Calibration method and apparatus

Abstract

The present application provides a calibration method and apparatus, which includes the steps of determining first channel state information and second channel state information, where the first channel state information is obtained through prediction by a first channel prediction model, and the second channel state information is determined based on a channel estimation for a downlink signal from an access network device, and, if a difference metric value between the first channel state information and the second channel state information is equal to or greater than a first threshold, transmitting the first information to the access network device, where the first information indicates the second channel state information.

Description

[Technical Field]

[0001] [CROSS-REFERENCE TO RELATED APPLICATIONS] This application claims priority to Chinese Patent Application No. 202111406622.0, entitled "CALIBRATION METHOD AND APPARATUS," filed with the State Intellectual Property Office of China on November 24, 2021, which is incorporated herein by reference in its entirety.

[0002] [Technical field] The present application relates to the field of communications technology, and in particular to calibration methods and apparatus. [Background technology]

[0003] In wireless communication networks, e.g., mobile communication networks, the services supported by the network are becoming increasingly diverse, and therefore the requirements that must be met are also becoming increasingly diverse. For example, networks must be able to support ultra-high speeds, ultra-low latency, and / or extremely large connections. This characteristic increases the complexity of network planning, network calibration, and / or resource scheduling. In addition, as network capabilities become increasingly powerful, energy conservation in networks has become a hot research topic, e.g., by supporting increasingly higher spectrum and new technologies such as high-order multiple-input multiple-output (MIMO) technology, beamforming, and / or beam management. These new requirements, scenarios, and capabilities pose unprecedented challenges to network planning, operation and maintenance, and efficient operation. To address this challenge, artificial intelligence (AI) technologies may be introduced into wireless communication networks to implement network intelligence. Based on this, how to effectively implement AI in networks is a worthy research topic. Summary of the Invention [Problem to be solved by the invention]

[0004] The present application provides a calibration method and apparatus for updating a channel prediction model and avoiding impact on communication performance. [Means for solving the problem]

[0005] According to a first aspect, the present application provides a calibration method. The method is applicable to a scenario in which channel state information is predicted based on a prediction model. The method is executed by a terminal device or a module in the terminal device. Here, an example in which the method is executed by a terminal device is used for explanation. The method includes: determining, by the terminal device, first channel state information and second channel state information, where the first channel state information is obtained through prediction using a first channel prediction model and the second channel state information is determined based on channel estimation for a downlink signal from an access network device; and transmitting, by the terminal device, the first information to the access network device, where the first information indicates the second channel state information, if a difference metric value between the first channel state information and the second channel state information is greater than or equal to a first threshold.

[0006] According to the above method, when the deviation between the predicted first channel state information and the actually measured second channel state information is large, the first channel prediction model can be updated to avoid affecting communication performance. In addition, the fed back channel state information can be further used for online training and updating of the channel prediction model to continuously improve the performance of the channel prediction model.

[0007] In a possible design, if a differential metric value between the first channel state information and the second channel state information is less than a second threshold, second information is sent to the access network device, where the second information indicates that the prediction result of the first channel prediction model is accurate and the second threshold is less than or equal to the first threshold.

[0008] In a possible design, the method further includes: training a target channel prediction model based on second channel state information to obtain a second channel prediction model, where an initial value of the target channel prediction model is the first channel prediction model; and updating the target channel prediction model to the second channel prediction model.

[0009] In a possible design, the information about the second channel prediction model includes variation information between the second channel prediction model and the first channel prediction model.

[0010] In a possible design, third information is sent to the access network device, where the third information indicates information related to the second channel prediction model.

[0011] In a possible design, transmitting the third information to the access network device includes transmitting the third information to the access network device when determining that variation information between the first channel prediction model and the second channel prediction model is greater than or equal to a third threshold.

[0012] According to the above method, when the variation information between the updated channel prediction model and the channel prediction model previously instructed to the access network device is less than a third threshold, the updated channel prediction model does not need to be instructed to the access network device, thereby avoiding frequent instructing of the updated channel prediction model to the access network device, reducing signaling overhead and improving system efficiency.

[0013] In one possible design, the method further includes updating the first channel prediction model to the second channel prediction model.

[0014] In a possible design, the method further includes receiving fourth information from the access network device, where the fourth information indicates information related to a third channel prediction model.

[0015] According to the above method, the access network device can ensure that communication efficiency is improved by updating the first channel prediction model, and can avoid communication interruptions caused by the access network device being unable to obtain an updated prediction model for a long period of time.

[0016] In one possible design, the method further includes updating the first channel prediction model to a third channel prediction model.

[0017] In a possible design, the information about the third channel prediction model includes variation information between the third channel prediction model and the first channel prediction model.

[0018] In a possible design, the first channel prediction model is configured by the access network device, and the first channel prediction model is identical to the channel prediction model in the access network device.

[0019] The same prediction model is configured in the access network device and the terminal device, so that the terminal device can determine whether the prediction model configured in the access network device is accurate based on downlink measurement information obtained through actual measurements, and when the prediction model is inaccurate, calibrate the prediction model of the access network device in real time.

[0020] In a possible design, the first information further indicates a time unit corresponding to the second channel state information.

[0021] According to a second aspect, the present application provides a calibration method. The method is applicable to a scenario in which channel state information is predicted based on a prediction model. The method is executed by an access network device or a module in the access network device. Here, an example in which the method is executed by the access network device is used for explanation. The method includes: an access network device transmitting a downlink signal to a terminal device; and receiving first information or second information from the terminal device, where the first information indicates second channel state information, and the second channel state information is determined based on channel estimation for the downlink signal; The second information indicates that the prediction result of the first channel prediction model is accurate.

[0022] In a possible design, when the first information is received, the method comprises: training the updated channel prediction model to obtain a third channel prediction model based on the second channel state information, where the initial value of the updated channel prediction model is the first channel prediction model; updating the update target channel prediction model to a third channel prediction model; Further includes:

[0023] In a possible design, the information about the third channel prediction model includes variation information between the third channel prediction model and the first channel prediction model.

[0024] In a possible design, fourth information is sent to the terminal device, where the fourth information indicates information related to a third channel prediction model.

[0025] In a possible design, transmitting the fourth information to the terminal device includes: and transmitting fourth information to the terminal device when it is determined that the variation information between the first channel prediction model and the third channel prediction model is equal to or greater than a third threshold.

[0026] In a possible design, the method further includes receiving third information from the terminal device, the third information indicating information related to the second channel prediction model, and updating the first channel prediction model to the second channel prediction model.

[0027] In a possible design, a differential metric value between the first channel state information and the second channel state information is greater than or equal to a first threshold, and the first channel state information is obtained through prediction using a first channel prediction model.

[0028] In a possible design, the first information further indicates a time unit corresponding to the second channel state information.

[0029] According to a third aspect, the present application provides a calibration method. The method is applicable to a scenario in which channel state information is predicted based on a prediction model. The method is executed by an access network device or a module within the access network device. Here, an example in which the method is executed by the access network device is used for explanation. The method includes: determining, by the access network device, first channel state information, where the first channel state information is obtained through prediction using the first channel prediction model; receiving, from a terminal device, first information, where the first information indicates second channel state information corresponding to a downlink signal; training the first channel prediction model based on the second channel state information to obtain the second channel prediction model when a difference metric value between the first channel state information and the second channel state information is equal to or greater than a first threshold; and updating the first channel prediction model to the second channel prediction model.

[0030] According to the above method, when the deviation between the predicted first channel state information and the actually measured second channel state information is large, the first channel prediction model can be updated to avoid affecting communication performance and continuously improve the performance of the channel prediction model.

[0031] In a possible design, the first information further indicates a time unit corresponding to the second channel state information.

[0032] According to a fourth aspect, the present application further provides a communication device. The communication device can implement any of the methods or implementations provided in the first aspect. The communication device may be implemented by hardware, software, or hardware executing corresponding software. The hardware or software may include one or more units or modules corresponding to the above-mentioned functions.

[0033] In a possible implementation, the communication device includes a processor configured to support the communication device in performing the method according to the first aspect. The communication device may further include a memory coupled to the processor and storing program instructions and data required by the communication device. Optionally, the communication device may further include an interface circuit configured to support the communication device in communicating with other communication devices.

[0034] In a possible implementation, the structure of the communication device includes a processing unit and a communication unit. These units may perform corresponding functions in the above-mentioned method examples. For details, please refer to the description of the method provided in the first aspect. Here, the details will not be described again.

[0035] According to a fifth aspect, the present application further provides a communication device. The communication device can implement any method or any implementation provided in the second or third aspect. The communication device may be implemented by hardware, software, or hardware executing corresponding software. The hardware or software may include one or more units or modules corresponding to the above-mentioned functions.

[0036] In a possible implementation, the communication device includes a processor. The processor is configured to support the communication device in performing the method according to the second or third aspect. The communication device may further include a memory. The memory is coupled to the processor and may store program instructions and data required by the communication device. Optionally, the communication device further includes an interface circuit, the interface circuit configured to support the communication device in communicating with other communication devices.

[0037] In a possible implementation, the structure of the communication device includes a processing unit and a communication unit. These units may perform the corresponding functions in the above-mentioned method examples. For details, please refer to the description of the method provided in the second aspect or the third aspect. Here, the details will not be described again.

[0038] According to a sixth aspect, there is provided a communications device including a processor and an interface circuit. The interface circuit is configured to receive a signal from another communications device other than the communications device and transmit the signal to the processor, or transmit a signal from the processor to another communications device other than the communications device. The processor is configured to execute a computer program or instructions stored in a memory to implement the method of any possible implementation of the first aspect. Optionally, the device further includes a memory, the memory storing the computer program or instructions.

[0039] According to a seventh aspect, there is provided a communications device including a processor and an interface circuit. The interface circuit is configured to receive a signal from another communications device other than the communications device and transmit the signal to the processor, or transmit a signal from the processor to another communications device other than the communications device. The processor is configured to execute a computer program or instructions stored in a memory to implement the method of any one of the second or third aspects and their possible implementations. Optionally, the device further includes a memory, the memory storing the computer program or instructions.

[0040] According to an eighth aspect, there is provided a computer-readable storage medium storing a computer program or instructions which, when executed on a computer, enables the computer to implement the method of any possible implementation of the first aspect.

[0041] According to a ninth aspect, there is provided a computer-readable storage medium storing a computer program or instructions which, when executed on a computer, enables the computer to implement the method of the second or third aspect and any one of its possible implementations.

[0042] According to a tenth aspect, there is provided a computer program product comprising computer-readable instructions that, when executed on a computer, enable the computer to implement the method of any possible implementation of the first aspect.

[0043] According to an eleventh aspect, there is provided a computer program product comprising computer readable instructions which, when executed on a computer, enable the computer to implement the method of the second or third aspect and any one of their possible implementations.

[0044] According to a twelfth aspect, there is provided a chip, the chip including a processor and optionally a memory, the processor being coupled to the memory and configured to execute computer programs or instructions stored in the memory to enable the chip to implement the method of any possible implementation of the first aspect.

[0045] According to a thirteenth aspect, there is provided a chip, the chip including a processor and optionally a memory, the processor being coupled to the memory and configured to execute computer programs or instructions stored in the memory to enable the chip to implement the method of any one of the second or third aspects and their possible implementations.

[0046] According to a fourteenth aspect, there is provided a communication system, the system including an apparatus (such as a terminal device) implementing the first aspect and an apparatus (such as an access network device) implementing the third aspect. [Brief explanation of the drawings]

[0047] [Figure 1] 1 is a diagram of the architecture of a communication system applicable to the present invention;

[0048] [Figure 2] FIG. 1 is a diagram of the hierarchical relationship of a neural network according to the present disclosure.

[0049] [Figure 3] FIG. 1 is an exemplary diagram of an application framework for AI in a communication system according to the present disclosure.

[0050] [Figure 4] FIG. 1 is an exemplary diagram of a network architecture to which the methods provided in the present disclosure may be applied.

[0051] [Figure 5] FIG. 1 is a diagram of a capability exchange according to the present disclosure.

[0052] [Figure 6] FIG. 1 is a diagram of the structure of a model according to the present disclosure.

[0053] [Figure 7] 1 is a schematic flow chart of a calibration method according to the present disclosure.

[0054] [Figure 8] FIG. 2 is a diagram of a downlink signal transmission according to the present disclosure.

[0055] [Figure 9] 1 is a schematic flowchart of a model updating method according to the present disclosure.

[0056] [Figure 10] 1 is a schematic flow chart of the calibration method of the present disclosure.

[0057] [Figure 11] 1 is a diagram of the structure of a communication device according to the present disclosure.

[0058] [Figure 12] 1 is a diagram of the structure of a communication device according to the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0059] Hereinafter, embodiments of the present application will be described in detail with reference to the accompanying drawings in this specification.

[0060] The technical solutions of the present disclosure may be applied to various communication systems, such as a long term evolution (LTE) system, a fifth generation (5G) mobile communication system, or a next generation mobile communication system, which is not limited herein. A 5G system may also be called a new radio (NR) system.

[0061] In this disclosure, an explanation is provided by using an example of an interaction between a terminal device and an access network device. The method provided in this disclosure can be applied not only to the interaction between a terminal device and a network side, but also to the interaction between any two devices. This is not limited in this disclosure.

[0062] To facilitate understanding of the present invention, a communication system applicable to the present disclosure will first be described in detail using the communication system shown in Figure 1 as an example. Figure 1 is a diagram of the architecture of a communication system to which the present disclosure can be applied. The communication system includes an access network device and a terminal device. The terminal device can establish a connection to the access network device and communicate with the access network device. Figure 1 is merely a diagram. The number of access network devices and terminal devices included in the communication system is not limited in the present disclosure.

[0063] To support artificial intelligence (AI) in wireless networks, a dedicated AI entity (also called an AI module) may be further introduced into the network. The AI ​​entity may correspond to an independent network element or may be located within a network element, such as a core network device, an access network device, or an operations, administration, and maintenance (OAM) device. For example, as shown in FIG. 1 , the AI ​​entity may be located outside the access network device and communicate with the access network device. The access network device may forward data related to the AI ​​model and reported by the terminal device to the AI ​​entity, which then performs operations such as training dataset construction and model training and forwards the trained artificial intelligence model to each terminal device via the access network device. In the present disclosure, the OAM is configured to operate, manage, and / or maintain the core network device (operation, administration, and maintenance of the core network device) and / or to operate, manage, and / or maintain the access network device (operation, administration, and maintenance of the access network device). For example, the present disclosure includes a first OAM and a second OAM. The first OAM is for operation, management, and maintenance of core network devices, and the second OAM is for operation, management, and maintenance of access network devices. The first OAM and / or the second OAM may include an AI entity. As another example, the present disclosure includes a third OAM, which is for operation, management, and maintenance of both core network devices and access network devices.

[0064] Optionally, the AI ​​entity may be integrated into the terminal or terminal chip to coordinate and support the AI.

[0065] Optionally, in the present disclosure, an AI entity may have another name and is mainly configured to implement an AI function (also referred to as an AI-related operation). The specific name of an AI entity is not limited in the present disclosure.

[0066] In this disclosure, an AI model is a specific method for implementing an AI function, and the AI ​​model represents a mapping relationship between the input and output of the model. The AI ​​model may be a neural network, a random forest model, a support vector machine model, a decision tree model, or another machine learning model. The AI ​​model may also be referred to as a model for short. The AI-related operations may include at least one of the following: data collection, model training, model information publication, model inference (also referred to as model inference or prediction), inference result publication, etc.

[0067] In this disclosure, a terminal device may be referred to as a terminal for short. A terminal device may communicate with one or more core networks via a radio access network (RAN).

[0068] In the present disclosure, a terminal device may be a device having wireless transmission and reception capabilities or a chip that can be disposed within a device. A terminal device may also be referred to as user equipment (UE), access terminal, subscriber unit, subscriber station, mobile station, remote station, remote terminal, mobile device, user terminal, wireless communication device, user agent, user equipment, etc. A terminal device in the present disclosure may be a mobile phone, a pad, a computer having wireless transmission and reception capabilities, a virtual reality (VR) terminal, an augmented reality (AR) terminal, a wearable device, a vehicle, an unmanned aerial vehicle, a helicopter, an aircraft, a ship, a robot, a robotic arm, a smart home device, etc. The terminal device of the present disclosure may be widely used in communications in various scenarios, including, but not limited to, at least one of the following scenarios: enhanced mobile broadband (eMBB), ultra-reliable low-latency communication (URLLC), device-to-device (D2D), vehicle-to-everything (V2X), machine-type communication (MTC), massive machine-type communication (mMTC), internet of things (IoT), virtual reality, augmented reality, industrial control, autonomous driving, telemedicine, smart grid, smart furniture, smart office, smart wearable, smart transportation, smart city, etc. The specific technology used by the terminal and the specific device form are not limited in this disclosure.

[0069] In the present disclosure, a device configured to implement the functions of a terminal may be a terminal, or may be a device capable of supporting a terminal in implementing the functions of, for example, a chip system, a hardware circuit, a software module, or a combination of a hardware circuit and a software module, and may be installed in or used in combination with a terminal.

[0070] The access network device may be a base station, a NodeB, an evolved NodeB (eNodeB), a transmission reception point (TRP), a next-generation NodeB (gNB) in a fifth-generation (5G) mobile communication system, an access network device in an open radio access network (O-RAN or open RAN), a next-generation base station in a sixth-generation (6G) mobile system, a base station in a future mobile communication system, an access node in wireless fidelity (Wi-Fi), or the like; or a module or unit that completes some functions of a base station, such as a central unit (CU), a distributed unit (DU), a central unit control plane (CU-CP) module, or a central unit user plane (CU-UP) module. The radio access network device may be a macro base station, a micro base station, an indoor base station, a relay node, a donor node, or the like. The specific technology used by the access network devices or the specific device configurations are not limited in this disclosure.

[0071] (1) Protocol layer structure

[0072] Communications between the access network device and the terminal conform to a specific protocol layer structure. The protocol layer structure may include a control plane protocol layer structure and a user plane protocol layer structure. For example, the control plane protocol layer structure may include protocol layer functions such as a radio resource control (RRC) layer, a packet data convergence protocol (PDCP) layer, a radio link control (RLC) layer, a media access control (MAC) layer, and a physical layer. For example, the user plane protocol layer structure may include protocol layer functions such as a PDCP layer, an RLC layer, a MAC layer, and a physical layer. In a possible implementation, a service data adaptation protocol (SDAP) layer may be further included above the PDCP layer.

[0073] Optionally, the protocol layer structure between the access network device and the terminal may further include an artificial intelligence (AI) layer used to perform the transmission of data related to AI functions.

[0074] (2) Central Unit (CU) and Distributed Unit (DU)

[0075] An access network device may include a CU and a DU. Multiple DUs may be centrally controlled by one CU. For example, the interface between a CU and a DU may be called an F1 interface. A control plane (CP) interface may be F1-C, and a user plane (UP) interface may be F1-U. Specific names of the interfaces are not limited in this disclosure. The CU and the DU may be divided based on the protocol layer of a wireless network. For example, the functions of the PDCP layer and protocol layers higher than the PDCP layer are located in the CU, and the functions of the protocol layers lower than the PDCP layer (e.g., the RLC layer and the MAC layer) are located in the DU. As another example, the functions of the protocol layers higher than the PDCP layer are located in the CU, and the functions of the PDCP layer and protocol layers lower than the PDCP layer are located in the DU. This is not a limitation.

[0076] The division of the processing functions of the CU and the DU based on protocol layers is merely exemplary, and the processing functions of the CU and the DU may alternatively be divided in other ways. For example, the CU or DU may be divided into functions with more protocol layers. In another example, the CU or DU may be further divided into several processing functions with protocol layers. In one design, some functions of the RLC layer and functions of protocol layers above the RLC layer are configured on the CU, and the remaining functions of the RLC layer and functions of protocol layers below the RLC layer are configured on the DU. In another design, the division of the functions of the CU or DU may alternatively be performed based on service type or other system requirements. For example, the division may be performed based on latency. Functions whose processing time must meet latency requirements are configured on the DU, and functions whose processing time does not need to meet latency requirements are configured on the CU. In another design, the CU may alternatively have one or more functions of a core network. For example, the CU may be located on the network side to facilitate centralized management. In another design, the radio unit (RU) of the DU is remotely located. Optionally, the RU may have radio frequency functionality.

[0077] Optionally, the DU and the RU may be distinguished by the physical layer (PHY). For example, the DU may implement upper layer functions of the PHY layer, and the RU may implement lower layer functions of the PHY layer. When the PHY layer is used for transmission, the PHY layer functions may include at least one of the following functions: cyclic redundancy check (CRC) bit, channel coding, rate matching, scrambling, modulation, layer mapping, precoding, resource mapping, physical antenna mapping, or radio frequency transmission. When the PHY layer is used for reception, the PHY layer functions may include at least one of the following: CRC check, channel decoding, rate de-matching, descrambling, demodulation, layer demapping, channel detection, resource demapping, physical antenna demapping, or radio frequency reception. The upper layer functions of the PHY layer may include some functions of the PHY layer, e.g., some functions are closer to the MAC layer. The lower layer functions of the PHY layer may include some other functions of the PHY layer, e.g., some other functions are closer to the radio frequency reception functions. For example, the upper layer functions of the PHY layer may include CRC code addition, channel coding, rate matching, scrambling, modulation, and layer mapping, and the lower layer functions of the PHY layer may include precoding, resource mapping, physical antenna mapping, and radio frequency transmission. Alternatively, the upper layer functions of the PHY layer may include CRC code addition, channel coding, rate matching, scrambling, modulation, layer mapping, and precoding. The lower layer functions of the PHY layer may include resource mapping, physical antenna mapping, and radio frequency transmission. For example, the upper layer functions of the PHY layer may include CRC check, channel decoding, rate removal matching, decoding, demodulation, and layer demapping, and the lower layer functions of the PHY layer may include channel detection, resource demapping, physical antenna demapping, and radio frequency reception.Alternatively, the upper layer functions of the PHY layer may include CRC checking, channel decoding, rate removal matching, decoding, demodulation, layer demapping and channel detection, and the lower layer functions of the PHY layer may include resource demapping, physical antenna demapping and radio frequency reception.

[0078] For example, the functions of the CU may be implemented by one entity or by different entities. For example, the functions of the CU may be further divided. In other words, the control plane and the user plane are separately implemented by different entities, which are a control plane CU entity (i.e., a CU-CP entity) and a user plane CU entity (i.e., a CU-UP entity). The CU-CP entity and the CU-UP entity may be coupled to the DU to jointly complete the functions of the access network device.

[0079] Optionally, any one of the DU, CU, CU-CP, CU-UP, and RU may be a software module, a hardware structure, or a combination of a software module and a hardware structure. This is not limited thereto. Different entities may exist in different forms. This is not limited thereto. For example, the DU, CU, CU-CP, and CU-UP are software modules, and the RU is a hardware structure. These modules and the methods performed by these modules are also within the scope of protection of the present disclosure.

[0080] In a possible implementation, the access network device includes a CU-CP, a CU-UP, a DU, and an RU. For example, the execution entity of the present disclosure includes a DU, or includes a DU and an RU, or includes a CU-CP, a DU, and an RU, or includes a CU-UP, a DU, and an RU. This is not limited. Methods performed by modules also fall within the scope of protection of the present disclosure.

[0081] In the present disclosure, an apparatus configured to implement the functions of an access network device may be an access network device, or may be an apparatus, such as a chip system, that can support an access network device to implement the functions. The apparatus may be installed in an access network device or used in combination with an access network device.

[0082] The present disclosure is applicable to the following scenario: An access network device uses an AI model to predict information that needs to be measured by a terminal device, thereby reducing the frequency of the access network device transmitting downlink signals and reducing the frequent measurement of information based on downlink signals by the terminal device, thereby reducing the resource overhead of the information and saving the computational resources of the terminal device.

[0083] For example, in a mobile communication system, obtaining accurate channel state information (CSI) of a wireless channel can significantly improve the transmission performance of the communication system. Based on the CSI, an access network device determines precoding of the corresponding wireless channel and performs precoding processing on the transmitted data to obtain transmit diversity gain and spatial beamforming gain, thereby reducing interference and improving transmission speed and spectral efficiency. The CSI of the transmission link from the access network device to the terminal device is called channel state information at the transmitter (CSIT). A typical method for obtaining CSIT is as follows: the access network device transmits a reference signal to the terminal device, and the terminal device estimates CSI based on the reference signal and feeds back the estimated CSI to the access network device. This method requires the receiver to periodically feed back the estimated CSI, which results in high air interface overhead.

[0084] To reduce the feedback overhead, a mathematical model-based CSI prediction solution may be used. The main principle of the mathematical model-based CSI prediction solution is to perform analytical processing on historical CSI using a time series processing algorithm in statistical signal processing theory, such as an auto-regression model (AR), and then perform mathematical operations such as weighting and interpolation on the historical CSI to predict the CSIT at a future time point. The overall CSI prediction algorithm uses historical CSI feedback as input and outputs a CSIT prediction value at a future time point. In the mathematical model-based CSI prediction solution, the terminal device does not need to feed back estimated CSI at the prediction time point. Therefore, the CSI feedback overhead on the air interface can be reduced.

[0085] However, the above-mentioned prediction solutions use a mathematical model to perform regression analysis on historical CSI feedback, and historical CSI at adjacent time points must be highly correlated to predict accurate CSI. In other words, good prediction performance can only be achieved for stable channels with smooth change trends. In unstable channels with variable channel change trends, the accurate CSI predicted by these prediction solutions is inaccurate, resulting in degradation of communication performance. In addition, because the prediction solutions are based on a unified mathematical formula, they are not designed specifically for different channel types. The change rules for different channel types vary significantly, making it difficult to adapt and update the prediction solutions to different channel types. As a result, the performance of the prediction solutions varies significantly depending on the channel environment, resulting in unstable performance in practical systems.

[0086] To solve the above problems, the present disclosure proposes a method for predicting channel state information by using machine learning techniques, thereby reducing air interface overhead and improving adaptability and robustness to actual channels. Specifically, an access network device inputs pieces of CSI at current and past times determined based on a reference signal into an artificial intelligence model, and predicts CSI at a future time through the artificial intelligence model. This may be understood as performing extrapolation in the time domain to estimate future CSI by using the current CSI and historical CSI.

[0087] Furthermore, in the present disclosure, the same artificial intelligence model may be deployed in the access network device and the terminal device, the terminal device may predict a predicted CSI at a future time point based on the artificial intelligence model, the terminal device may further obtain an actual CSI by measuring a reference signal, and if the difference metric value between the predicted CSI and the actual CSI is large, the access network device or the terminal device may retrain the artificial intelligence model to confirm the artificial intelligence model, as described in detail below.

[0088] Before describing the method in this disclosure, some relevant knowledge about artificial intelligence will be briefly explained first. Artificial intelligence enables machines to have human intelligence. For example, machines can simulate some intelligent human behavior by using computer software and hardware. To implement artificial intelligence, machine learning methods or many other methods can be used. For example, machine learning includes neural networks. A neural network (NN) is a specific implementation of machine learning. According to the general approximation theorem, a neural network can theoretically approximate any continuous function, thereby giving it the ability to learn any mapping. Therefore, a neural network can accurately perform abstract modeling for complex high-dimensional problems. The concept of a neural network comes from the neuron structure of brain tissue. Each neuron performs a weighted sum operation on its input values ​​and outputs the weighted sum result via an activation function. The neuron's inputs are

number

number

number

[0089] Neural networks generally have a multi-layer structure, with each layer containing one or more neurons. Increasing the depth and / or width of a neural network can improve the neural network's representational capabilities and provide more powerful information extraction and abstract modeling capabilities for complex systems. The depth of a neural network may refer to the number of layers contained in the neural network, and the number of neurons contained in each layer may be referred to as the layer width. Figure 2 illustrates the hierarchical relationship of a neural network. In one implementation, a neural network includes an input layer and an output layer. After performing neuronal processing on the received input, the input layer of the neural network forwards the results to the output layer, which obtains the neural network's output results. In another implementation, a neural network includes an input layer, a hidden layer, and an output layer. The input layer of the neural network performs neuronal processing on the received input and then forwards the results to an intermediate hidden layer. The hidden layer forwards the calculation results to the output layer or an adjacent hidden layer. Finally, the output layer obtains the neural network's output results. A neural network may include one hidden layer or multiple hidden layers connected in series. This is not a limitation. In the neural network training process, a loss function may be defined. The loss function describes the differential value or difference between the neural network output value and an ideal target value. The specific form of the loss function is not a limitation in this disclosure. The neural network training process is a process of adjusting neural network parameters, such as the number of layers and width of the neural network, neuron weights and / or neuron activation function parameters, so that the value of the loss function is equal to or less than a threshold or meets a target condition. In other words, the difference between the neural network output value and the ideal target value is minimized.

[0090] FIG. 3 is an exemplary diagram of a first application framework for AI in a communication system. In FIG. 3, a data source is used to store training data and inference data. A model training node (model training host) analyzes or trains training data provided by the data source to obtain an AI model, and then deploys the AI ​​model in a model inference node (model inference host). The AI ​​model represents a mapping relationship between the input and output of the model. Obtaining an AI model through learning by the model training node is equivalent to obtaining a mapping relationship between the input and output of the model through learning by the model training node using training data. The model inference node performs inference by using the AI ​​model based on inference data provided by the data source and obtains an inference result. This method can also be described as follows: the model inference node inputs inference data to the AI ​​model and obtains an output by using the AI ​​model. The output is an inference result. The inference result may indicate configuration parameters used (acted upon) by an action subject and / or an operation performed by the action subject. The inference results can be coordinated by actor entities and sent to one or more action subjects (e.g., network elements) for action.

[0091] In the present disclosure, the application framework shown in FIG. 3 may be deployed on the network element shown in FIG. 1. For example, the application framework of FIG. 3 may be deployed in the access network device or AI entity of FIG. 1. For example, in the access network device, a model training node may analyze or train training data provided by a data source to obtain a model. A model inference node may perform inference by using the model and inference data provided by the data source to obtain an output of the model. Specifically, the input of the model includes inference data, and the output of the model is an inference result corresponding to the model. For example, the model training node may be a CU, and the model inference node may be a CU or DU; or the model training node may be a DU, and the model inference node may be a quasi-real-time RIC (described later), and the model inference node may be a quasi-real-time RIC, CU, or DU. For example, in FIG. 3, a terminal device is considered to be a target of an action. The access network device may send inference data and / or inference results corresponding to the model to the terminal device, and the terminal device may perform corresponding operations based on the inference data and / or inference results.

[0092] With reference to FIGS. 4(a) to 4(d), the following describes a network architecture to which the communication solution provided in the present disclosure may be applied.

[0093] As shown in FIG. 4(a), in a first possible implementation, the access network device includes a near-real-time access network intelligent control (RAN intelligent controller, RIC) module configured to perform model training and / or inference. For example, the near-real-time RIC may obtain network-side and / or terminal-side information from at least one of the CU, DU, and RU, and may use the information as training data or inference data. Optionally, the near-real-time RIC may submit inference results to at least one of the CU, DU, and RU. Optionally, the CU and the DU may exchange inference results. Optionally, the DU and the RU may exchange inference results. For example, the near-real-time RIC may submit the inference results to the DU, and the DU may submit the inference results to the RU. For example, the near-real-time RIC may be configured to train an AI model, and inference is performed by using the AI ​​model.

[0094] As shown in FIG. 4(a), in a second possible implementation, a non-real-time RIC other than an access network device (optionally, the non-real-time RIC may be located in an OAM or core network device) is included and used for model training and inference. For example, the non-real-time RIC may obtain network-side and / or terminal-side information from at least one of a CU, a DU, and a RU. The information may be used as training data or inference data, and inference results may be submitted to at least one of a CU, a DU, and a RU. Optionally, the CU and the DU may exchange inference results. Optionally, the DU and the RU may exchange inference results. For example, the non-real-time RIC submits the inference results to the DU, and the DU submits the inference results to the RU. For example, the non-real-time RIC is configured to train an AI model, and inference is performed using the model.

[0095] As shown in FIG. 4(a), in a third possible implementation, the access network device includes a near-real-time RIC and a non-real-time RIC other than the access network device (optionally, the non-real-time RIC may be located in an OAM or core network device). As in the second possible implementation, the non-real-time RIC may be configured to perform model training and / or inference; and / or as in the first possible implementation, the near-real-time RIC may be configured to perform model training and / or inference; and / or the near-real-time RIC may obtain AI model information from the non-real-time RIC, obtain network-side and / or terminal-side information from at least one of the CU, DU, and RU, and obtain an inference result by using the information and the AI ​​model. Optionally, the near-real-time RIC may submit the inference result to at least one of the CU, DU, and RU. Optionally, the CU and the DU may exchange the inference result. Optionally, the DU and the RU may exchange the inference result. For example, the near-real-time RIC submits the inference result to the DU, and the DU submits the inference result to the RU. For example, a near-real-time RIC may be configured to train model A, and inference may be performed using model A. For example, a non-real-time RIC may be configured to train model B, and inference may be performed using model B. For example, a non-real-time RIC may be configured to train model C, and information about model C may be submitted to the near-real-time RIC, and the near-real-time RIC may perform inference using model C.

[0096] FIG. 4(b) is an exemplary diagram of a network architecture to which the method provided in the present disclosure can be applied. Compared with FIG. 4(a), FIG. b ), CU is separated into CU-CP and CU-UP.

[0097] 4(c) is an exemplary diagram of a network architecture to which the methods provided in the present disclosure can be applied. As shown in FIG. 4(c), optionally, the access network device includes one or more AI entities, the functions of which are similar to those of the near-real-time RIC. Optionally, the OAM includes one or more AI entities, the functions of which are similar to those of the non-real-time RIC. Optionally, the core network device includes one or more AI entities, the functions of which are similar to those of the non-real-time RIC. When the OAM and the core network device each include an AI entity, the models obtained by training by the AI ​​entities of the OAM and the core network device are different, and / or the models used for inference are different.

[0098] In the present disclosure, different models include, but are not limited to, at least one of the following differences: structural parameters of the model (e.g., the number of layers of the model, the connection relationships between the layers, the number of neurons included in each layer, the activation functions of the neurons, the weights and / or the offsets of the neurons), input parameters of the model, or output parameters of the model.

[0099] FIG. 4(d) is an exemplary diagram of a network architecture to which the methods provided in the present disclosure can be applied. Compared with FIG. 4(c), the access network device in FIG. 4(d) is separated into a CU and a DU. Optionally, the CU may include an AI entity, whose function is similar to that of a near-real-time RIC. Optionally, the DU may include an AI entity, whose function is similar to that of a near-real-time RIC. When the CU and the DU each include an AI entity, the models obtained by training by the AI ​​entities of the CU and the DU are different and / or the models used for inference are different. Optionally, the CU in FIG. 4(d) may be further separated into a CU-CP and a CU-UP. Optionally, one or more AI models may be deployed on the CU-CP. Optionally, one or more AI models may be deployed on the CU-UP. Optionally, in FIG. 4(c) or FIG. 4(d), the OAM of the access network device and the OAM of the core network device may be deployed separately and independently.

[0100] In the present disclosure, one parameter or multiple parameters may be obtained through inference using one model. The learning processes of different models may be deployed in different devices or nodes, or may be deployed in the same device or node. The inference processes of different models may be deployed in different devices or nodes, or may be deployed in the same device or node.

[0101] It may be understood that in this disclosure, an AI entity, an access network device, etc. may perform some or all of the steps in this disclosure. These steps or operations are merely exemplary. The present disclosure may also perform other operations or variations of various operations. In addition, steps may be performed in a different order than presented in this disclosure, and not all operations in this disclosure may necessarily be performed.

[0102] In various embodiments of the present disclosure, unless otherwise specified or there is no logical contradiction, the terms and / or descriptions in different embodiments are consistent and may be cross-referenced, and the technical features in different embodiments may be combined based on their internal logical relationships to form a new embodiment.

[0103] It can be understood that various numbers in the embodiments of the present disclosure are used only for distinction to facilitate description, and are not used to limit the scope of the embodiments of the present disclosure. The sequence numbers of the above processes do not imply an execution order, and the execution order of the processes should be determined based on the functions and internal logic of the processes.

[0104] The network architectures and service scenarios described in this disclosure are intended to more clearly explain the technical solutions in this disclosure and do not constitute limitations on the technical solutions provided in this disclosure. Those skilled in the art will know that with the development of network architectures and the emergence of new service scenarios, the technical solutions provided in this disclosure can also be applied to similar technical problems.

[0105] Optionally, before the method provided in the present disclosure is executed, capability exchange may further be performed between the terminal device and the access network device. Details may be shown in Figure 5. Figure 5 is a schematic flowchart of capability exchange according to the present disclosure, including the following steps:

[0106] Optionally, S501: A terminal device reports capability information to an access network device.

[0107] The capability information may include, but is not limited to, at least one of the following information: whether the terminal device supports the execution of an artificial intelligence model; the type of supported artificial intelligence model (e.g., a convolutional neural network (CNN), a recursive neural network (RNN), or a random forest model); the size of the memory space that can be used by the terminal device to store the artificial intelligence model; computing capability information of the terminal device, which may be the computing capability for executing the artificial intelligence model, such as information such as the computing speed of the processor and / or the amount of data that the processor can process; energy consumption information of the terminal device, such as the operating power consumption and / or battery capacity of the chip; or hardware information of the terminal device, which may include, but is not limited to, antenna configuration information (the number and / or polarization direction of antennas), the number of radio frequency channels, etc.

[0108] The terminal device may actively report the capability information, or may report the capability information upon receiving a request message from the access network device, which is not a limitation in the present disclosure.

[0109] S501 is an optional operation. For example, when the capabilities of the terminal device are agreed upon in a protocol, the method shown in Figure 5 does not need to include S501.

[0110] S502: The access network device sends information about the first channel prediction model to the terminal device.

[0111] The first channel prediction model may be an artificial intelligence model. The access network device may transmit information about the first channel prediction model when it determines that the terminal device supports execution of the artificial intelligence model. Alternatively, the access network device may not transmit information about the first channel prediction model. In other words, the first channel prediction model may be agreed upon in a protocol, or the access network device and the terminal device download the first channel prediction model via the same download address.

[0112] In a first possible implementation, the information about the first channel prediction model may be information including model parameters, model structure, cascade relationships, activation functions, etc. used to describe the first channel prediction model. The terminal device may determine the first channel prediction model based on the information about the first channel prediction model. In a second possible implementation, the information about the first channel prediction model may also be computer code describing the first channel prediction model, and the terminal device may perform compilation based on the computer code to obtain the first channel prediction model. In a third possible implementation, the information about the first channel prediction model may be a download address for downloading the first channel prediction model, and the terminal device may download the first channel prediction model based on the download address.

[0113] In the present disclosure, the first channel prediction model may be a trained artificial intelligence model, and the first channel prediction model may be used to predict channel state information of a downlink channel from an access network device to a terminal device.

[0114] For example, FIG. 6 is a diagram of the structure of an artificial intelligence model according to the present disclosure. The type of the artificial intelligence model may be a neural network model, a random forest model, a linear regression model, etc. This is not limited in the present disclosure. The artificial intelligence model includes a five-layer convolutional layer network and a two-layer fully connected layer network. This is merely an example and does not represent any limitation on the number of convolutional layer networks and fully connected layer networks. The convolutional layer network processes the acquired input and sends the input to the fully connected layer network. The fully connected layer network finally outputs the result. The input of the artificial intelligence model may be historical channel state information (i.e., channel state information acquired through measurements at the current time point and before the current time point), and the output of the artificial intelligence model is predicted channel state information. The propagation form of a wireless signal on a channel is Y = HX + N, where H is channel state information, X is a reference signal, N is noise, and Y is a received signal. The purpose of the channel prediction model is to predict channel state information at a future time point by using the historical channel state information H.

[0115] When the input historical channel state information is a channel frequency-domain response, the dimension of the input historical channel state information may be T*F*M*N. When the input historical channel state information is a channel time-domain response, the dimension of the input historical channel state information may be T*L*N, where T is the number of symbols of the historical channel state information in the time dimension, F is the number of subcarriers in the frequency-domain dimension, L is the number of sampling points of the time-domain channel, M is the number of antennas of the receiver (e.g., terminal device), and N is the number of antennas of the transmitter (e.g., access network device). The symbols in this specification may be orthogonal frequency division multiplexing (OFDM) symbols, etc.

[0116] When the output prediction channel state information is a channel frequency domain response, the dimension of the output prediction channel state information is T*F*M*N, and when the output prediction channel state information is a channel time domain response, the dimension of the output prediction channel state information is T*L*N.

[0117] In the present disclosure, the first channel prediction model may be trained by the access network device, or the pre-trained first channel prediction model may be obtained from a third-party network entity. For example, an AI entity performs the training, and the access network device obtains the trained first channel prediction model from the AI ​​entity. The specific training process of the first channel prediction model is not limited in the present disclosure. The training data required for training the channel prediction model is channel state measurement information, and the training data is provided by the Third Generation Partnership Project (the 3G Partnership Project). rd The channel model may be generated by using a channel model defined by the 3GPP (3rd Generation Partnership Project), collected in a real environment, or generated by another simulation platform such as a ray tracking channel simulation platform, which is not a limitation of this disclosure.

[0118] In the present disclosure, the same channel prediction model may be deployed in the access network device and the terminal device. In one implementation, the first channel prediction model may also be deployed in the access network device, or the same artificial intelligence model is deployed in the access network device and the terminal device. Deploying the same artificial intelligence model in the access network device and the terminal device may mean that the model parameters, model structure, cascade relationship, activation function, etc. of the artificial intelligence model deployed in the access network device are exactly the same as those of the artificial intelligence model deployed in the terminal device.

[0119] In another implementation, information such as the model structure of the AI ​​model in the access network device may be different from information of the AI ​​model in the terminal device. However, when the input information is the same, the output information is the same or the difference between the output information is less than a threshold. In other words, if the access network device and the terminal device agree on the same model input data format and input the same information, the same prediction results can be output through the respective AI models.

[0120] After the first channel prediction model is deployed in the terminal device, the access network device may transmit a downlink signal to the terminal device, so that the terminal device can input estimated downlink measurement information obtained by measuring the downlink signal into the first channel prediction model to determine whether the prediction result of the first channel prediction model is accurate. A detailed description is provided below.

[0121] 7 is a schematic flowchart of a calibration method according to the present disclosure. The procedure will be described by using an example of an interaction between a terminal device and an access network device. In the procedure shown in FIG. 7, the same channel prediction model may be deployed in the terminal device and the access network device, and the access network device may use the channel prediction model to predict the downlink signal with the highest received signal strength among the downlink signals received by the terminal device. The terminal device may check the prediction result of the deployed channel prediction model, and send feedback to the access network device when the prediction result is inaccurate. Specifically, the method includes the following steps:

[0122] Optionally, S700: The access network device sends indication information to the terminal device.

[0123] The indication information may indicate at least one of the following information:

[0124] (1) Input and output format information of the channel prediction model, indicating information such as the dimension of the input information and the dimension of the output information of the channel prediction model, where, for example, when the input history channel state information is a channel frequency domain response, the dimension of the input information may be T*F*M*N; when the input history channel state information is a channel time domain response, the dimension of the input information may be T*L*N; when the output prediction channel state information is a channel frequency domain response, the dimension of the output information is T*F*M*N; and when the output prediction channel state information is a channel time domain response, the dimension of the output information is T*L*N.

[0125] (2) Prediction time information indicating the interval between predicted channel state information and historical channel state information in the time domain, in other words, the number of symbols after the historical channel state information after which the channel state information is predicted.

[0126] (3) Channel prediction quality evaluation function information used to calculate a function of the differential metric value between the predicted channel state information and the actual channel state information, for example, a Euclidean distance function, where the predicted channel state information is H' and the actual channel state information estimated by the receiver is H, the Euclidean distance function used to calculate the differential metric value is e=||H-H'||.

[0127] (4) Threshold information indicating a first threshold, where the threshold information may further indicate a second threshold; optionally, the instruction information may not include threshold information; for example, the first threshold and the second threshold may be agreed upon in a protocol or set in another manner, and is not limited in this disclosure.

[0128] (5) Feedback indication information indicating whether the terminal device will feedback that the channel prediction model is trusted or that there is no indication signal or information indicating that the prediction result is accurate when the differential metric value between the predicted channel state information and the actual channel state information is less than a first threshold.

[0129] In addition, when the instruction information does not include any of the above information, the content indicated by the information may be agreed upon in a protocol. For example, the instruction information may not include format information of the input and output of the channel prediction model. In this case, the format information of the input and output of the channel prediction model may be agreed upon in a protocol.

[0130] S701: An access network device transmits a downlink signal to a terminal device, and in response, the terminal device receives the downlink signal from the access network device.

[0131] The access network device may periodically transmit a downlink signal to the terminal device based on a preset periodicity. For example, the preset periodicity may be P time units, where P is an integer greater than 1. In this case, the access network device transmits the downlink signal once every P time units. The access network device may also transmit the downlink signal in a non-periodic manner. This is not a limitation in this disclosure. The type of the downlink signal is not limited. The downlink signal may be a downlink reference signal, for example, a channel state information reference signal (CSI-RS). Alternatively, the downlink signal may be another type of signal. This is not a limitation in this disclosure.

[0132] For example, FIG. 8 is a diagram of downlink signal transmission according to the present disclosure. In FIG. 8, black grids represent resources for transmitting downlink reference signals, and white grids represent resources for not transmitting downlink reference signals. One resource includes one frequency unit in the frequency domain and one time unit in the time domain. In FIG. 8, an access network device transmits downlink signals in time unit 1 and time unit 3. Specifically, in time unit 1 and time unit 3, the downlink signals are transmitted separately via frequency unit 1, frequency unit 3, and frequency unit 5. The time unit may be a radio frame, a subframe, a slot, a symbol, etc., and the frequency unit may be a subcarrier interval, a bandwidth part (BWP), etc.

[0133] In the initial access stage of the terminal device, the terminal device may perform channel estimation for each of the first Q received downlink signals to determine one piece of channel state information, so that the terminal device may determine Q pieces of channel state information. The value of Q may be configured by the access network device or may be agreed upon in a protocol, and Q is an integer greater than 0. The terminal device may feed back the Q pieces of channel state information to the access network device.

[0134] For downlink signals after the Q downlink signals transmitted by the access network device, the terminal device may feedback channel state information corresponding to the received downlink signals, or may not feedback channel state information. If the terminal device does not feedback channel state information, the access network device may use one or more of the Q channel state information as historical channel state information, and the access network device may input the historical channel state information into a first channel prediction model to obtain predicted channel state information. In the present disclosure, when the access network device needs to predict channel state information corresponding to a downlink signal transmitted in a first time unit, the historical channel state information used to predict the channel state information may not only include channel state information acquired from the terminal device before the first time unit, but also channel state information predicted by the access network device before the first time unit. For ease of description, in the present disclosure, the historical channel state information input by the access network device into the channel prediction model for prediction is referred to as second historical channel state information, and the historical channel state information input by the terminal device into the channel prediction model for prediction is referred to as first historical channel state information.

[0135] For example, referring to FIG. 8 , at time t0 (the end of time unit 3), the access network device acquires second historical channel state information at time t≦t0, e.g., channel state information corresponding to the downlink signal transmitted in time unit 1 and channel state information corresponding to the downlink signal transmitted in time unit 3. The access network device may input one or more of the channel state information into a first channel prediction model to predict the channel state information after time t1 (the start time of time unit 5). The channel state information may be referred to as predicted channel state information. For the downlink signal after time t1, e.g., the downlink signal transmitted in time unit 5, the terminal device may not feed back channel state information. The access network device may input one or more channel state information before time unit 5 into the channel prediction model to continue predicting the channel state information of the downlink signal transmitted in time unit 5. Other cases can be inferred by analogy. In this manner, the terminal device does not feed back estimated channel state information, which may reduce air interface overhead and improve system efficiency.

[0136] Furthermore, the terminal device may predict channel state information at a future time point based on the first channel prediction model, the terminal device may further obtain actual channel state information by measuring downlink signals, and the access network device or the terminal device may calibrate the first channel prediction model based on the predicted channel state information and the actual channel state information. For details, see the description below.

[0137] S702: The terminal device determines first channel state information and second channel state information.

[0138] The first channel state information is obtained through prediction by a first channel prediction model, and the second channel state information is determined based on a channel estimate for a downlink signal from an access network device.

[0139] Specifically, when the terminal device needs to predict the first channel state information in the second time unit, the terminal device may use one or more pieces of channel state information obtained by performing channel estimation on one or more downlink signals before the second time unit as the first historical channel state information, and may input the first historical channel state information into a first channel prediction model and obtain the first channel state information based on the output of the first channel prediction model.

[0140] Furthermore, when the terminal device receives a downlink signal from the access network device in a second time unit, the terminal device may perform channel estimation on the downlink signal to obtain second channel state information.

[0141] The terminal device may determine, based on the second channel state information, whether the first channel state information predicted via the first channel prediction model is accurate.

[0142] S703: If the first channel state information and the second channel state information satisfy a first condition, the terminal device sends the first information to the access network device, and in response, the access network device receives the first information from the terminal device.

[0143] When the first channel state information and the second channel state information satisfy a first condition, the terminal device may consider that the prediction result of the first channel prediction model is inaccurate and that the first channel prediction model needs to be updated. In a possible implementation, the first information may indicate the second channel state information, and the information may be for updating the first channel prediction model. Specific details of how to update the first channel prediction model will be described in detail below. In another possible implementation, the first information may indicate that the first channel prediction model needs to be updated or that the first channel prediction model is not trusted. In addition, the first information may further indicate other information. For example, the first information may further indicate a time unit corresponding to the second channel state information, i.e., a second time unit.

[0144] The first condition may include that a differential metric value between the first channel state information and the second channel state information is greater than or equal to a first threshold, which may be configured by the access network device, agreed upon in a protocol, or determined by the terminal device.

[0145] In this disclosure, when "less than" or "greater than" is used for comparison, "less than" can be replaced with "equal to or less than" and "greater than" can be replaced with "equal to or greater than", and vice versa.

[0146] In the present disclosure, the differential metric value between the first channel state information and the second channel state information may satisfy the following form: e=||H−H′||. e represents the differential metric value, the first channel state information is H′, and the second channel state information is H.

[0147] The above are merely examples, and the differential metric values ​​may be obtained in other ways, and examples will not be described one by one here.

[0148] S704: If the first channel state information and the second channel state information satisfy a second condition, the terminal device sends the second information to the access network device, and in response, the access network device receives the second information from the terminal device.

[0149] When the first channel state information and the second channel state information satisfy the second condition, the terminal device may consider that the prediction result of the first channel prediction model is accurate and there is no need to update the first channel prediction model. In this case, the terminal device may not transmit the second information. In another possible implementation, the terminal device may transmit the second information. The second information may indicate that the first channel prediction model is trusted, may indicate that there is no need to update the first channel prediction model, or may indicate that the prediction result of the first channel prediction model is accurate.

[0150] The second condition may include any one of the following:

[0151] a differential metric value between the first channel state information and the second channel state information is less than a first threshold; or The differential metric value between the first channel state information and the second channel state information is less than a second threshold, where the second threshold is less than or equal to the first threshold, and the second threshold may be configured by the access network device, agreed upon in a protocol, or determined by the terminal device.

[0152] In the present disclosure, when the prediction result of the first channel prediction model is inaccurate, the terminal device may update the first channel prediction model, or the access network device may update the first channel prediction model, which will be described separately below.

[0153] When the terminal device updates the first channel prediction model, the following procedures may be performed.

[0154] S705: The terminal device trains a first channel prediction model based on the second channel state information to obtain a second channel prediction model.

[0155] The terminal device may use the first historical channel state information and the second channel state information as a group of training data to train a first channel prediction model, where the trained model is a second channel prediction model. The terminal device may update the first channel prediction model to the second channel prediction model. The specific method for updating the first channel prediction model to the second channel prediction model is not limited in the present disclosure and may be, for example, a back-propagation gradient update algorithm.

[0156] After obtaining the second channel prediction model, the terminal device may indicate the second channel prediction model to the access network device through the third information, ie, execute S706.

[0157] S706: The terminal device sends third information to the access network device, and in response, the access network device receives the third information from the terminal device.

[0158] Furthermore, the access network device may update the first channel prediction model to the second channel prediction model based on the third information.

[0159] In the present disclosure, in a possible implementation, the third information indicates a second channel prediction model. Specifically, the third information may include information including model parameters, model structure, cascade relationships, activation functions, etc. used to describe the second channel prediction model. The terminal device may determine the second channel prediction model based on the third information. The third information may include computer code describing the second channel prediction model, and the terminal device may compile the computer code to obtain the second channel prediction model.

[0160] In another possible implementation, the third information indicates information about a second channel prediction model, and the information about the second channel prediction model includes variation information between the second channel prediction model and the first channel prediction model. Specifically, the third information may include variation information between information including model parameters, model structure, cascade relationships, activation functions, etc. used to describe the second channel prediction model and information including model parameters, model structure, cascade relationships, activation functions, etc. used to describe the first channel prediction model.

[0161] In yet another possible implementation, the third information indicates information about the second channel prediction model, and the information about the second channel prediction model includes gradient information of the second channel prediction model. Specifically, the third information may include a gradient calculation value when updating the first channel prediction model to the second channel prediction model by using a gradient descent method.

[0162] In the present disclosure, in a first scenario, if the first channel prediction model is updated, the terminal device indicates the second channel prediction model to the access network device, that is, executes S706.

[0163] In a second scenario, when updating the first channel prediction model, the terminal device may determine whether variation information between the second channel prediction model and the first channel prediction model is greater than a third threshold. When the variation information is greater than or equal to the third threshold, the terminal device transmits the third information, and when the variation information is less than the third threshold, the terminal device does not transmit the third information. Without loss of generality, this disclosure will describe how the terminal device determines whether to transmit the third information by using the procedure shown in FIG. 9 as an example.

[0164] S901: A terminal device trains an updated channel prediction model to obtain an updated channel prediction model.

[0165] The initial value of the channel prediction model to be updated is the first channel prediction model, which is the channel prediction model currently deployed in the terminal device. The updated channel prediction model is obtained through training in S705. Here, if the channel prediction model to be updated is the first channel prediction model, the updated channel prediction model obtained in this case is the second channel prediction model.

[0166] S902: The terminal device determines variation information between the updated channel prediction model and the initially updated channel prediction model.

[0167] The initially updated channel prediction model may be the channel prediction model previously indicated to the access network device via the third information, and the initial value of the initially updated channel prediction model is the first channel prediction model.

[0168] Here, the initially updated channel prediction model is the first channel prediction model, assuming that the updated channel prediction model is the second channel prediction model.

[0169] S903: If the variation information is greater than or equal to a third threshold, the terminal device determines to send the third information to the access network device.

[0170] The third threshold may be configured by the access network device, or may be agreed upon in a protocol, or may be determined by the terminal device.

[0171] Furthermore, in this case, the updated channel prediction model may be used as the first updated channel prediction model. If the updated channel prediction model is the second channel prediction model, the first updated channel prediction model in S902 may be set to the second channel prediction model.

[0172] S904: If the fluctuation information is less than the third threshold, the terminal device determines not to send the third information.

[0173] Next time, when training the channel prediction model to be updated, the terminal device may perform the steps of S901 to S904 again.

[0174] Since the fluctuation information is less than the third threshold, the updated channel prediction model is significantly different from the channel prediction model before the update, and the performance of the updated channel prediction model is close to that of the channel prediction model before the update, and the prediction result of the updated channel prediction model can be considered to be close to that of the channel prediction model before the update. Therefore, the updated channel prediction model does not need to be indicated to the access network device.

[0175] Referring to the description of the procedure shown in Figure 9, in one example, it is assumed that the first channel prediction model is prediction model A and the second channel prediction model is prediction model B. The terminal device may perform the following steps.

[0176] Step 1: Train predictive model A to obtain predictive model B.

[0177] Step 2: Determine whether the fluctuation information between prediction model A and prediction model B is greater than or equal to the third threshold. If the fluctuation information between prediction model A and prediction model B is greater than or equal to the third threshold, proceed to step 3; if the fluctuation information between prediction model A and prediction model B is less than the second threshold, proceed to step 4.

[0178] Step 3: Send third information, where the third information in this step indicates information about the prediction model B.

[0179] Step 4: Skip sending the third information.

[0180] If the fluctuation information in step 2 is less than the second threshold, the terminal device does not send the third information this time.

[0181] Furthermore, if the terminal device needs to further train the prediction model B, the terminal device may perform the procedure shown in FIG. 9 again to obtain the prediction model C.

[0182] Step 5: Train predictive model B to obtain predictive model C.

[0183] Step 6: Determine whether the fluctuation information between prediction model C and prediction model A is greater than or equal to the third threshold. If the fluctuation information between prediction model C and prediction model A is greater than or equal to the third threshold, proceed to step 7; if the fluctuation information between prediction model C and prediction model A is less than the third threshold, proceed to step 8.

[0184] Prediction model C is obtained by training prediction model B, but prediction model A is deployed in the access network device and the terminal device does not present prediction model A to the access network device, so it is necessary to determine whether the variation information between the latest prediction model in the terminal device and the prediction model in the access network device is greater than or equal to a third threshold.

[0185] Step 7: Send third information, where the third information in this step indicates information about the prediction model C.

[0186] Step 8: Skip sending the third information.

[0187] According to the above procedure, when the variation information between the updated channel prediction model and the channel prediction model previously indicated to the access network device is less than a third threshold, the updated channel prediction model does not need to be indicated to the access network device, thereby avoiding frequently indicating the updated channel prediction model to the access network device, reducing signaling overhead and improving system efficiency.

[0188] In the present disclosure, the access network device may alternatively update the first channel prediction model. When the access network device updates the first channel prediction model, the following procedure may be performed.

[0189] S707: The access network device trains a first channel prediction model based on the second channel state information to obtain a third channel prediction model.

[0190] The access network device may use the first historical channel state information and the second channel state information as a group of training data to train a first channel prediction model, where the trained model is a third channel prediction model. The access network device may update the first channel prediction model to the third channel prediction model. The first historical channel state information may be indicated by the terminal device.

[0191] The access network device may also use the second historical channel state information and the second channel state information as a group of training data to train the first channel prediction model and obtain a third channel prediction model. The second historical channel state information includes one or more pieces of channel state information obtained by the access network device prior to the second time unit. The second historical channel state information may be the same as or different from the first historical channel state information. This is not a limitation in the present disclosure.

[0192] The second channel prediction model obtained by the terminal device through training in S705 and the third channel prediction model obtained by the access network device through training in S707 may be the same model. In this disclosure, for ease of explanation, the second channel prediction model and the third channel prediction model are distinguished, but this does not necessarily mean that they are different models.

[0193] S708: The access network device sends the fourth information to the terminal device, and in response, the terminal device receives the fourth information from the access network device.

[0194] Furthermore, the terminal device may update the first channel prediction model to a third channel prediction model based on the fourth information.

[0195] In the present disclosure, the fourth information indicates a third channel prediction model or indicates information related to the third channel prediction model, and the information related to the third channel prediction model includes variation information between the third channel prediction model and the first channel prediction model. For specific content of the fourth information, please refer to the description of the third information. Details will not be described again here.

[0196] In the present disclosure, in the first scenario, the access network device sends the fourth information to the terminal device on the condition that the first channel prediction model is updated, that is, S706 is executed.

[0197] In the second scenario, when updating the first channel prediction model, the access network device may determine whether the variation information between the third channel prediction model and the first channel prediction model is greater than a third threshold. If the variation information is greater than or equal to the third threshold, the access network device transmits the fourth information; if the variation information is less than the third threshold, the access network device does not transmit the fourth information. For details, see the procedure shown in FIG. 9, and the details will not be described again here.

[0198] According to the method provided in the present disclosure, the access network device and the terminal device may use the same channel prediction model, and the terminal device may test and monitor the channel prediction quality online. Furthermore, when the deviation between the predicted first channel state information and the actually measured second channel state information is large, the channel prediction model may be updated to avoid affecting communication performance. In addition, the fed-back channel state information may be further used for online training and updating of the channel prediction model to continuously improve the performance of the channel prediction model.

[0199] In the present disclosure, the channel prediction model may alternatively not be deployed in the terminal device but in the access network device. The access network device verifies the channel prediction model based on information fed back by the terminal device. Specifically, Figure 10 is a schematic flowchart of a calibration method according to the present disclosure. The method includes the following steps:

[0200] Optionally, S1000: The access network device sends indication information to the terminal device.

[0201] For the specific content of the instruction information, please refer to the explanation in S700, and the details will not be explained again here.

[0202] S1001: An access network device transmits L downlink signals to a terminal device.

[0203] S1002: The access network device receives first feedback information from the terminal device.

[0204] where L is an integer greater than 0, and the first feedback information may indicate second channel state information, where the second channel state information is determined based on a channel estimation for a first downlink signal, and the first downlink signal is one of the L downlink signals. The first downlink signal may be indicated by the access network device or determined by the terminal device from the L downlink signals.

[0205] In other words, the terminal device does not need to perform channel estimation for each downlink signal, but may select a portion of the downlink signals to perform channel estimation and obtain channel state information. The access network device may indicate downlink signals to be specifically selected by the terminal device to perform channel estimation. For example, the access network device transmits sampling information to the terminal device, where the sampling information indicates that the terminal device will feed back channel state information of one downlink signal at an interval of W downlink signals, where W is an integer greater than 0.

[0206] The first feedback information may further indicate a second time unit, where the second time unit represents a time unit in which the first downlink signal for determining the second channel state information is located.

[0207] S1003: If the first channel state information and the second channel state information satisfy a first condition, the access network device trains a first channel prediction model based on the second channel state information to obtain a third channel prediction model.

[0208] The first channel state information may be obtained by the access network device through prediction by the first channel prediction model. Specifically, the access network device may obtain second historical channel state information including one or more pieces of channel state information prior to a second time unit. The access network device may input the second historical channel state information into the first channel prediction model and obtain the first channel state information based on the output of the first channel prediction model.

[0209] When the first channel state information and the second channel state information satisfy a first condition, the access network device may consider that the prediction result of the first channel prediction model is inaccurate and the first channel prediction model needs to be updated.

[0210] The access network device may use the first historical channel condition information and the second channel condition information as a group of training data to train a first channel prediction model, the trained model being a third channel prediction model. Further, the access network device may update the first channel prediction model to the third channel prediction model.

[0211] In the present disclosure, the first condition may include: a differential metric value between the first channel state information and the second channel state information is greater than or equal to a first threshold.

[0212] For other details of S1003, please refer to the explanation of S703, and the details will not be explained again here.

[0213] S1004: If the first channel state information and the second channel state information satisfy a second condition, the access network device maintains the first channel prediction model without modification.

[0214] When the first channel state information and the second channel state information satisfy a second condition, the access network device may consider that the prediction result of the first channel prediction model is accurate and there is no need to update the first channel prediction model.

[0215] The second condition may include a differential metric value between the first channel state information and the second channel state information being less than a first threshold.

[0216] For other details of S1004, please refer to the explanation of S704, and the details will not be explained again here.

[0217] According to the method of the present disclosure, when it is determined that there is a large deviation between the predicted first channel state information and the actually measured second channel state information, the access network device may update the channel prediction model to avoid affecting communication performance and continuously improve the performance of the channel prediction model.

[0218] In the foregoing embodiments provided in the present application, the methods provided in the embodiments of the present application are described in terms of interactions between devices. To implement the functions of the methods provided in the embodiments of the present disclosure, the access network device or the terminal device may include a hardware structure and / or a software module, and may implement the functions in the form of a hardware structure, a software module, or a combination of a hardware structure and a software module. Whether a function among the above functions is implemented by using a hardware structure, a software module, or a combination of a hardware structure and a software module depends on the specific application and design constraints of the technical solution.

[0219] In this embodiment of the present application, the module division is merely an example and is merely a logical function division. In actual implementation, other division methods may be used. In addition, the functional modules in the embodiment of the present application may be integrated into one processor, may exist physically alone, or two or more modules may be integrated into one module. The integrated module may be implemented in the form of hardware or in the form of a software functional module.

[0220] Similar to the above concept, the embodiment of the present disclosure further provides a communication device 1100, as shown in FIG. 11, which implements the functions of the access network device or the terminal device in the above-mentioned method. The form of the communication device is not limited, and the communication device may be a hardware structure, a software module, or a combination of a hardware structure and a software module. For example, the device may be a software module or a chip system. In this embodiment of the present application, the chip system may include a chip, or may include a chip and separate components. The communication device 1100 may include a processing unit 1101 and a communication unit 1102.

[0221] In the embodiments of the present application, the communication unit may also be referred to as a transceiver unit, and may include a transmitting unit and / or a receiving unit configured to perform the transmitting steps and receiving steps, respectively, performed by the access network device, the AI ​​entity, or the terminal device in the embodiments of the aforementioned method.

[0222] Hereinafter, the communication device provided in the embodiment of the present application will 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, for the contents not described in detail, please refer to the method embodiment. For the sake of brevity, the details will not be described again here.

[0223] The communication unit may also be referred to as a transceiver. The processing unit may also be referred to as a processing module, a processing device, etc. Optionally, a device configured to implement a receiving function in the communication unit 1102 may be considered a receiving unit, and a device configured to implement a transmitting function in the communication unit 1102 may be considered a transmitting unit. In other words, the communication unit 1102 includes a receiving unit and a transmitting unit. The communication unit may be implemented as a pin, a transmitting / receiving machine, a transceiver, a transmitting / receiving circuit, etc. The processing unit may often be implemented as a processor, a processing board, etc. The receiving unit may often be implemented as a pin, a receiving machine, a receiver, a receiving circuit, etc. The transmitting unit may often be implemented as a pin, a transmitter, a transmitter, a transmitting circuit, etc.

[0224] When the communication device 1100 performs the function of the terminal device in the procedure shown in FIG. 7 in the above embodiment, the processing unit is configured to determine first channel state information and second channel state information, where the first channel state information is obtained through prediction by a first channel prediction model, and the second channel state information is determined based on channel estimation for a downlink signal from an access network device; The communication unit is configured to transmit first information to the access network device when a differential metric value between the first channel state information and the second channel state information is greater than or equal to a first threshold, where the first information indicates the second channel state information.

[0225] When the communication device 1100 performs the function of the terminal device in the procedure shown in FIG. 7 in the above embodiment, The processing unit is configured to transmit a downlink signal to a terminal device via a communication unit, and receive first information or second information from the terminal device, where the first information indicates second channel state information, and the second channel state information is determined based on a channel estimation for the downlink signal; The second information indicates that the prediction result of the first channel prediction model is accurate.

[0226] When the communication device 1100 performs the function of the terminal device in the procedure shown in FIG. 11 in the above embodiment, The processing unit is configured to determine first channel state information, where the first channel state information is obtained through prediction by a first channel prediction model; The communication unit is configured to receive first information from a terminal device, where the first information indicates second channel state information corresponding to the downlink signal; The processing unit is configured to, when a differential metric value between the first channel state information and the second channel state information is greater than or equal to a first threshold, train a first channel prediction model based on the second channel state information to obtain a second channel prediction model, and update the first channel prediction model to the second channel prediction model.

[0227] The above is just an example. The processing unit 1101 and the communication unit 1102 may further perform other functions. For more detailed descriptions, please refer to the relevant descriptions in the method embodiments shown in FIG. 7 or FIG. 11. Here, detailed descriptions are omitted.

[0228] Fig. 12 illustrates a communication device 1200 according to an embodiment of the present disclosure. The device illustrated in Fig. 12 may be a hardware circuit implementation of the device illustrated in Fig. 11. The communication device is applicable to the above-described flowcharts and performs the functions of a terminal device or an access network device in the above-described method embodiments. For ease of explanation, Fig. 12 illustrates only the main components of the communication device.

[0229] 12, the communication device 1200 includes a processor 1210 and an interface circuit 1220. The processor 1210 and the interface circuit 1220 are coupled to each other. It may be understood that the interface circuit 1220 may be a transceiver, a pin, an interface circuit, or an input / output interface. Optionally, the communication device 1200 may further include a memory 1230 configured to store instructions to be executed by the processor 1210, to store input data required by the processor 1210 to execute the instructions, or to store data generated after the processor 1210 executes the instructions. Optionally, some or all of the memory 1230 may be located within the processor 1210.

[0230] When the communication device 1200 is configured to implement the method shown in FIG. 7 or FIG. 11, the processor 1210 is configured to implement the functions of the processing unit 1101, and the interface circuit 1220 is configured to implement the functions of the communication unit 1102.

[0231] When the communication device is a chip used in a terminal device, the chip in the terminal device implements the functions of the terminal device in the above-mentioned method embodiments. The chip in the terminal device receives information from another module (e.g., a radio frequency module or an antenna) in the terminal device, where the information is transmitted to the terminal device by the access network device. Alternatively, the chip in the terminal device transmits information to another module (e.g., a radio frequency module or an antenna) in the terminal device, where the information is transmitted to the access network device by the terminal device.

[0232] When the communication apparatus is a chip used in an access network device, the chip in the access network device implements the functions of the access network device in the above-mentioned method embodiments. The chip in the access network device receives information from another module (e.g., a radio frequency module or an antenna) in the access network device, where the information is transmitted to the access network device by the terminal device. Alternatively, the chip in the access network device transmits information to another module (e.g., a radio frequency module or an antenna) in the access network device, where the information is transmitted to the terminal device by the access network device.

[0233] As described above, the method of Figure 7 uses the interaction between an access network device and a terminal device as an example. When the AI ​​entity is located outside the access network device and is an independent module or network element, the access network device or the terminal device may transfer second channel state information to the AI ​​entity, which may train a first channel prediction model based on the second channel state information, and the AI ​​entity may indicate information about the trained first channel prediction model to the access network device or the terminal device. Other processes are similar to those in Figure 5. Therefore, the above description of the access network device may also apply to the AI ​​entity.

[0234] It should be understood that a processor in this disclosure may be a central processing unit, or may be another general-purpose processor, a digital signal processor, an application specific integrated circuit, a field programmable gate array or other programmable logic device, a transistor logic device, a hardware component, or any combination thereof. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0235] Memory in the present disclosure may be random access memory, flash memory, read-only memory, programmable read-only memory, erasable programmable read-only memory, electrically erasable programmable read-only memory, registers, hard disk, removable hard disk or any other form of storage medium known in the art.

[0236] Those skilled in the art will understand that the present disclosure may be provided as a method, a system, or a computer program product. Thus, the present application may take the form of a hardware-only embodiment, a software-only embodiment, or an embodiment that combines software and hardware. Furthermore, the present application may take the form of a computer program product embodied in one or more computer-usable storage media (including, but not limited to, disk memory, optical memory, etc.) that contain computer-usable program code.

[0237] The present application will be described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the present application. It will be understood that computer program instructions may be used to implement each process and / or each block in the flowcharts and / or block diagrams, and combinations of processes and / or blocks in the flowcharts and / or block diagrams. These computer program instructions may be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or any other programmable data processing device to generate a machine, whereby the instructions executed by the processor of the computer or any other programmable data processing device generate an apparatus for implementing the particular function(s) of one or more processes in the flowcharts and / or one or more blocks in the block diagrams.

[0238] These computer program instructions may be stored in a computer-readable memory that can instruct a computer or any other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture that includes an instruction apparatus that implements a particular function in one or more processes of the flowcharts and / or one or more blocks of the block diagrams.

[0239] It is obvious that those skilled in the art can make various modifications and variations to this application without departing from the scope of this application, and this application is intended to cover these modifications and variations of this application, provided that they fall within the scope of protection defined by the following claims and their equivalent technologies.

Claims

1. A calibration method performed by a terminal device or a chip for said terminal device, comprising: determining first channel state information and second channel state information, the first channel state information and the second channel state information being channel state information of a downlink channel between an access network device and the terminal device, the first channel state information being obtained through prediction by a first channel prediction model, and the second channel state information being determined based on measurements of a downlink signal from the access network device to the terminal device; transmitting first information to the access network device when a differential metric value between the first channel state information and the second channel state information is greater than or equal to a first threshold, the first information indicating the second channel state information; A calibration method including:

2. and transmitting second information to the access network device when the difference metric value between the first channel state information and the second channel state information is less than a second threshold, the second information indicating that the prediction result of the first channel prediction model is accurate, and the second threshold is less than or equal to the first threshold. The method of claim 1.

3. The method comprises: training an updated channel prediction model based on the second channel state information to obtain a second channel prediction model, wherein an initial value of the updated channel prediction model is the first channel prediction model; updating the target channel prediction model to the second channel prediction model; transmitting third information to the access network device, the third information indicating information regarding the second channel prediction model; The method of claim 1 further comprising:

4. the information regarding the second channel prediction model includes variation information between the second channel prediction model and the first channel prediction model. The method of claim 3.

5. The method comprises: and further comprising updating the first channel prediction model to the second channel prediction model. The method of claim 3.

6. The step of transmitting third information to the access network device includes: transmitting the third information to the access network device when it is determined that variation information between the first channel prediction model and the second channel prediction model is equal to or greater than a third threshold; The method of claim 3, comprising:

7. The method comprises: receiving fourth information from the access network device, the fourth information indicating information related to a third channel prediction model; The method of claim 1.

8. The method comprises: and further comprising updating the first channel prediction model to the third channel prediction model. The method of claim 7.

9. the information regarding the third channel prediction model includes variation information between the third channel prediction model and the first channel prediction model. The method of claim 7.

10. the first channel prediction model is configured by the access network device, and the first channel prediction model is identical to a channel prediction model in the access network device; The method of claim 1.

11. the first information further indicates a time unit corresponding to the second channel state information. The method of claim 1.

12. A calibration method performed by an access network device or a chip for said access network device, comprising: transmitting a downlink signal to a terminal device; receiving first information or second information from the terminal device; Including, the first information indicates second channel state information, the second channel state information being channel state information of a downlink channel between the access network device and the terminal device, the second channel state information being determined based on measurements on the downlink signal; the second information indicates that the prediction result of the first channel prediction model is accurate; Calibration methods.

13. If the first information is received, the method further comprises: training an updated channel prediction model based on the second channel state information to obtain a third channel prediction model, wherein an initial value of the updated channel prediction model is the first channel prediction model; updating the target channel prediction model to the third channel prediction model; transmitting fourth information to the terminal device, the fourth information indicating information about the third channel prediction model; The method of claim 12 further comprising:

14. the information regarding the third channel prediction model includes variation information between the third channel prediction model and the first channel prediction model. The method of claim 13.

15. The step of transmitting fourth information to the terminal device includes: transmitting the fourth information to the terminal device when it is determined that variation information between the first channel prediction model and the third channel prediction model is equal to or greater than a third threshold; The method of claim 13.

16. The method comprises: receiving third information from the terminal device, the third information indicating information related to a second channel prediction model; updating the first channel prediction model to the second channel prediction model; The method of claim 12 further comprising:

17. a differential metric value between first channel state information and the second channel state information is equal to or greater than a first threshold, and the first channel state information is obtained through prediction using the first channel prediction model; The method of claim 12.

18. A calibration method performed by an access network device or a chip for said access network device, comprising: determining first channel state information, the first channel state information being channel state information of a downlink channel between the access network device and a terminal device, and the first channel state information being obtained through prediction by a first channel prediction model; receiving first information from the terminal device, the first information indicating second channel state information, the second channel state information being channel state information of a downlink channel between the access network device and the terminal device, the second channel state information being determined based on measurements on a downlink signal from the access network device to the terminal device; When a differential metric value between the first channel state information and the second channel state information is greater than or equal to a first threshold, training the first channel prediction model based on the second channel state information to obtain a second channel prediction model; updating the first channel prediction model to the second channel prediction model; A calibration method including:

19. the first information further indicates a time unit corresponding to the second channel state information.

20. The method of claim 18.

20. A communication device configured to implement the method of any one of claims 1 to 11.

21. A communication device configured to implement a method according to any one of claims 12 to 19.

22. 1. A communications device comprising: a processor, the processor coupled to a memory; A communications device, wherein the processor is configured to execute computer programs or instructions stored in the memory to enable the communications device to implement a method according to any one of claims 1 to 11.

23. 1. A communications device comprising: a processor, the processor coupled to a memory; A communications device, wherein the processor is configured to execute computer programs or instructions stored in the memory to enable the communications device to implement a method according to any one of claims 12 to 19.

24. 20. A computer readable storage medium storing a computer program or instructions, which when executed on a computer, enables the computer to implement a method according to any one of claims 1 to 11, or enables the computer to implement a method according to any one of claims 12 to 19.

25. 20. A computer program comprising instructions, which when executed on a computer, enable the computer to implement a method according to any one of claims 1 to 11, or enable the computer to implement a method according to any one of claims 12 to 19.

26. A communication system comprising a communication device configured to implement a method according to any one of claims 1 to 11 and a communication device configured to implement a method according to any one of claims 12 to 19.

Images