Communication method, terminal equipment and network equipment

Abstract

The invention provides a communication method, terminal equipment and network equipment. The method comprises: a terminal device sends first information and / or second information to a network device, the first information being used for indicating model-related capability information of the terminal device, and the second information being used for indicating applicability of a model configured by the network device. Therefore, the network equipment can configure the model used by the terminal equipment based on the first information and / or the second information to adapt to the terminal equipment, so that the model performance is improved.

Description

Technical Field

[0001] This application relates to the field of communication technology, and more specifically, to a communication method, terminal equipment, and network equipment. Background Technology

[0002] Introducing models into communication systems can effectively improve communication performance. For example, channel state information (CSI) feedback based on artificial intelligence (AI) / machine language (ML) models has been proposed. However, how network devices can configure these models to adapt to terminal devices remains a problem to be solved. Summary of the Invention

[0003] This application provides a communication method, a terminal device, and a network device. The various aspects covered by this application are described below.

[0004] In a first aspect, a communication method is provided, comprising: a terminal device sending first information and / or second information to a network device, wherein the first information is used to indicate model-related capability information of the terminal device, and the second information is used to indicate the applicability of a model configured by the network device.

[0005] In a second aspect, a communication method is provided, comprising: a network device receiving first information and / or second information sent by a terminal device, wherein the first information is used to indicate model-related capability information of the terminal device, and the second information is used to indicate the applicability of a model configured by the network device.

[0006] Thirdly, a terminal device is provided, comprising: a transceiver unit, configured to send first information and / or second information to a network device, wherein the first information is used to indicate model-related capability information of the terminal device, and the second information is used to indicate the applicability of a model configured by the network device.

[0007] Fourthly, a network device is provided, comprising: a transceiver unit, configured to receive first information and / or second information sent by a terminal device, wherein the first information is used to indicate model-related capability information of the terminal device, and the second information is used to indicate the applicability of a model configured by the network device.

[0008] Fifthly, a terminal device is provided, including a transceiver, a memory, and a processor, wherein the memory is used to store a program, and the processor is used to invoke the program in the memory and control the transceiver to receive or send signals so that the terminal device performs the method as described in the first aspect.

[0009] In a sixth aspect, a network device is provided, including a transceiver, a memory, and a processor, wherein the memory is used to store a program, and the processor is used to invoke the program in the memory and control the transceiver to receive or transmit signals so that the network device performs the method as described in the second aspect.

[0010] A seventh aspect provides an apparatus including a processor for calling a program from a memory to cause the apparatus to perform the method as described in the first or second aspect.

[0011] Eighthly, a chip is provided, including a processor for calling a program from memory to cause a device having the chip mounted to perform the method as described in the first or second aspect.

[0012] Ninth aspect, a computer-readable storage medium is provided having a program stored thereon that causes a computer to perform the method as described in the first or second aspect.

[0013] A tenth aspect provides a computer program product, including a program that causes a computer to perform the method as described in the first or second aspect.

[0014] Eleventhly, a computer program is provided that causes a computer to perform the method as described in the first or second aspect.

[0015] In this embodiment of the application, the terminal device can send first information and / or second information to the network device. The first information is used to indicate the terminal device's model-related capability information, and the second information is used to indicate the applicability of the model. In this way, the network device can configure the model used by the terminal device based on the first information and / or the second information to adapt to the terminal device, thereby improving the model performance. Attached Figure Description

[0016] Figure 1 This is a system architecture example diagram of a communication system applicable to embodiments of this application.

[0017] Figure 2 This is a flowchart illustrating the communication method according to an embodiment of this application.

[0018] Figure 3 for Figure 2 A schematic diagram of one possible implementation of the method shown.

[0019] Figure 4 for Figure 2 A schematic diagram of one possible implementation of the method shown.

[0020] Figure 5 This is a schematic diagram of the structure of a terminal device according to an embodiment of this application.

[0021] Figure 6 This is a schematic diagram of the structure of a network device according to an embodiment of this application.

[0022] Figure 7 This is a schematic diagram of a communication apparatus according to an embodiment of this application. Detailed Implementation

[0023] The technical solutions in this application will now be described with reference to the accompanying drawings.

[0024] Communication system Figure 1 This is an example diagram of the system architecture of a communication system 100 applicable to embodiments of this application. The communication system 100 may include a network device 110 and a terminal device 120. The network device 110 may be a device that communicates with the terminal device 120. The network device 110 can provide network coverage for a specific geographical area and can communicate with the terminal device 120 located within that coverage area. The terminal device 120 can access a network, such as a wireless network, through the network device 110. Optionally, the communication system 100 may also include other network entities such as a network controller and a mobility management entity; this application embodiment does not limit this.

[0025] It should be understood that the technical solutions of the embodiments of this application can be applied to various communication systems, such as: fifth-generation (5G) systems, new radio (NR), long-term evolution (LTE) systems, LTE frequency division duplex (FDD) systems, LTE time division duplex (TDD) systems, etc. The technical solutions provided in this application can also be applied to future communication systems, such as sixth-generation mobile communication systems, satellite communication systems, and so on.

[0026] In this application embodiment, the terminal device may also be referred to as user equipment (UE), access terminal, user unit, user station, mobile station, mobile station (MS), mobile terminal (MT), remote station, remote terminal, mobile device, user terminal, terminal, wireless communication device, user agent, or user apparatus. The terminal device in this application embodiment can be a device that provides voice and / or data connectivity to a user, and can be used to connect people, objects, and machines, such as a handheld device with wireless connectivity, vehicle-mounted device, etc. Terminal devices can also be mobile phones, tablets, laptops, PDAs, mobile internet devices (MIDs), wearable devices, virtual reality (VR) devices, augmented reality (AR) devices, wireless terminals in industrial control, self-driving, remote medical surgery, smart grids, transportation safety, smart cities, and smart homes. Optionally, terminal devices can act as base stations. For example, a terminal device can act as a dispatching entity, providing sidelink signals between terminal devices in vehicle-to-everything (V2X) or device-to-device (D2D) systems. For instance, cellular phones and cars communicate with each other using sidelink signals. Cellular phones and smart home devices communicate without relaying communication signals through base stations.

[0027] In this embodiment, the network device can be a device used to communicate with a terminal device. The network device can be an access network device or a wireless access network device. For example, the network device can be a base station. The term "base station" can broadly encompass various names as follows, or can be replaced by names such as: NodeB, evolved NodeB (eNB), next-generation NodeB (gNB), relay station, transmitting and receiving point (TRP), transmitting point (TP), master station (MeNB), secondary station (SeNB), multi-standard wireless (MSR) node, home base station, network controller, access node, wireless node, access point (AP), transmission node, transceiver node, baseband unit (BBU), remote radio unit (RRU), active antenna unit (AAU), remote radio head (RRH), central unit (CU), distributed unit (DU), positioning node, etc. A base station can be a macro base station, micro base station, relay node, donor node, or similar entity, or a combination thereof. A base station can also refer to a communication module, modem, or chip installed within the aforementioned equipment or apparatus. A base station can also be a mobile switching center, or an entity that performs base station functions in device-to-device (D2D), vehicle-to-everything (V2X), and machine-to-machine (M2M) communications, a network-side device in a 6G network, or an entity that performs base station functions in future communication systems. A base station can support networks using the same or different access technologies. The embodiments of this application do not limit the specific technologies or device forms used in the network equipment. In some deployments, the network equipment may include a CU or a DU; or, the network equipment may include both a CU and a DU. Optionally, the base station may include an AAU.

[0028] Furthermore, base stations can be fixed or mobile. For example, a helicopter or drone can be configured to act as a mobile base station, and one or more cells can move depending on the location of the mobile base station. In other examples, a helicopter or drone can be configured as a device to communicate with another base station.

[0029] Network devices and terminal devices can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; they can also be deployed on water; and they can also be deployed in the air on airplanes, balloons, and satellites. This application does not limit the scenario in which the network devices and terminal devices are located.

[0030] It should be understood that all or part of the functions of the communication device in this application can also be implemented by software functions running on hardware, or by virtualization functions instantiated on a platform such as a cloud platform.

[0031] Channel state information (CSI) is a key parameter describing channel characteristics in wireless communication, directly impacting base station optimization strategies such as beamforming and resource scheduling. Traditionally, CSI relies on terminal devices reporting it to network devices. However, with the development and maturation of AI technology, CSI prediction can be performed using AI models.

[0032] By predicting CSI using models, changes in channel state can be quickly inferred. This is especially suitable for scenarios such as high-speed mobility. Network devices and terminal devices can obtain CSI information earlier, capture channel changes in a timely manner, and rationally allocate resources. Combined with the predicted CSI, accurate beamforming and resource allocation can be performed in advance to improve spectrum efficiency.

[0033] When performing CSI prediction based on models, the models can learn from historical data and environmental features, maintaining accuracy even in occlusion or non-line-of-sight (NLOS) scenarios. Through training with massive amounts of data, the models can effectively adapt to the challenges of complex channel characteristics in millimeter-wave high-frequency bands and massive MIMO scenarios, thus providing key technical support for 5G evolution (5G-Advanced, 5G-A) and 6G systems. In the millimeter-wave band, high-frequency signals are susceptible to occlusion and atmospheric attenuation, requiring subwavelength-level spatial sampling to capture rapidly changing channel characteristics. Traditional modeling methods struggle to accurately characterize these time-varying multipath effects. For massive MIMO systems, as the antenna array size expands to hundreds or even thousands of elements, the channel matrix exhibits extremely high dimensionality, making manually designed compression feedback mechanisms extremely costly. Because the model is trained on massive amounts of real-world data using deep learning, it can automatically learn the sparsity and spatial correlation unique to millimeter-wave channels, as well as the structured characteristics of large-scale MIMO channels, enabling high-precision reconstruction of the complete channel state from finite pilots. This data-driven modeling approach not only adapts to complex propagation environments (e.g., NLOS link reconstruction, dynamic occlusion prediction), but also continuously optimizes through online learning, providing core algorithmic support for the dynamic control of reconfigurable intelligent surfaces (RIS) in 5G-A and terahertz communication in 6G, significantly improving the reliability and spectral efficiency of high-frequency systems.

[0034] In 5G-A and 6G systems, network devices achieve deep collaboration with terminal devices through intelligent task allocation mechanisms. Based on real-time assessments of terminal device capabilities, such as computing power, storage space, and power consumption, network devices can dynamically determine the execution location of AI computing tasks. For example, high-end terminal devices with strong local processing capabilities can deploy complete CSI prediction models for real-time processing; while resource-constrained IoT devices can adopt a "sensing + transmission" mode, requiring only the reporting of raw data. Network devices ensure that the distributed model parameters are fully compatible with the terminal device's AI framework through version verification mechanisms, including model format, quantization accuracy, and operator support, thereby avoiding execution failures due to hardware / software incompatibility.

[0035] Network devices achieve intelligent management of the collaborative operation of multiple terminal devices by constructing a comprehensive network situational awareness system. In millimeter-wave communication scenarios, network devices need to coordinate the outputs of beam prediction models from multiple terminal devices, using a two-dimensional "space-time" scheduling algorithm to avoid interference caused by beam direction overlap. Simultaneously, network devices establish a unified quality of service (QoS) policy framework, coordinating resource allocation for different service requirements through standardized parameters in AIConfig signaling (e.g., maximum latency, minimum throughput). For example, when services with different latency tolerances, such as ultra-reliable and low-latency communications (URLLC) and massive machine-type communication (mMTC), coexist, network devices can dynamically adjust the physical resource block (PRB) allocation ratio and reserve computing resources for high-priority services, maximizing overall system performance.

[0036] The "network-guided-terminal-execution" system achieves dynamic optimization through continuous performance feedback. Terminal devices periodically report model execution metrics to network devices, such as inference latency, accuracy, and power consumption. Based on this data, the network devices continuously adjust task allocation strategies and model parameters. Therefore, it is necessary to gradually improve relevant standards and signaling formats to standardize interoperability across vendors.

[0037] In order to achieve deep collaboration between network devices and terminal devices in terms of tasks, resources, processes, and other factors, network devices need to configure terminal devices according to models. In related technologies, terminal devices passively receive configurations from network devices. If the model configured by the network device is not applicable, the terminal device can only passively accept or abandon the corresponding resources.

[0038] In 5G-A and 6G networks, model execution in terminal devices requires deep collaboration with network devices to form a closed-loop optimization system of "network guidance - terminal execution". Furthermore, the terminal device's model can be guided by the network device to achieve collaborative inference. The network device allocates computational tasks based on the terminal device's information (e.g., determining whether CSI prediction should be executed by the terminal device's local model or uploaded to the network), avoiding the distribution of complex model inference tasks to terminal devices with lower computing power, and ensuring that the model parameters distributed by the network device match the terminal device's model.

[0039] Network devices (e.g., gNB, CU, etc.) have a comprehensive understanding of the distribution of terminal devices, interference maps, and service requirements across the entire cell, which can prevent resource conflicts caused by independent decisions from multiple terminal devices. For example, in millimeter-wave scenarios, network devices can coordinate the CSI prediction and / or beam prediction model outputs of various terminal devices to prevent beam direction overlap. Furthermore, ensuring that all terminal device models adhere to a unified QoS policy is crucial. If different terminal devices employ differentiated QoS policies (e.g., requiring low latency and pursuing high accuracy), it may lead to air interface resource allocation conflicts. For instance, UE_A's CSI prediction model requires a 1ms response time, consuming more uplink scheduling resources, while UE_B operates in an energy-saving model, allowing 10ms latency but continuously consuming computing resources. Network devices can optimize PRB allocation and computing resource scheduling through standardized QoS parameters in AIConfig (e.g., forcing all enhanced mobile broadband (eMBB) service models to have a latency ≤5ms).

[0040] To address this, the present application provides a corresponding solution whereby a terminal device sends first information and / or second information to a network device. The first information indicates the terminal device's model-related capability information, and the second information indicates the model's applicability. This enables the network device to configure the model used by the terminal device based on the first and / or second information, thereby adapting the model to the terminal device and improving model performance.

[0041] The following is combined Figure 2 The embodiments of this application will be described in detail below.

[0042] Figure 2 This is a flowchart illustrating the communication method provided in an embodiment of this application. Figure 2 The method 200 shown can be performed by a terminal device and a network device. The terminal device may be, for example, a terminal device... Figure 1 The terminal device 120 shown may be, for example, a network device. Figure 1 The network device 110 shown is shown. Figure 2 The method 200 shown may include some or all of the following steps.

[0043] In the following description, "model" can be replaced with terms such as "function". This application does not limit the inference task of the model. For example, the model can be a CSI prediction model used to predict CSI, or it can be a model used for beam management. The following description uses the CSI prediction model as an example. "CSI reporting" can also be called "CSI feedback". Reference signals used for CSI prediction include, but are not limited to, CSI-RS or SSB, etc. Hereinafter, "CSI-RS" will be used as an example, but "CSI-RS" can be replaced with other reference signals used for measurement.

[0044] In step 210, the terminal device sends first information and / or second information to the network device.

[0045] In step 220, the network device receives first information and / or second information sent by the terminal device.

[0046] The first information is used to indicate the model-related capability information of the terminal device. Optionally, the network device may send third information to the terminal device, the third information being used to request the terminal device to send the first information; correspondingly, the terminal device receives the third information.

[0047] As an example, the third piece of information is carried in `UECapabilityEnquiry`, and the first piece of information is carried in `UECapabilityInformation`. Specifically, when performing model negotiation or model alignment in a terminal device, a process such as capability negotiation (or capability reporting, capability alignment, and capability discovery) can be performed first. The network device can initiate a query through the `UECapabilityEnquiry` message at the RRC layer. In response, the terminal device sends `UECapabilityInformation` to the network device, which carries standardized capability fields (or AI capability fields) to carry capability information such as computing architecture (e.g., the terminal device's computing power), supported model frameworks, model task types (e.g., CSI prediction, beam management), and memory constraints (e.g., maximum loadable model size). This stage provides the network device with a comprehensive view of the terminal device's model-related capability information, laying the foundation for subsequent model configuration.

[0048] The terminal device's model-related capability information includes some or all of the following: the terminal device's hardware capabilities, whether the terminal device supports time alignment with network devices, whether the model supports specific algorithms, and the models supported by the terminal device.

[0049] As an example, when reporting capabilities, terminal devices can report their hardware capabilities, such as whether they support the ability to directly transfer CSI data to the NPU via direct memory access (DMA), and their computing power (e.g., a 4-TOPS NPU, indicating that the terminal device has a built-in neural processing unit (NPU) chip with a computing power of 4 TOPS). Assuming the terminal device has the ability to directly transfer CSI data to the NPU via DMA or has strong computing power, the network device can configure a model with shorter latency when configuring the model for the terminal device. For example, the latency of the entire process from CSI-RS measurement to CSI prediction result output may be less than a preset value. For instance, this preset value might be 1ms in a URLLC scenario and 5ms in an eMBB scenario.

[0050] For example, when reporting capabilities, terminal devices can report whether they support time alignment with network devices. Time alignment refers to synchronization between the terminal device and the network device. Since channel gain, phase, and multipath information in CSI change dynamically due to environmental variations, especially in high-speed or high-frequency scenarios, CSI prediction requires current or recent channel information. If the terminal device and network device are not synchronized or have poor synchronization, the CSI prediction result may be based on outdated channel information and may be inaccurate. Therefore, terminal devices need to report to the network device whether they support time alignment so that the network device can configure an appropriate model for them. Furthermore, collaborative inference also relies on time synchronization between base stations and terminal devices to coordinate resource allocation.

[0051] For example, when reporting capabilities, terminal devices can report the models they support, including model type, model identifier, model framework, and the inference tasks supported by the model (e.g., CSI prediction, beam management). Model type can include, for example, the type of neural network used by the model, such as recurrent neural network (RNN), convolutional neural network (CNN), or long short-term memory network (LSTM) for CSI prediction.

[0052] For example, when reporting capabilities, terminal devices can indicate whether a model supports a specific algorithm. For instance, some models supported by the terminal device rely on channel information obtained from continuous historical measurements. If the network device and the terminal device cannot synchronize or the synchronization is poor, it will lead to problems such as sequence breaks or phase shifts; for example, CSI prediction results after 100ms will be inaccurate. In this case, the network device can configure a suitable model for the terminal device. For example, if the terminal device cannot achieve strict time alignment with the network device, the CSI prediction results output by the model need to be based on channel information measured within a more recent time period. Similarly, some models supported by the terminal device rely on longer CSI-RS measurements; the network device needs to configure longer CSI-RS resources when configuring the model. Furthermore, some models supported by the terminal device are suitable for scenarios where channel information changes rapidly; the network device needs to configure smaller intervals between CSI-RS resources when configuring the model.

[0053] After capability negotiation, network devices can, for example, configure models for terminal devices based on the capability information of terminal devices, perform model activation, deactivation, inference, monitoring and other operations, and execute processes related to the second information.

[0054] The second information is used to indicate the suitability of the model configured by the network device. For example, the second information may include some or all of the following: model suitability flags, input data validity, time validity, environment fit, and fallback indication. The second information can be used by the network device to determine whether to trust the CSI prediction results output by the model, or to determine how to use the CSI prediction results output by the model. The second information can also be referred to as uplink AI information (UL AI information, UAI), etc.

[0055] The model suitability flag indicates whether the model matches the current channel conditions (or link conditions). For example, whether the model is suitable for the current channel conditions of the terminal device, such as frequency band, speed, direct path, SNR, and RSRP. Optionally, the model suitability flag may include multiple bits, each used to indicate whether it matches various channel conditions; for example, 1 indicates a match, and 0 indicates a mismatch.

[0056] Input data validity is used to indicate whether the input data for the model is valid. The terminal device can check whether the data it can obtain as model input (e.g., CSI-RS measurement results) is suitable for the current model. For example, for some models, if the input data is incomplete or contaminated, such as by interference or occlusion, the model cannot perform CSI prediction correctly (e.g., the CSI prediction results are inaccurate). In this case, the terminal device can report invalid input data to the network device. Another example is whether the resources configured on the network device for measuring reference signals such as CSI-RS are suitable. If not, the terminal device can report invalid input data to the network device.

[0057] Time validity indicates the duration of measurement data used to obtain prediction results, or simply, the valid time of the prediction results. For example, for a given CSI-RS measurement, the prediction data within the next 5ms is valid. When a network device configures a model for a terminal device, the terminal device needs to evaluate how long the model requires measurement data to accurately predict the CSI, and combine this with its own information, such as the amount of valid data it stores, the data acquisition time, and its own computing power, to determine whether it matches the model.

[0058] Environment matching score indicates the similarity between the current environment and the environment during model training. If the current environment has changed significantly from the environment during model training (e.g., channel information, or CSI-RS measurement results (e.g., RSRP)), the model is not suitable for the current environment, and the terminal device can report the environment matching score to the network device. Optionally, the environment matching score can include a value between 0 and 1, where 1 indicates a 100% similarity between the current environment and the environment during model training, and 0 indicates a 0% similarity. When the environment matching score is below a threshold or the CSI-RS measurement result is below a threshold, the network device can ignore the model's prediction results or reconfigure the model for the terminal device.

[0059] The fallback indication is used to indicate whether to fall back to a mechanism that does not use the model. In other words, if the model is not currently applicable, the terminal device can suggest that the network device fall back to the traditional CSI reporting mechanism, using beam scanning or specifying a backup beam.

[0060] Terminal devices can dynamically determine the applicability of a model based on multi-dimensional decision functions (e.g., network conditions, local resources, and QoS requirements corresponding to the model). For example, terminal devices can evaluate in real time some or all of the following factors: network conditions (e.g., wireless link conditions, such as whether the current signal meets the threshold), availability of computing power (or computing resources) of the terminal device (e.g., whether the remaining computing power or computing resources of the NPU are sufficient for the inference process of the model), battery status of the terminal device (e.g., whether it is in power-saving mode (PSM mode) and the remaining battery power), availability of the model (e.g., whether the model has been downloaded and whether the model needs to be updated), and hot loading latency of the model (e.g., the time from pulling the model into memory to completing the model initialization needs to be within 100ms). This will help determine whether the model configured by the network device matches the current state of the terminal device.

[0061] Optionally, the second information is triggered when the applicability of the model changes; or, the second information is triggered based on triggering conditions configured by the network device. These triggering conditions may include, for example, periodic triggering, semi-persistent triggering, or event triggering. Event triggering includes events such as channel state changes, channel state changes exceeding a threshold (e.g., RSRP mutation exceeding a threshold), or model confidence falling below a threshold. Furthermore, the network device can also trigger the terminal device to report the second information by sending a trigger message (e.g., DCI or MAC CE).

[0062] These triggering conditions can be configured by the network device. For example, the network device can be configured to allow the terminal device to report second information, or to report second information when the applicability of the model changes. Optionally, if the network device is configured to allow the terminal device to report second information, the terminal device can respond to it.

[0063] Of course, if the network device is not configured with triggering conditions, such as without relevant threshold information, then the terminal device can determine for itself whether to send the second message to the network device, or determine the threshold information itself. As an example, the terminal device can decide to send the second message when the RSRP mutation exceeds threshold A, so that the network device can adjust the configuration of the terminal device's model based on the second message sent by the terminal device.

[0064] Network devices can provide model configuration information through radio resource control (RRC) messages, such as RRC configuration messages or RRC reconfiguration messages, hereinafter referred to as RRC configuration / reconfiguration messages.

[0065] To achieve deep collaboration between network devices and terminal devices in terms of tasks, resources, and processes, network devices need to configure terminal devices according to models. In related technologies, terminal devices passively accept model configurations. If the model configured by the network device is inapplicable, the terminal device can only passively accept or abandon the resources. However, in this embodiment, the terminal device and network device can negotiate to complete the model configuration. For example, the network device performs an initial RRC configuration, and the terminal device provides feedback based on the initial RRC configuration. The network device can choose to perform a secondary RRC configuration or not, depending on the specific situation. For example, when the network device performs RRC configuration, it can complete the configuration through one RRC configuration or RRC reconfiguration; or, after one RRC configuration or RRC reconfiguration and negotiation with the terminal device, the model configuration can only be completed through a secondary RRC reconfiguration.

[0066] As an example, such as Figure 3 As shown, after one RRC configuration or RRC reconfiguration, a second RRC reconfiguration is performed based on the negotiation between the network device and the terminal device, thereby completing the configuration of the model.

[0067] In step 310, the network device sends a first RRC configuration / reconfiguration message to the terminal device.

[0068] In step 320, the terminal device sends a first RRC configuration / reconfiguration complete message to the network device.

[0069] For example, the terminal device can monitor the applicability of the model in real time, that is, determine whether the model configuration carried in step 320 matches the current state of the terminal device.

[0070] In step 330, the terminal device sends a second message to the network device to indicate the current applicability of the model.

[0071] It should be noted that the second piece of information can also be carried within the first RRC configuration / reconfiguration completion message, that is, Figure 3 Steps 320 and 330 shown can be combined into a single step; or, the second information can be reported independently of the first RRC configuration / reconfiguration completion message.

[0072] In step 340, the network device sends a second RRC reconfiguration message to the terminal device. For example, the network device may send a second RRC reconfiguration message to the terminal device when it determines that the model needs to be reconfigured.

[0073] If both the first RRC configuration / reconfiguration message and the second RRC reconfiguration message are sent, the message sent in step 310 is an RRC configuration message and the message sent in step 340 is an RRC reconfiguration message; or, the messages sent in both steps 310 and 340 are RRC reconfiguration messages.

[0074] First, the first RRC configuration / reconfiguration message is described in detail. Optionally, the network device may send the first RRC configuration / reconfiguration message to the terminal device. Accordingly, the terminal device receives the first RRC configuration / reconfiguration message sent by the network device. The first RRC configuration / reconfiguration message includes model configuration information. This configuration information may, for example, be determined based on capability information in the first information reported by the terminal device. Further, optionally, the terminal device may also send a first RRC configuration / reconfiguration completion message to the network device. Accordingly, the network device receives the first RRC configuration / reconfiguration completion message sent by the terminal device. The model configuration information described in this embodiment may also be referred to as model configuration or inference configuration, etc.

[0075] In some implementations, the configuration information of the model sent by the network device to the terminal device may include one or more of the following: the triggering condition of the second information, the validity period of the configuration information, the model parameters (e.g., inference-related parameters or inference parameter sets), the model version information, the model input information, the model output information, the information of time instances related to model-based channel measurement (e.g., the time position of CSI-RS measurement), the information of time instances related to model-based channel prediction (e.g., the time position of CSI prediction), the QoS requirements corresponding to the model, the activation condition of the model, and the model-based CSI reporting configuration (e.g., CSI-ReportConfig).

[0076] The triggering conditions may include, for example, some or all of the following: channel measurement results meeting a threshold; changes in channel measurement results meeting a threshold (e.g., CSI-RSRP changes exceeding 3 dB); the model's confidence level being lower than a threshold; abnormal input data of the model; or receiving a trigger from the network device to report the second information. For example, the terminal device reports the second information when the channel measurement result (e.g., RSRP) is greater than a threshold configured by the network device; or when the change in the channel measurement result (e.g., RSRP) is greater than a threshold; or when the model's confidence level is lower than a threshold; or when the network device instructs the terminal device to report the second information via DCI or MAC CE. Of course, the terminal device may also report the second information based on a period configured by the network device.

[0077] The model's input information may include some or all of the following: the model's input feature dimensions (e.g., a 12x16 compressed CSI matrix), the model's feature extraction method (e.g., using a compressed CSI matrix instead of the original channel estimate), and the model's input format.

[0078] The model's output information may include some or all of the following: the model's output quantization precision (e.g., FP16, i.e., 16-bit floating point), the model's latency information, and the model's output format.

[0079] The QoS requirements corresponding to the model include, for example, the accuracy of the model's output information and / or the model's latency information. For example, the accuracy information could be minimum accuracy, i.e., the worst-case scenario that the model's output CSI needs to meet.

[0080] The activation conditions for the model include, for example, conditions for activating local inference of the model, and / or conditions for activating collaborative inference of the model. Conditions for activating local inference of the model include, for example, channel measurement results meeting a threshold, and / or the terminal device's battery level meeting a threshold; conditions for activating collaborative inference of the model include, for example, NPU utilization, and / or the availability of edge computing for the network device.

[0081] Local inference can refer to the inference process of a model deployed in a terminal device being executed entirely within the terminal device. Collaborative inference can refer to the inference process being jointly completed between a model deployed in a terminal device and a model deployed in a network device. For example, a network device can receive CSIs reported by multiple terminal devices and comprehensively consider the CSIs reported by multiple terminal devices to determine the current channel state. In this case, the model deployed in the terminal device is allowed to output an incomplete CSI and report an incomplete CSI to the network device, and the model deployed in the network device assists in the inference to form a complete CSI.

[0082] Due to its collaborative mechanism applicable to AI / ML functions in 5G / 6G, it balances network control and terminal device autonomy. It retains the low-latency advantage of local computing on terminal devices while avoiding the "AI island" effect through global optimization of network devices, making it a key design for achieving native AI in 5G-A / 6G. Compared to fully distributed AI, network device guidance can reduce redundant signaling. For example, in the URLLC scenario defined by 3GPP, network device-guided model scheduling can effectively improve the end-to-end latency target of less than 1ms. Through network device coordination, the battery life of terminal devices can be effectively extended.

[0083] It should be noted that the latency information mentioned in the various embodiments of this application can be of various types, such as inference latency, scenario-based latency, hot-loading latency, and end-to-end latency. Inference latency, for example, refers to the time elapsed from the model's input to its output; scenario-based latency is scenario-dependent, for example, the latency for eMBB scenarios is ≤2ms, and the latency for URLLC scenarios is ≤1ms; hot-loading latency, for example, refers to the model's loading time; and end-to-end latency, for example, refers to the time elapsed from CSI-RS measurement to the model outputting the CSI prediction result.

[0084] In some implementations, the first RRC configuration / reconfiguration message sent by the network device may also include first indication information, which indicates whether model configuration updates based on second information are supported. If the first indication information indicates support for model configuration updates based on second information, the terminal device may send the second information to the network device when necessary, for example, by including the second information when sending the first RRC configuration / reconfiguration completion message.

[0085] Network devices can provide model configuration information through RRC messages, indicating whether terminal devices are allowed to report model applicability, and / or indicating whether terminal devices support proactive reporting and / or triggered reporting for model applicability (e.g., reporting second information when specific conditions are met, or reporting second information when the channel state changes, etc.).

[0086] In addition, the first RRC configuration / reconfiguration message may also include other information, such as, an indication of reference signal mapping information (i.e., resource configuration of the reference signal) used for measurement, which is configured via IE CSI-ResourceConfig; or, for example, the number of antenna ports that the terminal device needs to consider in beam prediction; or, for example, the number of prediction instances used for CSI prediction (e.g., N4) and / or the interval of prediction instances (e.g., d); or, for example, performance metrics used for model monitoring (e.g., SGCS).

[0087] The configuration information mentioned above can be carried by various information elements (IEs). For example, new RRC parameters can be introduced into the RRC message to integrate these IEs, so that the terminal device can perform a model suitability check based on the first RRC configuration / reconfiguration message.

[0088] As an example, network devices can carry key policy parameters through OtherConfig IEs, such as triggering conditions for carrying secondary information (e.g., CSI-RSRP change exceeding 3bB), validity timers (e.g., T320=5s), and model constraint information (e.g., model version, model type (e.g., neural network type)). The terminal device dynamically determines the applicability of the model based on factors such as network conditions, local resources, and the QoS requirements corresponding to the model.

[0089] For example, network devices can carry certain configuration information through the AIConfig IE in the first RRC configuration / reconfiguration message: AIConfig ::= SEQUENCE { modelID INTEGER(1..256), / / Standard model number inputPreprocess ENUMERATED {compressedCSI, rawIQ}, qosRequirement SEQUENCE { maxLatency INTEGER(1..100), / / Unit: ms minAccuracyINTEGER(70..100) / / Percentage } }

[0090] For example, network devices can set the activation threshold of the model through the ThresholdConfig IE in the first RRC configuration / reconfiguration message. An example of the triggering logic for CSI prediction is as follows: if (CSI_RSRP_variation > network_configured_threshold / / CSI-RSRP change > network device configuration threshold) and UE_battery_level>30%): / / UE battery level > 30% run_local_AI_inference() / / Run local AI inference.

[0091] For example, network devices can dynamically adjust task allocation based on load conditions: if (UE_NPU_utilization>80% / / NPU utilization>80%) and gNB_edge_compute_available): / / Base station edge computing is available Initiating Collaborative Reasoning else: Use local inference.

[0092] The second RRC configuration / reconfiguration message is described in detail below.

[0093] When performing CSI prediction based on a model, the terminal device can provide feedback to the network device on the model's current applicability through secondary information. This helps the network device determine whether to trust the model's output CSI prediction results or how to use them. The terminal device evaluates the applicability of its own model based on the model configured by the network device. For example, it assesses whether the current channel measurement resources are within the model's training range and whether the prediction results are reliable. For instance, the terminal device might use frequency intervals of 4 PRBs or 1 PRB for model training, but the network device's configured model uses sparser frequency intervals for inference, which may not meet QoS requirements, thus rendering the model unsuitable. Similarly, the terminal device might use random pilots during model training, which better matches the current scenario, but the network device's configured model uses equally spaced frequencies for inference, which may also fail to meet QoS requirements, making the model unsuitable.

[0094] After receiving the second information sent by the terminal device, the network device may optionally send a second RRC reconfiguration message to the terminal device. Accordingly, the terminal device receives the second RRC reconfiguration message sent by the network device. The second RRC reconfiguration message includes model configuration information. This configuration information is determined, for example, based on the applicability of the model indicated in the second information reported by the terminal device. Further, optionally, the terminal device may send a second RRC reconfiguration complete message to the network device. Accordingly, the network device receives the second RRC reconfiguration complete message sent by the terminal device.

[0095] As an example, such as Figure 4 As shown, in step 410, the network device sends third information to the terminal device to request the terminal device to report capability information related to the model.

[0096] In step 420, the terminal device sends first information to the network device to indicate the terminal device's model-related capability information.

[0097] In step 430, the network device sends a first RRC configuration / reconfiguration message to the terminal device, which carries configuration information of the model determined by the network device based on the first information.

[0098] In step 440, the terminal device sends a first RRC configuration / reconfiguration completion message to the network device. This first RRC configuration / reconfiguration completion message includes second information.

[0099] In step 450, the network device sends a second RRC reconfiguration message to the terminal device, which carries the configuration information of the model redefined by the network device based on the second information.

[0100] In step 460, the terminal device sends a second RRC reconfiguration complete message to the network device.

[0101] In step 470, the terminal device and the network device perform operations such as model activation, deactivation, model inference, and model monitoring.

[0102] Whether or not to send the second RRC reconfiguration message can be decided autonomously by the network device. That is, after receiving the second information sent by the terminal device, the network device can send the second RRC reconfiguration message to the terminal device based on the second information to update or adjust the model configured in the first RRC configuration / reconfiguration message; or, after receiving the second information sent by the terminal device, the network device may choose not to send the second RRC reconfiguration message to the terminal device for updating or adjusting the model.

[0103] The following examples illustrate how network devices can configure models.

[0104] Terminal devices and network devices exchange model-related capabilities and configurations via RRC signaling. Terminal devices can determine model suitability based on network conditions and local resources. Model configuration information may be distributed to terminal devices by network devices in stages (e.g., via a first RRC configuration / reconfiguration message and a second RRC reconfiguration message), supporting model optimization by network devices.

[0105] If the network device does not provide the model's configuration information in the first RRC configuration / reconfiguration message, the network device can send the model's configuration information (e.g., model parameters, model activation conditions, etc.) through the second RRC reconfiguration message after receiving the second information sent by the terminal device.

[0106] If the network device provides the model's configuration information in the first RRC configuration / reconfiguration message it sends, the network device determines whether to update or adjust the model's configuration information, such as whether to optimize model parameters or reconfigure the model to adapt to new scenarios or new states of terminal devices.

[0107] If the network device determines that it does not need to update or adjust the model's configuration information, it may not send a second RRC reconfiguration message. Optionally, the terminal device may determine the applicable configuration information from among the multiple configurations provided in the first RRC configuration / reconfiguration message based on the applicability of the model it monitors. Optionally, the second message or the first RRC configuration / reconfiguration message may indicate the configuration information of models that are available and / or unavailable to the terminal device (e.g., the model's reported configuration, the type of reported configuration, model parameters, etc.).

[0108] For example, if the network device provides configuration information for three models in the first RRC configuration / reconfiguration message, and the terminal device determines that the applicability of two of these configuration information sets can meet the requirements, then the terminal device considers these two configuration information sets to be usable and reports the usable and / or unusable configuration information to the network device through the second information.

[0109] For example, a network device provides configuration information for five models in its first RRC configuration / reconfiguration message. These five configurations include different CSI resource configurations, corresponding to five sets of CSI-RS resources. However, due to interference, beam blocking, or other limitations, the terminal device, after evaluation, determines that two of the CSI-RS resources are unavailable and reports the evaluation results—the available and / or unavailable CSI-RS resources—to the network device. For instance, this limitation is reported to the network device via a second message. If the network device decides not to reconfigure via the second RRC configuration / reconfiguration message, the terminal device defaults to using the remaining three available CSI-RS resources for measurement and prediction. If the network device does not send a second RRC reconfiguration message after receiving the second message from the terminal device, it implies acceptance of the terminal device's choice, essentially adopting the CSI-RS resources (or a subset of resources) recommended by the terminal device. Therefore, subsequent operations of the terminal device (e.g., beam management or CSI feedback) will only use the three verified CSI resources, while the remaining two CSI-RS resources will remain inactive unless subsequently reconfigured. This mechanism ensures backward compatibility and optimizes resource configuration by relying on the terminal device's real-time evaluation, reducing explicit reconfiguration signaling and thus significantly lowering signaling overhead.

[0110] For example, if the network device does not send a second RRC reconfiguration message to update or adjust the model's configuration information, the terminal device will independently determine which inference configurations are still applicable based on the previous evaluation results, and will continue to send the second message to maintain the network device's right to know its operating status.

[0111] For configuration information of models unavailable to the terminal device (e.g., inference parameter sets, invalid CSI-ReportConfig types, etc.), the network device will not configure it for the terminal device, and such configuration information (e.g., reported configuration of the model, type of reported configuration, etc.) can be considered invalid. Optionally, the terminal device can report available and / or unavailable configuration information through a second message. The terminal device and the network device exchange model applicability and configuration parameters through RRC signaling. The terminal device needs to determine the applicability of the model (or, the applicable function) based on network conditions and local resources. The model configuration information may be sent through a second RRC reconfiguration message to achieve network-side optimization.

[0112] If the configuration information of the model included in the first RRC configuration / reconfiguration message sent by the network device cannot meet the needs of the terminal device, that is, the applicability of the model cannot meet the current state of the terminal device, then the network device can send the configuration information of the model to the terminal device through the second RRC reconfiguration message after the terminal device reports the second information.

[0113] If the configuration information of the model included in the first RRC configuration / reconfiguration message sent by the network device can meet the needs of the terminal device, that is, if the model is suitable for the current state of the terminal device, then after receiving the second information reported by the terminal device, the network device can decide whether to update the configuration information based on the second information.

[0114] If the model configuration information included in the first RRC configuration / reconfiguration message sent by the network device meets the needs of the terminal device (i.e., the model is suitable for the current state of the terminal device), then the terminal device can send a first RRC configuration / reconfiguration completion message to the network device after receiving the first RRC configuration / reconfiguration message. If the network device decides not to reconfigure the model for the terminal device after receiving the first RRC configuration / reconfiguration message, it considers the first RRC configuration / reconfiguration message valid. Therefore, subsequent operations such as model activation, deactivation, model inference, and model monitoring can begin.

[0115] If the model configuration information included in the first RRC configuration / reconfiguration message sent by the network device meets the needs of the terminal device (i.e., the model is suitable for the current state of the terminal device), then the terminal device can send a first RRC configuration / reconfiguration completion message to the network device after receiving the first RRC configuration / reconfiguration message. If the network device decides to reconfigure the model for the terminal device after receiving the first RRC configuration / reconfiguration message, it considers the first RRC configuration / reconfiguration message invalid. In this case, it needs to wait until the model configuration information is reissued to the terminal device, for example, after the second RRC reconfiguration message or the second RRC reconfiguration completion message, before it can begin performing operations such as model activation, deactivation, model inference, and model monitoring.

[0116] The terminal device may not know whether the network device will update or adjust the model's configuration information via a second RRC reconfiguration message. That is, the terminal device may not know whether the network device will perform one RRC configuration / reconfiguration process (e.g., sending only the first RRC configuration / reconfiguration message, hereinafter also referred to as single configuration) or two RRC configuration / reconfiguration processes (e.g., sending both the first and second RRC reconfiguration messages, hereinafter also referred to as double configuration). This may lead to inconsistencies in the terminal device's and network device's understanding of the model's effective time. Therefore, embodiments of this application provide the following methods to solve this problem.

[0117] For example, a network device can carry corresponding indication information in the first RRC configuration / reconfiguration message to indicate whether the network device will send a second RRC reconfiguration message.

[0118] For example, after the terminal device indicates the applicability of the model through the second information, if the model is not applicable (e.g., the resources configured by the network device do not meet the resources required for model inference), the terminal device may assume that the network device needs to send a second RRC reconfiguration message.

[0119] For example, after the terminal device indicates the applicability of the model through the second information, if the model is not applicable (e.g., the resources configured by the network device do not meet the resources required for model inference), the terminal device may not send back a first RRC configuration / reconfiguration completion message in response to the first RRC configuration / reconfiguration message. In this case, optionally, the second information may not be carried in the first RRC configuration / reconfiguration completion message, but may be reported to the network device through other messages.

[0120] For example, the model may be configured to remain inactive for a predetermined time window following an RRC reconfiguration message. This means a time window is reserved to provide flexibility for network devices to reconfigure the model. During this time window, terminal devices will not report CSI, or in other words, will not perform CSI-RS measurements, or activate the model. This time window allows network devices time to perform RRC reconfiguration, send RRC reconfiguration messages, and for terminal devices to detect RRC reconfiguration messages and respond with corresponding RRC reconfiguration completion messages. The length of this time window can be configured by higher-layer signaling or agreed upon in the protocol. If no RRC reconfiguration message is received within the time window, the model can be activated after the time window ends; if RRC reconfiguration is performed within the time window, the model will be activated after the corresponding RRC reconfiguration completion message is received.

[0121] by Figure 4 For example, if the second RRC reconfiguration message is received within the time window following the first RRC configuration / reconfiguration message, the model's activation time is after the second RRC reconfiguration completion message corresponding to the second RRC reconfiguration message; and / or, if the second RRC reconfiguration message is not received within the time window following the first RRC configuration / reconfiguration message, the model's activation time is after the end of the time window.

[0122] For example, when a terminal device reports a CSI, it can include the timestamp of the associated CSI-RS measurement. The network device can then use this timestamp to determine whether the CSI-RS measurement or the CSI report is valid. As an example, using the CSI-RS measurement timestamps t1-t3 to predict the CSI at time t4, if the interval between t1-t3 and t4 is large, the CSI may be invalid; if the interval between t1-t3 and t4 is small, the CSI is valid.

[0123] For example, when a terminal device reports a CSI, the network device can determine whether the CSI or the CSI report is valid based on the time of the CSI report.

[0124] The activation of the model described in this application embodiment refers to the terminal device being able to start using the configured model to perform CSI prediction, which can also be called the activation of CSI reporting or the activation of CSI-RS measurement.

[0125] As mentioned earlier, the model activation time can be after the RRC reconfiguration completion message.

[0126] For example, without secondary configuration, the model is activated after the first RRC configuration / reconfiguration completion message. Specifically, for non-periodic and semi-persistent CSI reporting (i.e., the CSI reporting configuration carried in the model's configuration information indicates non-periodic and semi-persistent CSI reporting), the network device activates the model after the first RRC configuration / reconfiguration completion message. For periodic CSI reporting (i.e., the CSI reporting configuration carried in the model's configuration information indicates the CSI reporting period), the terminal device can activate the model after the first RRC configuration / reconfiguration completion message to perform CSI prediction and periodic CSI reporting.

[0127] For example, non-periodic and semi-persistent CSI reporting can be activated or triggered by the network device after RRC reconfiguration is completed. If the network device sends a second RRC reconfiguration message, the activation process will occur after the network device receives the RRC reconfiguration completion message. Alternatively, the terminal device can report available model configurations via a second message, and the network device, when activating the model, indicates the specific configuration item to be activated.

[0128] For example, in the case of secondary configuration, the model activation time is after the second RRC reconfiguration completion message. Specifically, for non-periodic and semi-persistent CSI reporting (i.e., the CSI reporting configuration carried in the model's configuration information indicates non-periodic and semi-persistent CSI reporting), the network device activates the model after the second RRC reconfiguration completion message. For periodic CSI reporting (i.e., the CSI reporting configuration carried in the model's configuration information indicates the CSI reporting period), the terminal device can activate the model after the second RRC reconfiguration completion message to perform CSI prediction and periodic CSI reporting. Since the first RRC configuration / reconfiguration completion message can also carry the second information, it can also be said that the model activation time is after the second information.

[0129] In other implementations, the model activation time can also occur after the second message; that is, the model is activated after the terminal device sends the second message. For example, for non-periodic or semi-persistent CSI reporting, the network device activates the model after the terminal device sends the first RRC configuration / reconfiguration message carrying the second information. As another example, for periodic CSI reporting, the terminal device can activate the model after sending the first RRC configuration / reconfiguration message carrying the second information, so as to perform CSI prediction and periodic CSI reporting through the model.

[0130] The network device can activate the model, for example, by sending RRC messages, DCI messages, MAC CE messages, etc., to the terminal device. For instance, the network device provides multiple sets of configuration information to the terminal device through a first RRC configuration / reconfiguration message. The terminal device can send a second message through the first RRC configuration / reconfiguration completion message to indicate the available and / or unavailable configuration information among the multiple configurations. The network device can then send a second RRC reconfiguration message to the terminal device, instructing the terminal device in the second RRC reconfiguration message to activate which set of available configuration information it reported.

[0131] In other implementations, the model activation time can also occur after the acknowledgment (e.g., ACK) message corresponding to the RRC configuration / reconfiguration message. For example, after the network device sends a second RRC reconfiguration message to the terminal device, the terminal device can send the corresponding acknowledgment message to the network device. The model can then be activated after this acknowledgment message.

[0132] Optionally, the first RRC configuration / reconfiguration message sent by the network device includes multiple sets of configuration information, and the second information is used to indicate the configuration information that is available and / or unavailable to the terminal device among the multiple sets of configuration information.

[0133] Optionally, the second information is used to indicate multiple sets of configuration information available to the terminal device, and the network device indicates that the configuration information activated by the terminal device belongs to the multiple sets of configuration information. That is, the network device can instruct the terminal device to activate one or more sets of configuration information among the multiple sets of configuration information it has available. Alternatively, the network device can also instruct the terminal device to activate a specific parameter in a set of available configuration information.

[0134] For example, the network device provides configuration information for three models in the first RRC configuration / reconfiguration message, assuming they are configuration information for model 1, model 2, and model 3, respectively. If the terminal device determines that the applicability of model 1 and model 2 can meet the requirements, then the terminal device considers the configuration information of model 1 and model 2 to be available, and indicates to the network device through the second message that the configuration information of model 1 and model 2 is available, or that model 1 and model 2 are available.

[0135] For example, if a terminal device indicates the configuration information of two available models to a network device through the second information, assuming they are configuration information of model 1 and configuration information of model 2 respectively, then when the network device activates or triggers a model in the terminal device, it can instruct the terminal device to activate model 1 (or the configuration information of model 1), activate model 2 (or the configuration information of model 2), or activate both model 1 and model 2 (or the configuration information of model 1 and model 2) simultaneously.

[0136] Optionally, the network device may also indicate the activation time of the aforementioned models that need to be activated.

[0137] It should be noted that, in the embodiments of this application, the first RRC configuration / reconfiguration message can also be replaced with other messages (e.g., RRC messages of other types or in other processes), and the second RRC reconfiguration message can also be replaced with other types of messages (e.g., RRC messages of other types or in other processes). The embodiments of this application do not limit this.

[0138] The above text combined Figures 1 to 4 The method embodiments of this application are described in detail below, in conjunction with... Figures 5 to 7 The present application provides a detailed description of the apparatus embodiments. It should be understood that the descriptions of the method embodiments correspond to the descriptions of the apparatus embodiments; therefore, any parts not described in detail can be found in the foregoing method embodiments.

[0139] Figure 5 This is a schematic diagram of the structure of a terminal device provided in an embodiment of this application. Figure 5 The terminal device 500 shown may include a transceiver unit 510. The transceiver unit 510 is used to send first information and / or second information to a network device, wherein the first information is used to indicate model-related capability information of the terminal device, and the second information is used to indicate the applicability of the model configured by the network device.

[0140] In some implementations, the capability information includes: the hardware capabilities of the terminal device; whether the terminal device supports time alignment with the network device; the models supported by the terminal device; and whether the models support specific algorithms.

[0141] In some implementations, the transceiver unit 510 is further configured to: receive third information sent by the network device, the third information being used to request the terminal device to send the first information.

[0142] In some implementations, the second information includes one or more of the following: a model suitability flag, used to indicate whether the model matches the current channel conditions; input data validity, used to indicate whether the input data of the model is valid; time validity, used to indicate the duration corresponding to the measurement data used to obtain the prediction result; environment matching degree, used to indicate the similarity between the current environment and the environment during the model training phase; and a rollback indication, used to indicate whether to roll back to a mechanism that does not use the model.

[0143] In some implementations, the second information is triggered when the applicability of the model changes; or, the second information is triggered based on triggering conditions configured by the network device.

[0144] In some implementations, the triggering conditions configured for the network device include periodic triggering, semi-persistent triggering, or event triggering.

[0145] In some implementations, the transceiver unit 510 is further configured to: the terminal device receive a first RRC configuration / reconfiguration message sent by the network device, the first RRC configuration / reconfiguration message including configuration information of the model, the configuration information being determined based on the capability information.

[0146] In some implementations, the configuration information includes one or more of the following: the triggering condition of the second information; the validity period of the configuration information; the model parameters; the model version information; the model input information; the model output information; information on time instances related to channel measurements based on the model; information on time instances related to channel prediction based on the model; the QoS requirements corresponding to the model; the activation condition of the model; and the channel state information (CSI) reporting configuration based on the model.

[0147] In some implementations, the triggering conditions include one or more of the following: the channel measurement result meets a threshold; the change in the channel measurement result meets a threshold; the confidence level of the model is lower than a threshold; the input data of the model is abnormal; or an instruction to trigger the reporting of the second information is received from a network device.

[0148] In some implementations, the input information of the model includes one or more of the following: the input feature dimension of the model; the feature extraction method of the model; and the input format of the model.

[0149] In some implementations, the output information of the model includes: the output quantization accuracy of the model; the time delay information of the model; and the output format of the model.

[0150] In some implementations, the QoS requirements corresponding to the model include: the accuracy of the model's output information; and / or, the model's latency information.

[0151] In some implementations, the activation conditions of the model include: conditions for activating local reasoning of the model; and / or, conditions for activating collaborative reasoning of the model.

[0152] In some implementations, the conditions for activating local inference of the model include channel measurement results meeting a threshold, and / or the battery level of the terminal device meeting a threshold; and / or, the conditions for activating collaborative inference of the model include the utilization rate of the neural network processor (NPU), and / or the availability of edge computing for the network device.

[0153] In some implementations, the first RRC configuration / reconfiguration message sent by the network device includes first indication information, which is used to indicate whether model configuration updates based on the second information are supported.

[0154] In some implementations, the transceiver unit 510 is further configured to: receive a second RRC reconfiguration message sent by the network device, the second RRC reconfiguration message including configuration information of the model updated based on the second information.

[0155] In some implementations, the transceiver unit 510 is further configured to: send an RRC reconfiguration complete message to the network device.

[0156] In some implementations, the RRC reconfiguration complete message includes: sending a first RRC configuration / reconfiguration complete message after the first RRC configuration / reconfiguration message; and / or sending a second RRC reconfiguration complete message after the second RRC reconfiguration message.

[0157] In some implementations, the model is activated after the RRC configuration / reconfiguration completion message; and / or, the model is activated after the second message; and / or, the model is activated after the confirmation message corresponding to the second RRC reconfiguration message.

[0158] In some implementations, the model is configured not to be activated within a predetermined time window following the RRC reconfiguration message.

[0159] In some implementations, if a second RRC reconfiguration message is received within the time window following the first RRC configuration / reconfiguration message, the activation time of the model is after the second RRC reconfiguration completion message corresponding to the second RRC reconfiguration message; and / or, if no second RRC reconfiguration message is received within the time window following the first RRC configuration / reconfiguration message, the activation time of the model is after the end time of the time window.

[0160] In some implementations, the first RRC configuration / reconfiguration message sent by the network device includes multiple sets of configuration information, and the second information is used to indicate the configuration information available to the terminal device among the multiple sets of configuration information.

[0161] In some implementations, the second information is used to indicate multiple sets of configuration information available to the terminal device, and the network device indicates that the configuration information activated by the terminal device belongs to the multiple sets of configuration information.

[0162] It is understood that the transceiver unit 510 may be, for example, a transceiver 730. Additionally, optionally, the terminal device 500 may also include a processor 710 and a memory 720, see details below. Figure 7 .

[0163] Figure 6 This is a schematic diagram of the network device provided in an embodiment of this application. Figure 6 The network device 600 shown may include a transceiver unit 610. The transceiver unit 610 is used to receive first information and / or second information sent by a terminal device, wherein the first information is used to indicate model-related capability information of the terminal device, and the second information is used to indicate the applicability of the model configured by the network device.

[0164] In some implementations, a first information and / or a second information sent by a terminal device is received, wherein the first information is used to indicate the model-related capability information of the terminal device, and the second information is used to indicate the applicability of the model configured by the network device.

[0165] In some implementations, the capability information includes: the hardware capabilities of the terminal device; whether the terminal device supports time alignment with the network device; the models supported by the terminal device; and whether the models support specific algorithms.

[0166] In some implementations, the transceiver unit 610 is further configured to: send third information to the terminal device, the third information being used to request the terminal device to send the first information.

[0167] In some implementations, the second information includes one or more of the following: a model suitability flag, used to indicate whether the model matches the current channel conditions; input data validity, used to indicate whether the input data of the model is valid; time validity, used to indicate the duration corresponding to the measurement data used to obtain the prediction result; environment matching degree, used to indicate the similarity between the current environment and the environment during the model training phase; and a rollback indication, used to indicate whether to roll back to a mechanism that does not use the model.

[0168] In some implementations, the second information is triggered when the applicability of the model changes; or, the second information is triggered based on triggering conditions configured by the network device.

[0169] In some implementations, the triggering conditions configured for the network device include periodic triggering, semi-persistent triggering, or event triggering.

[0170] In some implementations, the transceiver unit 610 is further configured to: send a first RRC configuration / reconfiguration message to the terminal device, the first RRC configuration / reconfiguration message including configuration information of the model, the configuration information being determined based on the capability information.

[0171] In some implementations, the configuration information includes one or more of the following: the triggering condition of the second information; the validity period of the configuration information; the model parameters; the model version information; the model input information; the model output information; information on time instances related to channel measurements based on the model; information on time instances related to channel prediction based on the model; the QoS requirements corresponding to the model; the activation condition of the model; and the channel state information (CSI) reporting configuration based on the model.

[0172] In some implementations, the triggering conditions include one or more of the following: the channel measurement result meets a threshold; the change in the channel measurement result meets a threshold; the confidence level of the model is lower than a threshold; the input data of the model is abnormal; or an instruction to trigger the reporting of the second information is received from a network device.

[0173] In some implementations, the input information of the model includes one or more of the following: the input feature dimension of the model; the feature extraction method of the model; and the input format of the model.

[0174] In some implementations, the output information of the model includes: the output quantization accuracy of the model; the time delay information of the model; and the output format of the model.

[0175] In some implementations, the QoS requirements corresponding to the model include: the accuracy of the model's output information; and / or, the model's latency information.

[0176] In some implementations, the activation conditions of the model include: conditions for activating local reasoning of the model; and / or, conditions for activating collaborative reasoning of the model.

[0177] In some implementations, the conditions for activating local inference of the model include channel measurement results meeting a threshold, and / or the battery level of the terminal device meeting a threshold; and / or, the conditions for activating collaborative inference of the model include the utilization rate of the neural network processor (NPU), and / or the availability of edge computing for the network device.

[0178] In some implementations, the first RRC configuration / reconfiguration message sent by the network device includes first indication information, which is used to indicate whether model configuration updates based on the second information are supported.

[0179] In some implementations, the transceiver unit 610 is further configured to: send a second RRC reconfiguration message to the terminal device, the second RRC reconfiguration message including configuration information of the model updated based on the second information.

[0180] In some implementations, the transceiver unit 610 is further configured to: receive an RRC reconfiguration complete message sent by the terminal device.

[0181] In some implementations, the RRC reconfiguration complete message includes: sending a first RRC configuration / reconfiguration complete message after the first RRC configuration / reconfiguration message; and / or sending a second RRC reconfiguration complete message after the second RRC reconfiguration message.

[0182] In some implementations, the model is activated after the RRC configuration / reconfiguration completion message; and / or, the model is activated after the second message; and / or, the model is activated after the confirmation message corresponding to the second RRC reconfiguration message.

[0183] In some implementations, the model is configured not to be activated within a predetermined time window following the RRC reconfiguration message.

[0184] In some implementations, if a second RRC reconfiguration message is received within the time window following the first RRC configuration / reconfiguration message, the activation time of the model is after the second RRC reconfiguration completion message corresponding to the second RRC reconfiguration message; and / or, if no second RRC reconfiguration message is received within the time window following the first RRC configuration / reconfiguration message, the activation time of the model is after the end time of the time window.

[0185] In some implementations, the first RRC configuration / reconfiguration message sent by the network device includes multiple sets of configuration information, and the second information is used to indicate the configuration information available to the terminal device among the multiple sets of configuration information.

[0186] In some implementations, the second information is used to indicate multiple sets of configuration information available to the terminal device, and the network device indicates that the configuration information activated by the terminal device belongs to the multiple sets of configuration information.

[0187] It is understood that the transceiver unit 610 may be, for example, a transceiver 730. Additionally, optionally, the network device 600 may also include a processor 710 and a memory 720, see details below. Figure 7 .

[0188] Figure 7 This is a schematic structural diagram of a communication apparatus according to an embodiment of this application. Figure 7 The dashed lines shown indicate that the unit or module is optional. The device 700 can be used to implement the methods described in the above method embodiments. The device 700 may be, for example, a chip, a terminal device, or a network device.

[0189] The apparatus 700 may include one or more processors 710. The processors 710 may support the apparatus 700 in implementing the methods described in the foregoing method embodiments. The processor 710 may be a general-purpose processor or a special-purpose processor. For example, the processor 710 may be a central processing unit (CPU). Alternatively, the processor 710 may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0190] The apparatus 700 may further include one or more memories 720. The memories 720 store programs that can be executed by the processor 710, causing the processor 710 to perform the methods described in the above method embodiments. The memories 720 may be independent of the processor 710, or they may be integrated into the processor 710.

[0191] The device 700 may also include a transceiver 730. The processor 710 can communicate with other devices or chips via the transceiver 730. For example, the processor 710 can send and receive data with other devices or chips via the transceiver 730.

[0192] This application also provides a communication system. The communication system includes the terminal device and network device described above. In some implementations, the system further includes other devices that interact with the terminal device and network device.

[0193] This application also provides a computer-readable storage medium for storing a program. This computer-readable storage medium can be applied to a terminal device or network device provided in this application, and the program causes a computer to execute the methods performed by the terminal device or network device in various embodiments of this application.

[0194] This application also provides a computer program product. The computer program product includes a program. This computer program product can be applied to a terminal device or network device provided in this application embodiment, and the program causes a computer to execute the methods performed by the terminal device or network device in the various embodiments of this application.

[0195] This application also provides a computer program. This computer program can be applied to the terminal device or network device provided in this application, and the computer program causes the computer to execute the methods performed by the terminal device or network device in the various embodiments of this application.

[0196] It should be understood that the terms "system" and "network" in the embodiments of this application can be used interchangeably. Furthermore, the terminology used in this application is only for explaining specific embodiments of this application and is not intended to limit this application. The terms "first," "second," "third," and "fourth," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. In addition, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion.

[0197] In the embodiments of this application, the term "instruction" can be a direct instruction, an indirect instruction, or an indication of a relationship. For example, A instructing B can mean that A directly instructs B, such as B being able to obtain information through A; it can also mean that A indirectly instructs B, such as A instructing C, so B can obtain information through C; or it can mean that there is a relationship between A and B.

[0198] In the embodiments of this application, "B corresponding to A" means that B is associated with A, and B can be determined based on A. However, it should also be understood that determining B based on A does not mean that B is determined solely based on A; B can also be determined based on A and / or other information.

[0199] In the embodiments of this application, the term "correspondence" can indicate a direct or indirect correspondence between two things, or an association between two things, or a relationship of instruction and being instructed, configuration and being configured, etc.

[0200] In this application embodiment, "predefined" or "preconfigured" can be implemented by pre-storing corresponding codes, tables, or other means that can be used to indicate relevant information in the device (e.g., including terminal devices and network devices). This application does not limit the specific implementation method. For example, predefined can refer to what is defined in the protocol.

[0201] In this application embodiment, the "protocol" may refer to a standard protocol in the field of communication, such as the LTE protocol, the NR protocol, and related protocols applied to future communication systems. This application does not limit this.

[0202] In the embodiments of this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0203] In the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0204] In the 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 devices or units may be electrical, mechanical, or other forms.

[0205] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0206] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0207] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in 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 the computer program 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. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can read or a data storage device such as a server or data center that integrates one or more available media. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., digital video discs (DVDs)), or semiconductor media (e.g., solid-state disks (SSDs)).

[0208] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A communication method, characterized in that, include: The terminal device sends first information and / or second information to the network device, wherein the first information is used to indicate the model-related capability information of the terminal device, and the second information is used to indicate the applicability of the model configured by the network device.

2. The method according to claim 1, characterized in that, The capability information includes: The hardware capabilities of the terminal device; Does the terminal device support time alignment with the network device? Models supported by the terminal device; Does the model support a specific algorithm? 3. The method according to claim 1 or 2, characterized in that, The method further includes: The terminal device receives third information sent by the network device, the third information being used to request the terminal device to send the first information.

4. The method according to any one of claims 1 to 3, characterized in that, The second information includes one or more of the following: A model suitability flag is used to indicate whether the model matches the current channel conditions; Input data validity, used to indicate whether the input data of the model is valid; Time validity is used to indicate the duration of the measurement data used to obtain the prediction results; Environment matching degree is used to indicate the similarity between the current environment and the environment during the model training phase; A rollback indicator is used to indicate whether to roll back to a mechanism that does not use the model.

5. The method according to any one of claims 1 to 4, characterized in that, The second information is triggered when the applicability of the model changes; or, The second information is triggered based on the triggering conditions configured in the network device.

6. The method according to claim 5, characterized in that, The triggering conditions configured for the network device include periodic triggering, semi-persistent triggering, or event triggering.

7. The method according to any one of claims 1 to 6, characterized in that, The method further includes: The terminal device receives a first Radio Resource Control (RRC) configuration / reconfiguration message sent by the network device. The first RRC configuration / reconfiguration message includes configuration information of the model, and the configuration information is determined based on the capability information.

8. The method according to claim 7, characterized in that, The configuration information includes one or more of the following: The triggering condition for the second information; The validity period of the configuration information; The model parameters of the model; The version information of the model; The input information of the model; The output information of the model; Information about time instances related to channel measurements based on the model; Information about time instances related to channel prediction based on the model; The Quality of Service (QoS) requirements corresponding to the model; The activation conditions of the model; Channel State Information (CSI) reporting configuration based on the aforementioned model.

9. The method according to claim 5 or 8, characterized in that, The triggering conditions include one or more of the following: The channel measurement results meet the threshold. The changes in the channel measurement results satisfy the threshold; The confidence level of the model is below the threshold; The model's input data is abnormal; Received an instruction from the network device to trigger the reporting of the second information.

10. The method according to claim 8 or 9, characterized in that, The input information of the model includes one or more of the following: The input feature dimension of the model; The feature extraction method of the model; The input format of the model.

11. The method according to any one of claims 8 to 10, characterized in that, The output information of the model includes: The output quantization accuracy of the model; The model's latency information; The output format of the model.

12. The method according to any one of claims 8 to 11, characterized in that, The QoS requirements corresponding to the model include: The accuracy of the model's output information; and / or, The time delay information of the model.

13. The method according to any one of claims 8 to 12, characterized in that, The activation conditions of the model include: Conditions used to activate local inference of the model; and / or, Conditions used to activate collaborative reasoning in the model.

14. The method according to claim 13, characterized in that, Conditions for activating local inference of the model include channel measurement results meeting a threshold, and / or the battery level of the terminal device meeting a threshold; and / or conditions for activating collaborative inference of the model include the utilization rate of the neural network processor (NPU), and / or the availability of edge computing for the network device.

15. The method according to any one of claims 1 to 14, characterized in that, The first RRC configuration / reconfiguration message sent by the network device includes first indication information, which is used to indicate whether model configuration updates based on the second information are supported.

16. The method according to any one of claims 1 to 15, characterized in that, The method further includes: The terminal device receives a second RRC reconfiguration message sent by the network device, the second RRC reconfiguration message including configuration information of the model updated based on the second information.

17. The method according to any one of claims 1 to 16, characterized in that, The method further includes: The terminal device sends an RRC reconfiguration complete message to the network device.

18. The method according to claim 17, characterized in that, The RRC reconfiguration completion message includes: Send a first RRC configuration / reconfiguration completion message after the first RRC configuration / reconfiguration message; and / or, A second RRC reconfiguration complete message is sent after the second RRC reconfiguration message.

19. The method according to any one of claims 1 to 18, characterized in that, The model is activated after the RRC configuration / reconfiguration completion message; and / or, The model is activated after the second information; and / or, The model is activated after the confirmation message corresponding to the second RRC reconfiguration message.

20. The method according to claim 19, characterized in that, The model is configured to remain inactive for a predetermined time window following the RRC reconfiguration message.

21. The method according to claim 20, characterized in that, If a second RRC reconfiguration message is received within the time window following the first RRC configuration / reconfiguration message, the activation time of the model is after the second RRC reconfiguration completion message corresponding to the second RRC reconfiguration message; and / or, If no second RRC reconfiguration message is received within the time window following the first RRC configuration / reconfiguration message, the activation time of the model is after the end of the time window.

22. The method according to any one of claims 1 to 21, characterized in that, The first RRC configuration / reconfiguration message sent by the network device includes multiple sets of configuration information, and the second information is used to indicate the configuration information available to the terminal device among the multiple sets of configuration information.

23. The method according to any one of claims 1 to 22, characterized in that, The second information is used to indicate multiple sets of configuration information available to the terminal device, and the network device indicates that the configuration information activated by the terminal device belongs to the multiple sets of configuration information.

24. A communication method, characterized in that, include: The network device receives first information and / or second information sent by the terminal device, wherein the first information is used to indicate the model-related capability information of the terminal device, and the second information is used to indicate the applicability of the model configured by the network device.

25. The method according to claim 24, characterized in that, The capability information includes: The hardware capabilities of the terminal device; Does the terminal device support time alignment with the network device? Models supported by the terminal device; Does the model support a specific algorithm? 26. The method according to claim 24 or 25, characterized in that, The method further includes: The network device sends a third message to the terminal device, the third message being used to request the terminal device to send the first message.

27. The method according to any one of claims 24 to 26, characterized in that, The second information includes one or more of the following: A model suitability flag is used to indicate whether the model matches the current channel conditions; Input data validity, used to indicate whether the input data of the model is valid; Time validity is used to indicate the duration of the measurement data used to obtain the prediction results; Environment matching degree is used to indicate the similarity between the current environment and the environment during the model training phase; A rollback indicator is used to indicate whether to roll back to a mechanism that does not use the model.

28. The method according to any one of claims 24 to 27, characterized in that, The second information is triggered when the applicability of the model changes; or, The second information is triggered based on the triggering conditions configured in the network device.

29. The method according to claim 28, characterized in that, The triggering conditions configured for the network device include periodic triggering, semi-persistent triggering, or event triggering.

30. The method according to any one of claims 24 to 29, characterized in that, The method further includes: The network device sends a first Radio Resource Control (RRC) configuration / reconfiguration message to the terminal device. The first RRC configuration / reconfiguration message includes the configuration information of the model, and the configuration information is determined based on the capability information.

31. The method according to claim 30, characterized in that, The configuration information includes one or more of the following: The triggering condition for the second information; The validity period of the configuration information; The model parameters of the model; The version information of the model; The input information of the model; The output information of the model; Information about time instances related to channel measurements based on the model; Information about time instances related to channel prediction based on the model; The Quality of Service (QoS) requirements corresponding to the model; The activation conditions of the model; Channel State Information (CSI) reporting configuration based on the aforementioned model.

32. The method according to claim 28 or 31, characterized in that, The triggering conditions include one or more of the following: The channel measurement results meet the threshold. The changes in the channel measurement results satisfy the threshold; The confidence level of the model is below the threshold; The model's input data is abnormal; Received an instruction from the network device to trigger the reporting of the second information.

33. The method according to claim 31 or 32, characterized in that, The input information of the model includes one or more of the following: The input feature dimension of the model; The feature extraction method of the model; The input format of the model.

34. The method according to any one of claims 31 to 33, characterized in that, The output information of the model includes: The output quantization accuracy of the model; The model's latency information; The output format of the model.

35. The method according to any one of claims 31 to 34, characterized in that, The QoS requirements corresponding to the model include: The accuracy of the model's output information; and / or, The time delay information of the model.

36. The method according to any one of claims 31 to 35, characterized in that, The activation conditions of the model include: Conditions used to activate local inference of the model; and / or, Conditions used to activate collaborative reasoning in the model.

37. The method according to claim 36, characterized in that, Conditions for activating local inference of the model include channel measurement results meeting a threshold, and / or the battery level of the terminal device meeting a threshold; and / or conditions for activating collaborative inference of the model include the utilization rate of the neural network processor (NPU), and / or the availability of edge computing for the network device.

38. The method according to any one of claims 24 to 37, characterized in that, The first RRC configuration / reconfiguration message sent by the network device includes first indication information, which is used to indicate whether model configuration updates based on the second information are supported.

39. The method according to any one of claims 24 to 38, characterized in that, The method further includes: The network device sends a second RRC reconfiguration message to the terminal device, the second RRC reconfiguration message including the configuration information of the model updated based on the second information.

40. The method according to any one of claims 24 to 39, characterized in that, The method further includes: The network device receives the RRC reconfiguration complete message sent by the terminal device.

41. The method according to claim 40, characterized in that, The RRC reconfiguration completion message includes: Send a first RRC configuration / reconfiguration completion message after the first RRC configuration / reconfiguration message; and / or, A second RRC reconfiguration complete message is sent after the second RRC reconfiguration message.

42. The method according to any one of claims 24 to 41, characterized in that, The model is activated after the RRC configuration / reconfiguration completion message; and / or, The model is activated after the second information; and / or, The model is activated after the confirmation message corresponding to the second RRC reconfiguration message.

43. The method according to claim 42, characterized in that, The model is configured to remain inactive for a predetermined time window following the RRC reconfiguration message.

44. The method according to claim 43, characterized in that, If a second RRC reconfiguration message is received within the time window following the first RRC configuration / reconfiguration message, the activation time of the model is after the second RRC reconfiguration completion message corresponding to the second RRC reconfiguration message; and / or, If no second RRC reconfiguration message is received within the time window following the first RRC configuration / reconfiguration message, the activation time of the model is after the end of the time window.

45. The method according to any one of claims 24 to 44, characterized in that, The first RRC configuration / reconfiguration message sent by the network device includes multiple sets of configuration information, and the second information is used to indicate the configuration information available to the terminal device among the multiple sets of configuration information.

46. ​​The method according to any one of claims 24 to 45, characterized in that, The second information is used to indicate multiple sets of configuration information available to the terminal device, and the network device indicates that the configuration information activated by the terminal device belongs to the multiple sets of configuration information.

47. A terminal device, characterized in that, include: A transceiver unit is used to send first information and / or second information to a network device, wherein the first information is used to indicate model-related capability information of the terminal device, and the second information is used to indicate the applicability of the model configured by the network device.

48. The terminal device according to claim 47, characterized in that, The capability information includes: The hardware capabilities of the terminal device; Does the terminal device support time alignment with the network device? Models supported by the terminal device; Does the model support a specific algorithm? 49. The terminal device according to claim 47 or 48, characterized in that, The transceiver unit is also used for: The third information sent by the network device is received, and the third information is used to request the terminal device to send the first information.

50. The terminal device according to any one of claims 47 to 49, characterized in that, The second information includes one or more of the following: A model suitability flag is used to indicate whether the model matches the current channel conditions; Input data validity, used to indicate whether the input data of the model is valid; Time validity is used to indicate the duration of the measurement data used to obtain the prediction results; Environment matching degree is used to indicate the similarity between the current environment and the environment during the model training phase; A rollback indicator is used to indicate whether to roll back to a mechanism that does not use the model.

51. The terminal device according to any one of claims 47 to 50, characterized in that, The second information is triggered when the applicability of the model changes; or, The second information is triggered based on the triggering conditions configured in the network device.

52. The terminal device according to claim 51, characterized in that, The triggering conditions configured for the network device include periodic triggering, semi-persistent triggering, or event triggering.

53. The terminal device according to any one of claims 47 to 52, characterized in that, The transceiver unit is also used for: The network device receives a first Radio Resource Control (RRC) configuration / reconfiguration message, the first RRC configuration / reconfiguration message including configuration information of the model, the configuration information being determined based on the capability information.

54. The terminal device according to claim 53, characterized in that, The configuration information includes one or more of the following: The triggering condition for the second information; The validity period of the configuration information; The model parameters of the model; The version information of the model; The input information of the model; The output information of the model; Information about time instances related to channel measurements based on the model; Information about time instances related to channel prediction based on the model; The Quality of Service (QoS) requirements corresponding to the model; The activation conditions of the model; Channel State Information (CSI) reporting configuration based on the aforementioned model.

55. The terminal device according to claim 51 or 54, characterized in that, The triggering conditions include one or more of the following: The channel measurement results meet the threshold. The changes in the channel measurement results satisfy the threshold; The confidence level of the model is below the threshold; The model's input data is abnormal; Received an instruction from the network device to trigger the reporting of the second information.

56. The terminal device according to claim 54 or 55, characterized in that, The input information of the model includes one or more of the following: The input feature dimension of the model; The feature extraction method of the model; The input format of the model.

57. The terminal device according to any one of claims 54 to 56, characterized in that, The output information of the model includes: The output quantization accuracy of the model; The model's latency information; The output format of the model.

58. The terminal device according to any one of claims 54 to 57, characterized in that, The QoS requirements corresponding to the model include: The accuracy of the model's output information; and / or, The time delay information of the model.

59. The terminal device according to any one of claims 54 to 58, characterized in that, The activation conditions of the model include: Conditions used to activate local inference of the model; and / or, Conditions used to activate collaborative reasoning in the model.

60. The terminal device according to claim 59, characterized in that, Conditions for activating local inference of the model include channel measurement results meeting a threshold, and / or the battery level of the terminal device meeting a threshold; and / or conditions for activating collaborative inference of the model include the utilization rate of the neural network processor (NPU), and / or the availability of edge computing for the network device.

61. The terminal device according to any one of claims 47 to 60, characterized in that, The first RRC configuration / reconfiguration message sent by the network device includes first indication information, which is used to indicate whether model configuration updates based on the second information are supported.

62. The terminal device according to any one of claims 47 to 61, characterized in that, The transceiver unit is also used for: The network device receives a second RRC reconfiguration message, the second RRC reconfiguration message including configuration information of the model updated based on the second information.

63. The terminal device according to any one of claims 47 to 62, characterized in that, The transceiver unit is also used for: Send an RRC reconfiguration complete message to the network device.

64. The terminal device according to claim 63, characterized in that, The RRC reconfiguration completion message includes: Send a first RRC configuration / reconfiguration completion message after the first RRC configuration / reconfiguration message; and / or, A second RRC reconfiguration complete message is sent after the second RRC reconfiguration message.

65. The terminal device according to any one of claims 47 to 64, characterized in that, The model is activated after the RRC configuration / reconfiguration completion message; and / or, The model is activated after the second information; and / or, The model is activated after the confirmation message corresponding to the second RRC reconfiguration message.

66. The terminal device according to claim 65, characterized in that, The model is configured to remain inactive for a predetermined time window following the RRC reconfiguration message.

67. The terminal device according to claim 20, characterized in that, If a second RRC reconfiguration message is received within the time window following the first RRC configuration / reconfiguration message, the activation time of the model is after the second RRC reconfiguration completion message corresponding to the second RRC reconfiguration message; and / or, If no second RRC reconfiguration message is received within the time window following the first RRC configuration / reconfiguration message, the activation time of the model is after the end of the time window.

68. The terminal device according to any one of claims 47 to 67, characterized in that, The first RRC configuration / reconfiguration message sent by the network device includes multiple sets of configuration information, and the second information is used to indicate the configuration information available to the terminal device among the multiple sets of configuration information.

69. The terminal device according to any one of claims 47 to 68, characterized in that, The second information is used to indicate multiple sets of configuration information available to the terminal device, and the network device indicates that the configuration information activated by the terminal device belongs to the multiple sets of configuration information.

70. A network device, characterized in that, include: The transceiver unit is used to receive first information and / or second information sent by the terminal device, wherein the first information is used to indicate the model-related capability information of the terminal device, and the second information is used to indicate the applicability of the model configured by the network device.

71. The network device according to claim 70, characterized in that, The capability information includes: The hardware capabilities of the terminal device; Does the terminal device support time alignment with the network device? Models supported by the terminal device; Does the model support a specific algorithm? 72. The network device according to claim 70 or 71, characterized in that, The transceiver unit is also used for: Send a third message to the terminal device, the third message being used to request the terminal device to send the first message.

73. The network device according to any one of claims 70 to 72, characterized in that, The second information includes one or more of the following: A model suitability flag is used to indicate whether the model matches the current channel conditions; Input data validity, used to indicate whether the input data of the model is valid; Time validity is used to indicate the duration of the measurement data used to obtain the prediction results; Environment matching degree is used to indicate the similarity between the current environment and the environment during the model training phase; A rollback indicator is used to indicate whether to roll back to a mechanism that does not use the model.

74. The network device according to any one of claims 70 to 73, characterized in that, The second information is triggered when the applicability of the model changes; or, The second information is triggered based on the triggering conditions configured in the network device.

75. The network device according to claim 74, characterized in that, The triggering conditions configured for the network device include periodic triggering, semi-persistent triggering, or event triggering.

76. The network device according to any one of claims 70 to 75, characterized in that, The transceiver unit is also used for: A first Radio Resource Control (RRC) configuration / reconfiguration message is sent to the terminal device. The first RRC configuration / reconfiguration message includes the configuration information of the model, and the configuration information is determined based on the capability information.

77. The network device according to claim 76, characterized in that, The configuration information includes one or more of the following: The triggering condition for the second information; The validity period of the configuration information; The model parameters of the model; The version information of the model; The input information of the model; The output information of the model; Information about time instances related to channel measurements based on the model; Information about time instances related to channel prediction based on the model; The Quality of Service (QoS) requirements corresponding to the model; The activation conditions of the model; Channel State Information (CSI) reporting configuration based on the aforementioned model.

78. The network device according to claim 74 or 77, characterized in that, The triggering conditions include one or more of the following: The channel measurement results meet the threshold. The changes in the channel measurement results satisfy the threshold; The confidence level of the model is below the threshold; The model's input data is abnormal; Received an instruction from the network device to trigger the reporting of the second information.

79. The network device according to claim 77 or 78, characterized in that, The input information of the model includes one or more of the following: The input feature dimension of the model; The feature extraction method of the model; The input format of the model.

80. The network device according to any one of claims 77 to 79, characterized in that, The output information of the model includes: The output quantization accuracy of the model; The model's latency information; The output format of the model.

81. The network device according to any one of claims 77 to 80, characterized in that, The QoS requirements corresponding to the model include: The accuracy of the model's output information; and / or, The time delay information of the model.

82. The network device according to any one of claims 77 to 81, characterized in that, The activation conditions of the model include: Conditions used to activate local inference of the model; and / or, Conditions used to activate collaborative reasoning in the model.

83. The network device according to claim 82, characterized in that, Conditions for activating local inference of the model include channel measurement results meeting a threshold, and / or the battery level of the terminal device meeting a threshold; and / or conditions for activating collaborative inference of the model include the utilization rate of the neural network processor (NPU), and / or the availability of edge computing for the network device.

84. The network device according to any one of claims 70 to 83, characterized in that, The first RRC configuration / reconfiguration message sent by the network device includes first indication information, which is used to indicate whether model configuration updates based on the second information are supported.

85. The network device according to any one of claims 70 to 84, characterized in that, The transceiver unit is also used for: A second RRC reconfiguration message is sent to the terminal device, the second RRC reconfiguration message including the configuration information of the model updated based on the second information.

86. The network device according to any one of claims 70 to 85, characterized in that, The transceiver unit is also used for: Receive the RRC reconfiguration complete message sent by the terminal device.

87. The network device according to claim 86, characterized in that, The RRC reconfiguration completion message includes: Send a first RRC configuration / reconfiguration completion message after the first RRC configuration / reconfiguration message; and / or, A second RRC reconfiguration complete message is sent after the second RRC reconfiguration message.

88. The network device according to any one of claims 70 to 87, characterized in that, The model is activated after the RRC configuration / reconfiguration completion message; and / or, The model is activated after the second information; and / or, The model is activated after the confirmation message corresponding to the second RRC reconfiguration message.

89. The network device according to claim 88, characterized in that, The model is configured to remain inactive for a predetermined time window following the RRC reconfiguration message.

90. The network device according to claim 89, characterized in that, If a second RRC reconfiguration message is received within the time window following the first RRC configuration / reconfiguration message, the activation time of the model is after the second RRC reconfiguration completion message corresponding to the second RRC reconfiguration message; and / or, If no second RRC reconfiguration message is received within the time window following the first RRC configuration / reconfiguration message, the activation time of the model is after the end of the time window.

91. The network device according to any one of claims 70 to 90, characterized in that, The first RRC configuration / reconfiguration message sent by the network device includes multiple sets of configuration information, and the second information is used to indicate the configuration information available to the terminal device among the multiple sets of configuration information.

92. The network device according to any one of claims 70 to 91, characterized in that, The second information is used to indicate multiple sets of configuration information available to the terminal device, and the network device indicates that the configuration information activated by the terminal device belongs to the multiple sets of configuration information.

93. A terminal device, characterized in that, The device includes a transceiver, a memory, and a processor. The memory stores a program, and the processor invokes the program in the memory and controls the transceiver to receive or send signals so that the terminal device performs the method according to any one of claims 1 to 23.

94. A network device, characterized in that, The device includes a transceiver, a memory, and a processor. The memory stores a program, and the processor invokes the program in the memory and controls the transceiver to receive or transmit signals so that the network device performs the method according to any one of claims 24 to 46.

95. An apparatus, characterized in that, Includes a processor for calling a program from memory to cause the apparatus to perform the method according to any one of claims 1 to 46.

96. A chip, characterized in that, Includes a processor for calling a program from memory, causing a device on which the chip is mounted to perform the method according to any one of claims 1 to 46.

97. A computer-readable storage medium, characterized in that, It contains a program that causes a computer to perform the method according to any one of claims 1 to 46.

98. A computer program product, characterized in that, Includes a program that causes a computer to perform the method according to any one of claims 1 to 46.

99. A computer program, characterized in that, The computer program causes the computer to perform the method according to any one of claims 1 to 46.

Images