UE capability reporting method and apparatus, communication node, and storage medium
By having the UE report AI/ML support capability information and applicability reports to the network-side nodes, the problem of limited UE computing power is solved, enabling the network side to effectively utilize the UE's AI/ML functions and improving the performance of the wireless communication system.
Patent Information
- Application Number
- PCT/CN2026/093794
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-08
- Filing Date
- 2026-04-29
- Publication Date
- 2026-11-12
AI Technical Summary
User equipment (UE) has limited computing power, resulting in insufficient support for artificial intelligence (AI)/machine learning (ML) functions. Therefore, effective capability reporting methods are needed to improve the performance of wireless communication systems.
The UE receives capability query information from the network-side node, sends AI/ML support capability information, and reports the applicability of supported AI/ML functions or models after receiving RRC reconfiguration information. The network-side node queries and configures the UE through corresponding signaling to obtain the specific support status of the UE.
This enables network-side nodes to fully utilize the AI/ML capabilities of the UE, thereby improving the performance of the mobile communication network.
Smart Images

Figure CN2026093794_12112026_PF_FP_ABST
Abstract
Description
UE capability reporting methods, devices, communication nodes, and storage media Technical Field
[0001] This application relates to the field of wireless communication, specifically to a UE capability reporting method, apparatus, communication node, and storage medium. Background Technology
[0002] With the development of Artificial Intelligence (AI) and Machine Learning (ML) technologies, there are proposals in the industry to apply AI / ML functions to mobile communications to improve the performance of wireless communication systems. However, AI / ML functions require significant computing resources, while the computing power of User Equipment (UE) is relatively limited. Therefore, the UE's support for AI / ML functions is also limited. To implement AI / ML functions on the UE side, it is necessary for the UE to report its support capabilities for AI / ML functions to the network side. Therefore, how to specifically report UE-side AI / ML functions is a problem that urgently needs to be solved. Summary of the Invention
[0003] In view of this, embodiments of this application aim to provide a UE capability reporting method, apparatus, communication node, and storage medium.
[0004] In a first aspect, embodiments of this application provide a UE capability reporting method, applied to a UE, including:
[0005] Receive UE capability query information sent by network-side nodes. The UE capability query information is used to query the AI / ML support capability information of the UE.
[0006] Send the UE's AI / ML support capability information to the network-side nodes;
[0007] Receive RRC reconfiguration information sent by the network-side node. The RRC reconfiguration information is used to request the UE to report the applicability of the supported AI / ML functions or models.
[0008] Send a report to the network-side node on the suitability of the UE for the supported AI / ML functions or models.
[0009] Secondly, embodiments of this application provide a UE capability reporting method, applied to a network-side node, including:
[0010] Send UE capability query information to the UE. The UE capability query information is used to query the AI / ML support capability information of the UE.
[0011] Receive AI / ML support capability information sent by the UE;
[0012] Send RRC reconfiguration information to the UE. The RRC reconfiguration information is used to request the UE to report the applicability of the supported AI / ML functions or models.
[0013] Receive the UE's applicability report for the supported AI / ML functions or models sent by the UE.
[0014] Thirdly, embodiments of this application provide a UE capability reporting device, comprising:
[0015] The receiving module is used to receive UE capability query information sent by the network-side node. The UE capability query information is used to query the AI / ML support capability information of the UE.
[0016] The sending module is used to send the UE's AI / ML support capability information to the network-side nodes;
[0017] The receiving module is also used to receive RRC reconfiguration information sent by the network-side node. The RRC reconfiguration information is used to request the UE to report the applicability of the supported AI / ML functions or models.
[0018] The sending module is also used to send a report to the network-side node on the suitability of the UE for the supported AI / ML functions or models.
[0019] Fourthly, embodiments of this application provide a UE capability reporting device, including:
[0020] The sending module is used to send UE capability query information to the UE. The UE capability query information is used to query the AI / ML support capability information of the UE.
[0021] The receiving module is used to receive AI / ML support capability information sent by the UE;
[0022] The sending module is also used to send RRC reconfiguration information to the UE. The RRC reconfiguration information is used to request the UE to report the applicability of the supported AI / ML functions or models.
[0023] The receiving module is also used to receive a report from the UE on the suitability of the UE for the supported AI / ML functions or models.
[0024] Fifthly, embodiments of this application provide a UE, including:
[0025] The memory is configured to store a program.
[0026] The processor is configured to execute a program, which, when executed, performs a UE capability reporting method as implemented in any of the first aspects.
[0027] Sixthly, embodiments of this application provide a network-side node, including:
[0028] The memory is configured to store a program.
[0029] The processor is configured to execute a program, which, when executed, performs a UE capability reporting method as implemented in any of the second aspects.
[0030] In a seventh aspect, embodiments of this application provide a non-volatile storage medium, the storage medium including a stored program, wherein the program executes a UE capability reporting method according to any implementation of the first aspect, or executes a UE capability reporting method according to any implementation of the second aspect when it runs. Attached Figure Description
[0031] Figure 1 is a flowchart illustrating a UE capability reporting method provided in an embodiment of this application;
[0032] Figure 2 is a schematic diagram of the AI / ML function or model information of the UE supporting data collection in the UE capability reporting method provided in the embodiments of this application;
[0033] Figure 3 is a schematic diagram of a UE capability information reporting method provided in the UE capability reporting method of this application;
[0034] Figure 4 is a schematic diagram of another UE capability information reporting method provided in the UE capability reporting method of this application;
[0035] Figure 5 is a flowchart illustrating another UE capability reporting method provided in an embodiment of this application;
[0036] Figure 6 is a schematic diagram of a UE capability reporting device provided in an embodiment of this application;
[0037] Figure 7 is a schematic diagram of another UE capability reporting device provided in an embodiment of this application;
[0038] Figure 8 is a schematic diagram of the structure of a UE provided in an embodiment of this application;
[0039] Figure 9 is a schematic diagram of the structure of a network side node provided in an embodiment of this application. Detailed Implementation
[0040] To make the purpose, technical solution and beneficial effects of this application clearer, the embodiments of this application will be described below with reference to the accompanying drawings. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this application can be arbitrarily combined with each other.
[0041] Figure 1 is a flowchart illustrating a UE capability reporting method provided in an embodiment of this application. As shown in Figure 1, the UE capability reporting method provided in this embodiment includes:
[0042] Step S110: Receive UE capability query information sent by the network-side node. The UE capability query information is used to query the AI / ML support capability information of the UE.
[0043] In one implementation, the UE capability query information sent by the network-side node includes a query indication for querying the UE AI / ML capability acquisition type to obtain information about the UE's AI / ML capabilities.
[0044] In another implementation, the UE capability query information sent by the network-side node includes a query indication for querying the UE NR capability acquisition type to obtain information about AI / ML UE capabilities.
[0045] The UE capability reporting method provided in this embodiment is applied to a UE in a mobile communication network. The UE can be any type of terminal device or node located on the user side of the mobile communication network. The UE can be, for example, a mobile terminal device such as a mobile phone or tablet computer, or other terminal devices with data acquisition and uplink data transmission capabilities. Since the UE capability reporting method provided in this embodiment is used to report the AI / ML capabilities of the UE, the UE must be a node that supports at least one AI / ML-related capability. The UE needs to report the supported AI / ML capability information to the network-side node. Only after the network-side node knows the AI / ML capability information supported by the UE can it fully utilize the UE's AI / ML capabilities to communicate with the UE. The network-side node can be any type of node providing access services, such as a base station or access point (AP). The network-side node provides corresponding services to the UE based on its capabilities.
[0046] First, the UE receives UE capability query information sent by the network-side node. This query information is used to query the UE's AI / ML support capabilities. The network-side node is the network-side node corresponding to the cell the UE accesses. If the network-side node does not know the UE's AI / ML support capabilities, directly instructing the UE to report complete support information for AI / ML functions or models would consume a lot of signaling resources. Therefore, the network-side node can first query the UE's basic support information for AI / ML capabilities, such as whether it supports AI / ML capabilities and what AI / ML capabilities it supports. In this way, the UE capability query information and the corresponding information reported by the UE do not consume excessive signaling resources.
[0047] Step S120: Send the UE's AI / ML support capability information to the network-side node.
[0048] After receiving the UE capability query information sent by the network-side node, the UE can send its AI / ML support capability information to the network-side node. The UE's AI / ML support capability information indicates whether the UE supports AI / ML capabilities and which AI / ML capabilities it supports.
[0049] In one embodiment, the UE's AI / ML support capability information may include at least one of the following: AI / ML function or model information supported by the UE; and UE support information for data collection functions. Specifically, the AI / ML function or model information supported by the UE may indicate whether the UE supports an AI / ML function or model, or a specific AI / ML function or model supported by the UE; the UE support information for data collection functions may indicate whether the UE supports a data collection function, or a specific data collection function supported by the UE. For example, the AI / ML function or model information supported by the UE may be represented as UE-AIML-Capability; and the UE support information for data collection functions may be represented as supportedDataCollection.
[0050] In one embodiment, the AI / ML function or model information supported by the UE includes at least one of the following: hardware support information for AI / ML functions or models of the UE; and support information of the UE for AI / ML-based functions or models. That is, the AI / ML function or model information supported by the UE may include the hardware support for AI / ML of the UE, or it may include the software support for AI / ML of the UE.
[0051] In one embodiment, the hardware support information for AI / ML functions or models in the UE includes at least one of the following: the maximum computing resources available to the UE for the AI / ML functions or models, including Central Processing Unit (CPU) resources and / or Accelerated Processing Unit (APU) resources; and the maximum storage resources available to the UE for the AI / ML functions or models. Wherein, computing resources represent the maximum computing resources available for the AI / ML functions / models, CPU resources can be represented as MaximumOcpu, and APU resources can be represented as MaximumOapu. Storage resources represent the maximum storage resources available for installing the AI / ML functions / models, and can be represented as MaximumOstorage.
[0052] In one embodiment, the UE's hardware support information for AI / ML functions or models includes at least one of the following: the UE's maximum available hardware resource information; the hardware resource occupancy value of each AI / ML function or model of the UE; wherein the UE's maximum available hardware resource information or the hardware resource occupancy value of each AI / ML function or model of the UE is based on any of the following granularity levels: each correlation between the input and output information of the AI / ML model, frequency band, AI / ML function characteristics. That is, the UE's hardware support for AI / ML can be either the UE's overall maximum available hardware resource information or the hardware resource occupancy value of each AI / ML function or model of the UE.
[0053] The UE's maximum available hardware resource information can be applied to activate all AI / ML functions or models in different use cases supported by the UE.
[0054] The hardware resource consumption of each AI / ML function or model in a UE can be categorized into three different granularity levels:
[0055] Level 1: The granularity of hardware resource utilization for each AI / ML function or model in the UE is each correlation between the input and output information of the AI / ML model. The maximum hardware resource utilization at this level refers to the total hardware resource utilization when all AI / ML models under this correlation are applied.
[0056] Level 2: The granularity of hardware resource usage for each AI / ML function or model of the UE is the frequency band. The maximum hardware resource usage at this level refers to the total hardware resource usage when all AI / ML functions or models in this frequency band are applied.
[0057] Level 3: The granularity of hardware resource consumption for each AI / ML function or model of the UE is based on AI / ML functional features. These AI / ML functional features include, for example, AI / ML-based spatial beam measurement prediction, AI / ML-based temporal beam measurement prediction, AI / ML-based Channel State Information (CSI) prediction, and AI / ML-based mobility. The maximum hardware resource consumption at this level refers to the total hardware resource consumption when all AI / ML models of the functional features are applied.
[0058] In one embodiment, the UE's support information for AI / ML-based functions or models includes whether the UE supports at least one of the following capabilities: AI / ML-based beam measurement prediction, AI / ML-based CSI feedback, AI / ML-based dynamic CSI prediction, and AI / ML-based mobility measurement and prediction; wherein, AI / ML-based beam measurement prediction includes AI / ML-based spatial beam measurement prediction and / or AI / ML-based temporal beam measurement prediction.
[0059] In one embodiment, for supported AI / ML-based beam measurement prediction, the UE includes at least one of the following capability information: input-related capability information of the supported AI / ML function or model, output-related capability information of the supported AI / ML function or model, correlation between input and output information of the supported AI / ML function or model, minimum time offset between input and output information of the supported AI / ML function or model, and inference capability related to CSI reports.
[0060] In one embodiment, the input-related capability information for each AI / ML function or model includes a list of the number of reference signals in dataset B; in another embodiment, the input-related capability information for each AI / ML function or model includes a list of the maximum number of reference signals in dataset B.
[0061] In one embodiment, the output-related capability information for each AI / ML function or model includes a list of the number of reference signals in dataset A. In another embodiment, the input-related capability information for each AI / ML function or model includes a list of the maximum number of CSIs or Synchronization Signaling Blocks (SSBs) in dataset A.
[0062] In one embodiment, the association between the input and output information of the supported AI / ML function or model can be implicitly represented as a one-to-one correspondence between items in the input list and items in the output list in sequence. In another embodiment, the association between the input and output information of the supported AI / ML function or model can be explicitly indicated as a list of associations between input and output information, where the input information can be a reference signal of dataset B and the output information can be a reference signal of dataset A.
[0063] In one embodiment, the minimum time offset between the input and output information of the supported AI / ML function or model identifies the minimum time requirement for model inference.
[0064] In one embodiment, the inference capability associated with CSI reporting can be the Reference Signal Receiving Power (RSRP) value or the maximum number of beams that can be reported in each CSI reporting instance.
[0065] In one embodiment, AI / ML-based beam measurement prediction includes one of the following: AI / ML-based beam measurement prediction for a frequency band; AI / ML-based beam measurement prediction for a combination of frequency bands; and AI / ML-based beam measurement prediction for a carrier.
[0066] For AI / ML-based beam measurement prediction for a frequency band, it means that the UE collects data to train models or functions at the frequency band or component carrier (CC) granularity, and trains AI / ML models or functions accordingly. In this sense, the network-side node cannot determine whether the serving cell belonging to the frequency band has any supported AI / ML functions or models. That is, if at least one AI / ML function or model is applicable to the frequency band, the network-side node can still trigger the CC applicability report.
[0067] For AI / ML-based beam measurement prediction for band-combination (BC), it means that the UE collects data to train a model or function, and trains the AI / ML model or function accordingly. If at least one AI / ML function or model is supported for the band combination to which the CC belongs, the network-side node must trigger the CC's applicability report.
[0068] In one embodiment, the supported AI / ML functions or models are placed in a Band-Combination, but unlike the traditional per-BC function, this means that if an AI / ML function / model is supported for a Band-Combination (e.g., BAND#1, BAND#2, BAND#3), then an AI / ML function / model is supported for BAND#1, BAND#2, and BAND#3 respectively.
[0069] In another embodiment, the supported AI / ML functions or models are placed in the root node of physical parameters or UE-NR-Capability, and a list of frequency bands is displayed for each supported AI / ML function or model.
[0070] For AI / ML-based beam measurement prediction for a carrier, it means that the UE collects data to train a model or function, and trains the AI / ML model or function accordingly. The capability reporting overhead may be greater than that for beam measurement prediction for a frequency band or a combination of frequency bands. However, if the CC does not have a supported AI / ML function or model, the network-side node does not need to trigger the CC's applicability report.
[0071] In one embodiment, conventional AI / ML-based beam measurement prediction capability information is placed into the downlink set (SetDownLinkperCC) of the function CC. The supported bandwidth (supportedBandWidthForAI) is introduced to indicate the bandwidth of the supported AI / ML function or model, and supportedBandWidthForAI is less than or equal to the supported downlink bandwidth (supportedBandWidthDL).
[0072] In one embodiment, the list of supported AI / ML functions or models is placed in the NR band (BandNR) or the MIMO-ParameterPerBand. Then, for each supported AI / ML function or model, there may be a list of supported CCs, and in each CC, there may be a supported AI bandwidth (BandWidthForAI) and / or a supported AI subcarrier spacing (SubcarrierSpacingForAI).
[0073] In one embodiment, the UE's support information for data collection functions indicates that the UE has at least one of the following capabilities: data collection capability, maximum data collection buffer size, AI / ML functions or model information that support data collection.
[0074] The data collection capability indicates whether the UE can record measurement results of data collected on the UE side or the network side. The maximum data collection buffer size indicates the maximum buffer size used to record data collected on the network side and / or the maximum buffer size used for data collection on the UE side.
[0075] The use cases for AI / ML functions or model information supporting data collection indicate the use cases or features that support data collection, such as AI / ML-based beam measurement prediction (including spatial beam measurement prediction or temporal beam measurement prediction), AI / ML-based CSI feedback, AI / ML-based CSI prediction, and AI / ML-based mobility measurement and prediction. For each AI / ML function or model use case or feature, it may include one of the following recording schemes and contents: Whether event-triggered data recording is supported: Indicates whether the UE supports event-triggered data recording for network-side data collection. Whether timestamp recording is supported: Indicates whether the UE supports recording timestamp information when recording data. Whether location information recording is supported: Indicates whether the UE supports recording current location information when recording data. Whether speed information recording is supported: Indicates whether the UE supports recording current UE speed information for the recorded data.
[0076] Figure 2 is a schematic diagram of the AI / ML function or model information supporting data collection of the UE in the UE capability reporting method provided in this application embodiment. As shown in Figure 2, the AI / ML function or model information supporting data collection of the UE (AIMLDataCollection) includes the maximum buffer size (MaximumASBufferSize), such as 64MB, 128MB, 256MB, 512MB, etc. (ENUMERATED(mb64, mb128, mb256, mb512…)). The AI / ML function or model information supporting data collection of the UE also includes a list of use cases supporting data collection (supportUseCaseList), for example, use case 1 supporting data collection (SupportUseCase1) includes use cases supporting data collection (SupportUse), event-based data logging (EventBasedDataLogging), timestamp (TimeStamp), location information (Locationinfo), UE speed (UESpeed), etc. Alternatively, the AI / ML function or model information supporting data collection of the UE may also include other information.
[0077] In one embodiment, the UE's support information for the data collection function is carried in the UE NR Capability information (UE-NR-Capability), or the UE's support information for the data collection function is carried in the UE-BasedPerfMeas-Parameter.
[0078] In one embodiment, the UE's AI / ML support capability information is reported via signaling dedicated to AI / ML capability reporting, or the UE's AI / ML support capability information is carried in the UE NR capability information.
[0079] In one implementation, the UE's AI / ML support capability information is reported via dedicated AI / ML capability reporting signaling (UE-AIML-Capability). This dedicated AI / ML capability reporting signaling may be located at the root node of the UE capability report, meaning that the UE's AI / ML-related capabilities and radio-related capabilities are separated. In this case, for AI / ML-based beam measurement prediction, the UE's AI / ML support capability information can be included in the AI / ML support function (SupportedAIMLFunction).
[0080] Figure 3 is a schematic diagram of UE capability information reporting information in the UE capability reporting method provided in this application embodiment. As shown in Figure 3, the UE capability information (UECapabilityInformation) includes UE NR capability information (UE-NR-Capability), UE Multi-Radio Dual Connectivity (MRDC) capability information (UE-MRDC-Capability), UE Evolved Universal Terrestrial Radio Access (EUTRA) capability information (UE-EUTRA-Capability), and UE AI / ML capability information (UE-AIML-Capability). Among them, UE-AIML-Capability includes, for example, maximum computing resources information (MaximumComputingResources), maximum storage resources information (MaximumStorageResources), and supported AI / ML functions (SupportedAIMLFunction).
[0081] In one implementation, the UE's AI / ML support capability information is carried in the UE NR capability information (UE-NR-Capability), which means that AI / ML related capabilities are merged into radio-related capabilities.
[0082] Figure 4 is a schematic diagram of another UE capability information reporting method provided in this application embodiment. As shown in Figure 4, the UE capability information (UECapabilityInformation) includes UE NR capability information (UE-NR-Capability), UE Multi-Radio Dual Connectivity (MRDC) capability information (UE-MRDC-Capability), and UE Evolved Universal Terrestrial Radio Access (EUTRA) capability information (UE-EUTRA-Capability). Among them, UE-NR-Capability includes UE AI / ML capability (UE-AIML-Capability), which includes, for example, maximum computing resources information (MaximumComputingResources), maximum storage resources information (MaximumStorageResources), and supported AI / ML functions (SupportedAIMLFunction).
[0083] Step S130: Receive RRC reconfiguration information sent by the network-side node. The RRC reconfiguration information is used to request the UE to report the applicability of the supported AI / ML functions or models.
[0084] After the network-side node learns the UE's AI / ML support capability information, it can send Radio Resource Control (RRC) information to the UE to configure specific AI / ML functions or models. The RRC reconfiguration information is used to request a suitability report from the UE for the supported AI / ML functions or models. The AI / ML support capability information reported by the UE only includes whether the UE's hardware or software supports AI / ML capabilities and whether it supports different AI / ML capabilities, but it does not include the specific support methods and capabilities of the UE for AI / ML. Therefore, the network-side node needs to further obtain a suitability report from the UE for the supported AI / ML functions or models based on the UE's AI / ML support capability information.
[0085] In one embodiment, the UE determines the activated AI / ML function or model based on the CSI reporting configuration sent by the network-side node. Specifically, for beam measurement prediction, the network-side node activates model inference by activating the CSI reporting configuration (CSI-ReportConfig). The CSI reporting configuration is indicated and / or activated by Downlink Control Information (DCI) or Media Access Control-Control Element (MAC CE), or by RRC signaling. The network-side node uses DCI or MAC CE to activate a semi-persistent, semi-fixed CSI reporting configuration, or uses RRC signaling to activate a periodic CSI reporting configuration.
[0086] In one embodiment, an activated AI / ML function or model may support multiple activated CSI-ReportConfigs across different serving cells. In this implementation, the model inference time of different CSI-ReportConfigs should be time-shifted to avoid model inference conflicts.
[0087] In one embodiment, an active AI / ML function or model may only support one of the active CSI-ReportConfigs in a serving cell.
[0088] Typically, when activating CSI-ReportConfig, there are two solutions to ensure that activating multiple CSI-ReportConfigs does not exceed resource limits. Solution 1: The hardware resource consumption value or maximum resource consumption value for each AI / ML function or model is included in the UE's AI / ML support capability information. Solution 2: The UE reports the hardware resource consumption value for each AI / ML function or model in its applicability report.
[0089] In one embodiment, the RRC reconfiguration information includes full model inference configuration information or partial model inference configuration information; the full model inference configuration information is reported through the CSI reporting configuration indication; the partial model inference configuration information is reported through other configuration indications in the RRC reconfiguration message.
[0090] In one implementation, an applicability report can be requested via an RRC reconfiguration message containing the complete model inference configuration. In this implementation, the complete model inference configuration can be configured via a CSI reporting configuration (CSI-ReportConfig). For example, a CSI-ReportConfig might include an indication (ApplicabilityAcquired) suggesting whether the applicability of this CSI-ReportConfig should be reported in the Applicability Reporting. In another implementation, an applicability report can be requested via an RRC reconfiguration message containing a partial model inference configuration. In this implementation, the partial inference configuration (CSI-ReportConfigAssistance) might be configured in another configuration (OtherConfig) within the RRC reconfiguration message.
[0091] Step S140: Send a report on the suitability of the UE for the supported AI / ML functions or models to the network-side node.
[0092] After receiving the RRC reconfiguration information sent by the network-side node, the UE can send a report on the applicability of the supported AI / ML functions or models to the network-side node, so that the network-side node can know the specific support methods and capabilities of the UE for AI / ML functions or models.
[0093] For triggering and sending applicability reports for AI / ML features or models:
[0094] In one implementation, the applicability report of the CSI Report Configuration (CSI-ReportConfig) may be triggered by one of the following conditions and embedded in the RRC Reconfiguration Complete message (RRCReconfigurationComplete):
[0095] The CSI-ReportConfig, which includes the associated ID and the applicability acquired, may be added via the RRC Reconfiguration message received by the UE.
[0096] The CSI-ReportConfig containing associatedId may be modified via RRCReconfiguration received by the UE.
[0097] The CSI-ReportConfig containing associatedId may be added via RRCReconfiguration received by the UE.
[0098] The serving cell configured with CSI-ReportConfig has been activated.
[0099] In one implementation, the applicability report of CSI ReportConfigAssistance may be triggered and embedded in UE Assistance Information (UAI) under one of the following conditions:
[0100] When the UE receives RRCReconfiguration, CSI-ReportConfigAssistance may exist in other configurations (otherConfig).
[0101] The applicability status of at least one CSI-ReportConfigAssistance has changed compared to the previous UAI applicability report.
[0102] The associated service cell configured with CSI-ReportConfigAssistance has been activated.
[0103] The applicability report in UAI and / or RRCReconfigurationComplete may contain one of the following information:
[0104] Each CSI-ReportConfigAssistance and / or CSI-ReportConfig has applicability information, where its configuration must include information containing the AssociatedId. In one implementation, the applicability information may be an octet string parameter, where the number of bits equals the maximum number of CSI-ReportConfigAssistances and / or CSI-ReportConfigs in the list. The first bit of the parameter corresponds to the first CSI-ReportConfigAssistance and / or the first CSI-ReportConfig used for AI model inference in the list, the second bit corresponds to the second CSI-ReportConfigAssistance and / or the CSI-ReportConfig used for AI model inference, and so on. When the bit value in the applicability information is 1 or 0, the corresponding CSI-ReportConfigAssistance and / or CSI-ReportConfig are applicable; otherwise, they are not applicable. When providing applicability information for CSI-ReportConfigAssistance and / or CSI-ReportConfig for multiple cells, they can be sorted by cell ID size. For example, first report the AI model inference applicability information for cells with serving cell ID value = 0, then report the AI model applicability information for cells with serving cell ID value = 1, and so on. For the AI model inference applicability information for each cell, they can be sorted in the aforementioned manner.
[0105] Resource occupancy information. In one implementation, resource occupancy information may exist as a sequence list type parameter, where the first entry corresponds to the first bit of the aforementioned applicability information list, the second entry corresponds to the second bit of the aforementioned applicability information list, and so on. In another implementation, resource occupancy information may exist as a sequence list type parameter, where the first entry corresponds to the first bit indicating applicability in the aforementioned applicability information list, the second entry corresponds to the second bit indicating applicability in the aforementioned applicability information list, and so on. In one implementation, each resource occupancy information entry may contain one or more CSI-ReportConfig Id / CSI-ReportConfigAssistance Ids and corresponding cell identifiers to indicate that one or more CSI-ReportConfig / CSI-ReportConfigAssistance entries may share the same AI / ML model / function.
[0106] The UE capability reporting method provided in this application is applied to a UE. After receiving UE capability query information sent by a network-side node, the UE first sends its AI / ML support capability information to the network-side node. Then, after receiving the RRC reconfiguration message generated and sent by the network-side node based on the UE's AI / ML support capability information, the UE sends a report on the applicability of the supported AI / ML functions or models to the network-side node. This allows the network-side node to know the specific support status of the accessed UE for AI / ML functions or models. In this way, the network-side node and the UE can communicate normally using AI / ML functions or models to improve the performance of the mobile communication network.
[0107] Figure 5 is a flowchart illustrating another UE capability reporting method provided in this application embodiment. As shown in Figure 5, the UE capability reporting method provided in this application embodiment includes:
[0108] Step S510: Send UE capability query information to the UE. The UE capability query information is used to query the AI / ML support capability information of the UE.
[0109] Step S520: Receive AI / ML support capability information sent by the UE.
[0110] Step S530: Send RRC reconfiguration information to the UE. The RRC reconfiguration information is used to request the UE to report the applicability of the supported AI / ML functions or models.
[0111] Step S540: Receive a report from the UE on the suitability of the supported AI / ML functions or models.
[0112] The UE capability reporting method provided in this embodiment is applied to the network-side node. Its implementation principle and technical effect are the same as those of the network-side node in the UE capability reporting method provided in the embodiment shown in Figure 1, and will not be repeated here.
[0113] Figure 6 is a schematic diagram of a UE capability reporting device provided in an embodiment of this application. The UE capability reporting device is disposed in the UE. As shown in Figure 6, the UE capability reporting device provided in this embodiment includes:
[0114] The receiving module 61 is used to receive UE capability query information sent by the network-side node, which is used to query the AI / ML support capability information of the UE; the sending module 62 is used to send the AI / ML support capability information of the UE to the network-side node; the receiving module 61 is also used to receive RRC reconfiguration information sent by the network-side node, which is used to request the UE to report the applicability of the supported AI / ML functions or models; the sending module 62 is also used to send the applicability report of the supported AI / ML functions or models of the UE to the network-side node.
[0115] The UE capability reporting device provided in this embodiment is used to implement the UE capability reporting method of the embodiment shown in Figure 1. Its implementation principle and technical effect are similar, and will not be described again here.
[0116] Figure 7 is a schematic diagram of another UE capability reporting device provided in this embodiment. This UE capability reporting device is disposed in a network-side node. As shown in Figure 7, the UE capability reporting device provided in this embodiment includes:
[0117] The sending module 71 is used to send UE capability query information to the user equipment (UE), which is used to query the AI / ML support capability information of the UE; the receiving module 72 is used to receive the AI / ML support capability information sent by the UE; the sending module 71 is also used to send RRC reconfiguration information to the UE, which is used to request the UE to submit an applicability report for the supported AI / ML functions or models; the receiving module 72 is also used to receive the UE's applicability report for the supported AI / ML functions or models.
[0118] The UE capability reporting device provided in this embodiment is used to implement the UE capability reporting method of the embodiment shown in Figure 5. Its implementation principle and technical effect are similar, and will not be described again here.
[0119] Figure 8 is a schematic diagram of the structure of a UE provided in an embodiment of this application. As shown in Figure 8, the UE includes a processor 81, a memory 82, a receiver 83, and a transmitter 84. The number of processors 81 in the UE can be one or more. Figure 8 shows one processor 81 as an example. The processor 81, memory 82, receiver 83, and transmitter 84 in the UE can be connected by a bus or other means. Figure 8 shows a connection via a bus as an example.
[0120] The memory 82, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules (receiving module 61 and sending module 62) corresponding to the UE capability reporting method in Figure 1 of this application. The processor 81 implements various functions of the UE and performs data processing by running the software programs, instructions, and modules stored in the memory 82, that is, it implements the UE capability reporting method described above.
[0121] The memory 82 may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function; the data storage area may store data created based on the UE's usage. Furthermore, the memory 82 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device.
[0122] The receiver 83 is any device / module with data receiving capability or a combination of multiple devices / modules with data receiving capability, and the transmitter 84 is any device / module with data transmitting capability or a combination of multiple devices / modules with data transmitting capability.
[0123] Figure 9 is a schematic diagram of the structure of a network side node provided in an embodiment of this application. As shown in Figure 9, the network side node includes a processor 91, a memory 92, a receiver 93, and a transmitter 94. The number of processors 91 in the network side node can be one or more. Figure 9 shows one processor 91 as an example. The processor 91, memory 92, receiver 93, and transmitter 94 in the network side node can be connected by a bus or other means. Figure 9 shows a connection via a bus as an example.
[0124] The memory 92, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules (sending module 71, receiving module 72) corresponding to the UE capability reporting method in Figure 5 of this application. The processor 91 implements various functions of the network-side node and performs data processing by running the software programs, instructions, and modules stored in the memory 92, that is, it implements the aforementioned UE capability reporting method.
[0125] The memory 92 may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function; the data storage area may store data created based on the use of network-side nodes, etc. Furthermore, the memory 92 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device.
[0126] The receiver 93 is any device / module with data receiving capability or a combination of multiple devices / modules with data receiving capability, and the transmitter 94 is any device / module with data transmitting capability or a combination of multiple devices / modules with data transmitting capability.
[0127] This application embodiment also provides a non-volatile storage medium, which includes a stored program. When the program runs, it executes a UE capability reporting method. The method includes: receiving UE capability query information sent by a network-side node, the UE capability query information being used to query the AI / ML support capability information of the UE; sending the AI / ML support capability information of the UE to the network-side node; receiving RRC reconfiguration information sent by the network-side node, the RRC reconfiguration information being used to request the UE to report the applicability of the supported AI / ML functions or models; and sending the applicability report of the UE to the supported AI / ML functions or models to the network-side node.
[0128] This application embodiment also provides a non-volatile storage medium, which includes a stored program. When the program runs, it executes a UE capability reporting method. The method includes: sending UE capability query information to the UE, the UE capability query information being used to query the UE's AI / ML support capability information; receiving the AI / ML support capability information sent by the UE; sending RRC reconfiguration information to the UE, the RRC reconfiguration information being used to request the UE to report the applicability of the supported AI / ML functions or models; and receiving the UE's applicability report of the supported AI / ML functions or models.
[0129] In this embodiment, the storage medium may include, but is not limited to, various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.
[0130] It will be understood by those skilled in the art that all or part of the steps in the methods disclosed above, and the functional modules / units in the systems and apparatuses, can be implemented as software, firmware, hardware, and suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all components may be implemented as software executed by a processor (such as a digital signal processor or microprocessor), or as hardware, or as integrated circuits, such as application-specific integrated circuits (ASICs). Such software may be distributed on a computer-readable medium, which may include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM (Electrically Erasable Programmable Read-Only Memory), flash memory or other memory technologies, CD-ROM (Compact Disc Read-Only Memory), DVD or other optical disc storage, magnetic cartridges, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.
Claims
1. A UE capability reporting method, applied to a user equipment (UE), comprising: Receive UE capability query information sent by the network-side node, wherein the UE capability query information is used to query the artificial intelligence / machine learning (AI / ML) support capability information of the UE; Send the UE's AI / ML support capability information to the network-side node; The UE receives Radio Resource Control (RRC) reconfiguration information sent by the network-side node. The RRC reconfiguration information is used to request the UE to report the applicability of the supported AI / ML functions or models. Send a report to the network-side node on the suitability of the UE for the supported AI / ML functions or models.
2. The method according to claim 1, wherein, The AI / ML support capability information of the UE includes at least one of the following: The AI / ML functions or model information supported by the UE; The UE's support information for data collection functions.
3. The method according to claim 2, wherein, The AI / ML functions or model information supported by the UE include at least one of the following: Information regarding the hardware support for AI / ML functions or models of the UE; The UE's functional support information is based on AI / ML functions or models.
4. The method according to claim 3, wherein, The hardware support information for AI / ML functions or models of the UE includes at least one of the following: The UE can be used for the maximum computing resources of AI / ML functions or models, including central processing unit (CPU) resources and / or accelerator processor (APU) resources. The UE can be used for the maximum storage resources of AI / ML functions or models.
5. The method according to claim 3, wherein, The hardware support information for AI / ML functions or models of the UE includes at least one of the following: The maximum available hardware resources of the UE; The hardware resource usage of the UE's AI / ML functions or models; The hardware resource usage of the UE's AI / ML function or model is divided according to any of the following granularity levels: each correlation between the input and output information of the AI / ML model, frequency band, and AI / ML function features.
6. The method according to claim 3, wherein, The UE's functional support information based on AI / ML functions or models includes whether the UE supports at least one of the following capabilities: AI / ML-based beam measurement prediction, AI / ML-based channel state information (CSI) feedback, AI / ML-based dynamic CSI prediction, and AI / ML-based mobility measurement and prediction. The AI / ML-based beam measurement prediction includes AI / ML-based spatial beam measurement prediction and / or AI / ML-based temporal beam measurement prediction.
7. The method according to claim 6, wherein, For AI / ML-based beam measurement prediction supported by the UE, the UE includes at least one of the following capability information: Information on the input capabilities of supported AI / ML functions or models, information on the output capabilities of supported AI / ML functions or models, the correlation between the input and output information of supported AI / ML functions or models, the minimum time offset between the input and output information of supported AI / ML functions or models, and the reasoning capabilities related to CSI reports.
8. The method according to claim 6, wherein, The AI / ML-based beam measurement prediction includes one of the following: AI / ML-based beam measurement prediction for frequency bands; AI / ML-based beam measurement prediction for frequency band combinations; AI / ML-based beam measurement prediction for carrier waves.
9. The method according to claim 2, wherein, The UE's support information for data collection functions indicates that the UE possesses at least one of the following capabilities: Data collection capabilities, maximum data collection cache size, and AI / ML functions or model information that support data collection.
10. The method according to claim 9, wherein, The UE's support information for data collection functions is carried in the UE's new radio NR capability information, or the UE's support information for data collection functions is carried in the UE's basic measurement parameters.
11. The method according to any one of claims 1 to 10, wherein, The AI / ML support capability information of the UE is reported through signaling dedicated to AI / ML capability reporting, or the AI / ML support capability information of the UE is carried in the UE NR capability information.
12. The method according to any one of claims 1 to 10, further comprising: The activated AI / ML function or model is determined based on the CSI reporting configuration sent by the network-side node; The CSI reporting configuration is indicated by downlink control information (DCI) or media access control layer control unit (MAC CE), or by RRC signaling.
13. The method according to any one of claims 1 to 10, wherein, The RRC reconfiguration information includes full model inference configuration information or partial model inference configuration information; The reporting of the complete model inference configuration information is accomplished through CSI reporting of configuration instructions. The partial model inference configuration information is indicated by other configurations in the RRC reconfiguration message.
14. A UE capability reporting method, applied to a network-side node, comprising: Send UE capability query information to the user equipment (UE), wherein the UE capability query information is used to query the artificial intelligence / machine learning (AI / ML) support capability information of the UE; Receive AI / ML support capability information sent by the UE; Send Radio Resource Control (RRC) reconfiguration information to the UE, the RRC reconfiguration information being used to request the UE to report the suitability of supported AI / ML functions or models; Receive the suitability report of the UE for the supported AI / ML functions or models sent by the UE.
15. The method according to claim 14, wherein, The AI / ML support capability information of the UE includes at least one of the following: The AI / ML functions or model information supported by the UE; The UE's support information for data collection functions.
16. The method according to claim 15, wherein, The AI / ML functions or model information supported by the UE include at least one of the following: Information regarding the hardware support for AI / ML functions or models of the UE; The UE's functional support information is based on AI / ML functions or models.
17. The method according to claim 16, wherein, The hardware support information for AI / ML functions or models of the UE includes at least one of the following: The UE can be used for the maximum computing resources of AI / ML functions or models, including central processing unit (CPU) resources and / or accelerator processor (APU) resources. The UE can be used for the maximum storage resources of AI / ML functions or models.
18. The method according to claim 16, wherein, The hardware support information for AI / ML functions or models of the UE includes at least one of the following: The maximum available hardware resources of the UE; The hardware resource usage of the UE's AI / ML functions or models; The hardware resource usage of the UE's AI / ML function or model is divided according to any of the following granularity levels: each correlation between the input and output information of the AI / ML model, frequency band, and AI / ML function features.
19. The method of claim 16, wherein, The UE's functional support information based on AI / ML functions or models includes whether the UE supports at least one of the following capabilities: AI / ML-based beam measurement prediction, AI / ML-based channel state information (CSI) feedback, AI / ML-based dynamic CSI prediction, and AI / ML-based mobility measurement and prediction. The AI / ML-based beam measurement prediction includes AI / ML-based spatial beam measurement prediction and / or AI / ML-based temporal beam measurement prediction.
20. The method according to claim 19, wherein, For AI / ML-based beam measurement prediction supported by the UE, the UE includes at least one of the following capability information: Information on the input capabilities of supported AI / ML functions or models, information on the output capabilities of supported AI / ML functions or models, the correlation between the input and output information of supported AI / ML functions or models, the minimum time offset between the input and output information of supported AI / ML functions or models, and the reasoning capabilities related to CSI reports.
21. The method according to claim 19, wherein, The AI / ML-based beam measurement prediction includes one of the following: AI / ML-based beam measurement prediction for frequency bands; AI / ML-based beam measurement prediction for frequency band combinations; AI / ML-based beam measurement prediction for carrier waves.
22. The method according to claim 15, wherein, The UE's support information for data collection functions indicates that the UE possesses at least one of the following capabilities: Data collection capabilities, maximum data collection cache size, and AI / ML functions or model information that support data collection.
23. The method according to claim 22, wherein, The UE's support information for data collection functions is carried in the UE's new radio NR capability information, or the UE's support information for data collection functions is carried in the UE's basic measurement parameters.
24. The method according to any one of claims 14 to 23, wherein, The AI / ML support capability information of the UE is reported through signaling dedicated to AI / ML capability reporting, or the AI / ML support capability information of the UE is carried in the UE NR capability information.
25. The method according to any one of claims 14 to 23, further comprising: Send CSI reporting configuration to the UE, the CSI reporting configuration being used to instruct the UE to determine the activated AI / ML function or model; The CSI reporting configuration is indicated by downlink control information (DCI) or media access control layer control unit (MAC CE), or by RRC signaling.
26. The method according to any one of claims 14 to 23, wherein, The RRC reconfiguration information includes full model inference configuration information or partial model inference configuration information; The reporting of the complete model inference configuration information is accomplished through CSI reporting of configuration instructions. The partial model inference configuration information is indicated by other configurations in the RRC reconfiguration message.
27. A user equipment (UE) capability reporting device, comprising: The receiving module is configured to receive UE capability query information sent by the network-side node. The UE capability query information is used to query the artificial intelligence / machine learning (AI / ML) support capability information of the UE. The sending module is configured to send the UE's AI / ML support capability information to the network-side node; The receiving module is further configured to receive Radio Resource Control (RRC) reconfiguration information sent by the network-side node, wherein the RRC reconfiguration information is used to request the UE to report the applicability of supported AI / ML functions or models. The sending module is also configured to send a report on the suitability of the UE for the supported AI / ML functions or models to the network-side node.
28. A UE capability reporting device, comprising: The sending module is configured to send UE capability query information to the user equipment (UE), wherein the UE capability query information is used to query the artificial intelligence / machine learning (AI / ML) support capability information of the UE; The receiving module is configured to receive AI / ML support capability information sent by the UE; The transmitting module is further configured to send Radio Resource Control (RRC) reconfiguration information to the UE, the RRC reconfiguration information being used to request the UE to report the suitability of supported AI / ML functions or models; The receiving module is also configured to receive a report from the UE on the suitability of the UE for supported AI / ML functions or models.
29. A user equipment (UE), comprising: The memory is configured to store a program. The processor is configured to execute a program, which, when executed, performs the UE capability reporting method as described in any one of claims 1 to 13.
30. A network side node, comprising: The memory is configured to store a program. The processor is configured to execute a program, which, when executed, performs the UE capability reporting method as described in any one of claims 14 to 26.
31. A non-volatile storage medium, the storage medium comprising a stored program, wherein the program, when executed, performs the UE capability reporting method according to any one of claims 1 to 13, or performs the UE capability reporting method according to any one of claims 14 to 26.
