a second correspondence, the second correspondence indicating the relationship between multiple numerical ranges and at least one data processing scheme for determining the augmented data, the multiple numerical ranges including a first numerical range, the value of M being within the first numerical range, the at least one data processing scheme including a first data processing scheme corresponding to the first numerical range, and M being a positive integer.

[0012] In some possible implementations, the first data processing scheme is determined based on first indication information, including: the first data processing scheme is determined based on a first model and / or a first function, M and a third correspondence, the third correspondence indicating the relationship between multiple numerical ranges and at least one model and / or at least one function and at least one data processing scheme for determining augmented data, the multiple numerical ranges including a first numerical range, the value of M being within the first numerical range, the at least one model and / or at least one function including the first model and / or the first function, the at least one data processing scheme including a first data processing scheme corresponding to the first numerical range and the first model and / or the first function, and M being a positive integer.

[0013] Based on the solution provided in the embodiments of this application, by determining a first data processing scheme applicable to the first model and / or the first function based on different correspondences (first correspondence, second correspondence, or third correspondence), it is possible to determine a data processing scheme applicable to different models / functions / tasks / scenarios, thereby improving the gains brought by using data augmentation methods.

[0014] In some possible implementations, M is determined based on a first model and / or a first function; and / or, M is indicated by first instruction information.

[0015] In some possible implementations, the value of M is within a first numerical range, including at least one of the following: the value of M is between a first threshold and a second threshold, the first threshold and the second threshold are not equal, and the first threshold and / or the second threshold is greater than 0; the value of M is greater than or equal to the first threshold, and the first threshold is greater than 0; or, the value of M is less than or equal to the second threshold, and the second threshold is greater than 0.

[0016] For example, the numerical range is determined based on a first threshold and / or a second threshold.

[0017] In some possible implementations, the first model and / or the first function is used for CSI processing, including at least one of the following: the first model and / or the first function is used for CSI compression; the first model and / or the first function is used for CSI decompression; or, the first model and / or the first function is used to determine the CSI of at least one time unit after at least one time unit based on the CSI of at least one time unit.

[0018] For example, the first model and / or the first function is used to predict the CSI of one or more time units after the at least one time unit based on the CSI of at least one time unit.

[0019] In some possible implementations, the first data processing scheme is used to determine augmented data, including any of the following: the augmented data is obtained by phase rotation of the original training data used to obtain the augmented data; the augmented data is obtained by phase rotation of the channel matrices of different time units in the same time series of the original training data used to obtain the augmented data by the same angle; or, the augmented data is obtained by phase rotation of the channel matrices of different time units in the same time series of the original training data used to obtain the augmented data by different angles.

[0020] In some possible implementations, the first data processing scheme is used to determine the augmented data, including any of the following: H′=H*e jθ,θ∈(0,2π), where H represents the original training data used to obtain augmented data, and H′ represents the augmented data; [H′1,H′2,H′3]=[H1,H2,H3]*e jθ ,θ∈(0,2π), where [H1,H2,H3] represents the original training data used to obtain augmented data, and [H′1,H′2,H′3] represents the augmented data; or, Where [H1,H2,H3] represents the original training data used to obtain augmented data, and [H′1,H′2,H′3] represents the augmented data.

[0021] In some possible implementations, the method further includes: receiving first request information, the first request information being used to request at least one of a training dataset, model parameters of a second model, or model parameters of at least one model corresponding to a second function, wherein the training dataset is the training dataset of the first model and / or at least one model corresponding to the first function, and the second model and / or the second function is used to train the first model and / or at least one model corresponding to the first function; and sending third indication information, the third indication information being used to indicate at least one of the training dataset, model parameters of the second model, or model parameters of at least one model corresponding to the second function.

[0022] For example, the second model may be different from the first model; or, the second model may belong to the same function / task / scenario as the first model, or have the same identity (ID); or, the second function may be the same as the first function, or have the same ID, or at least one model corresponding to the second function and at least one model corresponding to the first function may be completely the same or partially the same or completely different.

[0023] Based on the solution provided in the embodiments of this application, by sending instruction information indicating at least one of the following: the training dataset, the model parameters of the second model, or the model parameters of at least one model corresponding to the second function, on the one hand, it is possible to instruct the device receiving the instruction information to train the first model and / or at least one model corresponding to the first function using the training dataset; on the other hand, when the first model and / or at least one model corresponding to the first function are dual-end models, it is possible to achieve the docking and training of dual-end models; and furthermore, it is possible to instruct the device receiving the instruction information to train the first model and / or at least one model corresponding to the first function using the model parameters of the second model and / or the model parameters of at least one model corresponding to the second function, thereby improving the performance of the model obtained after training.

[0024] In some possible implementations, the model parameters of the second model and / or the model parameters of at least one model corresponding to the second function are used to adjust the parameters of the first model and / or at least one model corresponding to the first function.

[0025] In some possible implementations, the first communication device is a terminal device, a component in the terminal device, a logic module or software that can realize all or part of the functions of the communication device, or an AI entity on the terminal device side; the second communication device is a network device, a component in the network device, a logic module or software that can realize all or part of the functions of the communication device, or an AI entity on the network device side.

[0026] In some other possible implementations, the second communication device is a terminal device, a component in the terminal device, a logic module or software that can realize all or part of the functions of the communication device, or an AI entity on the terminal device side; the first communication device is a network device, a component in the network device, a logic module or software that can realize all or part of the functions of the communication device, or an AI entity on the network device side.

[0027] Secondly, a communication method is provided. This method can be executed by a second communication device, which can refer to a device on the terminal device side or network device side (e.g., a terminal device or a network device), or a component within the device (e.g., a communication module, processor, circuit, chip, or chip system), or a logic module or software capable of implementing all or part of the communication device's functions. The terminal device side can include at least one of a terminal device or an AI entity on the terminal device side. The AI ​​entity on the terminal device side can be the terminal device itself or an AI entity serving the terminal device, such as a server, such as an OTT server or a cloud server. The network device side can include at least one of a network device or an AI entity on the network device side. The AI ​​entity on the network device side can be the network device itself or an AI entity serving the network device, such as a RIC, OAM, or a server, such as an OTT server or a cloud server. Communication between servers can be achieved through a communication link between the terminal device and the network device, through forwarding via other communication devices outside the server, or through a wired link. For ease of description, the following explanation uses the execution of this method by a second communication device as an example.

[0028] The method includes: sending first indication information, the first indication information indicating a first model and / or a first function, the first model and / or the first function being used for channel state information (CSI) processing; receiving second indication information, the second indication information indicating a first data processing scheme, the first data processing scheme being determined based on the first indication information, the first data processing scheme being used to determine augmented data, the augmented data being used to train at least one model corresponding to the first model and / or the first function.

[0029] Based on the solution provided in the embodiments of this application, by receiving instruction information indicating a first data processing scheme, which is determined according to a first model and / or a first function, on the one hand, it is possible to determine augmented data through the first data processing scheme, thereby reducing the cost of data collection and the overhead of data transmission; on the other hand, it is possible to obtain the data processing scheme suitable for the first model and / or the first function determined by the device sending the instruction information through the instruction information, thereby improving the gain brought by the data augmentation method and ensuring the performance of the model obtained after training at least one model corresponding to the first model and / or the first function.

[0030] In some possible implementations, the method further includes: sending a first request message, the first request message being used to request at least one of a training dataset, model parameters of a second model, or model parameters of at least one model corresponding to a second function, wherein the training dataset is the training dataset of the first model and / or at least one model corresponding to the first function, and the second model and / or the second function is used to train the first model and / or at least one model corresponding to the first function; and receiving a third instruction message, the third instruction message being used to indicate at least one of the training dataset, model parameters of the second model, or model parameters of at least one model corresponding to the second function.

[0031] In some possible implementations, the method further includes at least one of the following: training a first model based on augmented data; training at least one model corresponding to a first function based on augmented data; training a first model based on a training dataset; training at least one model corresponding to a first function based on a training dataset; training a first model based on model parameters of a second model; training at least one model corresponding to a first function based on model parameters of the second model; training a first model based on model parameters of at least one model corresponding to a second function; or, training at least one model corresponding to a first function based on model parameters of at least one model corresponding to a second function.

[0032] For example, the training dataset, the model parameters of the second model, or the model parameters of at least one model corresponding to the second function are indicated by the third indication information.

[0033] Based on the solution provided in the embodiments of this application, by training the first model and / or at least one model corresponding to the first function through at least one of the augmented data, training dataset, model parameters of the second model, or model parameters of at least one model corresponding to the second function, the performance of the model obtained after training the first model and / or at least one model corresponding to the first function can be improved.

[0034] The second aspect and some of its implementation methods and their beneficial effects can be referred to in the relevant description of the first aspect, and will not be repeated here.

[0035] Thirdly, a communication device is provided for performing the method provided in the first aspect. The communication device can be a first communication device, or a component of the first communication device (e.g., a processor, chip, or chip system, such as circuitry or a chip responsible for communication functions in the first communication device (e.g., a modem chip, a baseband chip, or a system-on-chip (SoC) chip containing a modem core or a system-in-package (SIP) chip)). Alternatively, it can be a logic module or software capable of implementing all or part of the functions of the first communication device. For ease of description, the following description will use this device as the first device and a device that works with it to transmit / request training datasets as the second device.

[0036] The device includes: a transceiver unit, which receives first indication information, the first indication information indicating a first model and / or a first function, the first model and / or the first function being used for channel state information (CSI) processing; the transceiver unit is also used to transmit second indication information, the second indication information indicating a first data processing scheme, the first data processing scheme being determined based on the first indication information, the first data processing scheme being used to determine augmented data, the augmented data being used to train at least one model corresponding to the first model and / or the first function.

[0037] In some possible implementations, the apparatus further includes a processing unit for determining a first data processing scheme based on the first instruction information.

[0038] In some possible implementations, the transceiver unit is further configured to receive first request information, which requests at least one of the following: a training dataset, model parameters of a second model, or model parameters of at least one model corresponding to a second function. The training dataset is the training dataset of the first model and / or at least one model corresponding to the first function. The second model and / or the second function is used to train the first model and / or at least one model corresponding to the first function. The transceiver unit is further configured to send third indication information, which indicates at least one of the following: the training dataset, model parameters of the second model, or model parameters of at least one model corresponding to the second function.

[0039] In some possible implementations, the processing unit includes a processor.

[0040] In some possible implementations, the transceiver unit includes a transceiver, or an input / output interface. Optionally, the input / output interface can be input / output circuitry.

[0041] In some other possible implementations, the communication device may be a chip, chip system, or circuit, and the transceiver unit may be an input / output interface, interface circuit, output circuit, input circuit, pin, or related circuit on the chip, chip system, or circuit; the processing unit may be at least one processor, processing circuit, or logic circuit.

[0042] The third aspect and some of its implementation methods and their beneficial effects can be referred to in the relevant description of the first aspect, and will not be repeated here.

[0043] Fourthly, a communication device is provided for performing the method provided in the second aspect. The communication device can be a second communication device, or a component of a second communication device (e.g., a processor, chip, or chip system, such as circuitry or a chip responsible for communication functions in the second communication device (e.g., a modem chip, a baseband chip, or a system-on-chip (SoC) chip containing a modem core or a system-in-package (SIP) chip)). Alternatively, it can be a logic module or software capable of implementing all or part of the functions of the second communication device. For ease of description, the following description will use the device as the second device and the device that jointly implements the transmission / request of training datasets as the first device.

[0044] The device includes: a transceiver unit, which is configured to transmit first indication information, the first indication information indicating a first model and / or a first function, the first model and / or the first function being used for channel state information (CSI) processing; the transceiver unit is also configured to receive second indication information, the second indication information indicating a first data processing scheme, the first data processing scheme being determined based on the first indication information, the first data processing scheme being used to determine augmented data, the augmented data being used to train at least one model corresponding to the first model and / or the first function.

[0045] In some possible implementations, the transceiver unit is further configured to send a first request message, which requests at least one of the following: a training dataset, model parameters of a second model, or model parameters of at least one model corresponding to a second function. The training dataset is the training dataset of the first model and / or at least one model corresponding to the first function. The second model and / or the second function is used to train the first model and / or at least one model corresponding to the first function. The transceiver unit is further configured to receive a third indication message, which indicates at least one of the following: the training dataset, model parameters of the second model, or model parameters of at least one model corresponding to the second function.

[0046] In some possible implementations, the apparatus further includes a processing unit for at least one of the following: training a first model based on augmented data; training at least one model corresponding to a first function based on augmented data; training a first model based on a training dataset; training at least one model corresponding to a first function based on a training dataset; training a first model based on model parameters of a second model; training at least one model corresponding to a first function based on model parameters of the second model; training a first model based on model parameters of at least one model corresponding to a second function; or training at least one model corresponding to a first function based on model parameters of at least one model corresponding to a second function.

[0047] In some possible implementations, the processing unit includes a processor.

[0048] In some possible implementations, the transceiver unit includes a transceiver, or an input / output interface. Optionally, the input / output interface can be input / output circuitry.

[0049] In some other possible implementations, the communication device may be a chip, chip system, or circuit, and the transceiver unit may be an input / output interface, interface circuit, output circuit, input circuit, pin, or related circuit on the chip, chip system, or circuit; the processing unit may be at least one processor, processing circuit, or logic circuit.

[0050] The fourth aspect and some of its implementation methods and their beneficial effects can be referred to the relevant descriptions in the second aspect, and will not be repeated here.

[0051] Fifthly, this application provides a processor for executing the method provided by any implementation of the first and second aspects described above.

[0052] Unless otherwise specified, or if it does not contradict its actual function or internal logic in the relevant description, the transmission and acquisition / reception operations involved in the processor can be understood as processor output and reception, input and other operations, or as transmission and reception operations performed by radio frequency circuits and antennas. This application does not limit them in this regard.

[0053] A sixth aspect provides a computer-readable storage medium storing program code for execution by a device, the program code including a method for performing any of the implementations of the first and second aspects described above.

[0054] In a seventh aspect, a computer program product comprising instructions is provided, which, when run on a computer, causes the computer to perform the method provided by any implementation of the first and second aspects described above.

[0055] Eighthly, a chip is provided, the chip including one or more processors and a communication interface, wherein the processor reads a computer program or instructions stored in a memory through the communication interface and executes the method provided by any implementation of the first and second aspects described above.

[0056] Optionally, as one implementation, the chip also includes a memory storing computer programs or instructions, and a processor for executing the computer programs or instructions stored in the memory. When the computer programs or instructions are executed, the processor is used to perform the methods provided by any of the implementations of the first and second aspects described above.

[0057] Ninth aspect, a communication system is provided, including a first communication device for performing any implementation of the first aspect and a second communication device for performing any implementation of the second aspect.

[0058] The beneficial effects of the fifth to ninth aspects mentioned above can be referred to the first to fourth aspects mentioned above and any possible implementation methods, which will not be elaborated here. Attached Figure Description

[0059] Figure 1 is a schematic diagram of a communication system applicable to this application.

[0060] Figure 2 is a schematic diagram of another communication system applicable to this application.

[0061] Figure 3 is a schematic diagram of a possible application framework in a communication system applicable to embodiments of this application.

[0062] Figure 4 is a schematic diagram of another possible application framework in a communication system applicable to embodiments of this application.

[0063] Figure 5 is a schematic diagram of AI CSI feedback applicable to an embodiment of this application.

[0064] Figure 6 is a flowchart illustrating a communication method provided in an embodiment of this application.

[0065] Figure 7 is a schematic diagram of another communication method applicable to embodiments of this application.

[0066] Figure 8 is a schematic diagram of a CSI model / function / task / scenario application applicable to an embodiment of this application.

[0067] Figure 9 is a schematic diagram of another communication method provided in an embodiment of this application.

[0068] Figure 10 is a schematic diagram of another communication method provided in an embodiment of this application.

[0069] Figure 11 is a schematic block diagram of a communication device provided in an embodiment of this application.

[0070] Figure 12 is a schematic block diagram of another communication device provided in an embodiment of this application. Detailed Implementation

[0071] To facilitate understanding of the above embodiments provided in this application, the following points are made:

[0072] 1) In this application, unless otherwise specified or in case of logical conflict, the terms and / or descriptions of different embodiments are consistent and can be referenced by each other. The technical features of different embodiments can be combined to form new embodiments according to their inherent logical relationship.

[0073] 2) 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 exists alone, A and B exist simultaneously, or B exists alone, where A and B can be singular or plural. In the textual description of this application, the character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "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, and c can mean: a, or, b, or, c, or, a and b, or, a and c, or, b and c, or, a, b, and c. Here, a, b, and c can each be single or multiple.

[0074] 3) In this application, the terms "first," "second," and various numerical designations (e.g., #1, #2, etc.) indicate distinctions made for ease of description and are not intended to limit the scope of the embodiments of this application. For example, they may distinguish different messages, rather than describing a specific order or sequence. It should be understood that such descriptions can be interchanged where appropriate to describe solutions other than those in the embodiments of this application.

[0075] 4) In this application, descriptions such as “when…”, “under the circumstances of…” and “if” all refer to the fact that the device will make corresponding processing under certain objective circumstances. They are not time limits, nor do they require the device to make a judgment action when it is implemented, nor do they mean that there are other limitations.

[0076] 5) In this application, "instruction" or "for instruction" can include both direct and indirect instruction. When describing an instruction as being used to instruct A, it may include whether the instruction directly instructs A or indirectly instructs A, but does not necessarily mean that the instruction carries A.

[0077] The indication methods involved in the embodiments of this application should be understood to cover various methods that enable the party to be indicated to obtain the information to be indicated. The information to be indicated can be sent as a whole or divided into multiple sub-information and sent separately. Moreover, the sending period and / or sending time of these sub-information can be the same or different. This application does not limit the sending method, for example.

[0078] The "instruction information" in the embodiments of this application can be an explicit instruction, that is, a direct instruction through signaling, or an instruction obtained by combining other rules or parameters with the parameters indicated by the signaling, or by deduction. It can also be an implicit instruction, that is, an instruction obtained based on rules or relationships, or based on other parameters, or by deduction. This application does not specifically limit it in this regard.

[0079] 6) In this application, "protocol" can refer to a standard protocol in the field of communications, such as 5th generation (5G) protocols, new radio (NR) protocols, and related protocols applied to future communication systems. This application does not limit this term. "Predefined" can include predefined terms, such as protocol definitions. "Preconfiguration" can be implemented by pre-storing corresponding codes, tables, or other means that can be used to indicate relevant information in the device. This application does not limit the implementation method, for example.

[0080] 7) In this application, "communication" can also be described as "data transmission", "information transmission", "data processing", etc. "Transmission" includes "sending" and / or "receiving".

[0081] 8) In this application, "sending information to XX (device)" can be understood as the destination of the information being that device. This can include sending information directly or indirectly to that device. "Receiving information from XX (device), or receiving information from XX (device)" can be understood as the source of the information being that device, and can include receiving information directly or indirectly from that device. Information may undergo necessary processing between the source and destination, such as format changes, but the destination can understand the valid information from the source. Similar expressions in this application can be interpreted similarly, and will not be elaborated further here.

[0082] 9) In this application, the terms "exemplarily," "for example," "for instance," "as an example," etc., are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as an "example" in this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the term "example" is intended to present concepts in a concrete manner. In the embodiments of this application, "of," "corresponding, relevant," and "corresponding" may sometimes be used interchangeably, and it should be noted that their intended meanings are consistent unless their distinction is emphasized.

[0083] The technical solutions in this application will now be described with reference to the accompanying drawings.

[0084] The technical solutions of this application can be applied to various communication systems, such as: Long Term Evolution (LTE) systems, LTE Frequency Division Duplex (FDD) systems, LTE Time Division Duplex (TDD) systems, 5G systems, Wireless Local Area Network (WLAN) systems, satellite communication systems, or New Radio (NR) and 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.

[0085] Furthermore, the embodiments of this application are applicable to both homogeneous and heterogeneous network scenarios, and there are no restrictions on the transmission points. They can be applied to systems such as multi-point collaborative transmission between macro base stations, micro base stations, and macro base stations. The embodiments of this application are applicable to both low-frequency and high-frequency scenarios, including terahertz and optical communications.

[0086] In a communication system, a device can send signals to or receive signals from another device. These signals may include reference signals, information, signaling, or data. In this application, "device" can be replaced by an entity, network entity, communication device, communication module, node, communication node, network element, etc. For example, a network element in a communication system can send signals to or receive signals from another network element. This application uses a communication device as an example for illustration. For instance, a communication system may 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. The network device and the terminal device can communicate via a wireless link and / or multi-antenna technology.

[0087] Figure 1 is a schematic diagram of a communication system applicable to this application. As shown in Figure 1, the communication system 10 includes a radio access network (RAN) 100 and a core network (CN) 200. RAN 100 includes at least one RAN node (110a and 110b in Figure 1, collectively referred to as 110) and at least one terminal (120a-120j in Figure 1, collectively referred to as 120). RAN 100 may also include other RAN nodes, such as wireless relay devices and / or wireless backhaul devices (not shown in Figure 1). Terminal 120 is wirelessly connected to RAN node 110. RAN node 110 is wirelessly or wired connected to core network 200. The core network equipment in core network 200 and RAN node 110 in RAN 100 can be different physical devices, or they can be the same physical device integrating core network logical functions and radio access network logical functions.

[0088] RAN 100 can be a cellular system related to the 3rd Generation Partnership Project (3GPP), such as a 4G mobile communication system, a 5G mobile communication system, or a future-oriented evolution system. RAN 100 can also be an open RAN (O-RAN or ORAN), a cloud radio access network (CRAN), or a wireless fidelity (WiFi) system. RAN 100 can also be a communication system that integrates two or more of the above systems.

[0089] RAN node 110, sometimes also referred to as network equipment, access network equipment, radio access network equipment, RAN entity, or access node, constitutes part of the communication system and is used to help terminals achieve wireless access. Multiple RAN nodes 110 in communication system 10 can be of the same type or different types. In some scenarios, the roles of RAN node 110 and terminal 120 are relative. For example, network element 120i in Figure 1 can be a helicopter or drone, which can be configured as a mobile base station. For terminals 120j accessing RAN 100 through network element 120i, network element 120i is a base station; but for base station 110a, network element 120i is a terminal. RAN node 110 and terminal 120 are sometimes both referred to as communication devices. For example, network elements 110a and 110b in Figure 1 can be understood as communication devices with base station functions, and network elements 120a-120j can be understood as communication devices with terminal functions.

[0090] In one possible scenario, a RAN node can be a base station (BS), NodeB, evolved NodeB (eNodeB), next-generation NodeB (gNB), relay station, access point (AP), transmission point (TP), transmission reception point (TRP), master station, auxiliary station, motor slide retainer (MSR) node, home base station, network controller, access node, wireless node, transceiver node, positioning node, base station in a future mobile communication system, or access node in a WiFi system, etc. A RAN node can be a macro base station (as shown in Figure 1, 110a), a micro base station or indoor station (as shown in Figure 1, 110b), a relay node or donor node or the like, or a combination thereof, or a wireless controller in a CRAN scenario. Optionally, a RAN node can also be a server, wearable device, vehicle, or in-vehicle equipment, etc. For example, the access network equipment in vehicle-to-everything (V2X) technology can be a roadside unit (RSU).

[0091] In the embodiments of this application, the base station 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 according to 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.

[0092] In another possible scenario, multiple RAN nodes collaborate to assist the terminal in achieving wireless access, with different RAN nodes each implementing a portion of the base station's functions. For example, RAN nodes can be central units (CUs), distributed units (DUs), CU-control planes (CU-CPs), CU-user planes (CU-UPs), radio units (RUs), or CU-radio units (CU-RUs), etc. CUs and DUs can be configured separately or included in the same network element, such as a baseband unit (BBU). RUs can be included in radio equipment or radio units, such as remote radio units (RRUs), active antenna units (AAUs), or remote radio heads (RRHs).

[0093] For example, RAN nodes may include gNB-CU-CP, gNB-CU-UP, and / or gNB-DU.

[0094] In different systems, CU (including open CU-CP (O-CU-CP) and open CU-UP (O-CU-UP), DU, or RU may have different names, but those skilled in the art will understand their meaning. For example, in an ORAN system, CU can also be called an open central unit (O-CU), DU can also be called an open distributed unit (O-DU), CU-CP can also be called O-CU-CP, CU-UP can also be called O-CU-UP, and RU can also be called O-RU. For ease of description, this application uses CU, CU-CP, CU-UP, DU, and RU as examples. 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 and hardware modules.

[0095] 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 type of interface, relative to CPRI, some downlink and / or uplink baseband functions, such as, for downlink, precoding, digital beamforming (BF), or one or more of inverse fast Fourier transform (IFFT) / cyclic prefix addition (CP), are moved from the DU to the RU; and for uplink, one or more of digital beamforming (BF), or fast Fourier transform (FFT) / cyclic prefix removal (CP), are moved from the DU to the RU. In one possible implementation, the interface can be an enhanced common public radio interface (eCPRI). Under the eCPRI architecture, the segmentation between DU and RU differs, corresponding to different categories (Cat) of eCPRI, such as eCPRI Cat A, B, C, D, E, F.

[0096] 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, BF, or one or more functions in 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 one or more functions in de-RE mapping), while other functions following de-mapping (e.g., digital BF or fast Fourier transform (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.

[0097] 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. The terminal 120 can be a device or module that accesses the aforementioned communication system and has corresponding communication functions. The terminal can also be referred to as user equipment (UE), terminal, user device, access terminal, user unit, user station, mobile station, mobile station (MS), remote station, remote terminal, mobile device, user terminal, terminal unit, terminal station, terminal device, wireless communication equipment, user agent, or user device. The terminal typically contains communication modules, circuits, or chips that perform corresponding communication functions. The terminal may also be configured with program instructions for performing these corresponding communication functions.

[0098] The terminal in this application embodiment can be a device that provides voice / data, such as a mobile phone, a personal digital assistant (PDA) computer, a handheld computer, a laptop computer, a tablet computer, a mobile internet device (MID), a wearable device, a drone, a computer with wireless transceiver capabilities, a machine-type communication (MTC) terminal, a virtual reality (VR) terminal, an augmented reality (AR) terminal, an internet of things (IoT) terminal, a wireless terminal in industrial control, a wireless terminal in self-driving, a wireless terminal in remote medical care, a wireless terminal in a smart grid, a wireless terminal in transportation safety, a wireless terminal in a smart city, a wireless terminal in a smart home (e.g., game consoles, smart TVs, smart speakers, smart refrigerators, and fitness equipment), a transport vehicle with wireless communication capabilities, a communication module, or a roadside unit with terminal capabilities. Units (RSUs), cellular phones, cordless phones, session initiation protocol (SIP) phones, wireless local loop (WLL) stations, handheld devices with wireless communication capabilities, computing devices or other processing devices connected to a wireless modem, terminal devices in 5G networks or terminal devices in future evolved public land mobile networks (PLMNs), etc.

[0099] 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 worn directly on the body or integrated into a user's clothing or accessories. Wearable devices are not merely hardware devices; they achieve powerful functions through software support, data interaction, and cloud interaction. Broadly defined, wearable smart devices include those with comprehensive functions, large sizes, and the ability to perform complete or partial functions without relying on a smartphone, such as smartwatches or smart glasses. They also include devices focused on a specific application function that require the use of other devices, such as smart bracelets and smart jewelry for vital sign monitoring.

[0100] RAN 100 and terminal 120 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 aircraft, balloons, and satellites. The embodiments of this application do not limit the scenarios in which RAN 100 and terminal 120 are located.

[0101] Communication between access network devices and terminal devices follows a specific protocol layer structure. This protocol layer may include a control plane protocol layer and a user plane protocol layer. The control plane protocol layer may include at least one of the following: radio resource control (RRC) layer, packet data convergence protocol (PDCP) layer, radio link control (RLC) layer, media access control / medium access control (MAC) layer, or physical (PHY) layer, etc. The user plane protocol layer may include at least one of the following: service data adaptation protocol (SDAP) layer, PDCP layer, RLC layer, MAC layer, or physical layer, etc.

[0102] The correspondence between network elements and their achievable protocol layer functions in the ORAN system can be found in Table 1 below.

[0103] Table 1

[0104] CN 200 can be a 5G core network or an evolved 5G core network. Taking a 5G core network as an example, CN 200 includes access and mobility management (AMF) network elements responsible for mobility management and access management services; session management (SMF) network elements responsible for session management; user plane (UPF) network elements responsible for user plane packet routing and forwarding and quality of service (QoS) control; and policy control (PCF) network elements. These core network elements can operate independently or be combined to implement certain control functions; for example, AMF, SMF, and PCF can be combined into a single core network device.

[0105] 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, multiple-input multiple-output (MIMO) technology, beamforming, and / or 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.

[0106] Introducing artificial intelligence technology into communication networks can enable network intelligence.

[0107] The communication system provided in this application may also include artificial intelligence (AI) network elements to implement some or all AI-related operations. AI network elements can also be referred to as AI nodes, AI devices, AI entities, AI modules, AI models, or AI units, etc. The AI ​​network elements may be built into the network elements of the communication system. For example, an AI network element may be an AI module built into access network equipment, core network equipment, cloud servers, or operation, administration, and maintenance (OAM) systems to implement AI-related functions. The OAM system may function as the network management system for core network equipment and / or access network equipment. Alternatively, the AI ​​network element may be an independently configured network element within the communication system. Optionally, the terminal or its built-in chip may also include an AI entity to implement AI-related functions.

[0108] AI nodes can communicate with other devices in the communication system. These other devices can be one or more of the following: network devices, terminal devices, or network elements of the core network.

[0109] 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.

[0110] 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). The embodiments of this application do not limit the specific form of the AI ​​nodes described above.

[0111] It is understood that the first communication device in the embodiments of this application can be replaced by a first network element, and / or the second communication device can be replaced by a second network element, both of which execute the corresponding communication method 600 in the embodiments of this application. For example, the communication system may include at least one terminal device, at least one network device, and at least one network element. The network element is used to perform AI-related operations; for example, an AI network element is used to build a training dataset or train an AI model.

[0112] Figure 2 is a schematic diagram of another communication system applicable to this application.

[0113] In one possible implementation, as shown in Figure 2, the network device can send data related to the training of the AI ​​model to the AI ​​network element, which then constructs a training dataset and trains the AI ​​model. For example, the data related to the training of the AI ​​model may include data reported by the terminal device. The AI ​​network element can send the results of operations related to the AI ​​model to the network device, which then forwards them to terminal device #1 and / or terminal device #2. 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. Exemplarily, a portion of the trained AI model may be deployed on the network device, and another portion on the terminal device. Alternatively, the trained AI model may be deployed on the network device. Or, the trained AI model may be deployed on the terminal device.

[0114] It should be understood that AI network elements can be directly connected to network devices, or they can be connected to terminal devices. Alternatively, AI network elements can be connected to both network devices and terminal devices simultaneously. Alternatively, AI network elements can also be connected to network devices through third-party network elements. This application does not limit the connection relationship between AI network elements and other network elements.

[0115] In the embodiments of this application, the terminal device and / or network device may be a hardware device, or a software function running on dedicated hardware, or a software function running on general-purpose hardware, such as a virtualization function instantiated on a platform (e.g., a cloud platform), or an entity that includes dedicated or general-purpose hardware devices and software functions. This application does not limit the specific form of the terminal device and / or network device.

[0116] AI network elements can also be set as a module in network devices and / or terminal devices.

[0117] It should be noted that Figure 1 or Figure 2 is a simplified schematic diagram for ease of understanding. For example, the communication system may also include other devices, such as wireless relay devices and / or wireless backhaul devices.

[0118] Figure 3 is a schematic diagram of a possible application framework in a communication system applicable to embodiments of this application. As shown in Figure 3, network elements in the communication system are connected via interfaces (e.g., NG, Xn) or air interfaces. These network element nodes, such as core network equipment, access network nodes (RAN nodes), terminals, or one or more devices in OAM, are equipped with one or more AI modules (only one is shown in Figure 3 for clarity). The access network node can be a single RAN node or can include multiple RAN nodes, for example, including CU and DU. The CU and / 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 models are provided in CU-CP and / or CU-UP.

[0119] 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.

[0120] 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.

[0121] Figure 4 is a schematic diagram of another possible application framework in a communication system applicable to embodiments of this application. The network device can be a network device equipped with one or more AI modules. The network device can be one or more devices in the core network device, access network node (RAN node), or OAM shown in Figure 4. For example, the AI ​​module can be the RAN intelligent controller (RIC) shown in Figure 4, such as a near real-time RIC or a non-real-time RIC. For example, the near real-time RIC is set in the RAN node (e.g., in CU, DU), while the non-real-time RIC is set in the OAM, cloud server, core network device, or other network device. The RIC can obtain multiple subsets #1 from multiple terminal devices from the RAN node (e.g., CU, CU-CP, CU-UP, DU, and / or RU), reorganize them into a training dataset #2, and train based on the training dataset #2. Exemplarily, the near real-time RIC and the non-real-time RIC can also be set as separate network elements, and the network device can be a near real-time RIC or a non-real-time RIC.

[0122] As shown in Figure 4, the communication system includes a Resource Interchange (RIC). For example, the RIC can be the AI ​​module mentioned above, used to implement AI-related functions. The RIC includes near-real-time (near-RT) RICs and non-real-time (non-RT) RICs. Non-real-time RICs primarily process non-real-time information, such as data that is not sensitive to latency, with latency in the order of seconds. Real-time RICs primarily process near-real-time information, such as data that is relatively sensitive to latency, with latency in the order of tens of milliseconds.

[0123] The near real-time RIC is used for model training and inference. For example, it is 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 the inference results 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 near real-time RIC delivers the inference results to the DU, and the DU sends them to the RU.

[0124] 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., CU, CU-CP, CU-UP, DU, and / or RU) and / or terminals. This information can be used as training data or inference data, and the inference results can be delivered 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 non-real-time RIC delivers the inference results to the DU, which then forwards them to the RU.

[0125] 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.

[0126] It should be understood that the above naming is defined solely for the purpose of distinguishing different functions and should not constitute any limitation on this application. This application does not preclude the possibility of using other naming conventions in 5G networks and other future networks. For example, in future networks, some or all of the above-mentioned network elements may use the terminology from 5G, or they may use other names, etc.

[0127] It is understood that Figures 1, 2, 3, or 4 are merely examples and do not constitute a limitation on the scope of protection of this application. The communication method provided in the embodiments of this application may also involve network elements not shown in Figures 1, 2, 3, or 4. Of course, the communication method provided in the embodiments of this application may also include only some of the network elements shown in Figures 1, 2, 3, or 4.

[0128] To facilitate understanding of the embodiments of this application, the basic concepts involved in this application will be explained first.

[0129] 1. Machine learning (ML):

[0130] Machine learning is a crucial technological approach to achieving AI. AI endows machines with human-like intelligence, using computer hardware and software to simulate certain intelligent human behaviors, including machine learning and other methods. Machine learning refers to learning models or rules from raw data, such as neural networks, decision trees, and support vector machines. Machine learning can be categorized into supervised learning, unsupervised learning, and reinforcement learning.

[0131] Supervised learning, based on collected sample values ​​and labels, uses machine learning algorithms to learn the mapping relationship between sample values ​​and labels, and expresses this learned mapping relationship 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 mappings and nonlinear mappings. Based on the type of label, the learning task can be divided into classification tasks and regression tasks.

[0132] Unsupervised learning relies solely on collected sample values, using algorithms to discover inherent patterns within the samples. One type of unsupervised learning algorithm uses the samples themselves as supervisory signals; that is, the model learns the mapping relationship from sample to sample, which is called self-supervised learning. During training, model parameters are optimized by calculating the error between the model's predictions and the samples themselves. Self-supervised learning can be used for signal compression and decompression recovery applications; common algorithms include autoencoders and generative adversarial networks.

[0133] Reinforcement learning, unlike supervised learning, is a type of algorithm that learns problem-solving strategies through interaction with the environment. Unlike supervised and unsupervised learning, reinforcement learning problems do not have explicit "correct" action labels. The algorithm needs to interact with the environment to obtain reward signals from the environment, and then adjust its decision actions to obtain a larger reward signal value. For example, in downlink power control, the reinforcement learning model adjusts the downlink transmission power of each user based on the total system throughput feedback from the wireless network, aiming to achieve a higher system throughput. The goal of reinforcement learning is also to learn the mapping relationship between the environment state and the optimal decision action. However, because the label of the "correct action" cannot be obtained in advance, the network cannot be optimized by calculating the error between the action and the "correct action." Reinforcement learning training is achieved through iterative interaction with the environment.

[0134] Deep neural networks (DNNs) are a specific implementation of machine learning. According to the general approximation theorem, neural networks can theoretically approximate any continuous function, thus enabling them to learn arbitrary mappings. Traditional communication systems rely on extensive expert knowledge to design communication modules, while DNN-based deep learning communication systems can automatically discover hidden pattern structures from large datasets, establish mapping relationships between data, and achieve performance superior to traditional modeling methods.

[0135] Based on their construction method, DNNs can be divided into feedforward neural networks (FNNs), convolutional neural networks (CNNs), and recurrent neural networks (RNNs). FNNs can be neural networks where neurons in adjacent layers are completely connected pairwise, which makes FNNs typically require a large amount of storage space and have high computational complexity.

[0136] CNNs are neural networks specifically designed to process data with a grid-like structure. For example, time-series data (discrete sampling along the time axis) and image data (two-dimensional discrete sampling) can both be considered grid-like data. CNNs do not use all the input information at once for computation; instead, they use a fixed-size window to extract a portion of the information for convolution operations, which significantly reduces the computational cost of model parameters. Furthermore, depending on the type of information extracted by the window (such as people and objects in an image representing different types of information), each window can use different convolution kernels, allowing CNNs to better extract features from the input data.

[0137] Recurrent Neural Networks (RNNs) are a type of distributed neural network (DNN) that utilizes feedback time-series information. Their input includes the current input value and their own output value from the previous time step. RNNs are well-suited for acquiring temporally correlated sequence features, and are particularly applicable to applications such as speech recognition and channel coding / decoding.

[0138] AI models refer to function models that map inputs of a certain dimension to outputs of a certain dimension, and their parameters can be obtained through machine learning training. For example, f(X) = aX 2 +b represents a quadratic function model, which can be viewed as an AI model. a and b correspond to the model's parameters, which can be obtained through machine learning training. In machine learning, the data used for model training, validation, and / or testing can form a dataset or training dataset. The quantity and / or quality of data in the dataset or training dataset will affect the machine learning effect. Model training involves selecting an appropriate loss function (which measures the difference between the model's predictions and the true values) and using optimization algorithms to train the model parameters to minimize the loss function value. Model testing involves evaluating the model's performance using test data after training. Model application involves using the trained model to solve real-world problems.

[0139] A neural network, or artificial neural network, is a mathematical model that mimics the behavioral characteristics of animal neural networks to perform distributed parallel information processing. It is a special form of AI model.

[0140] 2. Reference signal (RS):

[0141] The reference signal can also be called a pilot, reference sequence, or reference signal. For consistency, it will be referred to as the reference signal below. The reference signal can be used for measurements, such as channel measurement or channel estimation.

[0142] The reference signals involved in this application include, but are not limited to: pilot reference signals (e.g., channel state information-reference signal (CSI-RS) and / or sounding reference signal (SRS)), synchronization signal block (SSB), demodulation reference signals (DMRS), user equipment specific reference signal (US-RS), tracking reference signal (TRS), phase tracking reference signal (PT-RS), positioning reference signal (PRS), or sensing reference signal (SeRS), cell reference signal (CRS), etc. Optionally, the pilot reference signal may be referred to as a pilot or pilot signal, wherein the pilot signal is used for channel measurement.

[0143] The reference signal in this application may also be any reference signal other than those listed above that can be carried in an orthogonal frequency division multiplexing (OFDM) symbol, which will not be described further here.

[0144] 3. Channel information:

[0145] Channel information can also be understood as measurement information, which refers to information about the path and / or the measured channel obtained by the device through channel measurement. Channel information can reflect channel characteristics and channel quality.

[0146] As an example, channel information includes at least one of the following: channel state information (CSI), time-domain channel information, or frequency-domain channel information. The following explanation primarily uses CSI as an example of channel information; however, it is understood that any information reflecting channel characteristics and channel quality is applicable to the embodiments of this application.

[0147] The meaning of CSI in this application is broader than that of traditional CSI, including but not limited to channel quality indication (CQI), precoding matrix indicator (PMI), rank indicator (RI), and CSI-RS resource indicator (CRI). It may also include one or more of the following: channel response information (such as channel response matrix, frequency domain channel response information, and time domain channel response information), weight information corresponding to the channel response, precoding matrix information corresponding to the channel response, reference signal receiving power (RSRP), reference signal receiving quality (RSRQ), and signal to interference plus noise ratio (SINR).

[0148] Taking the method of obtaining downlink CSI through uplink feedback from the terminal device side on the network device side as an example, specifically, the network device side sends a downlink reference signal to the terminal device side, and the terminal device side receives the downlink reference signal. Since the terminal device side knows the transmission information of the downlink reference signal, it can estimate (or measure) the downlink channel that the downlink reference signal has passed through based on the received downlink reference signal. Then, based on the measurement, the terminal device side can obtain the downlink channel matrix, generate CSI, and feed back the CSI to the network device side.

[0149] According to the LTE / NR protocol, at the physical layer, uplink communication includes the transmission of uplink physical channels and uplink signals. Uplink physical channels include the random access channel (PRACH), physical uplink control channel (PUCCH), and physical uplink shared channel (PUSCH), etc. Uplink signals include SRS, PUCCH de-modulation reference signal (PUCCH-DMRS), PUSCH-DMRS, PTRS, uplink positioning signal, etc. Downlink communication includes the transmission of downlink physical channels and downlink signals. The downlink physical channels include the physical broadcast channel (PBCH), the physical downlink control channel (PDCCH), and the physical downlink shared channel (PDSCH). The downlink signals include the primary synchronization signal (PSS) / secondary synchronization signal (SSS), the downlink control channel demodulation reference signal PDCCH-DMRS, the downlink data channel demodulation reference signal PDSCH-DMRS, PT-RS, CSI-RS, CRS, the time / frequency tracking reference signal (TRS), and the LTE / NR positioning signal (positioning RS).

[0150] In this embodiment of the application, CSI can be carried in uplink control information (UCI) and transmitted via PUCCH or PUSCH.

[0151] 4. In addition, the following CSI-related terms are involved:

[0152] Target CSI: Also known as the full CSI information, the uncompressed CSI information, the raw CSI, or the original CSI information. For ease of distinction and explanation, the target CSI will be represented as V in the following text.

[0153] CSI feedback information: also known as CSI feedback information, channel measurement result feedback information, channel information feedback information, compressed information, compressed channel information, compressed CSI information, compressed channel information, or compressed CSI, etc. For ease of distinction and explanation, CSI feedback information will be represented by "C" below.

[0154] In the embodiments of this application, during the training phase, CSI feedback information can refer to the information obtained after compressing the target CSI, which can be simply referred to as compressed CSI; or it can refer to the information obtained by compressing and quantizing the target CSI, which can be simply referred to as quantized CSI. During the model deployment / model application phase, CSI feedback information can be information sent from the terminal side to the network side. To save feedback overhead, CSI feedback information can refer to the information obtained by compressing and quantizing the measured CSI (also known as the target CSI), i.e., quantized CSI.

[0155] It is understandable that quantizing compressed CSI yields quantized CSI; dequantizing quantized CSI yields compressed CSI. The compressed CSI obtained by dequantizing quantized CSI is compressed CSI with quantization loss.

[0156] Reconstructed CSI: Also known as recovered CSI, reconstructed CSI, restored CSI, reconstructed channel information, decompressed CSI, decompressed channel information, etc. For ease of distinction and explanation, the term reconstructed CSI will be used in the following text. express.

[0157] 5. Model Training:

[0158] Model training involves selecting an appropriate function (such as a loss function) and using optimization algorithms to train the model parameters so that the difference between the model's predicted values ​​and the ground truth (or target values, labels) tends to be minimized.

[0159] For example, model training methods include, but are not limited to, supervised learning, self-supervised learning, and knowledge distillation.

[0160] 6. Model files and model parameters:

[0161] Model files and / or model parameters can be used to determine the model. Optionally, the model in this application may refer to the model itself, or it may refer to the model files and / or model parameters used to determine the model.

[0162] The model file can be used to indicate the model structure, which may include, but is not limited to, FNN, CNN, or RNN. The model file can have a fixed format, such as a standard predefined format, or a format pre-negotiated by both ends of the interface. 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.

[0163] 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 non-linear function. For example, the input of a neuron is x = [x0, x1, ..., x...]. N-1 The weights corresponding to the inputs are w = [w0, w1, ..., w] N-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 N is a positive integer, and n is a positive integer greater than or equal to 0 and less than or equal to (N-1). The weights of the weighted summation operation of neurons in a neural network and the nonlinear function are called the parameters of the neural network. The parameters of all neurons in a neural network constitute the parameters of the neural network.

[0164] A DNN typically 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. This application does not limit the structure and parameters used in the AI ​​model.

[0165] 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.

[0166] In this embodiment of the application, sending a model may refer to sending a model file and / or model parameters, and receiving a model may refer to receiving a model file and / or model parameters.

[0167] 7. MIMO system:

[0168] MIMO systems can significantly increase system capacity by configuring large-scale antenna arrays at the transceiver end to achieve spatial diversity gain. For example, the base station can use the same time and frequency resources to send data to multiple UEs simultaneously (i.e., multi-user MIMO (MU-MIMO)) or send multiple data streams to the same UE simultaneously (i.e., single-user MIMO (SU-MIMO)). The data of the multiple UEs or the multiple data streams of the same UE are spatially multiplexed.

[0169] In massive MIMO systems, base stations need to perform precoding on digital ports and select appropriate coding and modulation orders. Base stations can use precoding techniques to achieve spatial multiplexing between users or data streams, thus spatially isolating data between UEs or between different data streams of the same UE. This reduces interference between different UEs or different data streams and improves the received signal-to-interference-plus-noise ratio (SNR) at the UE.

[0170] The purpose of precoding is to better match the antenna (or beam) with the channel, ensuring better signal quality and less interference when transmitted data arrives at the terminal. A good modulation order and code rate maximize channel transmission capacity while ensuring reliable data transmission. The settings for precoding and the modulation coding scheme (MCS) need to be determined based on channel quality and channel response. A common method is for the base station to send a reference signal, the terminal to determine the channel based on the reference signal, and then feed back the corresponding channel state information, including precoding information, the number of transport streams (RI) supported by the channel, and the Channel Quality Information (CQI) used to provide feedback on the terminal's recommended MCS for the current channel quality. This process is called Channel State Information Feedback (CSI Feedback), and the terminal device can report CSI reports generated according to protocol predefined methods or base station configuration to the network device. Another method is to measure and obtain uplink channel information using an uplink reference signal, and then further obtain downlink channel information based on channel reciprocity.

[0171] To calculate the precoding matrix, the base station needs to obtain the downlink channel reference signal (CSI) and determine the precoding matrix based on the CSI. In an FDD system, the base station obtains the downlink CSI through uplink feedback from the UE. That is, the base station sends a downlink reference signal to the UE, and the UE receives this downlink reference signal. Since the UE knows the transmission information of the downlink reference signal, it can estimate (measure) the downlink channel traversed by the received downlink reference signal. Based on this measurement, the UE obtains the downlink channel matrix, generates the CSI, and feeds the CSI back to the base station.

[0172] In FDD systems, a crucial part of CSI feedback is PMI (Programmable Memory Interface), which uses 0-1 bits to quantize the channel matrix or precoding matrix in CSI. 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, the UE can feed back the corresponding codeword numbers and one or more weighting coefficients to the base station via PMI, which is used by the base station to recover the channel matrix or precoding matrix.

[0173] As the size of MIMO system antenna arrays 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 the UE to estimate (measure) the downlink channel, the overhead of the base station transmitting reference signals increases. Simultaneously, the error of approximating large-scale channel and precoding matrices with a finite number of predefined codewords increases. One method to improve channel recovery accuracy is to increase the number of codewords in the codebook; however, 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.

[0174] In summary, it is necessary to study 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 recover the channel based on feedback information.

[0175] Correlation exists between different elements in the downlink channel matrix between the base station and the UE. Furthermore, temporal correlation exists between downlink channel matrices in different time slots. For example, temporal correlation between different elements in the channel matrix implies the existence of a basis (represented by matrices U1, U2). Projecting the channel matrix H onto this basis yields a sparse equivalent channel, i.e., H' = U1 * H. H*U2 is a sparse matrix, where the superscript H denotes the conjugate transpose operation. Theoretically, the channel matrix H can be recovered simply by estimating the non-zero elements in H' through the reference signal transmission and feeding them back. Therefore, the overhead of the reference signal transmission and CSI feedback has room for compression. However, traditional CSI feedback schemes, such as codebook-based feedback methods 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 recover the channel based on feedback information.

[0176] The above description of the terminology is for ease of understanding only and does not limit the scope of protection of the embodiments of this application.

[0177] The above text, in conjunction with Figures 1 to 4, briefly introduces the scenarios in which the communication method provided in the embodiments of this application can be applied, as well as the basic concepts that may be involved in the embodiments of this application, and introduces compression and / or prediction involved in CSI feedback in the basic concepts.

[0178] In the process of terminal devices feeding back CSI to network devices, in order to reduce the overhead of CSI feedback, the terminal devices can compress the information to be fed back using an AI model; correspondingly, the network devices can decompress / restore the compressed information fed back by the terminal devices using an AI model. This compression and decompression process can also be called AI CSI feedback.

[0179] As shown in Figure 5, Figure 5 is a schematic diagram of AI CSI feedback applicable to an embodiment of this application.

[0180] In one possible implementation, the overhead of CSI feedback can be reduced by deploying autoencoder (AE) models on both the terminal device and network device sides. An AE model consists of two sub-models: an encoder model and a decoder model. AE can generally refer to a network structure composed of these two sub-models. AE models can also be called bilateral models, dual-end models, or collaborative models. The encoder and decoder models of an AE are usually trained together and can be used in a complementary manner. The reduced CSI feedback can be implemented based on the AI ​​model of the AE. For example, the terminal device compresses and quantizes the CSI using the encoder model and feeds the compressed and quantized CSI back to the network device. The network device then uses the decoder model to recover the fed-in CSI. For the network device, the input to the decoder model is the CSI fed back from the terminal device, and the output is the recovered CSI. Training the decoder model requires the CSI measured by the terminal device (i.e., the CSI before compression and quantization) as the ground truth label for the recovered CSI.

[0181] To support the integration and training of encoder and decoder models in two-way AI CSI feedback, training datasets are needed for training the encoder and decoder models on the terminal or network device side. Taking the encoder model as an example, the training dataset includes the encoder model's input and corresponding output, and may also include quantization methods. Similarly, taking the decoder model as an example, the training dataset includes the decoder model's input and corresponding output, and may also include quantization methods. The following sections detail possible implementations of the integration and training of encoder and decoder models in two-way AI CSI feedback, using scenarios #1 and #2 as examples.

[0182] Scenario #1: First, deploy an initial virtual encoder model and an initial virtual decoder model on the network device side, and jointly train the initial virtual encoder model and the initial virtual decoder model. After training, the virtual encoder model and the virtual decoder model are obtained. The network device side sends training dataset #1 to the terminal device side. Training dataset #1 includes the input of the virtual encoder model and the corresponding output of the input. The terminal device side trains the encoder model deployed on the terminal device side using training dataset #1.

[0183] Scenario #2: First, deploy an initial encoder model and an initial virtual decoder model on the terminal device side, and jointly train the initial encoder model and the initial virtual decoder model. After training, the encoder model and the virtual decoder model are obtained. The terminal device side sends training dataset #2 to the network device side. Training dataset #2 includes the input of the virtual decoder model and the corresponding output of the input. The network device side trains the decoder model deployed on the network device side using training dataset #2.

[0184] It is understandable that if training dataset #1 includes other information such as the quantization method for quantizing the virtual encoder model, the terminal device can also use this other information when training the encoder model; similarly, if training dataset #2 includes other information such as the quantization method for quantizing the virtual decoder model, the network device can also use this other information when training the decoder model. No restrictions are imposed on this.

[0185] In the process of terminal devices feeding back CSI to network devices, terminal devices can predict future CSI based on the measured current and / or historical CSI; and / or, network devices can predict future CSI based on the current and / or historical CSI fed back by terminal devices.

[0186] AI CSI prediction (which can also be understood as CSI prediction based on AI models) refers to inputting at least one historical and / or current CSI into an AI prediction model and outputting a predicted future CSI.

[0187] For example, by inputting the CSI at time t-5 and time t into the AI ​​prediction model, the AI ​​prediction model can output the CSI at time t+5, where time t is the current time, t-5 is a historical time, and t+5 is a future time.

[0188] For example, in AI beam prediction (which can also be understood as time-domain beam prediction based on an AI model), in an AI beam prediction scenario (i.e., the task of the AI ​​model is beam prediction), the input of the AI ​​model may include the RSRP of a first beam set, and the output of the AI ​​model may include the RSRP of a second beam set or the beam identity (ID) of at least one optimal beam in the second beam set. The first and second beam sets may be the same or different, and the second beam set output by the AI ​​model represents the prediction result obtained through the AI ​​model and the first beam set.

[0189] In AI CSI feedback and / or AI CSI prediction, the AI ​​model is obtained through training. Training the model requires acquiring a training dataset. However, the cost of collecting and / or transmitting data in the training dataset is significant. Therefore, reducing the cost of data collection and / or transmission is a pressing issue.

[0190] In AI CSI prediction, the input CSI can include either the target CSI or the reconstructed CSI. For example, when a terminal device predicts the CSI for a future time based on the measured CSI at the current and / or historical time, the input CSI can include the target CSI; when a network device predicts the CSI for a future time based on the reconstructed CSI at the current and / or historical time, the input CSI can include the reconstructed CSI.

[0191] In some possible implementations, data augmentation methods can be used to increase the amount of data used to train the AI ​​model before training it with the training dataset.

[0192] Data augmentation is a technique that uses algorithms to artificially expand a training dataset by generating more equivalent data from a limited dataset. It is an effective means of overcoming insufficient training data and is widely used in various fields of deep learning. Its effects can include the following:

[0193] 1) Increase the diversity of training data to avoid overfitting during training;

[0194] 2) Improve the robustness of AI models;

[0195] 3) Increase the amount of training data to improve the generalization ability of the AI ​​model;

[0196] 4) Avoid imbalanced training samples.

[0197] There are various data augmentation methods to obtain augmented data and increase the amount of training data. However, for CSI feedback and / or CSI prediction, not all data augmentation methods are suitable for specific scenarios. Using inappropriate data augmentation methods may result in negative gains.

[0198] To address the aforementioned problems, this application provides a communication method and apparatus to reduce data collection costs / overhead and transmission overhead.

[0199] In the embodiments of this application, CSI feedback may include at least one of the following: CSI compression based on non-AI methods, CSI decompression / reconstruction based on non-AI methods, CSI compression based on AI models, CSI prediction based on AI models, or CSI decompression / reconstruction based on AI models.

[0200] The communication method provided by the embodiments of this application will be described in detail below with reference to the accompanying drawings, and can be applied to the communication system shown in Figure 1 or Figure 2 above. It should be understood that the embodiments of this application can be applied to scenarios where the sending end and the receiving end communicate.

[0201] It should also be understood that the embodiments shown below do not specifically limit the 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 the code or program that records the method provided in the embodiments of this application. For example, the method provided in the embodiments of this application can be executed by a first communication device and a second communication device. Unless otherwise specified, the "first communication device" in this application can be a device on the terminal device side or network device side (e.g., a terminal device or a network device), or a component in the device (e.g., a communication module, processor, circuit, chip, or chip system, etc.), or a logic module or software that can implement all or part of the functions of the communication device; the "second communication device" in this application can be a device on the terminal device side or network device side (e.g., a terminal device or a network device), or a component in the device (e.g., a communication module, processor, circuit, chip, or chip system, etc.), or a logic module or software that can implement all or part of the functions of the communication device.

[0202] In this application embodiment, the chip system may be composed of chips, or it may include chips and other discrete devices. This application embodiment uses only a device for implementing the functions of a terminal device as an example for illustration, and does not constitute a limitation on the solution of this application embodiment.

[0203] The terminal device side may include at least one of a terminal device or an AI entity on the terminal device side. The AI ​​entity on the terminal device side can be the terminal device itself or an AI entity serving the terminal device, such as a server, like an over-the-top (OTT) server or a cloud server. The network device side may include at least one of a network device or an AI entity on the network device side. The AI ​​entity on the network device side can be the network device itself or an AI entity serving the network device, such as a radio access network (RAN) intelligent controller (RIC), operation administration and maintenance (OAM), or a server, such as an OTT server or a cloud server. Communication between servers can be achieved through a communication link between the terminal device and the network device, through forwarding via other communication devices outside the server, or through a wired link.

[0204] The terminal device in this application embodiment may also be referred to as "terminal device side", "UE side", or "UE part". The network device may also be referred to as "network device side", "network side", or "network part".

[0205] Figure 6 is a flowchart illustrating a communication method provided in an embodiment of this application. As shown in Figure 6, the method 600 includes the following steps.

[0206] S610, the first communication device receives the first indication information, and correspondingly, the second communication device sends the first indication information. The first indication information indicates the first model and / or the first function, and the first model and / or the first function is used for channel state information (CSI) processing.

[0207] For example, a function can be understood as a functional entity, and a model and / or function can be understood as a model, model parameters, a function implemented based on a model, or a task / scenario processed based on a model. This application does not impose any limitations on this.

[0208] For example, a function may correspond to at least one model.

[0209] For example, the first indication information indicating the first model and / or the first function may include the ID of the first model and / or the ID of the first function indicated by the first indication information; or, the first indication information indicating the first model and / or the first function may also be implemented according to the characteristics of the indicating terminal device (UE feature) and / or according to the capabilities of the indicating terminal device (UE capability).

[0210] For example, the first indication information indicating a first model can be interchanged with the first indication information indicating a first model among at least one models corresponding to a function, and express the same meaning; or, the first indication information indicating a first function can be interchanged with the first indication information indicating a model among at least one models corresponding to a first function, and express the same meaning. This application does not limit this.

[0211] In some possible implementations, the first model and / or the first function is used for CSI processing, including at least one of the following: the first model and / or the first function is used for CSI compression; the first model and / or the first function is used for CSI decompression; or, the first model and / or the first function is used to determine the CSI of at least one time unit after at least one time unit based on the CSI of at least one time unit.

[0212] For example, the first model and / or the first function is used for CSI compression, which can also be understood as the first model and / or the first function being used to compress (or compress and quantize) the target CSI to obtain CSI feedback information; the first model and / or the first function is used for CSI decompression, which can also be understood as the first model and / or the first function being used to reconstruct the CSI feedback information to obtain the reconstructed CSI; the first model and / or the first function is used to determine the CSI of at least one time unit after at least one time unit based on the CSI of at least one time unit, which can also be understood as the first model and / or the first function being used to determine the target CSI of at least one time unit after at least one time unit based on the target CSI of at least one time unit; the first model and / or the first function is used to determine the CSI of at least one time unit after at least one time unit based on the CSI of at least one time unit, which can also be understood as the first model and / or the first function being used to determine the reconstructed CSI of at least one time unit after at least one time unit based on the reconstructed CSI of at least one time unit.

[0213] In the embodiments of this application, CSI compression may include at least one of the following: joint compression of CSI in both spatial and frequency domains based on an AI model; or joint compression of CSI in three dimensions (spatial, frequency, and time domains) based on an AI model; or joint compression of CSI in both spatial and frequency domains using time-domain correlation based on an AI model; or joint compression of CSI in both spatial, frequency, and time domains using time-domain correlation based on an AI model; or joint compression of CSI in both spatial and frequency domains based on a non-AI method; or joint compression of CSI in both spatial, frequency, and time domains based on a non-AI method; or joint compression of CSI in both spatial and frequency domains using time-domain correlation based on a non-AI method; or joint compression of CSI in both spatial, frequency, and time domains using time-domain correlation based on a non-AI method, etc.

[0214] For example, the first model and / or the first function used to determine the CSI of at least one time unit after at least one time unit based on the CSI of at least one time unit can also be interchanged with, and express the same meaning, the first model and / or the first function used to predict the CSI of at least one time unit after at least one time unit based on the CSI of at least one time unit; the first model and / or the first function used to determine the CSI of at least one time unit after at least one time unit based on the CSI of at least one time unit can also be interchanged with, and express the same meaning, the first model and / or the first function used for CSI prediction. This application does not limit this aspect.

[0215] For example, CSI compression / decompression and CSI prediction can be used individually or in combination. For instance, using an AI model, the predicted CSI and the original CSI obtained from the original CSI prediction are compressed; the compressed data may include part or all of the predicted CSI and the original CSI. Alternatively, the compressed predicted CSI and the original CSI are decompressed; the predicted CSI is obtained based on the original CSI prediction. Or, the resulting compressed CSI is decompressed to obtain a decompressed CSI, and a predicted CSI is obtained based on the decompressed CSI prediction.

[0216] In the embodiments of this application, the compressed data / CSI may include the target CSI; the compressed CSI may include CSI feedback information; the decompressed CSI may include the reconstructed CSI; the predicted CSI or the predicted CSI for future moments may include the target CSI and / or the reconstructed CSI.

[0217] In some possible implementations, method 600 may also include:

[0218] S620, the first communication device determines the first data processing scheme according to the first instruction information.

[0219] The first communication device can determine the first data processing scheme in several ways. These are described below using methods #1, #2, and #3 respectively.

[0220] Method #1:

[0221] The first communication device determines a first data processing scheme based on a first model and / or a first function and a first correspondence. The first correspondence indicates the relationship between at least one model and / or at least one function and at least one data processing scheme for determining enhanced data; the at least one model and / or at least one function includes the first model and / or the first function; and the at least one data processing scheme includes a first data processing scheme corresponding to the first model and / or the first function.

[0222] Specifically, the first communication device determines, based on the first correspondence, a data processing scheme that satisfies the first correspondence with at least one model and / or at least one function in at least one data processing scheme, and uses it as the first data processing scheme.

[0223] For ease of description, different models / functions / tasks will be numbered as #1, #2, #3, #4, etc.; different data processing schemes will also be numbered as #1, #2, #3, etc. It should be understood that this numbering does not limit the embodiments of this application.

[0224] Take Table 2 as an example. Table 2 is one possible representation of the first correspondence. It is understood that the first correspondence can be other representations, or it can be a different correspondence than that shown in Table 2. This application does not limit this.

[0225] Table 2

[0226] For example, the first instruction information indicates model #1, and the first correspondence satisfies Table 2. The first communication device can select data processing scheme #1 corresponding to model #1 as the first data processing scheme according to Table 2; or, the first instruction information indicates function #4, and the first correspondence satisfies Table 2. The first communication device can select any one of data processing scheme #1, data processing scheme #2 and data processing scheme #3 corresponding to function #4 as the first data processing scheme according to Table 2.

[0227] Method #2:

[0228] The first communication device determines a first data processing scheme based on M and a second correspondence. The second correspondence indicates the relationship between multiple numerical ranges and at least one data processing scheme used to determine enhanced data. The multiple numerical ranges include a first numerical range, the value of M is within the first numerical range, and the at least one data processing scheme includes a first data processing scheme corresponding to the first numerical range.

[0229] Specifically, the first communication device determines, in at least one data processing scheme, a data processing scheme that satisfies the second correspondence with M, based on the second correspondence relationship, and uses it as the first data processing scheme.

[0230] For ease of description, the different numerical ranges will be numbered as #1, #2, #3, #4, etc., and the different data processing schemes will be numbered as #1, #2, #3, etc. It should be understood that this numbering does not limit the embodiments of this application.

[0231] Take Table 3 as an example. Table 3 is one possible representation of the second correspondence. It is understood that the second correspondence can be other representations, or it can be a different correspondence than that shown in Table 3. This application does not limit this.

[0232] Table 3

[0233] For example, M is indicated by the first indication information; or, M is determined according to the first model and / or the first function, and M is a positive integer.

[0234] For example, the first indication information indicates model #3, and the M corresponding to model #3 satisfies the numerical range #4 (for example, the value of M can be determined according to the ID of model #3; or, the ID of model #3 corresponds to a fixed value of M); or, the first indication information indicates the values ​​of model #3 and M, and M satisfies the numerical range #4. The second correspondence satisfies Table 3, and the first communication device can select any one of the data processing schemes #1, #2, and #3 corresponding to the numerical range #4 from Table 3 as the first data processing scheme.

[0235] Method #3:

[0236] The first communication device determines a first data processing scheme based on M, a first model and / or a first function, and a third correspondence. The third correspondence indicates the relationship between multiple numerical ranges and at least one model and / or at least one function and at least one data processing scheme used to determine enhanced data. The multiple numerical ranges include a first numerical range, the value of M is within the first numerical range, the at least one model and / or at least one function includes the first model and / or the first function, the at least one data processing scheme includes a first data processing scheme corresponding to the first numerical range and the first model and / or the first function, and M is a positive integer.

[0237] Specifically, the first communication device, based on the third correspondence, determines, in at least one data processing scheme, a data processing scheme that satisfies the third correspondence with M and the first model and / or the first function, as the first data processing scheme.

[0238] For ease of description, different models / functions / tasks / scenarios and numerical ranges will be numbered as #1, #2, #3, #4, etc.; different data processing schemes will also be numbered as #1, #2, #3, etc. It should be understood that this numbering does not constitute a limitation on the embodiments of this application.

[0239] Take Table 4 as an example. Table 4 is one possible representation of the third correspondence. It is understood that the third correspondence can be other representations, or it can be a different correspondence than that shown in Table 4. This application does not limit this.

[0240] Table 4

[0241] For example, M is indicated by the first indication information; or, M is determined according to the first model and / or the first function, and M is a positive integer.

[0242] For example, the first instruction information indicates the first model, and the first model corresponds to the value of M (or, the value of M can be determined based on the ID of the first model); or, the first instruction information indicates the values ​​of the first model and M. The values ​​of the first model and M satisfy model / function / task / scenario and numerical range #3. The third correspondence satisfies Table 4, and the first communication device can select either data processing scheme #1 or data processing scheme #3 corresponding to model / function / task / scenario and numerical range #3 from Table 4 as the first data processing scheme.

[0243] In the embodiments of the present application (for example, in Mode #2 and / or Mode #3), the numerical range can be an open interval (an open interval can be understood as the two values at the interval boundaries are not included in the numerical range), the numerical range can also be a closed interval (a closed interval can be understood as the two values at the interval boundaries are included in the numerical range), and the numerical range can also be a half-open and half-closed interval (a half-open and half-closed interval can be understood as the boundary value on the open interval side is not included in the numerical range, while the boundary value on the closed interval side is included in the numerical range). The embodiments of the present application do not limit this.

[0244] In the embodiments of the present application (for example, in Mode #2 and / or Mode #3), the representation form of the numerical range can be composed of the numerical values at both ends of the numerical range. Taking an open interval as an example, the numerical range #1 can be expressed as (A, B), where A represents any value less than B; the numerical range #2 can be expressed as (C, D), where C < D; the numerical range #3 can be expressed as (E, F), where E < F; the numerical range #4...... A, B, C, D, E, F are natural numbers. The remaining numerical ranges can be deduced by analogy and will not be elaborated here.

[0245] Exemplarily, M represents the number of compression time slots of the AI model compression CSI corresponding to the model / function / task / scenario.

[0246] For the model / function / task / scenario that utilizes the time-domain correlation of CSI, the first data processing scheme can be determined through the above-mentioned Mode 3 or Mode #3.

[0247] It is understood that the above-mentioned methods #1, #2, and #3 can be internal implementations of the first communication device and are not reflected in the communication between the first and second communication devices. The action of "determining the first data processing scheme" involved in methods #1, #2, and #3 may not be included in method 600. The above-mentioned method #1 can be interchanged with "the first data processing scheme is determined based on the first model and / or the first function and the first correspondence, the first correspondence indicating the relationship between at least one model and / or at least one function and at least one data processing scheme for determining enhanced data, the at least one model and / or at least one function includes the first model and / or the first function, and the at least one data processing scheme includes the first data processing scheme corresponding to the first model and / or the first function" and expresses the same meaning; the above-mentioned method #2 can be interchanged with "the first data processing scheme is determined based on M and the second correspondence, the second correspondence indicating the relationship between multiple numerical ranges and at least one data processing scheme for determining enhanced data, the multiple numerical ranges include the first numerical range, and the value of M is within the first numerical range". The above method #3 can be interchanged with "The first data processing scheme is determined based on the first model and / or the first function, M and the third correspondence, the third correspondence indicating the relationship between multiple numerical ranges and at least one model and / or at least one function and at least one data processing scheme used to determine the enhanced data, the multiple numerical ranges include the first numerical range, the value of M is within the first numerical range, at least one model and / or at least one function includes the first model and / or the first function, at least one data processing scheme includes the first data processing scheme corresponding to the first numerical range and the first model and / or the first function, M is a positive integer" and expresses the same meaning. This application does not limit this aspect.

[0248] In some possible implementations, the value of M is within a first numerical range, including at least one of the following: the value of M is between a first threshold and a second threshold, the first threshold and the second threshold are not equal, and the first threshold and / or the second threshold is greater than 0; the value of M is greater than or equal to the first threshold, and the first threshold is greater than 0; or, the value of M is less than or equal to the second threshold, and the second threshold is greater than 0.

[0249] For example, the numerical range may be determined based on a first threshold and / or a second threshold.

[0250] For example, the first threshold is N1, with a numerical range of [N1, ~) or (N1, ~), where ~ represents any value greater than N1; or, the second threshold is N2, with a numerical range of (~, N2] or (~, N2), where ~ represents any value less than N2; or, the first threshold is N1, the second threshold is N2, and the numerical range is (N1, N2) or [N1, N2) or (N1, N2] or [N1, N2].

[0251] For example, the first threshold and / or the second threshold may be predefined, preconfigured, or determined through negotiation between the first communication device and the second communication device.

[0252] In some possible implementations, the first data processing scheme is used to determine augmented data, including any of the following: the augmented data is obtained by phase rotation of the original training data used to obtain the augmented data; the augmented data is obtained by phase rotation of the channel matrices of different time units in the same time series of the original training data used to obtain the augmented data by the same angle; or, the augmented data is obtained by phase rotation of the channel matrices of different time units in the same time series of the original training data used to obtain the augmented data by different angles.

[0253] In some possible implementations, the first data processing scheme is used to determine the augmented data, including any of the following: H′=H*e jθ ,θ∈(0,2π), where H represents the original training data used to obtain augmented data, and H′ represents the augmented data; [H′1,H′2,H′3]=[H1,H2,H3]*e jθ ,θ∈(0,2π), where [H1,H2,H3] represents the original training data used to obtain augmented data, and [H′1,H′2,H′3] represents the augmented data; or, Where [H1,H2,H3] represents the original training data used to obtain augmented data, and [H′1,H′2,H′3] represents the augmented data.

[0254] For example, H′=H*e jθ θ∈(0,2π) can be a possible implementation of phase rotation of the original training data H to obtain the augmented data H′; [H′1,H′2,H′3]=[H1,H2,H3]*e jθ θ∈(0,2π) can be a possible implementation of performing phase rotation of the channel matrices H1, H2, H3 of different time units in the same time series of the original training data by the same angle to obtain the augmented data H′1, H′2, H′3. One possible implementation is to perform phase rotations of the channel matrices H1, H2, H3 at different time units in the same time series of the original training data at different angles to obtain augmented data H′1, H′2, H′3.

[0255] In some possible implementations, method 600 also includes:

[0256] S630, the first communication device sends a second instruction message, and correspondingly, the second communication device receives the second instruction message. The second instruction message indicates a first data processing scheme. The first data processing scheme is determined based on the first instruction message. The first data processing scheme is used to determine augmented data. The augmented data is used to train a first model and / or at least one model corresponding to a first function.

[0257] For example, the first data processing scheme may be determined by the above-described method #1, method #2, or method #3.

[0258] In some possible implementations, method 600 may also include:

[0259] S640, the first communication device receives the first request information, and correspondingly, the second communication device sends the first request information. The first request information is used to request at least one of the following: training dataset, model parameters of the second model, or model parameters of at least one model corresponding to the second function. The training dataset is the training dataset of the first model and / or at least one model corresponding to the first function. The second model and / or the second function is used to train the first model and / or at least one model corresponding to the first function.

[0260] Specifically, the first request information is used to request first information, which includes at least one of the following: training dataset; model parameters of the second model; or model parameters of at least one model corresponding to the second function.

[0261] The model parameters requested in the first request can include one or more of the model's multiple parameters.

[0262] In the embodiments of this application, the second model may be different from the first model; or, the second model may belong to the same function / task / scenario as the first model, or have the same ID; or, the second function may be the same as the first function, or have the same ID, or at least one model corresponding to the second function and at least one model corresponding to the first function may be completely the same or partially the same or completely different.

[0263] In some possible implementations, the model parameters of the second model and / or the model parameters of at least one model corresponding to the second function are used to adjust the parameters of the first model and / or at least one model corresponding to the first function.

[0264] For example, at least one model corresponding to the second model and / or the second function is one possible implementation of a trained virtual encoder model or a virtual decoder model.

[0265] In some possible implementations, method 600 may also include:

[0266] S650, the first communication device sends a third indication information, and correspondingly, the second communication device receives the third indication information. The third indication information is used to indicate at least one of the training dataset, the model parameters of the second model, or the model parameters of at least one model corresponding to the second function.

[0267] Specifically, the third instruction information is used to indicate the first information, which includes at least one of the following: the training dataset; the model parameters of the second model; or the model parameters of at least one model corresponding to the second function.

[0268] For example, the third instruction information indicates the first information, and may include a variety of possible implementations.

[0269] For example, the first information is stored in a third communication device, which is a different device from the first and second communication devices. The third indication information indicates the address of the third communication device. Furthermore, the third indication information can also indicate the ID of the first information.

[0270] For example, if the first information is already stored in the second communication device, the second communication device can determine the first information from the stored data based on the third instruction information.

[0271] For example, the third instruction information may include one or more of the first information.

[0272] For example, at least one model corresponding to the second model and / or the second function can be a virtual first model and / or at least one model corresponding to the first function deployed on the first communication device side. After the first communication device side completes the training of the virtual first model and / or at least one model corresponding to the first function, the first communication device can indicate to the second communication device the training dataset for training the virtual first model and / or at least one model corresponding to the first function, and / or the model parameters of the virtual first model and / or at least one model corresponding to the first function. The specific implementation of the second model and / or at least one model corresponding to the second function can also be referred to the description related to Figure 7, which will not be repeated here.

[0273] The third indication information indicates that the model parameters may include one or more of the model's multiple parameters.

[0274] In some possible implementations, method 600 may also include:

[0275] S660, the second communication device trains the first model and / or at least one model corresponding to the first function.

[0276] Specifically, the second communication device trains at least one model corresponding to the first model and / or the first function, including at least one of the following:

[0277] The second communication device trains the first model based on the augmented data;

[0278] The second communication device trains at least one model corresponding to the first function based on the enhanced data;

[0279] The second communication device trains the first model based on the training dataset;

[0280] The second communication device trains at least one model corresponding to the first function based on the training dataset.

[0281] The second communication device trains the first model based on the model parameters of the second model;

[0282] The second communication device trains at least one model corresponding to the first function based on the model parameters of the second model;

[0283] The second communication device trains the first model based on the model parameters of at least one model corresponding to the second function; or,

[0284] The second communication device trains at least one model corresponding to the first function based on the model parameters of at least one model corresponding to the second function.

[0285] In some possible implementations, the first communication device is a terminal device, a component within the terminal device, a logic module or software capable of implementing all or part of the communication device's functions, or an AI entity on the terminal device side; the second communication device is a network device, a component within the network device, a logic module or software capable of implementing all or part of the communication device's functions, or an AI entity on the network device side. Alternatively,

[0286] In some other possible implementations, the second communication device is a terminal device, a component in the terminal device, a logic module or software that can realize all or part of the functions of the communication device, or an AI entity on the terminal device side; the first communication device is a network device, a component in the network device, a logic module or software that can realize all or part of the functions of the communication device, or an AI entity on the network device side.

[0287] It is understood that method 600 can be used in either scenario #1 or scenario #2. For ease of description and understanding, the following text will use the example of method 600 in scenario #1 (which can also be understood as the second communication device being a terminal device, a component in the terminal device, a logic module or software that can realize all or part of the communication device functions, or an AI entity on the terminal device side; and the first communication device being a network device, a component in the network device, a logic module or software that can realize all or part of the communication device functions, or an AI entity on the network device side) for detailed explanation. The specific implementation of method 600 in scenario #2 is similar to that of method 600 in scenario #1, and can be referred to in the case of method 600 in scenario #1, which will not be repeated here.

[0288] For ease of description, the following embodiments use the example of a network device as the first communication device and a terminal device as the second communication device.

[0289] It is understood that the following embodiments also apply when the first communication device is a terminal device and the second communication device is a network device. For specific implementations where the first communication device is a terminal device and the second communication device is a network device, please refer to the specific implementations where the first communication device is a network device and the second communication device is a terminal device; these will not be repeated here.

[0290] Figure 7 is a schematic diagram of another communication method applicable to embodiments of this application.

[0291] As shown in Figure 7, an initial virtual encoder model and an initial virtual decoder model are deployed on the network device side, and jointly trained. After training, the virtual encoder model and the virtual decoder model are obtained. The network device side sends a training dataset to the terminal device side. The training dataset includes the input of the virtual encoder model and the corresponding output. The terminal device side trains the encoder model to be trained on the terminal device side using the training dataset.

[0292] The terminal device can determine the input and corresponding output of the encoder model to be trained based on the training dataset. The terminal device can use this input as the input to the encoder model to be trained, obtain the true output of the encoder model, and compare the difference between the true output and the corresponding output in the training dataset. If the difference between the true output and the corresponding output is greater than threshold #1, the encoder model is adjusted until the difference between the true output and the corresponding output is less than threshold #1. Alternatively, the terminal device can use this output as the output of the encoder model to be trained, obtain the true input of the encoder model, and compare the difference between the true input and the corresponding input in the training dataset. If the difference between the true input and the corresponding input is greater than threshold #2, the encoder model is adjusted until the difference between the true input and the corresponding input is less than threshold #2.

[0293] The values ​​of threshold #1 and / or threshold #2 can be set according to the requirements of the model, or should be as small as possible. This application embodiment does not impose any restrictions on this.

[0294] The network device can also send the model parameters and / or quantization method of the virtual encoder model to the terminal device. These model parameters and / or quantization method can also be used by the terminal device to train the encoder model to be trained.

[0295] For example, the terminal device requests a training dataset from the network device via a first request message; the network device instructs the terminal device to provide the training dataset via a third instruction message. The terminal device may also request model parameters and / or quantization methods for a virtual CSI compression model from the network device via the first request message; the network device instructs the terminal device to provide the model parameters and / or quantization methods for the virtual CSI compression model via the third instruction message.

[0296] The terminal device can also indicate the CSI model / function / task / scenario requested by the terminal device to the network device through the first indication information; the network device can select a data augmentation method according to the CSI model / function / task / scenario requested by the terminal device, and indicate the data augmentation method to the terminal device. The selected data augmentation method can be one possible implementation of the first data processing scheme. The encoder model to be trained or the CSI model / function / task / scenario can be one possible implementation of the first model and / or the first function.

[0297] For example, the CSI model can be a CSI prediction model for determining the predicted CSI; and / or, the CSI model can be a compression model for compressing (or compressing and quantizing) the target CSI; and / or, the CSI model can be a decompression / reconstruction model for reconstructing CSI feedback information / determining the reconstructed CSI. Alternatively, the CSI model can be used for at least one of the following: determining the predicted CSI, compressing the target CSI, reconstructing CSI feedback information, determining the reconstructed CSI, etc.

[0298] In the embodiments of this application, when the CSI model / function / task / scenario is used to determine / predict / reconstruct CSI feedback information, the CSI model / function / task / scenario can be interchanged with the CSI feedback model / feedback function / feedback task / feedback scenario and express the same meaning, without any limitation.

[0299] In the embodiments of this application, the CSI model / function / task / scenario used to determine / predict / reconstruct CSI feedback information may include at least one of the following: AI CSI compression model / function / task / scenario; or, AI CSI decompression model / function / task / scenario; or, AI CSI recovery model / function / task / scenario; or, AI CSI reconstruction model / function / task / scenario; or, AI CSI prediction model / function / task / scenario, etc.

[0300] In the embodiments of this application, the CSI model / function / task / scenario can be based on an AI model or on a non-AI method. Specific implementations of compression / feedback, prediction, reconstruction, etc., based on non-AI methods can be found in the detailed description of the CSI model / function / task / scenario based on the AI ​​model, and will not be repeated here.

[0301] The following section uses an AI-based CSI model / function / task / scenario as an example to describe in detail the possible implementations of the embodiments of this application. In the embodiments of this application, the AI-based CSI model / function / task / scenario can be interchanged with AI CSI model / function / task / scenario and expresses the same meaning, and there is no limitation thereto.

[0302] The network device instructs the terminal device to use a data enhancement method. This can be achieved by the network device sending instruction information to the terminal device, which indicates the data enhancement method (e.g., the instruction information indicates the index / ID of the data enhancement method; the correspondence between the index / ID and the data enhancement method can be predefined by the protocol, preconfigured, or negotiated between the terminal device and the network device, and is not limited thereto); or by the network device sending instruction information to the terminal device that includes the data enhancement method. This instruction information can be one possible implementation of a second instruction.

[0303] Furthermore, the AI ​​CSI model / function / task / scenario can correspond to the training dataset requested by the terminal device.

[0304] For example, the network device selects a data augmentation method based on the AI ​​CSI model / function / task / scenario requested by the terminal device. This can be achieved by the network device selecting one or more configurable data augmentation methods corresponding to the AI ​​CSI model / function / task / scenario requested by the terminal device. The AI ​​CSI model / function / task / scenario requested by the terminal device may correspond to a training dataset requested by the terminal device. One or more configurable data augmentation methods can serve as a possible implementation of at least one data processing scheme.

[0305] Several data augmentation methods are exemplified below. For ease of description, the methods are numbered #1, #2, and #3.

[0306] In the following text, data augmentation method #1, data augmentation method #2, and data augmentation method #3 can each be considered as a possible implementation of a data processing scheme; AI CSI model / function / task / scenario can be considered as a possible implementation of at least one model and / or at least one function.

[0307] The AI ​​CSI model / function / task / scenario includes one or more of the following: AI CSI spatial frequency domain compression, AI CSI spatial time frequency domain compression, AI CSI prediction, AI CSI multi-slot spatial time frequency domain compression, AI CSI prediction + single-slot CSI compression, or AI CSI prediction + multi-slot CSI compression. AI CSI spatial-frequency domain compression can be understood as the joint compression of CSI in both the spatial and frequency domains based on an AI model; AI CSI spatial-temporal-frequency domain compression can be understood as the joint compression of CSI in the spatial, frequency, and time domains based on an AI model, or the joint compression of CSI in the spatial and frequency domains based on an AI model utilizing time-domain correlation; AI CSI prediction can be understood as predicting CSI based on an AI model; AI CSI multi-slot spatial-temporal-frequency domain compression can be understood as the joint compression of CSI in the spatial, frequency, and time domains based on an AI model utilizing time-domain correlation; AI CSI prediction + single-slot CSI compression can be understood as predicting CSI based on an AI model and compressing the predicted and / or measured single-slot CSI; AI CSI prediction + multi-slot CSI compression can be understood as predicting CSI based on an AI model and compressing the predicted and / or measured multi-slot CSI. Multi-slot CSI compression and / or spatial-temporal-frequency domain compression can utilize the time-domain correlation of CSI.

[0308] Figure 8 is a schematic diagram of a CSI model / function / task / scenario application applicable to an embodiment of this application.

[0309] For example, the CSI is input to the CSI compression model on the terminal device side. The terminal device side CSI compression model processes the input CSI and outputs CSI feedback bits. The CSI feedback bits are used as input to the CSI reconstruction model on the network device side. The network device side CSI reconstruction model processes the input CSI feedback bits and outputs the reconstructed CSI. The CSI input to the terminal device side CSI compression model may include the measured CSI and / or the predicted CSI; the CSI feedback bits output by the terminal device side CSI compression model may include CSI feedback information; the reconstructed CSI output by the network device side CSI reconstruction model may include the recovery result of the CSI feedback information.

[0310] For example, in the case where the CSI compression model on the input terminal device side includes the predicted CSI, the predicted CSI can be obtained based on the CSI prediction model. As shown in Figure 8, the CSI (for RS measurement results H) of historical times (e.g., times tk to t-1 shown in Figure 8, k>1) and / or the current time (e.g., time t shown in Figure 8) is calculated. t-k ~H t-1 and / or H t The input is the CSI prediction model, which can output (predict) the CSI at one or more future times (e.g., predicting the measurement result H of RS at time t+1 in Figure 8). t+1 and / or the measurement result H of RS at time t+2 t+2 After outputting the predicted CSI, one or more CSI(s) can be compressed into CSI feedback bits using the CSI compression model on the terminal device side and reported to the network device side.

[0311] For example, if the CSI of the CSI compression model on the input terminal device side includes the measured CSI, the CSI compression model on the terminal device side can be used to compress one or more CSI(s) into CSI feedback bits and report them to the network device side.

[0312] For example, if the CSI compression model on the input terminal device side includes the measured CSI, the terminal device-side CSI compression model can also be used to predict the CSI at future times based on the measured CSI. After predicting the CSI at future times, the terminal device-side CSI compression model can compress one or more CSI(s) into CSI feedback bits and report them to the network device side. In this case, the CSI prediction model or the model's functionality can be deployed within the CSI compression model; or, the CSI compression model can have both compression and prediction capabilities.

[0313] The network device decompresses the feedback bit using the CSI reconstruction model. The reconstructed CSI obtained from the decompression can then be used in the design of the transmitter on the network device side.

[0314] For example, H t+1 (or H) t+1 The result V obtained through singular value decomposition (SVD) t+1 As input to the CSI compression model, the CSI compression model compresses H. t+1 (or V) t+1 The compressed result is then output as the input (CSI feedback bits) to the CSI reconstruction model. The CSI reconstruction model decompresses the compressed result and outputs the CSI, which includes... (or ).

[0315] For example, H t+1 and H t+2 The result V obtained after SVD t+1 V t+2 As input to the CSI compression model, the CSI compression model compresses V. t+1 V t+2 The CSI feedback information is then output, serving as the input (CSI feedback bits) to the CSI reconstruction model. After decompressing the CSI feedback information, the CSI reconstruction model outputs the reconstructed CSI, which includes...

[0316] The network device can also predict future CSI based on the reconstructed CSI. For example, the reconstructed CSI can be input into the network device's CSI prediction model to obtain the future CSI. The future CSI predicted by the network device can also be used for transmitter design.

[0317] The CSI prediction model or model functionality on the network device side used to predict CSI at future moments can be deployed within the CSI reconstruction model (e.g., the CSI reconstruction model has decompression and prediction functions), and the CSI prediction model or model functionality on the network device side used to predict CSI at future moments can also be deployed outside the CSI reconstruction model, without any restrictions.

[0318] The CSI compression model on the terminal device side and / or the CSI reconstruction model on the network device side in Figure 8 can be an AI model or a model based on a non-AI method, without any restrictions.

[0319] The following section introduces data augmentation methods using AI CSI models, functions, tasks, and scenarios.

[0320] For example, in AI CSI spatial frequency domain compression, to extend the model's generalization ability to handle sub-band phase changes and preserve the phase relative information between sub-bands, the original training sample is H, and the data H′ obtained through data augmentation method #1 satisfies: H′=H*e jθ ,θ∈(0,2π).

[0321] For example, in AI CSI spatiotemporal frequency domain compression, to extend the model's generalization ability to handle phase changes in different time slots and preserve the phase relative relationship information between different time slots, phase alignment between time slots is required. Therefore, data augmentation methods need to perform the same phase rotation on the channels of all time slots. The original training samples are [H1,H2,H3], and the data [H′1,H′2,H′3] obtained through data augmentation method #2 satisfies: [H′1,H′2,H′3]=[H1,H2,H3]*e jθ ,θ∈(0,2π).

[0322] For example, for AI CSI prediction, in order to extend the model's generalization ability to cope with phase changes in different time slots, the original training samples are [H1, H2, H3], and the data [H′1, H′2, H′3] obtained through data augmentation method #3 satisfies:

[0323] Data augmentation method #3 can simulate and adapt to phase changes in different time slots using AI models.

[0324] Data augmentation methods and their corresponding numbers can also be represented in tabular form. For example, Table 5 shows an example of the correspondence between the numbers of data augmentation methods and the data augmentation methods.

[0325] Table 5

[0326] It should be understood that Table 5 is only an example, and the correspondence between the data augmentation method number and the data augmentation method can also be in other forms, or the data augmentation method number and the data augmentation method can also satisfy other correspondences, which are not limited.

[0327] The following text uses data augmentation method numbers to indicate their corresponding data augmentation methods. The correspondence between data augmentation method numbers and data augmentation methods can be found in Table 5.

[0328] Table 6 illustrates an example of the correspondence between AI CSI models / functions / tasks / scenarios and data augmentation methods. Using Table 6 as an example, a network device can select a data augmentation method based on it. For instance, if the AI ​​CSI model / function / task requested by the terminal device is AI CSI spatiotemporal-frequency domain compression, the network device can select either data augmentation method #1 or data augmentation method #2 from the configurable data augmentation methods #1 and #2 corresponding to AI CSI spatiotemporal-frequency domain compression, and indicate the selected method to the terminal device.

[0329] Table 6

[0330] Table 6 can be considered as one possible implementation of the correspondence between Table 2 and the first correspondence.

[0331] Table 7 shows another example of the correspondence between AI CSI models / functions / tasks and data augmentation methods. Using Table 7 as an example, a network device can select a data augmentation method based on it. For instance, if the number of compressed time slots corresponding to the AI ​​CSI model / function / task requested by the terminal device is ≥Q, the network device can select either Data Augmentation Method #1 or Data Augmentation Method #2 from the configurable Data Augmentation Method #1 and Data Augmentation Method #2 corresponding to "Number of Compressed Time Slots ≥ Q", and indicate the selected method to the terminal device.

[0332] Table 7

[0333] Table 7 can serve as one possible implementation of the correspondence between Table 3 and the second table. The relationship between the number of compressed time slots and Q can serve as one possible implementation of whether M satisfies the numerical range. The number of compressed time slots can serve as one possible implementation of the value of M.

[0334] Table 8 shows another example of the correspondence between AI CSI models / functions / tasks and data augmentation methods. Using Table 8 as an example, a network device can select a data augmentation method based on it. For instance, if the AI ​​CSI model / function / task requested by the terminal device is AI CSI multi-slot spatiotemporal domain compression, and the number of compressed time slots is ≥Q, the network device can select either Data Augmentation Method #1 or Data Augmentation Method #2 from the configurable Data Augmentation Method #1 and Data Augmentation Method #2 corresponding to "AI CSI multi-slot spatiotemporal domain compression, number of compressed time slots ≥Q", and indicate the selected method to the terminal device.

[0335] Table 8

[0336] Table 8 can be considered as a possible implementation of the correspondence between Table 4 and the third correspondence.

[0337] In this embodiment, the number of compressed time slots can be indicated by the terminal device to the network device via indication information; alternatively, the number of compressed time slots can be determined based on the AI ​​CSI model / function / task / scenario requested by the terminal device. For example, the terminal device sends indication information to the network device, indicating that the number of compressed time slots is 3; or, if the number of compressed time slots corresponding to the AI ​​CSI model / function / task / scenario requested by the terminal device is 2, the network device can determine that the number of compressed time slots is 2 based on the AI ​​CSI model / function / task / scenario requested by the terminal device. This embodiment does not limit this aspect.

[0338] In this embodiment, Q can be predefined by the protocol, preconfigured, or negotiated between the terminal device and the network device. Both the number of compressed time slots and Q are integers greater than 1.

[0339] It should be understood that Tables 6, 7, or 8 are merely examples. The correspondence between AI CSI models / functions / tasks / scenarios and data augmentation methods can also take other forms, or the correspondence between AI CSI models / functions / tasks / scenarios and data augmentation methods can also satisfy other forms. No limitation is imposed on this.

[0340] There are multiple data augmentation use cases for AI CSI feedback. For example, based on the compressed information and whether historical information is used, the cases can be categorized as shown in Table 9.

[0341] Table 9

[0342] In the above AI CSI models / functions / tasks / scenarios, AI CSI spatial frequency domain compression corresponds to case 0; AI CSI spatial temporal frequency domain compression corresponds to case 2; AI CSI prediction + single-slot CSI compression corresponds to case 3; and AI CSI prediction + multi-slot CSI compression corresponds to case 3.

[0343] Figure 9 is a schematic diagram of another communication method provided in an embodiment of this application.

[0344] The communication between the first communication device and the second communication device was described above with reference to Figure 6. The first and / or second communication devices can be devices on the network device side and / or terminal device side, or they can be AI entities. The following description, with reference to Figure 9, exemplarily describes the method 900 provided in this application embodiment when the first communication device is an intelligent network element and the second communication device is an OTT server. The intelligent network element can be one possible implementation of the first network element; the OTT server can be one possible implementation of the second network element.

[0345] As shown in Figure 9, the method 900 includes the following steps.

[0346] S910, the OTT server sends request information #1 to the UE. Correspondingly, the UE receives request information #1 sent by the OTT server. Request information #1 is used to request the training data corresponding to model #5.

[0347] S920, the UE sends request information #1 to the base station, and the base station receives request information #1 sent by the UE.

[0348] S930, the base station sends request information #1 to the intelligent network element, and the corresponding intelligent network element receives request information #1 sent by the base station.

[0349] S940, the intelligent network element sends training data #1 to the base station, and the corresponding base station receives the training data #1 sent by the intelligent network element.

[0350] S950, the base station sends training data #1 to the UE, and the UE receives the training data #1 sent by the base station.

[0351] S960, the UE sends training data #1 to the OTT server, and the OTT server receives the training data #1 sent by the UE.

[0352] Method 900 also includes:

[0353] S970, the intelligent network element sends data enhancement method #4 to the base station, and the corresponding base station receives data enhancement method #4 sent by the intelligent network element.

[0354] S980, the base station sends data enhancement method #4 to the UE, and correspondingly, the UE receives data enhancement method #4 sent by the base station.

[0355] S990, the UE sends data enhancement method #4 to the OTT server, and the corresponding OTT server receives data enhancement method #4 sent by the UE.

[0356] Method 900 may also include:

[0357] S993, the OTT server performs data augmentation based on data augmentation method #4 and training data #1 to obtain augmented data.

[0358] Method 900 may also include:

[0359] S996, the OTT server trains model #5 based on augmented data and / or training data #1.

[0360] It is understood that, for any of the intelligent network elements, base stations, or UEs, data augmentation method #4 and training data #1 can be sent simultaneously or separately (i.e., S960 and S990 can be executed separately or simultaneously; or S950 and S980 can be executed separately or simultaneously; or S940 and S970 can be executed separately or simultaneously); and / or, data augmentation method #4 and training data #1 can be sent in the same data packet or in different data packets (i.e., the information sent in S960 and S990 can be sent in the same data packet or in different data packets; or the information sent in S950 and S980 can be sent in the same data packet or in different data packets; or the information sent in S940 and S970 can be sent in the same data packet or in different data packets). This application embodiment does not impose any limitations on this.

[0361] Data augmentation method #4 can be considered as a possible implementation of the first data processing scheme.

[0362] Training data #1 can be used as one possible implementation of the training dataset.

[0363] Model #5 can be considered as one possible implementation of the first model and / or the first function.

[0364] The determination of data augmentation method #4 can refer to method #1, method #2 or method #3.

[0365] The specific implementation of communication method 900 can be found in communication method 600, and will not be elaborated here.

[0366] Figure 10 is a schematic diagram of another communication method provided in an embodiment of this application.

[0367] As shown in Figure 10, the UE can establish a communication connection with the network device through the communication control module. The UE can receive signaling and / or reference signals sent by the network device through the antenna and receiver, and send CSI reports to the network device through the transmitter and antenna. The UE can also process the measurement results of each measurement through the processor and store the measurement results in the memory.

[0368] The various functions of the UE can be deployed in different parts of the chip.

[0369] The communication method embodiment of this application has been described in detail above with reference to Figures 6 to 9. The communication device embodiment of this application will now be described in detail with reference to Figures 11 and 12. It should be understood that the description of the device embodiment corresponds to the description of the method embodiment; therefore, any parts not described in detail can be referred to the preceding method embodiment.

[0370] Figure 11 is a schematic block diagram of a communication device 1000 provided in an embodiment of this application. As shown in Figure 11, the communication device 1000 includes a communication module 1020. The communication device 1000 can be a transmitting device, or a communication device applied to or used in conjunction with a transmitting device to implement a method executed by the transmitting device, such as a chip, chip system, or circuit; or, the communication device 1000 can be a receiving device, or a communication device applied to or used in conjunction with a receiving device to implement a method executed by the receiving device, such as a chip, chip system, or circuit.

[0371] The communication module can also be called a transceiver module, transceiver, transceiver unit, or transceiver device. The processing module can also be called a processor, processing board, processing unit, or processing device. Optionally, the communication module is used to execute the sending and receiving operations of the sending and receiving devices in the above method. The device in the communication module that implements the receiving function can be considered a receiving unit, and the device in the communication module that implements the sending function can be considered a sending unit; that is, the communication module includes a receiving unit and a sending unit.

[0372] Optionally, the communication device 1000 may further include a processing module 1010 for processing data.

[0373] Optionally, the communication device 1000 may also include a storage module for storing device program code and / or data.

[0374] In one example, when the communication device 1000 is applied to a first communication device (e.g., a terminal device or a network device), the processing module 1010 can be used to implement the processing function of the first communication device in the above embodiments, and the communication module 1020 can be used to implement the sending and receiving function of the first communication device in the above embodiments.

[0375] In another example, when the communication device 1000 is applied to a second communication device (e.g., a network device or a terminal device), the processing module 1010 can be used to implement the processing function of the second communication device in the above embodiments, and the communication module 1020 can be used to implement the sending and receiving function of the second communication device in the above embodiments.

[0376] Furthermore, it should be noted that the aforementioned communication module and / or processing 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. Alternatively, the processing module or communication module can also be implemented through physical devices, such as chips / circuits (e.g., integrated circuits or logic circuits). 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 is an integrated processor, microprocessor, or circuit (e.g., integrated circuits or logic circuits).

[0377] The module division in this application is illustrative and represents only one logical functional division. In actual implementation, other division methods are possible. Furthermore, the functional modules in the various examples of this application 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.

[0378] In one example, the functional unit in any of the above devices may be one or more integrated circuits configured to implement the above methods, such as: one or more application-specific integrated circuits (ASICs), or one or more central processing units (CPUs), one or more microcontroller units (MCUs), one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs), or a combination of at least two of these integrated circuit forms.

[0379] In one example, the storage module may include random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory and / or registers, etc.

[0380] Figure 12 is a schematic block diagram of a communication device 2000 provided in an embodiment of this application. Optionally, the communication device 2000 may be a chip or a chip system. Optionally, in this application, the chip system may be composed of chips or may include chips and other discrete devices.

[0381] As shown in Figure 12, the communication device 2000 can be used to implement the functions of any device (e.g., terminal device, network device) in the communication system described in the foregoing examples.

[0382] The communication device 2000 includes a communication interface 2030, through which the communication device 2000 can interact with other devices. For example, the communication interface 2030 can be a transceiver, circuit, bus, module, pin, or other type of communication interface.

[0383] The communication device 2000 may also include at least one processor 2010 (one processor 2010 is shown as an example in FIG12). Optionally, the processor 2010 is coupled to a memory, which may be located within the device, or the memory may be integrated with the processor, or the memory may be located outside the device.

[0384] The communication device 2000 may further include at least one memory 2020. The memory 2020 stores computer programs, computer programs or instructions and / or data necessary for implementing any of the above examples; the processor 2010 may execute the computer programs stored in the memory 2020 to perform the methods in any of the above examples.

[0385] When the communication device 2000 is a chip-type device or circuit, the communication interface 2030 in the device 2000 can also be an input / output circuit, which can input information (or receive information) and output information (or send information). The processor 2010 is an integrated processor, microprocessor, integrated circuit or logic circuit, etc. The processor can determine the output information based on the input information.

[0386] In one example, when the communication device 2000 is applied to a first communication device (e.g., a terminal device or a network device), the processor 2010 can be used to implement the processing function of the first communication device in the above embodiments, and the communication interface 2030 can be used to implement the sending and receiving function of the first communication device in the above embodiments.

[0387] In another example, when the communication device 2000 is applied to a second communication device (e.g., a network device or a terminal device), the processor 2010 can be used to implement the processing function of the second communication device in the above embodiments, and the communication interface 2030 can be used to implement the sending and receiving function of the second communication device in the above embodiments.

[0388] The coupling in this application refers to indirect coupling or communication connection between devices, units, or modules, which can be electrical, mechanical, or other forms, used for information exchange between devices, units, or modules. The processor 2010 may operate in conjunction with the memory 2020 and the communication interface 2030. This application does not limit the specific connection medium between the processor 2010, the memory 2020, and the communication interface 2030.

[0389] Optionally, as shown in FIG12, the processor 2010, the memory 2020, and the communication interface 2030 are interconnected via a bus 2040. Optionally, the bus may include buses of the types such as address bus, data bus, and control bus. Furthermore, for ease of illustration, FIG12 shows one bus 2040, but does not indicate that there is only one bus or only one type of bus.

[0390] It should be understood that the processor mentioned in the embodiments of this application can be one of the following devices or a portion of the circuitry used for processing functions: a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. Some or all steps of the communication method in the embodiments of this application can be implemented by a graphics processing unit (GPU), or by a GPU in conjunction with other processors.

[0391] It should also be understood that the memory mentioned in the embodiments of this application can be volatile memory and / or non-volatile memory. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM). For example, RAM can be used as an external cache. By way of example and not limitation, RAM includes the following forms: static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (SLDRAM), and direct rambus RAM (DR RAM).

[0392] It should be noted that when the processor is a general-purpose processor, DSP, ASIC, FPGA, or other programmable logic device, discrete gate or transistor logic device, or discrete hardware component, the memory (storage module) can be integrated into the processor.

[0393] It should also be noted that the memory described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0394] This application also provides a chip, which includes one or more processors and a communication interface. The processor reads computer programs or instructions stored in the memory through the communication interface and executes the methods executed by the communication device (e.g., a first communication device and / or a second communication device) in the above-described method embodiments.

[0395] This application also provides a computer-readable storage medium storing computer instructions for implementing the methods executed by a communication device (e.g., a first communication device and / or a second communication device) in the above-described method embodiments.

[0396] This application also provides a computer program product comprising instructions which, when executed by a computer, implement the methods described above as being performed by a communication device (e.g., a first communication device and / or a second communication device).

[0397] This application also provides a communication system, which includes the first communication device and / or the second communication device described in the above embodiments.

[0398] The explanations and beneficial effects of the relevant contents in any of the devices provided above can be found in the corresponding method embodiments provided above, and will not be repeated here.

[0399] In the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0400] This application will present various aspects, embodiments, or features relating to systems that may include multiple devices, components, modules, etc. It should be understood and appreciated that individual systems may include additional devices, components, modules, etc., and / or may not include all the devices, components, modules, etc. discussed in conjunction with the accompanying drawings. Furthermore, combinations of these approaches are also possible.

[0401] In this application, examples may reference each other without logical contradiction. For example, methods and / or terms between method embodiments may reference each other, functions and / or terms between device embodiments may reference each other, and functions and / or terms between device examples and method examples may reference each other.

[0402] It should be understood that the above embodiments are mainly illustrated using devices in existing network architectures as examples, and the specific form of the devices is not limited in the embodiments of this application. For example, any device that can achieve the same function in the future is applicable to the embodiments of this application.

[0403] 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.

[0404] 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 described again here.

[0405] 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.

[0406] 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.

[0407] 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.

[0408] 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 existing solutions, 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, external hard drives, ROM, RAM, magnetic disks, or optical disks.

[0409] 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 by comprising: Comprising: receiving first indication information, the first indication information indicating a first model and / or a first function, the first model and / or the first function being used for channel state information, CSI, processing; sending second indication information, the second indication information indicating a first data processing scheme, the first data processing scheme being determined according to the first indication information, the first data processing scheme being used for determining enhancement data, the enhancement data being used for training at least one model corresponding to the first model and / or the first function.

2. The method of claim 1, wherein, The first data processing scheme is determined according to the first indication information, comprising: The first data processing scheme is determined according to the first model and / or the first function and a first correspondence relationship, the first correspondence relationship indicating a relationship between at least one model and / or at least one function and at least one data processing scheme used for determining enhancement data, the at least one model and / or the at least one function comprising the first model and / or the first function, the at least one data processing scheme comprising the first data processing scheme corresponding to the first model and / or the first function.

3. The method of claim 1, wherein, The first data processing scheme is determined according to the first indication information, comprising: The first data processing scheme is determined according to M and a second correspondence relationship, the second correspondence relationship indicating a relationship between a plurality of numerical value ranges and at least one data processing scheme used for determining enhancement data, the plurality of numerical value ranges comprising a first numerical value range, the value of M being located in the first numerical value range, the at least one data processing scheme comprising the first data processing scheme corresponding to the first numerical value range, M being a positive integer, The M is determined according to the first model and / or the first function; and / or, The M is indicated by the first indication information.

4. The method of claim 1, wherein, The first data processing scheme is determined according to the first indication information, comprising: The first data processing scheme is determined according to the first model and / or the first function, M and a third correspondence relationship, the third correspondence relationship indicating a relationship between a plurality of numerical value ranges, at least one model and / or at least one function and at least one data processing scheme used for determination of enhancement data, the plurality of numerical value ranges comprising a first numerical value range, the value of the M being located in the first numerical value range, the at least one model and / or the at least one function comprising the first model and / or first function, the at least one data processing scheme comprising the first data processing scheme corresponding to the M and the first model and / or the first function, M being a positive integer.

5. The method according to claim 3 or 4, characterized in that, The value of the M is located in the first numerical value range, comprising at least one of: The value of the M is located between a first threshold value and a second threshold value, the first threshold value and the second threshold value being unequal, the first threshold value and / or the second threshold value being greater than 0; The value of the M is greater than or equal to a first threshold value, the first threshold value being greater than 0; Or, The value of the M is less than or equal to a second threshold value, the second threshold value being greater than 0.

6. The method according to any one of claims 1 to 5, characterized in that, The first model and / or the first function are used for CSI processing, including at least one of the following: The first model and / or the first function are used for CSI compression. The first model and / or the first function are used for CSI decompression; or The first model and / or the first function are used for determining CSI of at least one time unit after at least one time unit based on the CSI of the at least one time unit.

7. The method according to any one of claims 1 to 6, characterized in that, The first data processing scheme is used to determine enhancement data, including any of the following: The enhancement data is obtained by phase rotation on original training data used to obtain the enhancement data; The enhancement data is obtained by phase rotation on channel matrices of different time units in the same time sequence in original training data used to obtain the enhancement data at the same angle; or The enhancement data is obtained by phase rotation on channel matrices of different time units in the same time sequence in original training data used to obtain the enhancement data at different angles.

8. The method according to any one of claims 1 to 7, characterized in that, The first data processing scheme is used to determine enhancement data, including any of the following: H' = H * e jθ , θ ∈ (0, 2π), wherein the H represents original training data used to derive the augmented data, and the H' represents the augmented data; [H'1, H'2, H'3] = [H1, H2, H3] * e jθ , θ ∈ (0, 2π), wherein the [H1, H2, H3] represents original training data used to obtain the augmented data, and the [H'1, H'2, H'3] represents the augmented data; or, Wherein, [H1, H2, H3] represents original training data used to obtain the enhancement data, and [H'1, H'2, H'3] represents the enhancement data.

9. The method according to any one of claims 1 to 8, characterized in that, The method further comprises: receiving first request information, the first request information being used to request at least one of a training data set, model parameters of a second model, or model parameters of at least one model corresponding to a second function, the training data set being a training data set of at least one model corresponding to the first model and / or the first function, the second model and / or the second function being used to train at least one model corresponding to the first model and / or the first function; sending third indication information, the third indication information being used to indicate at least one of the training data set, the model parameters of the second model, or the model parameters of at least one model corresponding to the second function.

10. A communication method characterized by comprising: including: sending first indication information, the first indication information indicating a first model and / or a first function, the first model and / or the first function being used for CSI processing; receiving second indication information, the second indication information indicating a first data processing scheme, the first data processing scheme being determined according to the first indication information, the first data processing scheme being used to determine enhancement data, the enhancement data being used to train at least one model corresponding to the first model and / or the first function.

11. The method of claim 10, wherein, The first data processing scheme is determined according to the first indication information, including: The first data processing scheme is determined according to the first model and / or the first function and a first correspondence relationship, the first correspondence relationship indicating a relationship between at least one model and / or at least one function and at least one data processing scheme used for determining enhancement data, the at least one model and / or the at least one function including the first model and / or the first function, and the at least one data processing scheme including the first data processing scheme corresponding to the first model and / or the first function.

12. The method of claim 10, wherein, The first data processing scheme is determined according to the first indication information, including: The first data processing scheme is determined according to M and a second correspondence relationship, the second correspondence relationship indicating a relationship between a plurality of numerical value ranges and at least one data processing scheme used for determining enhancement data, the plurality of numerical value ranges including a first numerical value range, a value of the M being located in the first numerical value range, the at least one data processing scheme including the first data processing scheme corresponding to the first numerical value range, and the M being a positive integer, The M is determined according to the first model and / or the first function; and / or, The M is indicated by the first indication information.

13. The method of claim 10, wherein, The first data processing scheme is determined according to the first indication information, including: The first data processing scheme is determined according to the first model and / or the first function, M and a third correspondence relationship, the third correspondence relationship indicating a relationship between a plurality of numerical value ranges, at least one model and / or at least one function and at least one data processing scheme used for determining enhancement data, the plurality of numerical value ranges including a first numerical value range, a value of the M being located in the first numerical value range, the at least one model and / or the at least one function including the first model and / or the first function, and the at least one data processing scheme including the first data processing scheme corresponding to the first numerical value range and the first model and / or the first function, the M being a positive integer.

14. The method according to claim 12 or 13, characterized in that, The value of the M is located in the first numerical value range, including at least one of the following: The value of the M is located between a first threshold value and a second threshold value, the first threshold value and the second threshold value being different, and the first threshold value and / or the second threshold value being greater than 0; The value of the M is greater than or equal to a first threshold value, the first threshold value being greater than 0; Or, The value of the M is less than or equal to a second threshold value, the second threshold value being greater than 0.

15. The method according to any one of claims 10 to 14, characterized in that, The first model and / or the first function are used for CSI processing, including at least one of the following: The first model and / or the first function are used for CSI compression; The first model and / or the first function are used for CSI decompression; or The first model and / or the first function are used for determining CSI of at least one time unit after at least one time unit based on the at least one time unit.

16. The method according to any one of claims 10 to 15, characterized in that, The first data processing scheme is used for determining enhancement data, including any one of the following: The enhancement data is obtained by phase rotation on original training data used to obtain the enhancement data; The enhanced data is obtained by performing phase rotation of the same angle on channel matrices of different time units in the same time sequence in original training data used to obtain the enhanced data; or The enhanced data is obtained by performing phase rotation of different angles on channel matrices of different time units in the same time sequence in original training data used to obtain the enhanced data.

17. The method according to any one of claims 10 to 16, characterized in that, The first data processing scheme is used to determine enhanced data, including any one of the following: H' = H * e jθ , θ ∈ (0, 2π), where H represents original training data used to derive the augmented data, and H' represents the augmented data. [H'1, H'2, H'3] = [H1, H2, H3] * e jθ , θ ∈ (0, 2π), wherein the [H1, H2, H3] represents original training data used to obtain the augmented data, and the [H'1, H'2, H'3] represents the augmented data; or, Wherein, [H1, H2, H3] represents original training data used to obtain the enhanced data, and [H'1, H'2, H'3] represents the enhanced data.

18. The method according to any one of claims 10 to 17, characterized in that, The method further comprises: sending first request information, the first request information being used to request at least one of a training data set, model parameters of a second model, or model parameters of at least one model corresponding to a second function, the training data set being a training data set of the first model and / or at least one model corresponding to the first function, the second model and / or the second function being used to train the first model and / or at least one model corresponding to the first function; receiving third indication information, the third indication information being used to indicate at least one of the training data set, the model parameters of the second model, or the model parameters of at least one model corresponding to the second function.

19. The method according to any one of claims 10 to 18, characterized in that, The method further comprises at least one of the following: training the first model according to the enhanced data; training at least one model corresponding to the first function according to the enhanced data; training the first model according to a training data set; training at least one model corresponding to the first function according to a training data set; training the first model according to model parameters of a second model; training at least one model corresponding to the first function according to model parameters of a second model; training the first model according to model parameters of at least one model corresponding to a second function; or training at least one model corresponding to the first function according to model parameters of at least one model corresponding to a second function.

20. A communications device, characterized by The computer program product comprises a computer program, and a computer readable storage medium storing the computer program.

21. A communications device, characterized by The computer readable storage medium is used to store a computer program, when the computer program is run on a computer, so that the method of any one of claims 1 to 19 is executed.

22. The communication apparatus according to claim 21, wherein The communication apparatus further comprises a memory, the memory being used to store the computer program or instructions; and / or The communication apparatus further comprises a communication interface, the communication interface being coupled with at least one of the processors, the communication interface being used to input and / or output information.

23. A computer-readable storage medium, characterized in that, The computer readable storage medium is used to store a computer program, when the computer program is run on a computer, so that the method of any one of claims 1 to 19 is executed. The computer readable storage medium is used to store a computer program, when the computer program is run on a computer, so that the method of any one of claims 1 to 19 is executed.

24. A computer program product, characterised in that, comprising computer program or instructions which, when executed by a processor, cause the method of any one of claims 1 to 19 to be performed.

Images