Communication methods and related apparatuses

Abstract

This application provides a communication method and related apparatus, exemplarily applied in the field of wireless communication technology for transmitting CSI reports based on AI models. A terminal device sends a Channel State Information (CSI) report to a network device. The CSI report includes physical anchor points and self-test results. The self-test results are the terminal device's detection results for a first model and a second model. The first model is deployed on the terminal device, the second model is a proxy model for a third model, and the third model is deployed on the network device. The CSI report is used by the network device to locate the root cause of faults. Faults include data obtained based on the third model not meeting requirements. Multiple dimensions are used to determine the root cause of the fault, which helps the network device to determine the root cause with high accuracy.

Description

Technical Field

[0001] This application relates to the field of wireless communication technology, and in particular to a communication method and related apparatus. Background Technology

[0002] Channel state information reports are mainly used by terminals to feed back channel state information to the network, which is beneficial for the network to dynamically optimize transmission strategies.

[0003] With the development of artificial intelligence models, using these models to compress channel state information to save transmission resources is a growing trend. Correspondingly, the network reconstructs the compressed channel state information based on the model.

[0004] Currently, in scenarios where channel state information is transmitted based on a model, it is impossible to pinpoint the cause of the fault when the accuracy of the reconstructed channel state information observed by the network decreases. Summary of the Invention

[0005] In view of the above, this application provides a communication method and related apparatus to solve at least some of the aforementioned problems, and the disclosed technical solution is as follows:

[0006] Firstly, a communication method is provided. This method can be executed by a terminal device, or by a component (such as a circuit, chip, or chip system) configured in the terminal device, or by a logic module or software capable of implementing all or part of the functions of the terminal device. This application does not limit this approach. The following description uses a terminal device as an example.

[0007] The method includes: a terminal device sending a Channel State Information (CSI) report to a network device. The CSI report includes physical anchor points and self-test results. The self-test results are the terminal device's test results for a first model and a second model. The first model is deployed on the terminal device, the second model is a proxy model for a third model, and the third model is a model deployed on the network device. The CSI report is used by the network device to locate the root cause of a fault, including faults such as data obtained based on the third model not meeting requirements.

[0008] It is evident that the CSI report sent by the terminal device can lay the foundation for the network device to determine the root cause of the model failure. Furthermore, because the CSI report indicates physical anchor points, it can also lay the foundation for the network device to determine the root cause of the model failure based on the physical anchor points. Moreover, because the CSI report also indicates self-test results, the network device can also determine the root cause of the failure based on the self-test results of the terminal device. All of the above dimensions are used to determine the root cause of the failure, which is conducive to the high accuracy of the root cause determined by the network device.

[0009] In some implementations, before sending the Channel State Information (CSI) report to the network device, the process includes: the terminal device receiving first information from the network device, which indicates the initiation of a mechanism to locate the root cause of a fault. Initiating the fault location mechanism based on the first information allows the network to flexibly control its activation.

[0010] In some implementations, before sending the Channel State Information (CSI) report to the network device, the process includes: the terminal device receiving second information from the network device. This second information includes at least one of the following: self-test parameter configuration information, physical anchor point configuration information, and self-test result reporting mode configuration information. The self-test parameter configuration information is used by the terminal device to obtain the self-test result; the physical anchor point configuration information is used by the terminal device to obtain the physical anchor point; and the self-test result reporting mode configuration information is used by the terminal device to send the CSI report. This second information facilitates flexible network control over the terminal's self-testing.

[0011] In some implementations, the self-test parameter configuration information indicates the parameter type, baseline threshold, and threshold dynamic adjustment parameters. The parameter type indicates the type of parameter used as the basis for self-testing, while the baseline threshold and threshold dynamic adjustment parameters are used to obtain the threshold used by the terminal device for self-testing. Besides facilitating flexible network control of terminal self-testing, the self-test parameter configuration information, along with the baseline threshold and threshold dynamic adjustment parameters, allows the terminal to adaptively obtain the threshold used for self-testing based on channel conditions, thereby improving the accuracy of the self-test results.

[0012] In some implementations, the configuration information for physical anchors indicates the physical characteristics of the downlink channel that can serve as physical anchors, as well as the amount of data occupied by the physical anchor in the CSI report. Besides facilitating flexible network selection of physical anchors, the configuration information for physical anchors also helps to balance transmission overhead and decoding accuracy, because the amount of data occupied by the physical anchor in the CSI report relates to transmission overhead and the likelihood of accurately decoding the physical anchor.

[0013] In some implementations, the self-test result reporting mode configuration information indicates the reporting mode and the amount of data occupied by the self-test result in the CSI report. The reporting mode includes reporting the self-test result only when the self-test result indicates that it is unhealthy, which helps to save transmission overhead. It can also adaptively configure the amount of data occupied by the self-test result in the CSI report based on the decoding capability, thereby taking into account the decoding accuracy.

[0014] In some implementations, after sending the Channel State Information (CSI) report to the network device, the method further includes: the terminal device receiving third information from the network device. This third information is used to adjust the first model and / or the second model. Therefore, the third information is used to instruct the terminal device to adjust the model. For example, the third information is obtained based on the root cause located by the network device, which is beneficial for eliminating faults.

[0015] In some implementations, the third information includes at least one of the following: Hybrid Automatic Repeat Request (HARQ) information, indications for checking, restarting, or retraining the model, backoff indications, indications for adjusting the Modulation and Coding Scheme (MCS) or frequency points, indications for updating self-test rules, and indications for increasing the amount of data occupied by physical anchors in the CSI report. The model includes at least one of the first model and the second model. These indications are determined based on different fault root causes, which helps the terminal equipment to eliminate faults in a targeted manner.

[0016] In some implementations, before sending the Channel State Information (CSI) report to the network device, the process further includes: the terminal device determining a self-test result based on the relationship between parameter values ​​and a threshold used for self-testing. The parameter values ​​are calculated based on a first feature and a second feature. The first feature is a characteristic of the original channel, and the second feature is a characteristic output by the second model to the CSI compressed data. The CSI compressed data is the data output by the first model based on the characteristics of the original channel. This self-test process for the terminal device helps to obtain more accurate self-test results.

[0017] Secondly, a communication method is provided, which can be executed by a network device, or by a component (such as a circuit, chip, or chip system) configured in the network device, or by a logic module or software capable of implementing all or part of the functions of the network device. This application does not limit this. The following description uses a network device (such as a satellite) as an example.

[0018] The method includes: a network device receiving a Channel State Information (CSI) report from a terminal device. The CSI report includes physical anchor points and self-test results. The self-test results are the detection results of the terminal device on a first model and a second model. The first model is deployed on the terminal device, the second model is a proxy model of the third model, and the third model is a model deployed on the network device. The CSI report is used by the network device to locate the root cause of the fault, including the fault being that the data obtained based on the third model does not meet the requirements.

[0019] In some implementations, after receiving the Channel State Information (CSI) report from the terminal device, the network device further determines the root cause of the decreased accuracy of the third model output based on the CSI compressed data, physical anchor points, and self-test results in the CSI report. Determining the root cause of the decreased accuracy of the third model output from three dimensions (CSI compressed data, physical anchor points, and self-test results) helps to obtain more accurate results.

[0020] In some implementations, the root cause of the decreased output accuracy of the third model is determined based on the CSI compressed data, physical anchors, and self-test results in the CSI report. This includes: sequentially performing a first-level check, a second-level check, and a third-level check based on the block error rate. For any one of the first and second-level checks, if a first condition is met, the next check is performed. If the check fails, the root cause is determined based on at least one of the check results and the self-test results. If the block error rate exceeds the block error rate threshold, the root cause is determined based on the self-test results. The first-level check includes at least one of CRC and valid manifold check, and the second-level check includes a check based on physical anchors. The first condition includes a successful check. This serial three-level check process saves resources, is easy to implement, and can quickly locate the root cause of the fault.

[0021] In some implementations, the first condition also includes: the arbitration condition is not met, and the arbitration condition includes: the self-test result indicates a fault in the first model and / or the second model. Combining arbitration with triple verification helps to further improve the accuracy of the identified root cause.

[0022] Some implementations also include: if the valid manifold verification passes and the arbitration conditions are met, determining the root cause based on the similarity between the CSI compressed data and the CSI data from the terminal device. Passing the valid manifold verification and meeting the arbitration conditions indicates a situation where there is a significant conflict between the terminal device's self-test results and the network device's verification results. In this case, determining the root cause based on the similarity between the CSI compressed data and CSI data (such as CSI data not compressed by the AI ​​model) has higher accuracy.

[0023] In some implementations, when the block error rate exceeds a threshold, the root cause is determined based on the self-test results. This includes: when the block error rate exceeds the threshold and the self-test results indicate a healthy state, the root cause is determined based on the similarity between CSI compressed data and CSI data from the terminal device. A block error rate exceeding the threshold and a healthy self-test result represent another situation where there is a significant conflict between the self-test results of the terminal device and the verification results of the network device. The root cause determined based on the similarity between CSI compressed data and CSI data (such as CSI data without AI model compression) has higher accuracy.

[0024] The second aspect is the implementation on the network device side, which corresponds to the first aspect. The explanations, supplements, and descriptions of the beneficial effects of the first aspect also apply to the second aspect, and will not be repeated here.

[0025] Thirdly, a communication device is provided, comprising a transceiver module. The transceiver module is used to send a Channel State Information (CSI) report to a network device. The CSI report includes physical anchor points and self-test results. The self-test results are the detection results of the terminal device on a first model and a second model. The first model is deployed on the terminal device, and the second model is a proxy model for a third model deployed on the network device. The CSI report is used by the network device to locate the root cause of a fault, including a fault where the data obtained based on the third model does not meet requirements.

[0026] In some implementations, the communication device also includes a processing module for the terminal device to perform self-tests, etc.

[0027] Fourthly, a communication device is provided, comprising a transceiver module. The transceiver module is used to receive a Channel State Information (CSI) report from a terminal device. The CSI report includes physical anchor points and self-test results. The self-test results are the detection results of the terminal device on a first model and a second model. The first model is deployed on the terminal device, and the second model is a proxy model for a third model deployed on a network device. The CSI report is used by the network device to locate the root cause of a fault, including a fault where the data obtained based on the third model does not meet requirements.

[0028] In some implementations, the communication device also includes a processing module for determining the root cause of the decrease in the output accuracy of the third model (which may be caused by at least one fault of the first model, the second model, and the third model) based on the CSI compressed data, physical anchors, and self-test results in the CSI report.

[0029] The third and fourth aspects are the implementation on the device side, which correspond to the first and second aspects. The explanations, supplements, and descriptions of the beneficial effects of the first and second aspects also apply to the third and fourth aspects, and will not be repeated here.

[0030] Fifthly, a communication device is provided, including a processor. The processor is coupled to a memory and can be used to execute instructions or data in the memory to implement the method in any possible implementation of the first aspect described above. Optionally, the communication device further includes a memory. Optionally, the communication device further includes a communication interface, and the processor is coupled to the communication interface.

[0031] In one implementation, the communication interface may be a transceiver, or an input / output interface.

[0032] In another implementation, the communication device is a chip configured in a terminal device. When the communication device is a chip configured in a terminal device, the communication interface can be an input / output interface.

[0033] In a sixth aspect, a communication device is provided, including a processor. The processor is coupled to a memory and can be used to execute instructions or data in the memory to implement the method in any possible implementation of the second aspect described above. Optionally, the communication device further includes a memory. Optionally, the communication device further includes a communication interface, and the processor is coupled to the communication interface.

[0034] In one implementation, the communication interface may be a transceiver, or an input / output interface.

[0035] In another implementation, the communication device is a chip configured in a network device. When the communication device is a chip configured in a satellite, the communication interface can be an input / output interface.

[0036] In a seventh aspect, a processor is provided, comprising: an input circuit, an output circuit, and a processing circuit. The processing circuit is configured to receive signals through the input circuit and transmit signals through the output circuit, causing the processor to execute a method in any possible implementation of any aspect.

[0037] In specific implementation, the processor can be one or more chips, the input circuit can be input pins, the output circuit can be output pins, and the processing circuit can be transistors, gate circuits, flip-flops, and various logic circuits. The input signal received by the input circuit can be received and input by, for example, but not limited to, a receiver, and the signal output by the output circuit can be, for example, but not limited to, output to and transmitted by a transmitter. Furthermore, the input circuit and the output circuit can be the same circuit, which is used as both the input circuit and the output circuit at different times. This application does not limit the specific implementation of the processor and various circuits.

[0038] Eighthly, a communication device is provided, including a processor and a memory. The processor is used to read instructions stored in the memory, receive signals via a receiver, and transmit signals via a transmitter to execute the method in any possible implementation of any of the preceding aspects.

[0039] Optionally, the processor may be one or more, and the memory may be one or more.

[0040] Ninthly, a computer program product is provided, the computer program product comprising: a computer program (also referred to as code or instructions) that, when the computer program is run, causes a computer to perform a method in any possible implementation of any of the above aspects.

[0041] In a tenth aspect, a computer-readable storage medium is provided that stores a computer program (also referred to as code or instructions) that, when run on a computer, causes the computer to perform the methods in any possible implementation of any of the preceding aspects.

[0042] Eleventhly, embodiments of this application provide a chip system including one or more processors for calling and executing instructions stored in memory, causing the methods in any of the above aspects or possible implementations to be executed. The chip system may be composed of chips or may include chips and other discrete devices.

[0043] The chip system may include input circuits or interfaces for transmitting information or data, and output circuits or interfaces for receiving information or data.

[0044] In a twelfth aspect, a communication system is provided, including the aforementioned terminal device and network device. Optionally, the communication system may further include other devices that communicate with the terminal device and / or network device. Attached Figure Description

[0045] Figure 1 This is a schematic diagram of a communication system used in an embodiment of this application;

[0046] Figure 2 This is an example diagram of CSI data transmission based on a model;

[0047] Figure 3 A flowchart illustrating a communication method provided in an embodiment of this application;

[0048] Figure 4 A flowchart illustrating the process of locating the root cause of a network device fault in a communication method provided in this application embodiment;

[0049] Figure 5 This is a schematic diagram of the structure of a communication device provided in an embodiment of this application;

[0050] Figure 6 This is a schematic diagram of another communication device provided in an embodiment of this application. Detailed Implementation

[0051] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. The terminology used in the following embodiments is for the purpose of describing specific embodiments only and is not intended to be a limitation of this application. As used in the specification and appended claims of this application, the singular expressions "a," "an," "the," "the," "the," and "this" are intended to also include expressions such as "one or more," unless the context clearly indicates otherwise. It should also be understood that in the embodiments of this application, "one or more" refers to one, two, or more; "and / or" describes the relationship between related objects, indicating that three relationships may exist; for example, A and / or B can represent: A alone, A and B simultaneously, or B alone, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship.

[0052] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0053] The "multiple" mentioned in the embodiments of this application refers to two or more. It should be noted that in the description of the embodiments of this application, terms such as "first" and "second" are used only for the purpose of distinguishing descriptions and should not be construed as indicating or implying relative importance, nor should they be construed as indicating or implying order.

[0054] The technical solutions provided in this application can be applied to various communication systems, such as: Global System for Mobile Communications (GSM) systems, General Packet Radio Service (GPRS), Wireless Local Area Network (WLAN), Long Term Evolution (LTE) systems, LTE Frequency Division Duplex (FDD) systems, LTE Time Division Duplex (TDD) systems, sidelink communication systems, Universal Mobile Telecommunication System (UMTS), Worldwide Interoperability for Microwave Access (WiMAX) communication systems, non-terrestrial network (NTN) communication systems, 5th generation (5G) mobile communication systems, or new radio access technology (NR). Among these, 5G mobile communication systems can include non-standalone (NSA) and / or standalone (SA) networking. The technical solutions provided in this application can also be applied to future communication systems. This application does not limit the scope of these applications.

[0055] Figure 1 This is a schematic diagram of a communication system used in an embodiment of this application. The communication system may include network devices, such as... Figure 1 The network device 1 shown. This communication system may also include terminal devices, such as... Figure 1 The terminal device 2 shown. Network device 1 and terminal device 2 can communicate via a wireless link.

[0056] Figure 1 An exemplary network device 1 and a terminal device 2 are shown. Optionally, the communication system may also include multiple network devices and / or multiple terminal devices.

[0057] The network equipment in this application can be network-side equipment such as access network and core network equipment. Access network equipment is sometimes also called access node. Access network equipment has wireless transceiver capabilities and is used to communicate with terminals. Access network equipment includes, but is not limited to, base stations, evolved NodeBs (eNodeBs), transmission reception points (TRPs) in the above-mentioned communication systems, next-generation NodeBs (gNBs) in 5G mobile communication systems, access network equipment or modules of access network equipment in open RAN (ORAN) systems, satellites in NTN communication systems, base stations in future mobile communication systems, or access nodes in WiFi systems. Access network equipment can also be modules or units capable of implementing some of the functions of a base station. Access network equipment can be macro base stations, micro base stations, or indoor stations, relay nodes or donor nodes, or wireless controllers in cloud radioaccess network (CRAN) scenarios. Optionally, access network equipment can also be servers, wearable devices, or vehicle-mounted equipment, etc. For example, the access network equipment in vehicle-to-everything (V2X) technology can be a roadside unit (RSU). Multiple access network devices in a communication system can be base stations of the same type or different types. Base stations can communicate with terminals directly or via relay stations. Terminals can communicate with multiple base stations using different access technologies. The embodiments of this application do not limit the specific technology or device form used in the access network equipment. In this application, the access network equipment is referred to as a network device.

[0058] In this application, the means for implementing the functions of a network device can be a network device itself, or a means capable of supporting the network device in implementing those functions, such as a processor, circuit, chip, or chip system. This means can be installed in or connected to the network device. In the technical solutions provided in this application, the example of a network device being used to implement the functions of a network device is used to describe the technical solutions provided in this application.

[0059] The terminal device in this application can be a wireless terminal device capable of receiving network device scheduling and instruction information. The wireless terminal device can be a device providing voice and / or data connectivity to a user, a handheld device with wireless connectivity, or other processing devices connected to a wireless modem. For example, the terminal device can communicate with one or more core networks or the Internet via a radio access network (RAN). The terminal device can also be referred to as a terminal, user equipment (UE), mobile station, mobile terminal, etc. Terminal devices can be widely used in various scenarios, such as device-to-device (D2D), vehicle-to-everything (V2X) communication, machine-type communication (MTC), Internet of Things (IoT), ultra-reliable low-latency communication (URLLC), virtual reality, augmented reality, industrial control, autonomous driving, telemedicine, smart grids, smart furniture, smart offices, smart wearables, smart transportation, smart cities, or satellite communication, etc. The terminal can be a mobile phone, tablet computer, computer with wireless transceiver capabilities, wearable device, vehicle, aircraft (such as drone, helicopter, airplane), hot air balloon, ship, robot, robotic arm, or smart home device, etc. The embodiments of this application do not limit the form of the terminal device.

[0060] In this application, the apparatus for implementing the functions of a terminal device can be the terminal device itself, or any apparatus capable of supporting the terminal device in implementing those functions, such as a processor, circuit, chip, or chip system. This apparatus can be installed in or connected to the terminal device. In the technical solutions provided in this application, the example of a terminal device being used to implement the functions of a terminal device is used to describe the technical solutions provided in this application.

[0061] Access network equipment and / or terminal equipment can be fixed or mobile. They can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; on water; or in the air on aircraft, balloons, and satellites. This application does not limit the application scenarios of the access network equipment and terminal equipment. They can be deployed in the same or different scenarios; for example, both can be deployed on land simultaneously; or the access network equipment can be deployed on land while the terminal equipment is deployed on water, etc., and so on.

[0062] In practical applications, multiple network devices can collaborate to assist terminals in achieving wireless access, with different network devices each implementing a portion of the base station's functions. For example, network devices can be central units (CUs), distributed units (DUs), CUs (control planes, CPs), CUs (user planes, UPs), or radio units (RUs), etc. CUs and DUs can be set up separately or included in the same network element, such as a baseband unit (BBU). RUs can be included in radio frequency equipment or radio frequency units, such as remote radio units (RRUs), active antenna units (AAUs), or remote radio heads (RRHs).

[0063] In different systems, CU (or CU-CP and CU-UP), DU, or RU may have different names, but those skilled in the art will understand their meaning. For example, in an ORAN system, CU can also be called O-CU (Open CU), DU can also be called O-DU, CU-CP can also be called O-CU-CP, CU-UP can also be called O-CU-UP, and RU can also be called O-RU. Any of the units among CU (or CU-CP, CU-UP), DU, and RU in this application can be implemented through software modules, hardware modules, or a combination of software and hardware modules. CU (or CU-CP and CU-UP), DU, and RU can implement different protocol layer functions.

[0064] To facilitate understanding of the embodiments of this application, the terms involved in this application will be briefly explained first.

[0065] Channel state information (CSI) is defined as a set of measurements of the physical characteristics of the radio channel by the user equipment (UE), used to quantify downlink or uplink channel conditions. Its core function is to provide feedback to the base station (gNB) to assist in adjusting transmission parameters to optimize radio link performance.

[0066] Alternatively, the explanations of some terms can also be found in the 3rd Generation Partnership Project (3GPP) standard protocol.

[0067] To reduce the overhead of feedback channel state information, artificial intelligence (AI) models are deployed on both the terminal and the network. Specifically, a two-sided model is deployed in the communication system, with an encoder deployed on the terminal side and a decoder deployed on the network side.

[0068] by Figure 2 For example, the terminal inputs the measured CSI data into the encoder to obtain the CSI compressed data output by the encoder. The terminal sends the CSI compressed data to the access network device to save transmission resources. The network side inputs the CSI compressed data into the decoder to obtain the CSI reconstructed data output by the decoder.

[0069] The reasons for the decrease in accuracy of CSI reconstructed data include:

[0070] Model failure refers to a situation where the encoder deployed on the terminal or the decoder deployed on the network is inconsistent with or damaged compared to the trained model.

[0071] Link failure: refers to the introduction of uncorrectable bit errors during air interface transmission, resulting in a mismatch between the received compressed bits and the transmitted compressed bits.

[0072] Data drift refers to a mismatch between the current channel environment statistical characteristics, such as the distribution of input data, and the distribution of the training dataset used for model training.

[0073] The inventors discovered that when the network side observes a decrease in the accuracy of CSI reconstructed data, it is often difficult to determine the specific cause of the fault, and therefore cannot carry out targeted processing.

[0074] To address the aforementioned technical problems, embodiments of this application provide a communication method in which a terminal device sends a Channel State Information (CSI) report to a network device. The CSI report includes physical anchor points and self-test results. The self-test results are the terminal device's detection results for a first model and a second model. The first model is deployed on the terminal device, and the second model is a proxy model for a third model deployed on the network device. The CSI report is used by the network device to locate the root cause of a fault, where the fault includes data obtained based on the third model not meeting requirements. Therefore, the terminal device can provide the network device with a basis for locating the root cause of a fault, which is beneficial for the network device to locate the cause of the deployed model's failure. Furthermore, providing evidence from multiple dimensions, including the self-test structure and physical anchor points, enables more accurate location of the root cause.

[0075] Correspondingly, the network device receives a Channel State Information (CSI) report from the terminal device. The CSI report includes physical anchor points and self-test results. The self-test results are the detection results of the terminal device on the first model and the second model. The first model is deployed on the terminal device, and the second model is a proxy model of the third model. The third model is a model deployed on the network device. The CSI report is used by the network device to locate the root cause of the fault. The fault includes data obtained based on the third model that does not meet the requirements, thereby laying the foundation for the network device to locate the cause of the deployed model failure.

[0076] The solution provided in this application will be described in detail below with reference to the corresponding flowcharts. It is understood that the illustrative flowcharts provided in this application primarily use different devices (e.g., terminal devices, network devices) as examples of the execution subjects of this interactive illustration to illustrate the method, but this application does not limit the execution subjects of the interactive illustrations. For example, the devices (e.g., terminal devices, network devices) in the illustrative flowcharts can also be chips, chip systems, or processors that support the implementation of this method on the device, or logic modules or software that can implement all or part of the functions of the device.

[0077] As a general statement, the message or signaling interactions involved in the interaction process of this application embodiment can be standard messages or signaling or newly introduced messages or signaling. This application embodiment does not make specific limitations on this.

[0078] Figure 3 This is a schematic diagram illustrating a communication method according to an embodiment of this application. It can be understood that... Figure 3 The terminal device in the middle can be Figure 1 Any terminal device or Figure 2 Any terminal in the context can also refer to a device within a terminal device (such as a processor, chip, or chip system). Network devices can be... Figure 1 Any access network device or Figure 2 The term "access network equipment" can also refer to devices within the access network equipment (such as processors, chips, or chip systems).

[0079] like Figure 3 As shown, the method includes the following steps:

[0080] S101. The network device sends radio resource control (RRC) signaling, and the terminal device receives the RRC signaling accordingly.

[0081] For example, RRC signaling carries CSI report configuration (CSI-ReportConfig) information, which includes a newly added IE (Information Provider) used to indicate collaborative diagnosis configuration information. This newly added IE is referred to as the first IE, and may also be called the CSI-AIDiagnosis-Config IE.

[0082] The collaborative diagnostic configuration information (also known as the second information) includes: self-test parameter configuration information, physical anchor point configuration information, and self-test result reporting mode configuration information.

[0083] Self-test parameters are parameters used by terminal devices for self-testing, including: parameter type, baseline threshold, and threshold dynamic adjustment parameters. The self-test parameter configuration information indicates the parameter type, baseline threshold, and threshold dynamic adjustment parameters. The parameter type indicates the type of parameter used as the basis for self-testing. The baseline threshold and threshold dynamic adjustment parameters are used to obtain the threshold used by the terminal device for self-testing. For example, the parameter type includes at least one of squared generalized cosine similarity (SGSC) and mean squared error (MSE), and the baseline threshold includes a first baseline threshold T. high_base Second reference threshold T low_base The threshold dynamic adjustment parameters include: the offset coefficient and the signal-to-noise ratio (SNR) of the reference signal, denoted as SNR. ref The function and usage of each parameter will be explained in detail in subsequent steps.

[0084] For example, in the newly added IE in CSI-ReportConfig, the self-test parameter configuration information indicates that the parameter types SGSC and MSE are optional, and indicates that the default parameter type is SGSC. The terminal can select SGSC or MSE based on the self-test parameter configuration information.

[0085] For example, in the newly added IE in CSI-ReportConfig, the self-test parameter configuration information also indicates information on the optional first benchmark threshold (such as an identifier) ​​and the second benchmark threshold (such as an identifier). The terminal device can learn about the first benchmark threshold and the second benchmark threshold through the identifier based on this indication.

[0086] For example, in the newly added IE in CSI-ReportConfig, the self-test parameter configuration information also indicates information (such as an identifier) ​​about the offset coefficient and the signal-to-noise ratio of the reference signal. The terminal device can learn about the offset coefficient and the signal-to-noise ratio of the reference signal based on this indication and the identifier.

[0087] A physical anchor point is a physical characteristic that fully represents the channel; that is, a physical anchor point is a strong physical characteristic among the channel's physical features. Physical characteristics that can serve as physical anchor points include at least one of the following:

[0088] Spatial beam characteristics, such as at least one of the following: main beam index, angle spread, wideband precoding matrix indicator (PMI), and rank indicator (ri);

[0089] Delay domain features, such as at least one of delay spread and strongest path delay;

[0090] Frequency domain characteristics, such as at least one of coherence bandwidth and subband CQI;

[0091] Energy characteristics, such as at least one of the reference signal received power (RSRP) and channel norm.

[0092] The configuration information of physical anchors indicates the physical characteristics that can be used as physical anchors. For example, the configuration information of physical anchors includes the identifier of the main beam index, that is, the configuration information of physical anchors indicates that the main beam index is used as a physical anchor.

[0093] For example, the configuration information of the physical anchor also indicates the amount of data occupied by the physical anchor in the CSI report. For example, the configuration information of the physical anchor indicates a low amount of data, such as 3-6 bits, to save transmission resources.

[0094] For example, the new IE in CSI-ReportConfig indicates that various physical features can be selected as physical anchors, and also indicates information such as the number of bits occupied by the physical anchor (e.g., identifier). The terminal device can optionally select at least one of the indicated physical features as a physical anchor, and know the number of bits occupied by the physical anchor in the CSI report based on the bit count information.

[0095] The self-test result reporting mode configuration information is used to indicate the mode for reporting self-test results, including the reporting mode and the amount of data occupied by the self-test results in the CSI report. The details will be explained in detail in subsequent steps.

[0096] For example, in the newly added IE in CSI-ReportConfig, the self-test result reporting mode configuration information indicates that the reporting mode is only reported when there is an anomaly. That is, the self-test result is reported when the terminal device's self-test result is abnormal (i.e., unhealthy), and not reported when it is normal (i.e., healthy), in order to save costs. The self-test result reporting mode configuration information also indicates that the self-test result occupies 2 bits in the CSI report. It can be understood that 2 bits can represent 4 types of self-test results. Combined with reporting only when there is an anomaly, there may be 3 types of self-test results reported.

[0097] For example, after the terminal device completes the access procedure to the cell, it sends RRC signaling. For example, the RRC signaling is an RRC configuration message or an RRC reconfiguration message.

[0098] S102. After the network device detects a performance degradation in the decoder, it sends a mechanism activation instruction (also known as the first message), and correspondingly, the terminal device receives a mechanism activation instruction.

[0099] For example, network devices determine whether to activate a performance degradation root cause identification and response mechanism based on CSI compressed data, CSI raw data, and CSI reconstructed data.

[0100] For example, the terminal device sends CSI compressed data and CSI raw data to the network device. The CSI raw data is CSI data obtained by the terminal device through measurement without undergoing encoding and compression processing by an encoder. The CSI compressed data is the data obtained by the terminal device by inputting the CSI raw data into an encoder. The network device inputs the CSI compressed data into a decoder to obtain the CSI reconstructed data output by the decoder.

[0101] For example, decoder performance degradation includes decoder failures in network device deployments, such as the accuracy of CSI reconstructed data obtained based on the decoder not meeting requirements.

[0102] For example, the network device calculates at least one of the following parameters:

[0103] SGSC of CSI raw data and CSI reconstructed data;

[0104] Generalized cosine similarity (GSC) between CSI original data and CSI reconstructed data.

[0105] Normalized mean squared error (NMSE) between raw CSI data and reconstructed CSI data.

[0106] MSE of raw CSI data and reconstructed CSI data;

[0107] The structural similarity index (SSIM) between the original CSI data and the reconstructed CSI data.

[0108] The calculation methods for each of the above parameters can be found in existing technologies, and will not be repeated here.

[0109] For example, if all the above parameters meet the corresponding threshold conditions, the network device determines that both the model deployed by the terminal device and the model deployed by the network device meet the requirements, and the performance degradation root cause identification and response mechanism is not activated. If any of the above parameters does not meet the corresponding threshold conditions, the performance degradation root cause identification and response mechanism is activated. Meeting the corresponding threshold conditions can be understood as meeting the relationship with the threshold value, for example, being greater than the threshold value. The corresponding threshold value is a threshold value that is adapted to the parameter.

[0110] The mechanism startup indication is used to inform the system for identifying and addressing the root causes of performance degradation (i.e., the mechanism for locating the root cause of the fault). The mechanism startup indication is also known as the first message.

[0111] For example, the mechanism initiation instruction is transmitted within a medium access control (MAC) control element (CE). That is, the network device sends a MAC CE carrying the mechanism initiation instruction.

[0112] S103. The terminal device initiates the mechanism and sends a CSI report, and the network device receives the CSI report accordingly.

[0113] The CSI report includes: CSI compressed data, physical anchor points extracted from the terminal device, and the terminal device's self-test results.

[0114] For example, the terminal device extracts the value of the physical feature indicated by the physical anchor point configuration information in the RRC signaling as the physical anchor point. For example, the configuration information of the physical anchor point indicates the rank indicator and the main beam index are optional. The terminal device performs a lightweight traditional algorithm on the downlink channel H, such as peak finding based on discrete Fourier transform (DFT), to extract the main beam index, denoted as the physical anchor point P. anchor .

[0115] The self-test result is obtained by the terminal device based on the self-test parameter configuration information in the RRC signaling and the proxy decoder. The proxy decoder is obtained by training a model in advance by the terminal device based on training data indicated by the training data information in the RRC signaling. The process of the terminal device training the proxy decoder will be described in detail in the following embodiments.

[0116] For example, the process by which a terminal device obtains a self-test result includes: determining the self-test result based on the relationship between the value of a parameter and the threshold used for self-testing. The value of the parameter is calculated based on a first feature and a second feature. The first feature is a feature of the original channel, and the second feature is a feature output by the second model to the CSI compressed data. The CSI compressed data is the data output by the first model to the feature of the original channel. The self-test process will be illustrated in detail below.

[0117] The terminal device inputs the measurement parameters (such as the feature matrix, i.e., the first feature) of the original downlink channel H into the encoder to obtain the compressed feature vector Z. AI (CSI compressed data), reconstructing the downlink channel (i.e., the second feature) using a proxy decoder: H local =proxy decoder(Z) AI (1), that is, equation (1) can be regarded as a simulation of the CSI reconstruction of the decoder deployed on the network device to obtain the CSI reconstruction data.

[0118] The terminal device calculates the value of the parameter type indicated by the self-test parameter configuration information, for example, the squared generalized cosine similarity SGSC=Corr(H, H). local ')(2).

[0119] The terminal device generates a self-test result (HI) based on the similarity S. The bits occupied by HI are indicated by the self-test result reporting mode configuration information in the RRC signaling. For example, HI occupies 2 bits.

[0120] If S>T high If the self-test result of the terminal device indicates that the model is working normally, then HI is set to 00 (indicating healthy).

[0121] If T low <S<T high If the self-test result of the terminal device is data drift or the encoder and the proxy decoder of the terminal device are not matched, that is, the physical characteristics of the channel are normal but the accuracy is reduced, then HI is set to 01 (indicating unfit).

[0122] If S <T low If the self-test result of the terminal device indicates that at least one of the encoder and the proxy decoder has failed, then HI is set to 10.

[0123] The value of HI and the relationship (S>T) high T low <S<T high S <T lowThe correspondence between the network devices and the terminal devices can be pre-configured in the terminal devices. The terminal devices set the HI value based on this correspondence. The meaning of the HI value (such as 00 indicating health) can be pre-configured in the network devices and terminal devices.

[0124] For example, T high and T low The method of obtaining it is:

[0125] T high =T high_base +6×(SNR) current -SNR ref (3),

[0126] T low =T low_base +6×(SNR) current -SNR ref (4).

[0127] In equations (3) and (4), T high_base The first baseline threshold, T, is indicated by the self-test parameter configuration information carried in the RRC signaling. low_base The second baseline threshold is indicated by the self-test parameter configuration information carried in the RRC signaling; 6 is the offset coefficient indicated by the self-test parameter configuration information carried in the RRC signaling; SNR ref The signal-to-noise ratio (SNR) of the reference signal indicated by the self-test parameter configuration information carried in the RRC signaling. current The current signal-to-noise ratio (SNR) measured for the terminal device. It can be seen that the terminal device can adjust the signal-to-noise ratio based on channel quality (determined by SNR). current (Reflection), based on the benchmark threshold, adjustments are made to obtain the threshold T. high and T low When the offset coefficient is greater than 0, it is possible to achieve a threshold T at low signal-to-noise ratio. high and T low Threshold T at lower and higher signal-to-noise ratios high and T low The higher accuracy eliminates interference from environmental noise, effectively avoids false alarms and missed detections, and ensures the accuracy of diagnosing the health status of AI models.

[0128] As mentioned earlier, the self-test result reporting mode configuration information carried in the RRC signaling indicates the mode in which the terminal device reports the self-test result. For example, if the self-test result reporting mode configuration information indicates that the HI field is not reported when HI is 00, in order to save resources. Based on this, if the self-test result HI is not 00, the terminal device must report the HI field.

[0129] For example, the terminal device sends a CSI report by sending a physical uplink control channel (PUCCH) or a physical uplink shared channel (PUSCH).

[0130] For example, a CSI report is in the form of {Payload:Z AI , Info:[HI, P anchor ]}, where Payload represents the load and Info represents the information. As mentioned earlier, the configuration information of the physical anchor point indicates P anchor In the CSI report, the self-test result reporting mode configuration information indicator HI occupies 3-6 bits, while Z occupies 2 bits in the CSI report. AI CSI data is compressed, and HI may not be included in the CSI report. As can be seen, the data volume of the CSI report is small, which can save transmission resources.

[0131] For example, the Rel-15 Type I PMI field is reused to carry P anchor The Rel-15 Type I PMI field is the bit structure and encoding method defined for the Type I codebook (Type I CSI) in the 3GPP 5G NR Release 15 standard (Release 15 is the version number). The Rel-15 Type I PMI field was originally used to transmit wideband PMI or RI. Here, the bit structure and encoding method are reused to report physical anchors, which include, but are not limited to, wideband PMI and RI. Therefore, this reuse does not occupy the original information but rather extends it.

[0132] It is understandable that terminal devices initiate the CSI report based on the mechanism, but if the network device does not detect a performance degradation, the terminal device will not report the CSI report, which can further save resources.

[0133] S104. Network devices perform triple verification and arbitration, and, in conjunction with CSI reports, locate the root cause of the fault.

[0134] In this embodiment, the network device performs verification in multiple dimensions, namely triple verification, and based on the results of triple verification and the self-test results of the terminal device in the CSI report, the root cause of the fault is determined comprehensively, which has high accuracy.

[0135] The specific procedures for S104 will be combined with Figure 4 Please provide a detailed explanation.

[0136] S105. Based on the root cause of the fault, the network device performs a first adjustment operation and / or sends an adjustment instruction, and correspondingly, the terminal device receives the adjustment instruction and / or performs a second adjustment operation.

[0137] For example, the operations performed or adjustment instructions sent based on different fault causes are shown in Table 1:

[0138] Table 1

[0139]

[0140]

[0141] For example, the terminal device can determine the root cause of the fault and corresponding countermeasures by referring to Table 1 based on the results of triple verification, arbitration, and the self-test result HI of the terminal device in the CSI report. Specifically, this will be combined with... Figure 4 The process is illustrated below. The lookup table method is easy to implement and highly efficient, facilitating rapid identification and resolution of faults.

[0142] The following section, in conjunction with Table 1, provides a detailed explanation of the process for locating the root cause of network device failures and corresponding countermeasures. Figure 4 As shown, it includes the following steps:

[0143] First layer of verification (CRC + valid manifold check):

[0144] S201. Perform a cyclic redundancy check (CRC). If the CRC passes, proceed to S202; otherwise, proceed to S203.

[0145] For methods of CRC verification, please refer to existing technologies, which will not be elaborated here.

[0146] S202. The root cause of the fault is determined to be an explicit link fault.

[0147] An explicit link failure can be understood as a link failure that corrupts the data, i.e., a physical layer transmission error, resulting in a corrupted data packet. This causes the CRC calculated by the network device (generated by the terminal device's RINT) to not match the CRC carried in the CSI compressed data. In this case, subsequent checks cannot be performed.

[0148] Based on the characteristics of CRC, as long as CRC fails, a visible link failure is determined regardless of the self-test result of the terminal device. Therefore, it is not necessary to combine the CSI report to determine the cause of the failure in this step.

[0149] Referring to Table 1, S201-S202 correspond to case number 1 in Table 1 ( Figure 4(represented by (1) in S202), correspondingly, the network device discards the data packet received from the terminal device and triggers the terminal device to retransmit the data packet, for example, triggering a hybrid automatic repeat request (HARQ) retransmission.

[0150] S203. Perform valid manifold verification. If the valid manifold verification passes, proceed to S205; otherwise, proceed to S204.

[0151] An effective manifold refers to a structure where the compressed vectors output by a well-trained AI encoder are not randomly distributed in high-dimensional space, but rather clustered on a specific geometric surface (manifold). If the data deviates from this surface, the data structure is corrupted. Data drift does not alter the effective manifold.

[0152] For example, calculating Z in the CSI report AI The distance to the pre-trained "standard manifold" is used to determine the validity of the manifold. If the distance D satisfies D < the first threshold (threshold1), the validity of the manifold is successfully verified. If the distance D satisfies D > threshold1, the validity of the manifold is not verified.

[0153] The pre-trained "standard manifold" is saved during the training of the encoder-decoder model pair. Essentially, the "standard manifold" is a mathematical summary of "what shape the AI ​​model's output should be" during the offline training phase. It is generated using statistical patterns of feature vectors in the training set or auxiliary models while training the decoder, and preloaded into the network device's inference engine.

[0154] For example, network devices compute Z using a lightweight autoencoder or a one-class support vector machine (SVM). AI The distance to the "standard manifold".

[0155] S204. Based on the self-test results of the terminal equipment, determine the root cause of the fault.

[0156] For example, when HI=00 (i.e., the terminal device passes the self-test; for example, the CSI report does not include the value of HI in this case), it is an implicit link failure or malicious interference, corresponding to case number 7 in Table 1. An implicit link failure refers to a special transmission error in which data is damaged and becomes garbled during transmission, but accidentally passes the physical layer CRC due to reasons such as CRC collision (i.e., CRC missed detection). Ultimately, only manifold inspection can detect that the data is no longer usable.

[0157] The corresponding countermeasures are: packet loss and / or adjustment of modulation and coding scheme (MCS) (or frequency point). For example, the received data packet in this instance (a single occurrence) is discarded. If the case number 7 occurs consecutively, the MCS (or frequency point) is adjusted.

[0158] For example, when HI=10 (self-test result indicates encoder or proxy decoder failure), it indicates an encoder failure, corresponding to case number 2 in Table 1. This means the terminal device failed the self-test, and the output data failed the manifold check. The corresponding response is as follows: the network device instructs the terminal device to use the traditional CSI report reporting mode. The traditional reporting mode reports the CSI report using the bit structure and encoding method without encoder compression of the CSI data, and the reported content is a Type I codebook index (the traditional CSI report reporting mode can be simply referred to as Type I codebook). Alternatively, the network device instructs the terminal device to restart the encoder.

[0159] For example, when HI=01 (self-test result is decreased accuracy), it is an encoder failure, corresponding to case number 6 in Table 1. In this case, the terminal device considers the accuracy to be decreased, but the data sent is completely unusable (valid manifold verification fails). Therefore, it is also handled as an encoder failure, that is, the response measures are the same as those for case number 2: return to traditional mode or restart the encoder.

[0160] S205. Determine whether the self-test result HI of the terminal device is 10. If yes, execute S211; otherwise, execute S206.

[0161] The inventors discovered that if the valid manifold verification passes and HI=10, it indicates a significant conflict between the self-test result of the terminal device and the judgment result of the network device. In this case, the "gold standard" is introduced for arbitration, as detailed in S211.

[0162] Second layer of verification (physical anchor point alignment verification):

[0163] S206. Determine if the physical anchor points are aligned. If yes, proceed to S208; otherwise, proceed to S207.

[0164] The physical anchor point in the CSI report is denoted as P. anchor The physical anchor point extracted by the network device is denoted as P. recon For example, a network device uses a decoder to convert Z... AI Reconstruct the channel H' and extract its physical anchor P from H'. recon .

[0165] If P recon equals Panchor Then it is assumed that the physical anchor points are aligned.

[0166] If P recon Not equal to P anchor If so, it is considered that the physical anchor points are misaligned.

[0167] S207. Determine the root cause of the fault and corresponding countermeasures based on local testing.

[0168] For example, the network device might assume that the physical anchor misalignment is caused by encoder-decoder mismatch or decoder failure. To further pinpoint the root cause, the network device performs a local test, which involves loading a set of locally stored, known "standard feature vectors" Z. ref (its corresponding physical anchor point P) ref (Known), then use the currently deployed decoder to Z ref Decode and extract the physical anchor point to obtain P.

[0169] If P is not equal to P ref If the local test fails, and HI=00, the root cause of the fault is considered to be a decoder failure, i.e., case number 9 in Table 1. The corresponding measures are: reload the decoder parameters or switch to a backup decoder. For example, the switch is a "hot" switch, that is, switching to a backup decoder (instance) without notifying the terminal device.

[0170] If P equals P ref If the local test is successful, then HI=00 indicates that the decoder and encoder are mismatched (i.e., mismatched), which is the case number 8 in Table 1. The corresponding countermeasures for this case are: the network device instructs the terminal device to synchronize the training dataset and model. For example, the network device sends the training dataset information to the terminal device, and the terminal device selects to retrain the proxy decoder and encoder according to the dataset ID in the training dataset information. Alternatively, the terminal device switches the proxy decoder and encoder corresponding to the dataset ID. Or, the network device does not send the training dataset information, but instructs the terminal device to switch the encoder to the target encoder, which is the encoder trained on the dataset corresponding to the dataset ID.

[0171] For example, if the physical anchor points are misaligned, and HI=01, this test can be skipped. This is considered a serious data drift, corresponding to case number 5 in Table 1. In other words, the terminal device's self-test result indicates a decrease in accuracy, while the network device finds that although the channel structure still resembles CSI, the physical direction has deviated. This may be because the channel characteristics have exceeded the training domain, and the model can no longer correctly represent the current channel. The countermeasure is to trigger the CSI data acquisition process, such as triggering the terminal device to collect CSI data and send it to the network device, so that the network device can retrain the encoder-decoder model pair.

[0172] Third verification:

[0173] S208. Determine if the error rate of the business is high. If yes, execute S209. If no, end the process.

[0174] Block Error Rate (BLER) refers to the ratio of the number of erroneous data blocks received by the receiving end (i.e., network device) to the total number of data blocks sent. That is, Block Error Rate = Number of erroneous data blocks / Total number of data blocks transmitted.

[0175] Errored data blocks are identified as follows: if a data block fails the CRC check at the receiving end, it is recorded as an error (i.e., a "mistaken block"). A higher error rate indicates more severe packet loss and more unstable communication.

[0176] For example, the block error rate can be compared with a corresponding threshold to determine whether the block error rate is high.

[0177] S209. Determine whether the self-test result HI of the terminal device is 00. If yes, execute S212; otherwise, execute S210.

[0178] The inventors discovered that the actual business error rate is high and HI=00, which indicates a significant conflict between the self-test results of the terminal equipment and the judgment results of the network equipment. In this case, the "gold standard" is introduced for arbitration, as detailed in S212.

[0179] S210. Determine the root cause of the fault based on the self-test results of the terminal device.

[0180] For example, if HI=01, the root cause of the fault is determined to be slight data drift or quantization noise, i.e., case number 4 in Table 1. In other words, the terminal device considers the accuracy insufficient, but the network device considers the structure valid and the physical direction (anchor point) correct. The corresponding response is to continue observation or fine-tuning without interrupting the service. If this situation occurs frequently, the terminal device will be triggered to fine-tune the encoder parameters online.

[0181] S211. Use the "gold standard" for arbitration.

[0182] For example, the network device sends a MAC CE to the terminal device. The MAC CE carries a standard information acquisition instruction and receives CSI data from the terminal device. This CSI data is uncompressed CSI data (or low-precision CSI data), which serves as the "gold standard." Understandably, the "gold standard" reflects the characteristics of the actual channel measured by the terminal device. The uncompressed CSI data is the Type I codebook index calculated based on the traditional 3GPP Rel-15 standard, which has lower precision but stronger robustness. Using low-precision feedback as a reference benchmark ("gold standard") in the arbitration mechanism aims to verify the effectiveness of the proxy decoder's monitoring and accurately locate the root cause of the fault when a conflict arises between the terminal device's self-test result and the network device's verification result. This is achieved by comparing the similarity between the "gold standard" and the network device's reconstructed channel.

[0183] For example, the network device obtains CSI reconstructed data through the decoder and compares the CSI reconstructed data with the "gold standard". If the similarity between the two is high, combined with HI=10, it can be known that the terminal device's self-test has failed. However, the network device's task data structure is complete (valid manifold verification has passed). It is possible that the terminal device's monitoring threshold is too strict or the proxy decoder parameters are outdated. Corresponding to case number 3 in Table 1, it is considered that the proxy decoder monitoring has a slight false alarm.

[0184] The countermeasures for minor false alarms detected by the proxy decoder are as follows: the network device instructs the terminal device to relax the self-test threshold or update the proxy decoder parameters.

[0185] For example, the network device may consider the reason for the proxy decoder detecting a minor false alarm to be that the terminal device calculates T. high and T low This is not entirely accurate. For example, if the settings are too strict relative to the current channel conditions, the proxy decoder may misjudge (identifying normal compressed data samples that meet the preset reconstruction accuracy requirements as abnormal samples). Therefore, the network device sends a dynamic threshold bias to the terminal device. This dynamic threshold bias is not an absolute threshold, but an offset. The terminal device adjusts the threshold based on this dynamic threshold bias, i.e., T high =T high -Bias, T low =T low -Bias. In this embodiment, this method is referred to as parameter fine-tuning, and the level is divided into Level 1.

[0186] If the similarity between the two is low, the decoder is considered to have made an incorrect decision. The network device then checks the decoder, corresponding to case number 12 in Table 1. The corresponding measures are: verifying the model ID to resolve the pairing mismatch issue, or performing model reloading and reset after confirming a decoder software fault. For example, first check the model pairing status. The network device checks whether the currently used decoder model ID matches the encoder model ID reported by the terminal device. If the IDs are inconsistent (i.e., pairing mismatch), the network device reconfigures or switches the decoder to ensure synchronization with the encoder version of the terminal device (re-pairing). Next, check the decoder's functional integrity. If the IDs match, the network device further checks whether the decoder software module is malfunctioning (e.g., corrupted weights, memory errors). If a functional fault in the decoder is confirmed, the network device performs a decoder reset, model reload, or reports to the OAM system for repair, while temporarily instructing the terminal device to revert to Type I reporting mode.

[0187] S212. If the similarity between the CSI reconstructed data and the "gold standard" is low, it is determined to be a missed detection by the proxy decoder. If the similarity between the CSI reconstructed data and the "gold standard" is high, it is determined to be a non-AI model failure.

[0188] If the CSI reconstructed data has a high similarity to the "gold standard", it indicates that the AI ​​model itself is working normally. An excessively high BLER is caused by non-AI model failures (such as external interference, improper scheduling strategies, etc.).

[0189] The low similarity between the CSI reconstructed data and the "gold standard" indicates a successful diagnosis, and the BLER may have recovered or fluctuated. This corresponds to case number 10 in Table 1. The corresponding strategy is for the network device to cease troubleshooting the model's root cause and instead investigate interference or scheduling issues. For example, a successful diagnosis means that the self-test result reported by the terminal device (e.g., the value of the HI field) indicates health (e.g., HI=00), and the results of the cyclic redundancy check, valid manifold check, and physical anchor comparison performed by the network device are all passed or matched. In other words, the terminal device's self-test and the network device's first and second checks both passed.

[0190] The CSI reconstructed data shows a high degree of similarity to the "gold standard," indicating that the diagnosis passed smoothly. However, the error rate is high, as shown in Table 1, which is case number 11. This indicates that the proxy decoder deployed on the terminal device missed detection, suggesting that the current monitoring method is not sensitive enough. The corresponding strategy is to instruct the terminal device to increase the self-test threshold or increase the number of anchor bits, or to perform a light calibration on the terminal device.

[0191] Lightweight calibration is classified as Level 2, which addresses situations where accuracy deviations or missed detection rates increase, but the system is not completely ineffective.

[0192] For example, a network device trains a very small linear adapter or bias vector parameter (at the KB level) and sends this parameter to the terminal device via RRC signaling or MAC CE. The terminal device then adjusts the proxy decoder based on this adoption count: H' local =W×Proxy Decoder(Z AI )+b, where W and b are linear transformation matrices or Bias vectors.

[0193] For example, this embodiment also sets up a three-level (Level 3) mechanism: the last resort to be taken after trying Level 1 (parameter fine-tuning) or Level 2 (light calibration) strategies at least three times and still failing to adjust to normal, that is, the gap between the output of the proxy decoder of the terminal device and the "gold standard" (Type I low precision feedback) is still too large (that is, the proxy decoder has completely failed or the calibration is ineffective).

[0194] The measures to address the complete failure or ineffective calibration of the proxy decoder (Level 3) are as follows: disable the self-test function of the terminal device, fall back to network-side single-sided monitoring, and mark the terminal device's dataset as needing to be re-collected for future model training.

[0195] The methods provided in the above embodiments have the following advantages:

[0196] 1. Low overhead: Utilizes HI's silent mechanism (reporting only in abnormal states 01 or 10), saving bit overhead; physical anchors reuse existing fields, eliminating the need for a large amount of new load.

[0197] 2. Fine-grained diagnosis: Through "triple verification" (CRC + manifold check, physical anchor alignment, and data drift judgment), it can specifically distinguish whether the data drift is caused by hardware failure, damaged model parameters, model mismatch, or environmental changes.

[0198] 3. Strong robustness: Combined with SNR to dynamically adjust the self-test threshold, it reduces false alarms and false negatives.

[0199] 4. Self-healing capability: It provides a hierarchical management strategy (Level 1-3) for the terminal-side monitoring model, which can correct the deviation of the monitoring model online.

[0200] 5. Unlike relying solely on blind testing on the network side, this solution deploys a lightweight proxy decoder on the terminal side to generate a health status (HI) and introduces physical feature anchors. The network side combines manifold distance detection (to determine the validity of the data structure) with anchor comparison (to determine the correctness of physical features), achieving precise isolation of fault sources (terminal side, network side, link, or data environment).

[0201] 6. For scenarios where there is a conflict between terminal self-testing and network-side decision-making (e.g., the terminal reports a failure but the network is fine), Type I low-precision feedback is introduced as the "gold standard" for arbitration. Based on the arbitration result, the terminal-side proxy decoder is managed in a hierarchical manner: Level 1 performs dynamic threshold bias fine-tuning, Level 2 performs lightweight calibration through a linear transformation matrix, and Level 3 disables functions and rolls back the model, thereby maintaining the reliability of the diagnostic mechanism itself.

[0202] The following is the process for a terminal device to build a proxy decoder, including the following steps:

[0203] 1. Obtain the training dataset.

[0204] For example, the RRC signaling transmitted in S101 also carries information about training data, which is used by the terminal to train the proxy decoder. The proxy decoder can be understood as a copy of the decoder deployed on the network side.

[0205] For example, the training data information includes, but is not limited to: the training dataset of the proxy decoder, the identifier (referred to as ID) of the training dataset of the proxy decoder, and the SNR corresponding to the ID of the training dataset of the proxy decoder. current Information such as...

[0206] 2. Construct a mixed training dataset.

[0207] For example, the mixed training dataset includes a set of positive samples and a set of abnormal samples.

[0208] For example, the normal sample set is a sample set that indicates the information of the training data. For instance, the normal sample set includes a training dataset or subset (referred to as target CSI) collected by the terminal device and reported to the network device for training the codec model pair, and a dataset or subset (referred to as CSI feedback) generated by the network device during the training of the codec model pair.

[0209] For example, the abnormal sample set is a set of samples generated by the terminal device by applying specific perturbations or setting parameters to zero to the normal sample set, used to simulate fault scenarios. For instance, the compressed feature vector Z... AIAdd Gaussian white noise or randomly set it to zero.

[0210] 3. Based on the mixed training dataset and loss function, train the AI ​​model to obtain the proxy decoder.

[0211] For example, the loss function is the composite loss function L. total =α×L MSE +β×L Classification .

[0212] L MSE To mitigate reconstruction errors, the goal is to ensure reconstruction accuracy, meaning the proxy decoder (CSI feedback) should be approximately equal to the decoder (CSI feedback) deployed on the network device.

[0213] L Classification For classification loss (such as cross-entropy), the SGCS value of the proxy decoder output must be higher than the high threshold T used in the self-testing process of the terminal device when the input is a "normal sample". high When faced with an "abnormal sample" input, the threshold T is lower than the low threshold T used in the self-test process of the terminal device. low That is, when the input is "normal sample", L Classification When the input is 1 and the input is "abnormal sample", L Classification The value is 0. α and β are the reconstruction weight and classification weight, respectively, and can be pre-configured.

[0214] For example, performance metrics are evaluated before the proxy decoder training is complete: that is, the following criteria are met on the test set (constructed in the same way as the hybrid training dataset described above):

[0215] False Alarm Rate (FAR) < When the input is a normal sample, the SGCS value of the proxy decoder output and the input is lower than the high threshold T in the decision threshold. high The probability is extremely low.

[0216] False negative rate (MDR) < When the input is an anomalous sample, the SGCS value of the proxy decoder output and the input is higher than the lower threshold T in the decision threshold. low The probability is extremely low.

[0217] and The network device can issue commands based on training requirements, or it can obtain the terminal device's capability information from the terminal and determine the network device's capabilities based on that information. For example, the capability information includes the computing power of the decoder deployed on the terminal device. Based on this computing power, the network device can determine the network device's capabilities. and .

[0218] After the proxy decoder is trained, the terminal device collects real (raw) channel data and, together with the proxy decoder, uses the aforementioned false alarm rate (False Alarm Rate) and false negative rate (False Negative Rate) as training constraints. It then uses the aforementioned steps (constructing a hybrid training dataset, constructing a composite loss function, and performance evaluation) to adversarially train the encoder. Only after meeting the false alarm rate and false negative rate tests can the device be deployed for performance evaluation. Subsequent steps involving the encoder, proxy decoder, and decoder deployed on network devices all use models that have been trained and passed the false alarm rate and false negative rate tests.

[0219] The training method provided in this embodiment constructs a mixed dataset containing normal and abnormal samples during the training phase of the proxy decoder. Through a composite loss function (reconstruction error + classification loss), it ensures that the proxy decoder can not only recover the data but also accurately and sensitively identify abnormal states in the input data (such as distinguishing between normal samples and attack / faulty samples). Furthermore, it introduces false alarm rate and false negative rate constraints to train the proxy decoder and dynamically adjusts the judgment threshold in conjunction with SNR, reducing false alarms and false negatives, thus exhibiting strong robustness.

[0220] The above embodiments can be applied to the following two scenarios:

[0221] 1. Semantic communication and image / video compression transmission scenarios:

[0222] In 6G semantic communication or low-bandwidth video surveillance scenarios, the sending end typically extracts the "semantic features" of images / videos for compression and transmission, while the receiving end uses generative AI to reconstruct the image.

[0223] Examples of application logic for the above embodiments in this scenario include:

[0224] The self-check at the sending end includes: the sending end (such as a camera) roughly reconstructs the image using a lightweight proxy model, compares it with the original image, and determines whether the semantic feature extraction network (encoder) is working properly.

[0225] Extracting physical anchor points at the sending end includes: extracting traditional physical features of the image as anchor points (such as histogram statistical features, key edge detection operator results, or extremely low-resolution thumbnails), and sending them along with semantic features.

[0226] The receiver's identification of the root cause of the fault includes comparing the reconstructed image with the physical anchor point. If the AI ​​model (decoder) reconstructs "a cat", but the anchor point shows that the area should be "blue background" (color histogram mismatch), it can be determined as data drift (input image outside the model training domain) or model illusion, rather than simple link noise.

[0227] 2. Predictive maintenance scenarios for industrial IoT devices:

[0228] Industrial sensors (such as vibration and temperature sensors) upload raw waveform data to the cloud for fault diagnosis using AI-compressed data. Examples of application logic for the above embodiments in this scenario include:

[0229] The sender's self-check includes: the sensor uses a local lightweight model to check the manifold structure of the compressed data to prevent the cloud from misjudging the device failure due to the sending of junk data caused by sensor aging or encoding module failure.

[0230] The extraction of physical anchor points at the transmitting end includes: the sensor extracting time-domain / frequency-domain statistics of the original signal (such as RMS, peak factor, and main frequency component) as anchor points.

[0231] The receiver's identification of the root cause of the fault includes: after cloud decoding, if the waveform reconstruction is normal but the RMS value is seriously inconsistent with the anchor point, it can quickly identify whether it is a compression model failure or a real abnormal vibration of the equipment, thus avoiding false shutdown and maintenance alarms.

[0232] As can be seen, the two scenarios above demonstrate that the sending end can also perform self-testing based on the manifold structure, as long as the receiving end pre-configures the parameters for self-testing. The receiving end can also perform other types of verification adapted to the application scenario.

[0233] It should be understood that Figures 1 to 4 The flowcharts or scene diagrams shown are for illustrative purposes only and are not intended to limit the embodiments of this application to the examples illustrated. In fact, those skilled in the art can interpret the embodiments based on... Figures 1 to 4 The examples in the document can be transformed into equivalent ways to obtain more implementations.

[0234] The above text combined Figures 1 to 4 This document describes in detail the communication method provided in the embodiments of this application. The following will combine... Figures 5 to 6 The device embodiments of this application are described in detail below. It should be understood that the communication device of this application embodiment can execute the various communication methods of the foregoing embodiments of this application, that is, the specific working processes of the various products below can be referred to the corresponding processes in the foregoing method embodiments.

[0235] In the embodiments described above, the terminal device may execute some or all of the steps in each embodiment; the network device may execute some or all of the steps in each embodiment. These steps or operations are merely examples, and the embodiments of this application may also perform other operations or variations thereof. Furthermore, the steps may be executed in different orders as presented in the embodiments, and it is not necessary to execute all the operations in the embodiments of this application. Moreover, the sequence number of each step does not imply the order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0236] Figure 5This is a schematic block diagram of a communication device provided in an embodiment of this application. Figure 5 As shown, the communication device 100 may include a communication module 120. The communication module 120 can implement corresponding communication functions, which can be internal communication functions of the communication device 100 or communication functions between the communication device 100 and other devices. Optionally, the communication module 120 may also be referred to as a communication interface or transceiver module. Optionally, the communication device 100 also includes a processing module 110. The processing module 110 can implement corresponding processing functions.

[0237] Optionally, the communication device 100 further includes a storage module, which can be used to store instructions and / or data; the processing module 110 can read the instructions and / or data in the storage module so that the communication device 100 can implement the aforementioned method embodiments.

[0238] In one possible design, the communication device 100 may correspond to the terminal device in the above method embodiments, or a component (such as a circuit, chip, or chip system) configured in the terminal device. The communication device 100 can be used to perform the steps or processes performed by the terminal device in any of the above method embodiments.

[0239] For example, the communication module 120 is used to send a Channel State Information (CSI) report to a network device. The CSI report includes physical anchors and self-test results. The self-test results are the detection results of the terminal device on a first model and a second model. The first model is deployed on the terminal device, and the second model is a proxy model of a third model. The third model is a model deployed on the network device. The CSI report is used by the network device to locate the root cause of a fault in the third model. The fault includes data obtained based on the third model that does not meet requirements.

[0240] In some implementations, the communication module 120 is also used to receive first information from the network device, the first information indicating the initiation of a mechanism to locate the root cause of a fault.

[0241] In some implementations, the communication module 120 receives the first information from the network device by receiving a Media Access Control (MAC) element (CE) from the network device, wherein the MAC CE carries the first information.

[0242] In some implementations, the communication module 120 is further configured to: receive second information from the network device before sending a Channel State Information (CSI) report to the network device, the second information including at least one of: self-test parameter configuration information, physical anchor point configuration information, and self-test result reporting mode configuration information; the self-test parameter configuration information is used by the terminal device to obtain the self-test result; the physical anchor point configuration information is used by the terminal device to obtain the physical anchor point; and the self-test result reporting mode configuration information is used by the terminal device to send the CSI report.

[0243] In some implementations, the self-test parameter configuration information indicates parameter type, baseline threshold, and threshold dynamic adjustment parameter. The parameter type indicates the type of parameter used as the basis for self-testing. The baseline threshold and the threshold dynamic adjustment parameter are used to obtain the threshold used by the terminal device for self-testing.

[0244] In some implementations, the configuration information of the physical anchor point indicates the physical characteristics of the downlink channel that can serve as the physical anchor point, as well as the amount of data occupied by the physical anchor point in the CSI report.

[0245] In some implementations, the self-test result reporting mode configuration information indicates the reporting mode and the amount of data occupied by the self-test result in the CSI report. The reporting mode includes reporting the self-test result only when the self-test result indicates that the self-test result is unhealthy.

[0246] In some implementations, the communication module 120 receives the second information from the network device by receiving Radio Resource Control (RRC) signaling from the network device, wherein the second information is carried in a New Information Element (IE) in the RRC signaling.

[0247] In some implementations, the communication module 120 is further configured to receive third information from the network device after sending a Channel State Information (CSI) report to the network device, the third information being used to adjust the first model and / or the second model.

[0248] In some implementations, the third information includes at least one of the following: Hybrid Automatic Repeat Request (HARQ) information, indication information for checking, restarting, or retraining the model, backoff indication information, indication information for adjusting the modulation and coding scheme (MCS) or frequency point, indication information for updating self-test rules, and indication information for increasing the amount of data occupied by the physical anchor in the CSI report. The model includes at least one of the first model and the second model.

[0249] In some implementations, the communication module 120 receives third information from the network device by receiving a MAC CE from the network device, wherein the MAC CE carries the third information.

[0250] In some implementations, the processing module 110 is used to determine the self-test result based on the relationship between the value of a parameter and the threshold used for self-test before the communication module 120 sends the Channel State Information (CSI) report to the network device. The value of the parameter is calculated based on a first feature and a second feature. The first feature is a feature of the original channel, and the second feature is a feature output by the second model to the CSI compressed data. The CSI compressed data is the data output by the first model to the feature of the original channel.

[0251] The above are merely examples; for detailed steps or procedures, please refer to the descriptions in the foregoing embodiments.

[0252] In one possible design, the communication device 100 may correspond to the network device in the above method embodiments, or to a component (such as a circuit, chip, or chip system) configured in the network device. The communication device 100 can be used to perform the steps or processes performed by the network device in any of the above method embodiments.

[0253] For example, the communication module 120 is used to receive a Channel State Information (CSI) report from a terminal device. The CSI report includes physical anchors and self-test results. The self-test results are the detection results of the terminal device on a first model and a second model. The first model is deployed on the terminal device, and the second model is a proxy model of a third model. The third model is a model deployed on the network device. The CSI report is used by the network device to locate the root cause of a fault, including a fault where the data obtained based on the third model does not meet the requirements.

[0254] In some implementations, the communication module 120 is further configured to: send first information to the terminal device before receiving a Channel State Information (CSI) report from the terminal device, the first information indicating a mechanism for initiating a fault root cause in the localization model.

[0255] In some implementations, the communication module 120 sends the first information to the terminal device by sending a Media Access Control (MAC) control element (CE) to the terminal device, wherein the MAC CE carries the first information.

[0256] In some implementations, the communication module 120 is further configured to: send second information to the terminal device before receiving the Channel State Information (CSI) report from the terminal device, the second information including at least one of: self-test parameter configuration information, physical anchor point configuration information, and self-test result reporting mode configuration information; the self-test parameter configuration information is used by the terminal device to obtain the self-test result; the physical anchor point configuration information is used by the terminal device to obtain the physical anchor point; and the self-test result reporting mode configuration information is used by the terminal device to send the CSI report.

[0257] In some implementations, the self-test parameter configuration information indicates parameter type, baseline threshold, and threshold dynamic adjustment parameter. The parameter type indicates the type of parameter used as the basis for self-testing. The baseline threshold and the threshold dynamic adjustment parameter are used to obtain the threshold used by the terminal device for self-testing.

[0258] In some implementations, the configuration information of the physical anchor point indicates the physical characteristics of the downlink channel that can serve as the physical anchor point, as well as the amount of data occupied by the physical anchor point in the CSI report.

[0259] In some implementations, the self-test result reporting mode configuration information indicates the reporting mode and the amount of data occupied by the self-test result in the CSI report. The reporting mode includes reporting the self-test result only when the self-test result indicates that the self-test result is unhealthy.

[0260] In some implementations, the communication module 120 sends the second information to the terminal device by sending Radio Resource Control (RRC) signaling to the terminal device, wherein the second information is carried in the newly added information element (IE) in the RRC signaling.

[0261] In some implementations, the communication module 120 is further configured to: after receiving the Channel State Information (CSI) report from the terminal device, send third information to the terminal device, the third information being used to adjust the first model and / or the second model.

[0262] In some implementations, the third information includes at least one of the following: Hybrid Automatic Repeat Request (HARQ) information, indication information for checking, restarting, or retraining the model, backoff indication information, indication information for adjusting the modulation and coding scheme (MCS) or frequency point, indication information for updating self-test rules, and indication information for increasing the amount of data occupied by the physical anchor in the CSI report. The model includes at least one of the first model and the second model.

[0263] In some implementations, the processing module 110 is used to determine the root cause of the decrease in the output accuracy of the third model based on the CSI compressed data in the CSI report, the physical anchor point, and the self-test result after the communication module 120 receives the Channel State Information (CSI) report from the terminal device.

[0264] In some implementations, the processing module 110 determines the root cause of the decrease in the output accuracy of the third model based on the CSI compressed data in the CSI report, the physical anchor point, and the self-test result. This includes: sequentially performing a first verification, a second verification, and a third verification based on the block error rate; for any one of the first and second verifications, if a first condition is met, performing the next verification; if the verification fails, determining the root cause based on at least one of the verification result and the self-test result; and if the block error rate is greater than a block error rate threshold, determining the root cause based on the self-test result. The first verification includes at least one of CRC and valid manifold check; the second verification includes verification based on the physical anchor point; the third verification includes verification based on the service's block error rate; and the first condition includes verification passing.

[0265] In some implementations, the first condition further includes: not meeting the arbitration condition, wherein the arbitration condition includes: the self-test result indicates that the first model and / or the second model is faulty.

[0266] In some implementations, the processing module 110 is further configured to: determine the root cause based on the similarity between the CSI compressed data and the CSI data from the terminal device, provided that the valid manifold verification passes and the arbitration conditions are met.

[0267] In some implementations, when the error rate is greater than the error rate threshold, the processing module 110 determines the root cause based on the self-test result by: when the error rate is greater than the error rate threshold and the self-test result indicates health, determining the root cause based on the similarity between the CSI compressed data and the CSI data from the terminal device.

[0268] The above are merely examples; for detailed steps or procedures, please refer to the descriptions in the foregoing embodiments.

[0269] Figure 6 This is another schematic block diagram of the communication device 200 provided in the embodiments of this application. The communication device 200 may be a chip, chip system, or processor, etc., in a terminal device or network device that implements the above-described methods. The communication device 200 can be used to implement the methods described in the above-described method embodiments; for details, please refer to the descriptions in the above-described method embodiments.

[0270] like Figure 6 As shown, the communication device 200 may include one or more processors 210, which may also be referred to as processing units or processing modules, and can implement certain control functions. The processor 210 may be a general-purpose processor or a dedicated processor, such as a baseband processor or a central processing unit. The baseband processor can be used to process communication protocols and communication data, while the central processing unit can be used to control the communication device 200 (e.g., a base station, baseband chip, user, user chip), execute software programs, and process data from the software programs.

[0271] In an alternative design, the processor 210 may also store instructions and / or data that can be executed by the processor 210 to cause the communication device 200 to perform the methods described in the above method embodiments.

[0272] In another alternative design, the communication device 200 may include a communication interface 220 for implementing receiving and transmitting functions. For example, the communication interface 220 may be a transceiver circuit, interface, interface circuit, or transceiver. The transceiver circuit, interface, interface circuit, or transceiver for implementing receiving and transmitting functions may be separate or integrated. The aforementioned transceiver circuit, interface, interface circuit, or transceiver may be used for reading and writing code / data, or it may be used for transmitting or relaying signals.

[0273] Optionally, the communication device 200 may include one or more memories 230, which may store instructions that can be executed on the processor 210, causing the communication device 200 to perform the methods described in the above method embodiments. Optionally, the memories 230 may also store data. Optionally, the processor 210 may also store instructions and / or data. The processor 210 and the memories 230 may be provided separately or integrated together.

[0274] It should be understood that, in one possible design, the steps in the method embodiments provided in this application can be implemented by integrated logic circuits in the processor's hardware or by instructions in software form. The steps of the methods disclosed in the embodiments of this application can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules in the processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method. To avoid repetition, detailed descriptions are not provided here.

[0275] In one implementation, the communication device 200 may correspond to the terminal device in the above method embodiments and may be used to execute the various steps and / or processes executed by the terminal device in the above method embodiments. The processor 210 may be used to execute instructions stored in the memory 230, and when the processor 210 executes the instructions stored in the memory, the processor 210 is used to execute the various steps and / or processes of the above method embodiments corresponding to the terminal device.

[0276] In another implementation, the communication device 200 may correspond to the network device in the above method embodiments and may be used to execute the various steps and / or processes executed by the network device in the above method embodiments. The processor 210 may be used to execute instructions stored in the memory 230, and when the processor 210 executes the instructions stored in the memory, the processor 210 is used to execute the various steps and / or processes of the above method embodiments corresponding to the network device.

[0277] It should be understood that the aforementioned processing device can be one or more chips. For example, the processing device can be a field-programmable gate array (FPGA), an application-specific integrated circuit (ASIC), a system-on-chip (SoC), a central processor unit (CPU), a network processor (NP), a digital signal processor (DSP), a microcontroller unit (MCU), a programmable logic device (PLD), or other integrated chips.

[0278] It is understood that the memory in the embodiments of this application can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (SLDRAM), and direct rambus RAM (DR RAM). It should be noted that the memory used in the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0279] According to the method provided in the embodiments of this application, this application also provides a chip system, which includes one or more processors for calling and executing instructions stored in memory, thereby causing the method described in the embodiments of this application to be executed. The chip system may be composed of chips or may include chips and other discrete devices.

[0280] The chip system may include input circuits or interfaces for transmitting information or data, and output circuits or interfaces for receiving information or data.

[0281] According to the method provided in the embodiments of this application, this application also provides a communication system, which includes the aforementioned network device and terminal device.

[0282] According to the method provided in the embodiments of this application, this application also provides a computer program product, which includes: computer program code, which, when run on a computer, causes the computer to execute the various steps or processes executed by the network device or terminal device in any of the foregoing method embodiments.

[0283] According to the method provided in the embodiments of this application, this application also provides a computer-readable storage medium storing program code, which, when run on a computer, causes the computer to execute the various steps or processes executed by the network device or terminal device in any of the foregoing method embodiments.

[0284] The computer-readable storage medium may be the aforementioned volatile memory or non-volatile memory, or it may include both volatile memory and non-volatile memory.

[0285] In the embodiments of this application, the terms and English abbreviations are exemplary examples given for ease of description and should not be construed as limiting the application in any way. This application does not preclude the possibility of defining other terms that can achieve the same or similar functions in existing or future agreements.

[0286] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions. When these computer instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated.

[0287] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0288] It should be understood that in the various embodiments of this application, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0289] In summary, the above description is merely a preferred embodiment of the technical solution of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A communication method, characterized in that, include: The terminal device sends a Channel State Information (CSI) report to the network device. The CSI report includes physical anchors and self-test results. The self-test results are the terminal device's test results for a first model and a second model. The first model is deployed on the terminal device, and the second model is a proxy model for a third model deployed on the network device. The physical anchors are the values ​​of the physical characteristics of the channel extracted by the terminal device. The CSI report is used by the network device to locate the root cause of a fault. The fault includes data obtained based on the third model not meeting requirements. Locating the root cause of the fault includes verification based on the physical anchors.

2. The method according to claim 1, characterized in that, Before sending the Channel State Information (CSI) report to the network device, the following is also included: The terminal device receives first information from the network device, the first information indicating the activation of a mechanism to locate the root cause of the fault.

3. The method according to claim 2, characterized in that, The receipt of the first information from the network device includes: The network device receives a Media Access Control (MAC) control element (CE), in which the MAC CE carries the first information.

4. The method according to claim 1, characterized in that, Before sending the Channel State Information (CSI) report to the network device, the following is also included: The terminal device receives second information from the network device, the second information including at least one of: self-test parameter configuration information, physical anchor point configuration information, and self-test result reporting mode configuration information; The self-test parameter configuration information is used by the terminal device to obtain the self-test result; The configuration information of the physical anchor point is used by the terminal device to obtain the physical anchor point; The self-test result reporting mode configuration information is used by the terminal device to send the CSI report.

5. The method according to claim 4, characterized in that, The self-test parameter configuration information indicates the parameter type, the baseline threshold, and the threshold dynamic adjustment parameter. The parameter type indicates the type of parameter used as the basis for self-testing. The baseline threshold and the threshold dynamic adjustment parameter are used to obtain the threshold used by the terminal device for self-testing.

6. The method according to claim 4, characterized in that, The configuration information of the physical anchor point indicates the physical characteristics of the downlink channel that can serve as the physical anchor point, as well as the amount of data occupied by the physical anchor point in the CSI report.

7. The method according to claim 4, characterized in that, The self-test result reporting mode configuration information indicates the reporting mode and the amount of data occupied by the self-test result in the CSI report. The reporting mode includes reporting the self-test result only when the self-test result indicates that it is unhealthy.

8. The method according to any one of claims 4-7, characterized in that, The receipt of the second information from the network device includes: The network device receives Radio Resource Control (RRC) signaling, and the second information is carried in the New Information Element (IE) in the RRC signaling.

9. The method according to any one of claims 1-7, characterized in that, After sending the Channel State Information (CSI) report to the network device, the following is also included: The terminal device receives third information from the network device, and the third information is used to adjust the first model and / or the second model.

10. The method according to claim 9, characterized in that, The third information includes at least one of the following: The model includes at least one of the first model and the second model. The model includes information such as HARQ (Automatic Repeat Request) information, indication information for checking, restarting or retraining the model, backoff indication information, indication information for adjusting the modulation and coding scheme (MCS) or frequency point, indication information for updating self-test rules, and indication information for increasing the amount of data occupied by the physical anchor in the CSI report.

11. The method according to claim 9, characterized in that, The receipt of third information from the network device includes: Receive a MAC CE from the network device, the MAC CE carrying the third information.

12. The method according to any one of claims 1-7, characterized in that, Before sending the Channel State Information (CSI) report to the network device, the following is also included: The terminal device determines the self-test result based on the relationship between the parameter value and the threshold used for self-testing. The parameter value is calculated based on a first feature and a second feature. The first feature is a feature of the original channel, and the second feature is a feature output by the second model for CSI compressed data. The CSI compressed data is data output by the first model for the feature of the original channel.

13. A communication method, characterized in that, include: The network device receives a Channel State Information (CSI) report from a terminal device. The CSI report includes physical anchors and self-test results. The self-test results are the terminal device's detection results for a first model and a second model. The first model is deployed on the terminal device, and the second model is a proxy model for a third model deployed on the network device. The physical anchors are the values ​​of the physical characteristics of the channel extracted by the terminal device. The CSI report is used by the network device to locate the root cause of a fault. The fault includes data obtained based on the third model not meeting requirements. Locating the root cause of the fault includes verification based on the physical anchors.

14. The method according to claim 13, characterized in that, Before receiving the Channel State Information (CSI) report from the terminal device, the method further includes: The network device sends a first message to the terminal device, the first message indicating the activation of a mechanism to locate the root cause of the fault.

15. The method according to claim 14, characterized in that, Sending the first information to the terminal device includes: The terminal device is sent a Media Access Control (MAC) control element (CE), in which the MAC CE carries the first information.

16. The method according to claim 13, characterized in that, Before receiving the Channel State Information (CSI) report from the terminal device, the method further includes: The network device sends a second message to the terminal device, the second message including at least one of the following: self-test parameter configuration information, physical anchor point configuration information, and self-test result reporting mode configuration information; The self-test parameter configuration information is used by the terminal device to obtain the self-test result; The configuration information of the physical anchor point is used by the terminal device to obtain the physical anchor point; The self-test result reporting mode configuration information is used by the terminal device to send the CSI report.

17. The method according to claim 16, characterized in that, The self-test parameter configuration information indicates the parameter type, the baseline threshold, and the threshold dynamic adjustment parameter. The parameter type indicates the type of parameter used as the basis for self-testing. The baseline threshold and the threshold dynamic adjustment parameter are used to obtain the threshold used by the terminal device for self-testing.

18. The method according to claim 16, characterized in that, The configuration information of the physical anchor point indicates the physical characteristics of the downlink channel that can serve as the physical anchor point, as well as the amount of data occupied by the physical anchor point in the CSI report.

19. The method according to claim 16, characterized in that, The self-test result reporting mode configuration information indicates the reporting mode and the amount of data occupied by the self-test result in the CSI report. The reporting mode includes reporting the self-test result only when the self-test result indicates that it is unhealthy.

20. The method according to any one of claims 16-19, characterized in that, Sending second information to the terminal device includes: The terminal device is sent a Radio Resource Control (RRC) signaling message, and the second information is carried in the newly added information element (IE) in the RRC signaling message.

21. The method according to any one of claims 13-18, characterized in that, After receiving the Channel State Information (CSI) report from the terminal device, the method further includes: The network device sends third information to the terminal device, the third information being used to adjust the first model and / or the second model.

22. The method according to claim 21, characterized in that, The third information includes at least one of the following: The model includes at least one of the first model and the second model. The model includes information such as HARQ (Automatic Repeat Request) information, indication information for checking, restarting or retraining the model, backoff indication information, indication information for adjusting the modulation and coding scheme (MCS) or frequency point, indication information for updating self-test rules, and indication information for increasing the amount of data occupied by the physical anchor in the CSI report.

23. The method according to any one of claims 13-19, characterized in that, After receiving the Channel State Information (CSI) report from the terminal device, the method further includes: The network device determines the root cause of the decrease in the output accuracy of the third model based on the CSI compressed data in the CSI report, the physical anchor point, and the self-test results.

24. The method according to claim 23, characterized in that, The determination of the root cause of the decrease in the output accuracy of the third model, based on the CSI compressed data in the CSI report, the physical anchor points, and the self-test results, includes: The system sequentially performs a first-level check, a second-level check, and a third-level check based on the block error rate. For any one of the first-level or second-level checks, if a first condition is met, the next level of check is performed. If the check fails, the root cause is determined based on at least one of the check result and the self-test result. If the block error rate is greater than the block error rate threshold, the root cause is determined based on the self-test result. The first-level check includes at least one of CRC and valid manifold check. The second-level check includes a check based on the physical anchor point. The first condition includes passing the check.

25. The method according to claim 24, characterized in that, The first condition also includes: The arbitration conditions are not met, including the self-test result indicating that the first model and / or the second model is faulty.

26. The method according to claim 25, characterized in that, Also includes: If the valid manifold verification passes and the arbitration conditions are met, the root cause is determined based on the similarity between the CSI compressed data and the CSI data from the terminal device.

27. The method according to claim 24, characterized in that, When the error rate is greater than the error rate threshold, determining the root cause based on the self-test results includes: If the error rate is greater than the error rate threshold and the self-test result indicates that the system is healthy, the root cause is determined based on the similarity between the CSI compressed data and the CSI data from the terminal device.

28. A communication device, characterized in that, The device includes at least one processor coupled to a memory storing a program or instructions, the processor executing the program or instructions to cause the device to perform the method as described in any one of claims 1 to 27.

29. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, When the computer program or instructions are executed, they cause the computer to perform the method as described in any one of claims 1 to 27.

30. A chip system, characterized in that, The chip system includes one or more processors, which are configured to retrieve and execute instructions stored in memory, such that the method as described in any one of claims 1 to 27 is performed.

Images