Communication method and related apparatus

Abstract

Embodiments of the present application provide a communication method and related apparatus, applied to a UE, the communication method comprising: receiving a first message and / or a second message sent by a network side, the first message comprising first configuration information about AI-related information, the second message being used for requesting the UE to report the AI-related information; and sending a reporting message to the network side, the reporting message comprising the AI-related information and / or UE auxiliary information.

Description

Technical Field

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

[0002] With the development of computer and communication technologies, artificial intelligence and machine learning (AI / ML) have been introduced into communication systems to improve communication performance. However, current data collection schemes between user terminals and the network side, including the reporting of AI-based applicable functionality, have a series of problems that need to be improved or enhanced. Summary of the Invention

[0003] This application provides a communication method and a communication device that can enhance or improve the reporting of AI information.

[0004] To achieve the above objectives, this application adopts the following technical solution: The first aspect of this application provides a communication method applied to a UE, comprising: receiving a first message sent by a network side, and / or a second message, wherein the first message includes first configuration information about AI-related information, and the second message is used to request the UE to report the AI-related information; A reporting message is sent to the network side, the reporting message including AI-related information and / or UE assistance information.

[0005] A second aspect of this application also provides a communication method for use on the network side, comprising: sending a first message and / or a second message to a UE, wherein the first message includes first configuration information related to AI, and the second message is used to request the UE to report the AI-related information; and receiving a reporting message sent by the UE, wherein the reporting message includes the AI-related information and / or UE auxiliary information.

[0006] A third aspect of this application also provides a wireless communication device, including: a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory to perform the method as described in any of the preceding embodiments.

[0007] A fourth aspect of this application also provides a computer-readable storage medium, the computer-readable storage medium including instructions that, when executed, cause the method described in any of the preceding claims to be implemented. Attached Figure Description

[0008] Figure 1A flowchart of one possible existing communication method; Figure 2 A flowchart illustrating a possible communication method provided in an embodiment of this application; Figure 2a A flowchart illustrating another possible communication method provided in the embodiments of this application; Figure 2b A flowchart illustrating another possible communication method provided in the embodiments of this application; Figure 2c A flowchart illustrating another possible communication method provided in the embodiments of this application; Figure 2d A flowchart illustrating another possible communication method provided in the embodiments of this application; Figure 2e A flowchart illustrating another possible communication method provided in the embodiments of this application; Figure 2f A flowchart illustrating another possible communication method provided for embodiments of this application. Figure 3 A flowchart illustrating another possible communication method provided in the embodiments of this application; Figure 3a A flowchart illustrating another possible communication method provided in the embodiments of this application; Figure 3b A flowchart illustrating another possible communication method provided in the embodiments of this application; Figure 3c A flowchart illustrating another possible communication method provided in the embodiments of this application; Figure 4 This is a storage diagram of a possible wireless communication device provided in this application. Detailed Implementation

[0009] For ease of understanding, the relevant technologies involved in the embodiments of this application will be described below.

[0010] The technical solutions of the embodiments of this application will now be described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0011] It should be understood that the term "and / or" in this document 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, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship. It should be noted that the naming of the parameters in this document is for ease of description; other names may be used in practice, and this application does not impose any restrictions on their specific use.

[0012] The messages described in this article include frames, instructions, commands, etc., and the names of device or functional entities, process names, frames, fields, etc. are not unique and are only used to assist in the description of functions, methods, behaviors, information, etc.

[0013] To facilitate a better understanding of this plan, the following will provide explanations for some of the terms that may appear in this plan.

[0014] UE Assistance Information (UAI) refers to information proactively provided to the network by the User Equipment (UE). Its purpose is to help the network better understand the UE's needs, capabilities, and current state, thereby optimizing network performance and user experience. For example, if a UE is configured by the network to support reporting Overheating Assistance Information, it will report this information to the network when it detects that it is overheating, or report that its state has returned to normal when it detects that it is no longer overheating. The network configures this through the `otherConfig IE` in the RRC Reconfiguration Message, and then transmits the UE's assistance information through the `UAIMessage`. It's important to note that the UAI message is not sent periodically, but rather based on trigger conditions. The Channel State Information (CSI) Framework is a mechanism in 5G NR used to report radio channel state information, primarily for optimizing downlink transmission performance. It helps the base station (gNB) understand the UE's radio channel state, enabling scheduling and beamforming operations. The RRC Reconfiguration Message configures CSI through CSI-MeasConfig, which has three report types: 1) Periodic CSI: The UE periodically reports channel state information to the gNB; 2) Aperiodic CSI: The gNB triggers the UE to report CSI as needed; 3) Semi-persistent CSI: Used to reduce signaling overhead. CSI reports are uploaded via PUSCH (Physical Uplink Shared Channel) and PUCCH (Physical Uplink Control Channel). The content and frequency of the CSI reports are configured by the RRC layer.

[0015] The term "network" (NW) or "network side" can refer to different network elements, depending on the context. Generally, the network side can include two main parts: base stations (gNB) and the core network. Base stations are part of the Radio Access Network (RAN) and are responsible for wireless communication with the UE. Base stations include centralized and distributed units that jointly handle the radio interface protocol stack, including the physical layer, data link layer, and part of the network layer. The core network is the control plane and data plane of the wireless communication network, responsible for handling user data and signaling, as well as connections with other networks. The core network includes multiple network functions (NFs), such as AMF (Access and Mobility Management Function), SMF (Session Management Function), UPF (User Plane Function), and AI-related functional network elements.

[0016] AI-related information: This refers to information related to AI technology or information generated when performing AI-related operations, such as AI functionality, AI model, AI feature, AI feature group, or AI operations. AI functionality and / or AI model can include one or any combination of these, such as: applicablefunctionalities / AI models, functionalities / AI models, supportedfunctionalities / AI models, activated applicable functionalities / AI models, deactivated applicable functionalities / AI models, activated functionalities / AI models, and / or deactivated functionalities / AI models, etc.

[0017] Inference: Used to represent AI technology-related models or to execute logical functions based on AI models, and to obtain one or more AI-based prediction results by executing or activating the functions.

[0018] - L3 signaling: The L3 signaling can be an RRC reconfiguration completion message, a UAI message, or other (newly defined) RRC messages; SetA and SetB: Resource set information, typically referring to resource sets used for Channel State Information (CSI) reporting. These resource sets are used by the UE (User Equipment) for measurement and reporting during beam management. The following are the specific definitions and explanations of setA and setB: Set A: Set A typically refers to a set of resources used for beam management, which are used by the UE to perform CSI measurements. These measurements are typically used to evaluate beam quality and performance. The resources in Set A are used by the UE to perform channel state measurements, and these measurement results can be used for beam selection and optimization. As in CSI-ReportConfig, the configuration of Set A typically includes the resource set identifier (e.g., CSI-RS resources) and the associated measurement parameters. 5 Set B: Set B is an optional set of resources typically used to assist measurements in set A. It can provide additional reference information to help the UE perform more accurate beam management. Set B can be used to provide auxiliary measurements for the UE, which can be used to verify the measurement results of set A or provide additional contextual information. The configuration of Set B can be a CSI-RS resource or other types of reference signal resources.

[0019] In some cases, SetB can be a subset of SetA, meaning that resources in SetB are part of resources in SetBA.

[0020] By analyzing existing technologies, the basic process of reporting applicable functionality of UE-side BM in AI-based beam management (BM) use cases, such as time-based BM prediction and spatial-based BM prediction, is described. Figure 1 As shown. Specifically, in step 3, before reporting the applicable functionality, the inference configuration is configured via CSI-reportConfig signaling. The associated ID is included in the inference configuration. Subsequently, the UE determines the final applicable functionalities based on the received associated ID, its own additional conditions (memory, battery level, CPU, GPU), and available AI models. Then, in step 4a, the UE reports the initial applicable functionality via RRC ReconfigurationComplete, and subsequently reports updated applicable functionalities via UAI. These applicable functionalities are used for inference configuration.

[0021] Based on current conclusions and progress, some existing solutions only discuss signaling without specific details, or they do not discuss signaling flow and container design. Furthermore, most existing methods are designed to solve AI-based BM prediction use cases, and their CSI architecture is not suitable for AI-based mobility use cases. Therefore, it is necessary to further address the reporting of applicable functionality in AI-based mobility use cases, including but not limited to the following issues: Question 1: In the existing mechanism, the relationship between the Inference Configuration procedure and the applicable functionality reporting procedure is unclear. In the current scheme, the inference configuration is configured using the current CSI reporting configuration, CSI-ReportConfig. This inference configuration allows configuration of reporting content, reporting methods, etc. The reported content must follow the CSI architecture and be reported via PHY signaling. Furthermore, the inference configuration carries associated ID information, which assists the UE in determining or deciding on applicable functionality. Applicable functionality is reported via L3 messages (RRC reconfiguration completion messages and / or UAI messages). However, the traditional CSI mechanism does not support reporting information using L3 signaling. That is, although the associated ID is based on the CSI configuration, this ID is mainly used for the UE's internal process of determining applicable functionalities. In other words, applicable functionalities reporting and inference configuration and reporting are two separate mechanisms. Under the existing mechanism, when the network is configured with multiple inference configurations, the UE may generate multiple applicable functionalities, or a single inference configuration may also generate multiple applicable functionalities. Therefore, when the network receives these applicable functionalities, it may not be able to determine their specific Inference Configuration, which in turn makes it impossible for the network side to effectively manage the applicable functionalities.

[0022] Question 2: Existing mechanisms are unsuitable for applicable functionality reporting in AI-based mobility use cases. For current AI-based Machine Learning (BM) use cases, the network side can configure inference-related applicable functionality information for BM through the CSI mechanism. However, this is not applicable to AI-based mobility use cases because AI-based mobility typically uses an RRM architecture, and directly reusing the CSI mechanism is clearly inappropriate. Therefore, further research is needed on the specific configuration and reporting mechanisms for applicable functionality reporting in AI-based mobility use cases. Currently, there is no relevant research or discussion on this topic in the context of AI-based mobility use cases.

[0023] Question 3: Existing mechanisms suffer from excessive signaling overhead. Regardless of whether it's physical layer AI use cases (such as AI-based BM prediction) or higher-layer AI use cases (such as AI-based mobility management), the applicable functionality reporting is designed separately for each use case. This leads to excessive signaling overhead, and too many mechanisms also increase the burden on the UE and network sides. Therefore, we will further explore a unified applicable functionality reporting mechanism, that is, on-demand request reporting of the applicable functionality. This mechanism can be applied to different use cases and can significantly reduce the overhead and resource issues caused by signaling interaction.

[0024] To address the aforementioned issues, this application provides various communication methods, which are mainly described in two aspects: Scenario 1, the network side configures the reporting of AI-related information, and the UE reports based on the configuration; 2, the network side triggers the UE to report AI-related information as needed.

[0025] For scenario 1, please refer to Figure 2 This application provides a possible communication method, including but not limited to at least one of the following steps: 201. The network side sends the first message to the UE; 202. The network side sends a third message to the UE; The network sends a first message to the UE, which includes first configuration information related to AI. The content configured in the first configuration information includes, but is not limited to, at least one of the following: AI-related information-specific measurement resource information, and / or its identification information, used to identify AI-related information resources; AI-related reporting configuration identifier information is used to represent the identifier information of the AI-related information reporting configuration. AI-related information reporting type, which indicates the specific form of reporting AI-related information, can be periodic, event-triggered, semi-static, and / or irregular; The resource identification information configured for inference is used to represent the identification information of the measurement resources configured by the base station with the inference configuration information; The first identification information of the inference reporting configuration is used to represent the identification information of the channel state related (CSI) inference reporting configuration; Association identifier, used to represent the identification information of the network side information; The reported content is used to indicate the specific information that the UE needs to report, and the content must include at least AI-related information; The result information of the inference configuration is used to represent the prediction results obtained based on AI-related technologies, and / or the inference results, along with the predicted beam, the predicted target cell, the predicted RSRP, etc. The bandwidth information configured for inference is used to represent the relevant bandwidth information configured during inference. The measurement object information of AI-related information can be used to represent the granularity of AI-related information, that is, in which granularity / range it is available. The measurement object information can be UE identifier, cell identifier, cell list, area information identifier, AI model information, inference configuration signaling identifier such as RRC ID, base station identifier information, AMF identifier information, PLMN information identifier, AI task identifier, third-party server identifier, AI management network element identifier, and / or, AI service identifier. Measurement identifier, used to indicate the measurement object of the inference and the associated identifier information configured in the inference reporting report; The identification information of the measurement object in the inference is used to indicate the unique identifier of the measurement object that performs the network-side configuration; The second identifier information configured in the reasoning report; AI-related configuration association information is used to represent the identification information of the inference reporting configuration for the cell measurement or beam measurement (RRM measurement); The first associated information in the AI-related configuration is used to associate the measurement object information related to AI and the reporting configuration related to AI. The second associated information for AI-related configurations is used to associate AI-related information and associated identifiers, and / or the first identifier information for inference and reporting configurations; Cell information, which refers to the cell information configured during the inference configuration, can be cell information or cell group information, or the serving cell or neighbor cell information of the inference object; Reporting priority rules are used to determine the signaling type that carries the reported AI-related information; Signaling reporting rules; rules for determining AI information.

[0026] It should be noted that any of the information included in the first configuration information above may also be carried in other information such as request instruction signaling and sent separately, and this application does not limit the specifics.

[0027] The network sends a third message to the UE, which includes inference configuration information based on AI algorithms. This inference configuration information can be configured by CSI-ReportConfig, MeasConfig, or other IE configurations, and carries additional network information (represented by associated ID). The associated ID can be any perconfiguration, per gNB, per cell, per area, per AMF, per OAM, per PLMN, or per third party, etc., and is not specifically limited here.

[0028] Optionally, the third message may include an association identifier.

[0029] Optionally, in practical applications, the inference configuration information and the first configuration information can be sent in two separate messages, i.e., the first message and the third message are different messages, or they can be sent in one message, i.e., the first message and the third message are the same message. The specifics are not limited here.

[0030] In this embodiment, the network side can configure the aforementioned inference configuration information and first configuration information based on different architectures or scenarios, including but not limited to the following architectures or combinations thereof: L3 architecture (framework), CSI architecture, RRM architecture, and UAI architecture. For ease of understanding, various examples will be given below of how the network side configures inference configuration information and first configuration information based on different architectures. In practical applications, other architectures or scenarios can also be used, and this application does not limit the specifics.

[0031] A. Based on the CSI architecture, configure inference configuration information and first configuration information.

[0032] Option 1: Enhance CSI-Measconfig.

[0033] In this solution, a new AI-specific CSI-measconfig configuration can be introduced, or a configuration for reporting AI-related information can be introduced into the CSI-measconfig.

[0034] Specifically, the configuration includes at least one of the following: Measurement resource information and / or its identification information specific to AI-related information. This resource information is used to identify resources related to AI, and may include time-frequency resource information, computational resource information, or other information. Optionally, it may include resource group information.

[0035] AI-related reporting configuration is used to configure the reporting information related to AI, assuming it is named "aiInfo-ReportConfig".

[0036] The AI-related reporting configuration shall include at least one of the following: The configuration identifier information "ai-ReportConfigID" is reported. The reported configuration identifier is used to represent the identifier information for configuring and reporting AI-related information. It can be a list or a group of lists. The identifier information can be configured by the network. The reporting type "ai-ReportType" indicates the specific form in which AI-related information is reported, which can be periodic or event-triggered. Optionally, it can also be configured as semi-continuous or intermittent. Resource identification information for inference configuration is used to identify the resource information configured when configuring AI inference; in some cases, it can be a set of resources, with each set represented by resource set identification information. The resources can be measurement resources or reporting resources; measurement resources are used when performing the measurement process, and reporting resources are used to transmit reporting information. The resources can be AI-related dedicated resources or non-AI-related dedicated resources. The first identifier information of the inference reporting configuration, "CSI-ReportconfigID", is used to represent the identifier of the reporting configuration configured when configuring AI inference reporting; The associated identifier, also referred to as the "associated ID" in this article, is used to represent network-side information. Here, the associated ID is the identification information of the network information carried by the inference configuration configured on the network side. The reported content "ai-Reportquantity" is used to represent the specific information that the UE needs to report. The content includes at least AI-related information, and optionally, in some cases, it may also include an associated ID; optionally, in some cases, it may also include identification information of the inference reporting configuration and resource identification information of the inference configuration.

[0037] Optionally, in some cases, the results of the inference configuration may also be included, such as the predicted beam information; Optionally, in some cases, bandwidth information for inference configuration, such as BWP-ID, may also be included.

[0038] Option 2: Enhance CSI-ReportConfig.

[0039] In this solution, a new AI-specific CSI-reportConfig configuration can be introduced, or configuration information for reporting AI-related information can be introduced into CSI-Reporconfig.

[0040] AI-related reporting configuration; the aforementioned reporting configuration information, namely the first configuration information, is used to represent the reporting information for configuring AI-related information, assuming it is named "aiInfo-CSI-ReportConfig".

[0041] The first configuration information includes at least one of the following: reporting configuration identifier information, reporting type, resource identifier information of inference configuration, first identifier information of inference reporting configuration, association identifier, and reporting content. Optionally, in some cases, the first configuration information may also include the result information of the inference configuration, such as the predicted beam information, etc. Optionally, in some cases, the first configuration information may also include bandwidth information for inference configuration, such as BWP-ID.

[0042] Here's a configuration example: In the AI-specific CSI-reportConfig configuration, define the reporting content, which can be any AI-related information. The reporting content can also include an associated ID, which is the network identification information carried in the inference configuration configured on the network side. When the UE determines the applicable functionality according to the rules, or determines that the conditions for reporting applicable functionality are met, the UE reports the applicable functionality to the NW via an L3 message. When the NW receives the applicable functionality and the associated ID, it can determine which inference corresponds to the applicable functionality based on the associated ID. Here's another configuration example: aiInfo-CSI-ReportConfig::=SEQUENCE { ai-ReportConfigIDAI-ReportConfigIDoptional ai-ReportTypeAI-ReportTypeoptional ai-Reportquantity Ai-Reportquantityoptional Ai-ReportquantityCHOICE { noneNULL, applicable functionalityoptional associated IDoptional CSI-ReportconfigID optional }, Option 3: Enhance reportQuantity In this solution, the reporting volume can be increased, such as by introducing a new AI-related reporting volume, such as "ai-reportQuantity", and defining its related content, or by adding new reporting content to reportQuantity. The reporting content includes at least one of the following information: AI-related information, resource identification information of inference configuration, first identification information of inference reporting configuration, and association identification.

[0043] Optionally, in some cases, the reported content may also include the results of the inference configuration, such as the predicted beam information.

[0044] Optionally, in some cases, the reported content may also include bandwidth information of the inference configuration, such as BWP-ID.

[0045] The above content and definitions are the same as those in Scheme 1; In particular, in some cases, a reporting type can be defined, which may be the same as or different from the inference configuration.

[0046] Option 4: Enhance reportQuantity In this solution, the reporting volume can be increased, such as by adding configuration information for AI-related messages to reportQuantity. The information includes at least one of the following: measurement resource information dedicated to AI-related information, and / or its identification information, reporting configuration identification information "ai-ReportConfigID", and AI-related information.

[0047] Here is a configuration example: In some cases, the inference report content may also include ai-ReportConfigID, which is the reporting configuration identifier of the AI-related applicable functionalities configured on the network side; when the UE reports the inference result to the NW according to the rules, when the NW receives the inference result and ai-ReportConfigID, it can know which one or more applicable functionalities the inference corresponds to based on ai-ReportConfigID. CSI-ReportConfigforAIinference ::=SEQUENCE { reportConfigIdCSI-ReportConfigId, carrierServCellIndexOPTIONAL,-- Need S resourcesForChannelMeasurementCSI-ResourceConfigId, reportQuantity for AI inferenceCHOICE { Predicted beam index OPTIONAL, ai-ReportConfigIDOPTIONAL ...... }, Specifically, in some cases, such as in each periodic reporting configuration, a unique PUCCH resource index (pucch-ResourceIndicator) is assigned in the RRC signaling. The network side directly maps the PUCCH resource location (time slot, symbol, PRB, etc.) of the received report to the corresponding inference, such as the reportConfigId configured in the inference.

[0048] Optionally, the configuration does not require configuring time-frequency resources for reporting AI-related information. In some examples, however, the configuration may include time-frequency resources for reporting AI-related information. Optionally, the configuration can also be associated with time-frequency resources for inference / measurement corresponding to AI Inference, such as those included in the CSI-ReportConfigforAIfunctionality configuration; Optionally, the configuration can also be associated with the time-frequency resources corresponding to the AI ​​Inference result reporting, such as those included in the CSI-ReportConfigforAIfunctionality configuration; Optionally, the configuration can also further configure the reporting type of AI-related information, such as periodic reporting, event-triggered reporting, etc.

[0049] In particular, any one or more of the above schemes can be combined.

[0050] For better understanding, this application also provides a communication method applicable to schemes 1 to 3 above. Please refer to [link / reference]. Figure 2a This includes, but is not limited to, at least one of the following steps: Step 1: The UE receives the inference configuration information sent by the base station. The inference configuration information is configured through CSI-ReportConfig and carries at least one of the following information: association identifier, reporting content, or reporting type, etc. The UE can determine the AI-related information it needs to report, such as applicable functionality, based on the received inference configuration information.

[0051] Step 2: The UE reports applicable functionality to the base station via L3 signaling (such as RRC message). The applicable functionality can be one or more, which is optional. It can also carry an association identifier and / or infer and report the configured first identifier information, etc. Step 3: Optionally, the base station sends an activation message to the UE to activate inference, or activate functionality, or activate the inference corresponding to the functionality. The activation message can be MAC CE or DCI. Optionally, the activation message can carry functionality information, such as functionality ID. Step 4: The UE reports the inference results via PHY messages.

[0052] After receiving the inference configuration information sent by the base station, the UE reports the inference result based on the inference configuration information. For example, the PHY message can be a UCI message.

[0053] This application also provides a communication method applicable to scheme 4 above; please refer to [link / reference]. Figure 2b This includes, but is not limited to, at least one of the following steps: Step 1: The UE receives the inference configuration information sent by the base station. The inference configuration information is configured through CSI-ReportConfig and carries at least one of the following information: association identifier, reporting content, or reporting type, etc. The UE can determine the AI-related information it needs to report, such as applicable functionality, based on the received inference configuration information.

[0054] Step 2: The UE reports applicable functionality to the base station via L3 signaling (such as RRC message). The applicable functionality can be one or more. Step 3: The UE reports the inference results via PHY messages.

[0055] After receiving the inference configuration information sent by the base station, the UE reports the inference result based on the inference configuration information. The inference result information may carry the reporting configuration identifier information (ai-ReportConfigID) when reporting AI-related information. For example, this PHY message can be a UCI message.

[0056] Case 2: When AI-related information equals the inference configuration.

[0057] In some cases, certain AI-related information (such as applicable functionality) is equivalent to inference configuration; or, one of the factors determining AI-related information is inference configuration. In such cases, the UE may report AI-related information in the following two possible ways: Method 1: The UE receives the inference configuration, which may include one of the following information: association ID, resource information, reporting content, or reporting method. The UE parses the received information and, based on the information received and / or some internal information and / or available model information, determines AI-related information, such as applicable functionality, and reports it, specifically according to the reporting message format. The NW receives the AI-related information sent by the UE and can perform related operations, such as activating the information, or the inference configuration corresponding to the information, or the inference configuration itself.

[0058] Method 2: The UE receives the inference configuration, which may include one of the following information: association ID, resource information, reporting content, or reporting method. The UE parses the received information and, based on the received information and / or some internal information and / or available model information, determines the AI-related information, such as applicable functionality. In this method, it can be reported to the NW along with inference reporting messages, such as UCI messages. The reporting type of the AI-related messages is consistent with the reporting type in the inference reporting configuration. In this method, in some cases, if the inference configuration is configured as intermittent or semi-persistent, the UE needs to receive an activation message from the NW before it can report. The NW, upon receiving the information from the UE, can determine which AI-related information is currently being used.

[0059] In this approach, the inference report can include AI-related information.

[0060] B. Based on the combination of L3 architecture and CSI architecture, configure inference configuration information and primary configuration information.

[0061] Given that the L3 architecture includes RRC Reconfiguration, UAI, or RRM, it can be further discussed in the following cases.

[0062] B1.RRC Reconfiguration and CSI Architecture In this scheme, the network side, such as the base station, carries the first configuration information in the RRC Reconfiguration message, that is, the configuration information reported by AI-related information, and then configures the inference configuration information InferenceConfiguration through CSI-ReportConfig, where Inference Configuration carries the association identifier.

[0063] Correspondingly, after receiving the first configuration information and the inference configuration information, the UE can report the inference results to the base station based on L1 signaling and report AI-related information to the base station based on L3 signaling. Optionally, the UE can also report AI-related information to the base station based on L1 signaling, which is not limited in this application.

[0064] Specifically, in this embodiment, a first configuration information is added to the RRCReconfiguration message to represent the reporting information of AI-related information. The first configuration information is exemplarily named "AIInfo-ReportConfig". In practical applications, other naming methods can also be used, and this application does not limit this. The method of adding the first configuration information to the RRCReconfiguration message can be as follows: RRCReconfiguration -radioBearerConfigRadioBearerConfig -measConfigMeasConfig -otherConfigOtherConfig - AIInfo-ReportConfigAIInfo-ReportConfig The first configuration information includes, but is not limited to, at least one of the following: measurement resource information specific to AI-related information, and / or its identification information; AI-related reporting configuration; the AI-related reporting configuration information is used to represent the reporting information configured for AI-related information, assuming it is named "aiInfo-CSI-ReportConfig".

[0065] The AI-related reporting configuration information shall include at least one of the following: reporting configuration identification information, reporting type, resource identification information of inference configuration, first identification information of inference reporting configuration, associated identification, and / or reporting content (which may be referred to as AIInfo-ReportquantityConfig).

[0066] Optionally, in some cases, the first configuration information may also include the result information of the inference configuration, such as the predicted beam information, etc. Optionally, in some cases, the first configuration information may also include bandwidth information for inference configuration, such as BWP-ID.

[0067] For example, to facilitate a better understanding of this solution, the following will provide examples of possible configuration implementations in the first configuration information.

[0068] For example, the first configuration information AIInfo-ReportConfig includes the associated ID, the first identifier information of the inference reporting configuration, also known as CSI-reportConfigId or reportConfigId, and / or the reporting content AIInfo-ReportquantityConfig. When AIInfo-ReportquantityConfig is AI-related information, the possible configuration methods are as follows: AIInfo-ReportConfig::=SEQUENCE { aiInforeportConfigIdAIInfo-ReportConfigId,OPTIONAL, associated IDAssociatedId,OPTIONAL, reportConfigIdCSI-ReportConfigId, OPTIONAL, aiInfo-reportquantityConfigCHOICE { noneNULL, AIInforeportConfigCHOICE{ applicable functionalityOPTIONAL, applicable model OPTIONAL, ..., ...} } Alternatively, the first configuration information AIInfo-ReportConfig includes AIInfo-ReportConfigID, the first identifier information of the inference reporting configuration, and / or AI-related information reporting content. The reporting content includes AI-related information and may also include associated ID. Its configuration method can be as follows: AIInfo-ReportConfig::=SEQUENCE { aiInforeportConfigIdAIInfo-ReportConfigId,OPTIONAL reportConfigIdCSI-ReportConfigId,OPTIONAL, aiInfo-quantityConfigCHOICE { noneNULL, AIInfo-quantityConfigCHOICE{ applicable functionalityOPTIONAL, applicable model OPTIONAL, ...} associated IDAssociatedId, ...} . } The above provides an example of how to configure the first configuration information. Please refer to [link / reference]. Figure 2c The flowchart illustrates a possible communication method based on CSI Reconfiguration and CSI architecture, as provided in this application embodiment, including: Step 1: The UE receives an RRC reconfiguration message sent by the base station. The RRC reconfiguration message carries inference configuration information, which is configured by CSI-ReportConfig. It carries associated information, such as associated ID. The RRC reconfiguration message also carries first configuration information, which is the configuration information of AI-related information. It can be configured through newly defined information elements. The first configuration information includes at least one of the following: reporting content, reporting type of AI-related information, and first identifier information of inference reporting configuration, etc. Step 2: The UE reports AI-related information via L3 messages (such as RRC messages), such as applicablefunctionality, associated identifiers, and / or the first identifier information configured for inference reporting, where applicablefunctionality can be one or more.

[0069] Specifically, the UE receives inference configuration information, such as associated ID, sent by the base station, and can determine the AI-related information it needs to report based on the received information. In addition, the UE can use the first configuration information sent by the base station to judge or measure whether the reporting conditions for AI-related information are met. If the reporting conditions are met, the UE reports AI-related information, such as applicable functionalities, through RRC messages. Step 3: Optionally, the UE receives an Inference activation message sent by the base station to activate the AI ​​inference function; The Inference activation message can be a MAC CE or a DCI message. Optionally, the Inference activation message may carry information about one or more applicable functionalities, or priority information of applicable functionalities; optionally, the activation message may also activate one or more applicable functionalities simultaneously.

[0070] Step 4: The UE reports the inference result to the base station, which can be carried in the PHY message.

[0071] B2: UAI architecture + CSI architecture When based on the UAI architecture + CSI architecture, the base station can report the first configuration information, i.e. AI-related information, through the UAI framework, and configure the Inference Configuration through CSI-ReportConfig. In other words, CSI-ReportConfig refers to Inference Configuration, where Inference Configuration carries associated ID information.

[0072] Correspondingly, after receiving the first configuration information and the inference configuration information, the UE uses L1 signaling to report the inference results to the base station and uses L3 signaling to report AI-related information to the base station. In particular, under certain circumstances, L1 signaling can also be used to report AI-related information to the base station.

[0073] Specifically, the base station can add first configuration information to the OtherConfig message. Specifically, otherConfig introduces a newly defined element, exemplarily named AI-related information reporting configuration AIInfo-ReportConfig, indicating that the UE can / is capable of reporting AI-related information. This element AIInfo-ReportConfig can have multiple representations, including but not limited to one of the following: Form 1: A simple on / off switch, such as AIInfo-ReportConfigENUMERATED {true}. If it exists, it means enabled; if it does not exist, it may be disabled by default. Format 2: A parameter-carrying format, such as `aiInfo-ReportConfigSetupRelease {AIInfo-ReportConfig}`. For example, when set to `Setup`, it indicates that the network side enables the reporting of AI-related information, requiring the UE to report AI-related information. If the AI-related information is set to `Release`, the function of reporting AI-related information is released or disabled. Possible configuration methods are as follows: RRCReconfiguration -otherConfigOtherConfig ... - OtherConfig ::=SEQUENCE { ul-GapFR2-PreferenceConfig-r17ENUMERATED {true}OPTIONAL, -- Need R musim-GapAssistanceConfig-r17SetupRelease {MUSIM-GapAssistanceConfig-r17}OPTIONAL, -- Need M musim-LeaveAssistanceConfig-r17SetupRelease {MUSIM-LeaveAssistanceConfig-r17}OPTIONAL, -- Need M aiInfo-ReportConfigAIInfo-ReportConfigOPTIONAL, Optionally, the first configuration information may also include the reporting frequency of AI-related information, which can be represented by `aiInfo-ReportProhibitTimer`. This element can be defined as an enumeration type, including specific numerical values, to configure the time interval at which the UE is prohibited from reporting AI-related information again after it has already reported it. This can prevent the UE from frequently reporting AI-related information, reduce signaling overhead, and protect network resources.

[0074] Optionally, the first configuration information may also include, but is not limited to, at least one of the following: reporting content, which may be AI-related information or types of AI-related information; optionally, it may be represented in the form of enumeration, inference configuration resource identification information, inference reporting configuration first identification information, and / or association identification.

[0075] For example, in this embodiment, the configuration method of the first configuration information can be as follows, wherein the first configuration information is carried in OtherConfig, including aIInfo-ReportConfig, aIInfo-ReportProhibitTimer, reporting type / aIInforeportinfoType, associated ID, and / or reportConfigId.

[0076] RRCReconfiguration { otherConfig::=SEQUENCE { aIInfo-ReportConfigSetupRelease {AIInfo-ReportConfig:}OPTIONAL, AIInfo-ReportConfig ::= SEQUENCE { aIInfo-ReportProhibitTimerENUMERATED {s0, s1, s2, s5, s10, s20, s30, s60, s90, s120, s300, s600, spare3, spare2, spare1} aIInfo-reportinfoTypeENUMERATED {applicable functionality, applicablemodel non-applicable functionality, spare1} ....} associated IDAssociatedId,OPTIONAL, reportConfigIdCSI-ReportConfigId, OPTIONAL, } The above provides an example of how to configure the first configuration information. Please refer to [link / reference]. Figure 2d The flowchart illustrates a possible communication method based on UAI and CSI architectures, as provided in this application embodiment. Step 1: The UE receives the first configuration information and inference configuration information sent by the base station; The first configuration information is configured by otherConfig, which includes, but is not limited to, aIInfo-ReportConfig, aIInfo-ReportProhibitTimer, associated ID, and / or reportConfigId, and inference configuration information. Configured by CSI-ReportConfig, which carries network-side information such as the associated ID; Step 2: The UE reports AI-related information via RRC messages, such as applicable functionality, associated representation and / or the first identifier information (reportConfigID) of the inference reporting configuration, where applicable functionality can be one or more.

[0077] Specifically, the UE receives inference configuration information, such as associated ID, from the base station and can determine the AI-related information it needs to report based on the received information. In addition, the UE can use the first configuration information sent by the base station to judge or measure whether the reporting conditions for AI-related information are met. If the reporting conditions are met, the UE reports AI-related information, such as applicable functionalities, through an RRC message. It should be noted that the UE can also report AI-related information through a UAI message. This UAI message can carry AIInfo-reportType, and optionally, it can also carry associated ID and / or reportConfig ID. The specifics are not limited here.

[0078] Step 3: Optionally, the UE receives an Inference activation message sent by the base station to activate the AI ​​inference function; The Inference activation message can be a MAC CE or a DCI message. Optionally, the Inference activation message may carry information about one or more applicable functionalities, or priority information of applicable functionalities; optionally, the activation message may also activate one or more applicable functionalities simultaneously.

[0079] Step 4: The UE reports the inference result to the base station, which can be carried in the PHY message.

[0080] B3: RRM framework + CSI architecture When based on RRM and CSI architectures, the base station can configure the first configuration information, i.e., AI-related information, through the RRM framework and report it, and configure the Inference Configuration through CSI-ReportConfig. The InferenceConfiguration can carry associated ID information.

[0081] Correspondingly, after receiving the first configuration information and the inference configuration information, the UE uses L1 signaling to report the inference results to the base station and uses L3 signaling to report AI-related information to the base station.

[0082] In the RRM architecture, the base station can add first configuration information, referred to as AIInfo-ReportConfig, to the MeasConfig message. For example, adding AIInfo-ReportConfig to the MeasConfig message can be done as follows: MeasConfig ::=SEQUENCE { measObjectToRemoveListMeasObjectToRemoveListOPTIONAL,-- Need N measObjectToAddModListMeasObjectToAddModListOPTIONAL,-- Need N reportConfigToRemoveListReportConfigToRemoveListOPTIONAL,-- Need N reportConfigToAddModListReportConfigToAddModListOPTIONAL,-- Need N measIdToRemoveListMeasIdToRemoveListOPTIONAL,-- Need N aiInfo-ReportConfigAIInfo-ReportConfigOPTIONAL,-- Need N The first configuration information may also include at least one of the following: reporting configuration identifier "ai-ReprotConfigID", reporting type "ai-ReportType", reporting content "ai-Reprotquantity", associated identifier "associated ID", resource identifier information of inference configuration, and / or first identifier information of inference reporting configuration.

[0083] Optionally, when configuring multiple associated IDs or multiple reportConfigIds, the first configuration information may also include AI-related configuration second association information, or AI ID#2, which is used to associate AI-related information with associated IDs and / or reportConfigIds. The AI ​​ID#2 is used to manage multiple sets of configurations and reports.

[0084] Optionally, in some cases, the first configuration information may also include the result information of the inference configuration, such as the predicted beam information, etc. Optionally, in some cases, the first configuration information may also include bandwidth information for inference configuration, such as BWP-ID.

[0085] For example, in this embodiment, the configuration method of the first configuration information can be as follows, wherein the first configuration information is carried in MeasConfig, including aIInfoReportConfigID, associated ID, reportConfigId, and aiInfo-quantityConfig, wherein AI-related information is contained in aiInfo-quantityConfig.

[0086] MeasConfig ::=SEQUENCE { aIInforeportConfigIdAIInfo-ReportConfigId,OPTIONAL, associated IDAssociatedId,OPTIONAL, reportConfigIdCSI-ReportCIInfo-reportConfig aiInfo-quantityConfigAiInfo-QuantityConfig, OPTIONAL aIIdListAIIDList, OPTIONAL --------------------------------------------------------------------- AiInfo-QuantityConfigCHOICE { noneNULL, AIInforeportConfigCHOICE{ applicable functionalityOPTIONAL, applicable modelOPTIONAL, ...} ---------------------------------------------------------------------AIInfo-reportConfig::=SEQUENCE { reportTypeCHOICE { periodicalPeriodicalReportConfig, eventTriggeredEventTriggerConfig, ...} aiInfo-QuantityConfigOPTIONAL, ... } It should be noted that the content included in the first configuration information in scenarios B1, B2, and B3 above is exemplary. In actual applications, different configurations can be made according to specific circumstances. Therefore, this application does not limit the content of the first configuration information in each scenario. Optionally, in the above three schemes, the first configuration information may also include the identification information of the measurement resource. The identification signaling of the measurement resource implicitly associates with the Associated ID or reportconfigID, so that the network can further determine which inference configuration corresponds to the reported AI-related information based on the identification information.

[0087] C. Based on the RRM architecture, configure inference configuration information and primary configuration information.

[0088] In this embodiment, the existing RRM framework mechanism can be enhanced to associate AI-related information (such as applicable functionality) reported by the UE with Inference, enabling the network side, such as the NW, to distinguish which Inference Configuration corresponds to one or more reported AI-related information, especially when multiple Inference Configurations are configured. Furthermore, the NW can also manage AI-related operations more effectively, such as activating functionality, switching functionality, or reselecting functionality, thereby improving NW management efficiency.

[0089] Specifically, it can be divided into two cases, including Case 1c and Case 2c, as detailed below: Case 1c: When AI-related information is not configured for inference.

[0090] In this scheme, configuration information for AI-related information reporting is introduced into MeasConfig. This configuration information, denoted as "aiInfo-measConfig", indicates how to configure AI-related information reporting and / or its content. Furthermore, the base station configures AI inference or related parameters through MeasConfig, and this configuration may carry associated ID information. The UE uses L3 signaling (such as a measurement report message) to report inference results to the base station, and also uses L3 signaling to report AI-related information to the base station. Specifically, in some cases, L1 signaling can also be used to report AI-related information to the base station. See the specific scheme below: Option 1: Enhance Measconfig.

[0091] In this solution, a new AI-specific measconfig configuration can be introduced, or a configuration for reporting AI-related information can be introduced into the measconfig. Possible configurations include: Format 1: AIInfo-measConfigSEQUENCE { aiInfo-measobjectAiInfo-measobjectOPTIONAL aiInfo-reprotConfigAIInfo-reprotConfigOPTIONAL, ... Form 2: MeasConfig ::=SEQUENCE { measObjectToAddModListMeasObjectToAddModListOPTIONAL,-- Need NreportConfigToAddModListReportConfigToAddModListOPTIONAL,-- Need N measIdToAddModListMeasIdToAddModListOPTIONAL,-- Need N aiInfo-measConfigAIInfo-measConfigOPTIONAL,-- Need N ... The configuration information includes any one or a combination of the following: The measurement object information of AI-related information can be used to represent the granularity of AI-related information, i.e., in which granularity / range it is available. The measurement object information can be UE identifier, cell identifier, cell list, area information identifier, AI model information, inference configuration signaling identifier such as RRC ID, base station identifier information, AMF identifier information, PLMN information identifier, AI task identifier, third-party server identifier, AI management network element identifier, and / or AI service identifier. The configuration for reporting AI-related information; the aforementioned configuration information is used to represent the configuration of AI-related information reporting information, assuming it is named "aiInfo-ReprotConfig"; The first associated information in the AI-related configuration is “AI ID#1”. The AI ​​ID#1 is used to associate the measurement object information related to AI and the AI-related reporting configuration. When multiple sets of AI-related measurement objects and multiple sets of AI-related reporting configurations are configured, they can be associated through AI ID#1 so that the network can distinguish them. The AI ​​ID can be configured on the network side. The configuration identifier information "ai-ReprotConfigID" is reported. The reported configuration identifier is used to represent the identifier information of the configuration reporting of AI-related information. It can be a list or a group of lists. The identifier information can be configured by the network. The reported content (aiInfo-quantityReport) refers to the specific AI-related information indicators that the UE needs to pay attention to in the process of determining AI-related parameters, such as applicable functionality or applicable model. The reported content can be any one or more AI-related information. The reporting type identifier "ai-ReprotType" indicates the specific form of the AI-related information being reported, which can be periodic or event-triggered. Optionally, it can also be configured as semi-continuous or intermittent. The associated identifier "associated ID" is network-side information identification information, which can be carried through inference configuration or other information. The measurement identifier "measID" refers to a measurement object (such as a frequency band or neighboring cell) and a reporting configuration (such as reporting conditions and triggering events) configured on the network side. The measurement object (such as a frequency band or neighboring cell) and the reporting configuration mentioned here are related to AI, and can be further described as an AI-related measurement object or reporting configuration. The identification information of the measurement object in the inference, such as "MeasObjectId", can represent: a unique identifier of the measurement object configured on the network side. The measurement object can also be further understood as an object that performs AI-related operations. The measurement object can be an AI-related measurement object or a non-AI-related measurement object. The second identifier information of the inference reporting configuration, such as "ReportConfigtId", is used to distinguish different report configurations. The report configuration can also be understood as the reporting configuration for performing AI-related operations. The report configuration can be an AI-related report configuration or a non-AI-related report configuration.

[0092] Optionally, in some cases, it may include resource information for measurement configuration. The resource representation may be the identification information of the measurement resources configured for inference configuration on the network side. The measurement resources may be AI-related dedicated resources or non-AI-related dedicated resources. Optionally, in some cases, the configuration information may also include resource identification information reported by the measurement. The resource identification information may be the resource identification information when the network side configures the inference configuration reporting information. The resource may be AI-related dedicated resource or non-AI-related dedicated resource.

[0093] Optionally, in some cases, the results of the inference configuration may also be included, such as the predicted RSRP. Optionally, in some cases, it may also include cell information, beam information, etc., for inference configuration.

[0094] Method 1: AIInfo-measConfig contains associated ID and / or measID, and the reported content is AI-related information. An example is as follows: aIInfo-MeasConfig::=SEQUENCE { associated IDAssociatedId,OPTIONAL, measIDMeasId,OPTIONAL measObjectIdMeasObjectId,OPTIONAL, reportConfigIdReportConfigId, OPTIONAL, aI-quantityConfigAI-QuantityConfig, OPTIONAL ... AI-QuantityConfigCHOICE { noneNULL, applicable functionalityOPTIONAL, applicable model OPTIONAL, ..., } Method 2: AIInfo-measConfig includes ainfo-measconfig, aIInfo-reprotConfig, and / or reports containing AI-related information, and may also include associated IDs, etc. An example configuration is shown below: aIInfo-MeasConfig::=SEQUENCE { aiInfo-measObjectAIInfo-MeasObject,OPTIONAL, aIInfo-reprotConfigAIInfo-reprotConfigOPTIONAL, aI-quantityConfigAI-QuantityConfig, OPTIONALAI-QuantityConfigCHOICE { noneNULL, applicable functionalityOPTIONAL, applicable model OPTIONAL, ..., associated IDAssociatedId,OPTIONAL, measIDMeasId,OPTIONAL measObjectIdMeasObjectId,OPTIONAL, reportConfigIdReportConfigId, OPTIONAL, ...} . } Option 2: Enhance ReportConfig.

[0095] In this solution, a new AI-specific `reportConfig` configuration can be introduced, or configuration information for reporting AI-related information can be imported into `ReportConfig`. The aforementioned reporting configuration information is used to represent the reporting information configured for AI-related information, assuming it is named "aiInfo-ReportConfig". The reporting configuration information includes at least, but is not limited to, at least one of the following: reporting configuration identifier, reporting type identifier, measurement identifier "measID", inference report configuration identifier, inference report configuration identifier, association identifier, and / or reporting content. Optionally, in some cases, the reported configuration information may also include the results of the inference configuration, such as the predicted RSRP. Optionally, in some cases, the reported configuration information may also include cell information, beam information, etc., for inference configuration. In particular, under certain circumstances, the reported configuration information can also carry resource identification information for inferred configuration.

[0096] Option 3: Enhance reportQuantity.

[0097] In this solution, the reported content can be enhanced, such as by introducing a new AI-related reported content, such as "ai-reportQuantity", and defining its related content, or by adding new reported content to reportQuantity. The reported content includes, but is not limited to, at least one of the following: AI-related information, measurement identifier "measID", inference report configuration identifier information, inference report configuration identifier information, and association identifier.

[0098] Optionally, in some cases, the reported content may also include the results of the inference configuration, such as the predicted RSRP; Optionally, in some cases, the reported content may also include cell information, beam information, etc., related to inference configuration.

[0099] Optionally, in some cases, resource identification information for inference configuration may also be carried.

[0100] Case 2c: When AI-related information is configured for inference.

[0101] Method 1: The UE receives the inference configuration, which may include one of the following information: association ID, resource information, reporting content, or reporting method. The UE parses the received information and, based on the received information and / or some internal information and / or available model information, determines AI-related information, such as applicable functionality, and reports it, specifically according to the reporting message format. The NW receives the AI-related information sent by the UE and can perform related operations, such as activating the information, or the inference configuration corresponding to the information, or the inference configuration itself. Method 2: The UE receives the inference configuration, which may include one of the following information: association ID, resource information, reporting content, or reporting method. The UE parses the received information and, based on the received information and / or some internal information and / or available model information, determines the AI-related information, such as applicable functionality. In this method, it can be reported to the NW along with inference reporting messages, such as RRC messages. The reporting type of the AI-related messages is consistent with the reporting type in the inference reporting configuration. In this method, in some cases, activation is not required; reporting can only proceed after the configured reporting conditions are met. The NW, upon receiving the information from the UE, can determine which AI-related information is currently being used.

[0102] It should be noted that in this scheme, the content reported through reasoning can include AI-related information.

[0103] D. Based on the UAI architecture and RRM architecture, configure inference configuration information and first configuration information.

[0104] In this embodiment, by combining two levels of signaling, the NW can distinguish which inference configuration corresponds to one or more AI-related information reported. Specifically, by enhancing the existing UAI architecture and designing new elements or indication information, the AI-related information (such as applicable functionality) reported by the UE is associated with the inference configuration configured in the RRM architecture. This enables the NW to manage AI-related operations more effectively, such as activating functionality, switching functionality, reselecting functionality, activating inference, and starting performance monitoring, thereby improving NW management efficiency.

[0105] Specifically, in this scheme, the base station adds first configuration information to OtherConfig, namely configuration information related to AI. Specifically, a newly defined element, assuming "AIInfo-ReportConfig", is introduced into otherConfig, indicating that the UE can / is capable of reporting AI-related information. It should be noted that the representation, configuration method, and configuration content of this element AIInfo-ReportConfig are similar to the relevant description in sub-scheme B2 of Scheme B above, and will not be repeated here. Optionally, the first configuration information may include, but is not limited to, at least one of the following: the reporting frequency of AI-related information, the reporting content, the association identifier, the measurement identifier "measID", the identifier information of the measurement object in the inference, the second identifier information of the inference report configuration, the resource information of the measurement configuration, the resource identifier information of the measurement report, the result information of the inference configuration, such as the predicted RSRP, and / or the cell information, beam information, or spectrum information of the inference configuration; wherein, the cell information can be the serving cell or the neighboring cell. The reporting frequency of AI-related information can be represented by "aiInfo-ReprotProhibitTimer", which can be an enumerated type that configures the time interval at which the UE is prohibited from reporting AI-related information again after reporting it. Its main function is to avoid the UE frequently reporting AI-related information, reduce signaling overhead, and protect network resources.

[0106] In this scheme, by combining two levels of signaling, the network side, such as the NW, can distinguish the inference configuration information corresponding to one or more AI-related information reported.

[0107] 203. UE determines AI-related information; After receiving the first configuration information and / or inference configuration information, the UE determines AI-related information based on the first configuration information and / or inference configuration information and / or other information.

[0108] In this step, the following issues need to be further considered: Issue a: In some scenarios, a configuration message can contain multiple associated IDs. Therefore, a single configuration will generate AI-related information that satisfies different network-side conditions. For example, in the current BM inference configuration, if the configurations set A and set B are associated with two different associated IDs, how does the UE determine the applicable functionalities to be reported based on the associated IDs? Issue b: When the UE determines that multiple applicable functionalities exist after decision evaluation, how does it determine the final activated applicable functionalities and the applicable functionalities used? These two issues involve how to determine the final applicable functionality in different scenarios; therefore, relevant determination rules can be further designed. In this embodiment, a relevant solution is provided, which enables the UE to clearly define and select the AI-related information to be activated or used, such as applicable functionality, based on rules. This can reduce the occurrence of abnormal situations and improve the stability and reliability of the network. Furthermore, it can also reduce the number of applicable functionality activations, thereby reducing signaling overhead and UE power consumption.

[0109] Regarding the situation in question a where the configuration corresponds to multiple applicable functionalities, this embodiment provides the following solutions, including but not limited to: In the specific BM inference configuration, the following methods can be used: 1. The UE determines applicable functionalities based on the associated ID in set A; 2. The UE determines applicable functionalities based on the associated ID in set B; 3. The UE determines applicable functionalities separately based on the different associated IDs in set A and set B, and these applicable functionalities are all considered as the applicable functionalities ultimately determined by the UE; 4. The UE determines the corresponding applicable functionalities separately based on the different associated IDs in set A and set B, and the overlapping applicable functionalities among these applicable functionalities can be considered as the applicable functionalities ultimately determined by the UE.

[0110] In addition, the following methods can be used to achieve the above-mentioned UE behaviors and / or operations: Method I: NW configures network condition usage indication information. The NW will configure usage indication information for a network condition (such as an associated ID). This usage indication information instructs the UE which associated ID to use to determine applicable functionalities. The indication information is associated with the associated ID. Optionally, this usage indication information can be represented as a switch. If a configuration carries multiple associated IDs, and the usage indication information corresponding to the associated ID is enabled, it means the UE needs to use this associated ID to determine applicable functionalities.

[0111] Optionally, if the usage indication information corresponding to multiple associated IDs is all in an enabled state, then NW can further indicate or define which applicable functionalities are the final applicable functionalities. This can be done through network-side configuration rules or system default rules to select the final applicable functionalities.

[0112] For example, if the AI-related information enabled for associated ID#1 is #1 and #2, and the AI-related information enabled for associated ID#2 is #1 and #3, then #1, #2, and #3 can all be considered valid and are ultimately applicable AI-related information; or only if both associated ID#1 and associated ID#2's corresponding AI-related information #1 are satisfied, is it considered ultimately applicable AI-related information.

[0113] Method II: Fixed Scheme. The protocol defaults or is fixed so that the UE needs to use the associated ID associated with the AI-related reporting configuration, or the UE needs to use the associated ID associated with the AI-related measurement configuration. The UE determines the applicable functionalities according to the default rules, and the indication information is associated with the associated ID to further determine the applicable functionalities.

[0114] Method III: UE Autonomous Selection. In some cases, the UE can select any associated ID to determine applicable functionalities. When reporting the determined applicable functionalities to the network side, the UE can also include the associated ID.

[0115] Furthermore, regarding the situation in question b where the UE has multiple applicable functionalities, i.e., the UE determines that there are multiple applicable functionalities based on certain decision conditions, such as internal information on the UE side, additional conditions on the network side, and model availability, this application provides several solutions for how the UE can determine which applicable functionality to use based on the determination rules of AI information, including but not limited to: Option 1: Network-side instructions.

[0116] The UE can report AI-related information, such as applicable functionalities, that meets the conditions to the network side. Optionally, it can also report other auxiliary information. The network side selects one or more applicable functionalities based on the reported ones. Further, after selection, an activation message can be sent to the UE to activate the selected applicable functionalities. Specifically, the network side can select the final applicable functionalities through various possible methods or a combination of methods.

[0117] Method 1: The UE reports all applicable functionalities that meet the conditions, and the network randomly selects one or more applicable functionalities; or Method 2: The UE reports all applicable functionalities that meet the conditions, and reports the status information of the applicable functionalities. The network side selects one or more applicable functionalities based on the status information. It should be noted that status information can be used to represent the stability of applicable functionalities. The determination of stability can be based on the fact that it does not change over a period of time; or it can be based on the fact that it does not change under different scenarios; or it can be based on the fact that it does not change under different configurations, etc., and this application does not limit the specifics. The status information can be indicated by level information, such as "high, medium, low" or first level, second level, etc., where high can indicate that the applicable functionality is stable. Each applicable functionality corresponds to a status information, and the network side can select one or more applicable functionalities based on the status information corresponding to the applicable functionality. Optionally, the network side can select the applicable functionality with the highest stability. Method 3: The UE reports all applicable functionalities that meet the conditions, and reports the number of models corresponding to each applicable functionality. For example, applicable functionality#1 corresponds to 2 models, applicable functionality#2 corresponds to 10 models, applicable functionality#3 corresponds to 3 models, and applicable functionality#4 corresponds to 1 model. Since the number of models is related to the device's capabilities, such as computing power and operational power, the network side can select one or more applicable functionalities based on the number of models corresponding to the applicable functionalities and the device's capabilities.

[0118] Method 4: The UE reports all applicable functionalities that meet the conditions and the generalization capability of the applicable functionalities. The network side selects one or more applicable functionalities according to the generalization capability corresponding to the applicable functionalities. The generalization capability means that the applicable functionalities are available for different configurations or scenarios. Optionally, in some scenarios, the generalization capability is equivalent to or similar to the state information.

[0119] Method 5: The UE reports all applicable functionalities that meet the conditions, and reports the device capabilities required to run the applicable functionalities, including computing power information or memory information, as well as the capability information or memory information required by the applicable functionalities, so that the network side can select one or more applicable functionalities.

[0120] For easier understanding, please refer to Figure 2e The flowchart below illustrates a possible communication method provided in this application embodiment, used by the NW to select the final AI-related information based on the AI-related information reported by the UE. The specific steps include the following: Step 0: The UE determines the AI-related information that meets the conditions, which includes applicable functionalities; Step 1: The UE reports AI-related information and / or auxiliary information to the NW. The AI-related information includes at least one applicable functionalities or applicable models. Step 2: NW selects an AI-related information from at least one AI-related information, such as an applicable functionality or an applicable model; Step 3: NW performs AI-related operations based on the selected AI-related information, including function activation and function deactivation.

[0121] Option 2: Network side configuration rules.

[0122] In this scheme, the UE can receive AI information selection rules configured on the network side. These rules instruct the UE to select one or more applicable functionalities from multiple AI-related information sources, such as multiple applicable functionalities, for reporting. This reduces signaling overhead in the UE's reporting messages and conserves uplink resources. The AI ​​information selection rules include, but are not limited to, at least one of the following rules: Rule 1: The UE reports applicable functionalities based on the configured status information. Specifically, the network side configures thresholds or conditions for status information. Only when the applicable functionalities on the UE side meet the thresholds or conditions will the UE report them to the network side. Rule 2: The UE reports applicable functionalities based on the number of models. Specifically, the network side configures a threshold or condition regarding the number of models. Only when the applicable functionalities on the UE side meet the threshold or condition will the UE report them to the network side. For example, the number of models is a positive integer, such as threshold 2. The threshold can be the minimum value (min) of the corresponding number of models, such as 2, or the maximum value (max) of the corresponding number of models, such as 4, or a range, such as N to M models. It should be noted that the condition for the number of models can be models supported by the NW, models supported by the UE, or models jointly supported by the UE and the NW. The specific details are not limited here.

[0123] Rule 3: The UE reports applicable functionalities based on the degree of generalization. Specifically, the network side configures a threshold or condition related to the degree of generalization. Only when the applicable functionalities on the UE side meet the threshold or condition will the UE report them to the network side.

[0124] Rule 4: The UE reports applicable functionalities based on the device capabilities required to run applicable functionalities, such as computing power information or memory information. Specifically, the network side configures thresholds or conditions for general computing power information or memory information. Only when the applicable functionalities on the UE side meet the thresholds or conditions will the UE report them to the network side.

[0125] It should be noted that the selection rules for this AI information can be predefined by the network or fixed by a standard.

[0126] Optionally, in some implementations, the AI-related information selected by the UE according to the AI ​​information selection rules can be activated by default; For easier understanding, please refer to Figure 2f This is a flowchart illustrating a possible communication method provided in this application embodiment, used by the UE to select the final AI-related information for reporting based on the AI ​​information selection rules configured by the NW. The specific steps include the following: Step 0: The UE determines AI-related information, which includes at least one applicable functionality; Step 1: The NW sends configuration information to the UE to configure the selection rules for reporting AI information; Step 2: Based on the selection rules, the UE selects an AI-related information from at least one AI-related information, such as an applicable functionality or an applicable model; Step 3: The UE reports the relevant information of the AI ​​to the NW; Step 4: NW performs AI-related operations based on the selected AI-related information, including function activation and function deactivation.

[0127] Option 3: UE selects independently.

[0128] Optionally, in this embodiment, after the UE determines a series of AI-related information that meets the configuration / conditions, it can also select one or more AI-related information through at least one of the following: its own memory, computing power, power consumption, number of models, resource status, or status information of AI-related information. For example, if there are 10 applicable functionalities that meet the configuration / conditions, the UE selects to report 3 applicable functionalities to the network side after evaluation. After receiving these 3 applicable functionalities, the network side performs related management work.

[0129] 204. The UE sends a reporting message to the network side; After determining the AI-related information to be reported, the UE sends a reporting message to the network side. This reporting message may carry AI-related information and / or auxiliary information. In this application, the type of reporting message can be various, including but not limited to the following formats: Type 1: The reported message is L3 signaling.

[0130] In this embodiment, when the reported message is L3 signaling, it can be any of the following messages: 1. The reported message is an RRC Reconfiguration Complete message. This reported message may carry a reported quantity, which is AI-related information. Optionally, the reported message may also carry at least one of the following information: measurement resource information dedicated to AI-related information, and / or its identification information, association identifier, AI-related configuration association information, reporting configuration identifier information, measurement identifier, identifier information of the measurement object of inference, first identifier information of the inference report configuration, second identifier information of the inference report configuration, AI ID#1, AI ID#2, resource information of measurement configuration, and / or resource information of measurement reporting. 2. The reported message is a UAI message. Specifically, the reported message can be a UAI completion message. The information it carries can be referred to the information that can be carried when the reported message is an RRC reconfiguration completion message, which will not be repeated here.

[0131] 3. The reported message is UEInformationResponse. Specifically, this reported message can be a UEInformationResponse completion message. The information it carries can be found in the information that can be carried when the reported message is an RRC reconfiguration completion message, as described above. It will not be repeated here. 4. Other L3 message: Specifically, the reported message can be a UEInformationResponse completion message, and the information it carries can refer to the information that can be carried when the reported message is an RRC reconfiguration completion message, which will not be repeated here. Taking UAI as an example of the reported message, a possible message structure can be configured as follows, where the reported message includes: associated ID, measID, reportconfigID, and AI information such as Applicable functionality and / or Applicable model.

[0132] UAI ::=SEQUENCE { aiInfoAiInfoOPTIONAL, AssociatedIDCSI-ReportConfigIdOPTIONAL, ReportConfigIDINTEGER (1..32)OPTIONAL, measIDMeasIdOPTIONAL, AiInfo ::=SEQUENCE { Applicable functionalityOPTIONAL Applicable modelOPTIONAL Type 2: The reported message is MAC CE signaling.

[0133] Optionally, in practical applications, the reporting message can also be carried through a MAC CE, such as in scenarios requiring rapid reporting of AI-related information to the network side. The MAC CE includes at least one of the following: measurement resource information specific to AI-related information, and / or its identification information, association identifier, AI-related configuration association information, reporting configuration identifier information, measurement identifier, identifier information of the measurement object for inference, first identifier information of the inference report configuration, second identifier information of the inference report configuration, AI ID#1, AI ID#2, resource information of the measurement configuration, and / or resource information of the measurement reporting; Optionally, it may also carry LCID and / or -LCG; where LCID is used to represent the priority of the logical channel, which can represent the logical channel priority for transmitting AI-related information; LCG is used to represent the priority group identifier of the logical channel, which can represent the logical channel priority group information for transmitting AI-related information; Type 3: The reported message is L1 signaling.

[0134] Optionally, in practical applications, the reported message can also be carried through L1 signaling. For example, in scenarios where it is necessary to quickly report AI-related information to the network side, the L1 signaling can be UCI. The L1 signaling includes, but is not limited to, at least one of the following: measurement resource information dedicated to AI-related information, and / or its identification information, association identifier, AI-related configuration association information, reporting configuration identifier information, measurement identifier, identification information of the measurement object for inference, first identifier information of the inference report configuration, second identifier information of the inference report configuration, AI ID#1, AI ID#2, resource information of measurement configuration, and / or resource information of measurement reporting.

[0135] In addition, in practical applications, there are scenarios where multiple reporting messages (such as L3 messages, MAC CE messages, UCI messages, etc.) can report AI-related content. To achieve flexible reporting, this application also designs a reporting priority rule for reporting messages, which allows different reporting messages to be used in different situations. Specifically, at least one of the following reporting rules can be adopted: Reporting Rule 1: The network side configures the tagging information, or Flag information, of the reported messages. This Flag information indicates to the UE whether to send AI-related information using L3 signaling, MAC CE, or L1 signaling (such as UCI). For example, Flag (indication of L3, indication of MAC CE, indication of L1). This Flag information can be included in the configuration of AI-related information reporting.

[0136] Reporting Rule 2: Network configuration priority rules, and / or predefined priority rules. Optionally, the reported message can be selected based on real-time requirements; for example, if the network urgently needs to report AI-related information, the UE will prioritize reporting messages with high real-time requirements, such as UCI messages or MAC CE messages. Alternatively, the reported message can be selected based on data volume; for example, if the amount of AI-related information to be reported is small, but certain reporting conditions are met, the UE can prioritize reporting UCI messages or MAC CE messages.

[0137] Reporting rule 3: Other default or implicit rules.

[0138] Optionally, the reporting rules are related to the type of AI-related information request message. For example, if the request message is a DCI message, it can be understood that the network urgently needs AI-related information, and a UCI message can be used to reply; if the request message is an RRC message, it can be understood that the network is not sensitive to the latency of the requested AI-related information, and an RRC message can be used to reply; if the request message is a MAC CE message, it can be understood that the network is not sensitive to the latency of the requested AI-related information, and a MAC CE message can be used to reply. Optionally, the reporting rules are related to the reporting type of AI-related information, implicitly representing the reporting type of different messages. For example, if the UE is configured to report messages of a semi-persistent or intermittent type, UCI can be used; if the reporting type of the message is event-triggered, RRC messages are used.

[0139] Optionally, the UE can also dynamically select the type of request message to send. For example, if the amount of AI-related information is relatively small, the UE can preferentially choose to send AI-related information using UCI. In some cases, the UE can also use more than one signaling method to send AI-related information.

[0140] It should be noted that the dynamic selection rule can be network-configured, AI-based, or autonomously selected by the UE based on the current situation. The dynamic selection rule can consider any of the following factors: time-frequency resources, UE power consumption, UE link state, UE energy, the size / scale of AI-related information, and / or the latency requirements of AI-related information, etc. The dynamic selection rule can be composed of one or more of these factors.

[0141] The information mentioned above can be carried in configuration messages related to AI, in request messages, or in other messages; this application does not specify which.

[0142] Additionally, it's important to note that, generally, an RRC reconfiguration complete message is sent only once for the RRC reconfiguration message; there is no periodic, event-triggered, non-continuous, or irregular sending. UAI, however, can be configured for periodic or event-triggered sending. Without mechanism enhancements, there might be overlap between the applicable functionality (1, 2, 3) reported in the RRC reconfiguration complete message and the applicable functionality (1, 2, 3) reported in the UAI, leading to signaling redundancy. Furthermore, if applicable functionality needs updating, the existing mechanism cannot send the updated applicable functionality via the UAI message. This is because the UAI message and the RRCReconfigurationComplete message are two independent messages, and their sending has no temporal relationship. The UAI message can be sent at any time, such as before or after the RRCReconfigurationComplete message, depending on the UE's triggering conditions and network requirements.

[0143] Therefore, regarding the mechanism currently under discussion, for the reporting of applicable functionality, this application can adopt a method where the initial report of applicable functionality uses a first-type message, such as an RRC reconfiguration completion message, and subsequent update reports of applicable functionality use a second-type message, such as a UAI message. This application provides several methods for this purpose, including through specific configurations and / or triggering conditions, enabling the UE to first send initial AI-related information to the base station via a first-type message, and then send updated AI-related information to the base station via a second-type message. This reduces excessive signaling overhead caused by the UE reporting duplicate information and improves the efficiency of UE message reporting. For ease of description, this application uses the RRC reconfiguration completion message as the first-type message and the UAI message as the second-type message as an example. It should be noted that in practical applications, the first and second types of messages can also be other messages. This application mainly aims to implement how to use different types of messages to separately carry the initial and update reports of the reported message, thereby solving the problems existing in the prior art. Specifically, this includes the following methods: Method A: Based on displayed instructions.

[0144] A new indication message, the signaling reporting rule indication, or Indication for aiInforeporting rule, is defined. This indication tells the UE that when reporting AI-related information, it must first use the RRCReconfigurationComplete message to carry the AI-related information. After the RRCReconfigurationComplete message is sent, the updated AI-related information is then sent via the UAI message. This signaling reporting rule indication can be configured in the RRCReconfiguration message.

[0145] It should be noted that the signaling reporting rule indication can be represented in the following ways: 1. Using a Boolean value (e.g., True or False) to indicate whether the signaling reporting rule indication is configured for the UE. For example, if the signaling reporting rule indication is set to False, it means that the function of the signaling reporting rule indication is not enabled; if the signaling reporting rule indication is set to True, it means that the function of the signaling reporting rule indication is enabled; 2. Using a bitmap to indicate whether the signaling reporting rule indication is configured for the UE. Assuming an N-bit indication is used, if N=1, the value of the indication information is 0 or 1; for example, if the value of the signaling reporting rule indication is set to 0, it means that the function of the signaling reporting rule indication is not enabled; if the value of the signaling reporting rule indication is set to 1, it means that the function of the signaling reporting rule indication is enabled.

[0146] It should be noted that the indication method of this signaling reporting rule is optional and can be configured to save signaling overhead. Optionally, the signaling reporting rules indicate a correlation with AI-related information, and this correlation can be implemented through the same IE configuration below. An example configuration is as follows: aIInfo-ReportConfigSetupRelease {AIInfo-ReportConfig:}OPTIONAL, AIInfo-ReportConfig ::= SEQUENCE { aIInfo-ReportProhibitTimerENUMERATED {s0, s1, s2, s5, s10, s20, s30, s60, s90, s120, s300, s600, spare3, spare2, spare1} OPTIONAL, aIInfo-reportTypeENUMERATED {applicable functionality, applicablemodel non-applicable functionality, spare1} OPTIONAL, Indication for aiInfo reporting rule,OPTIONAL, } Option B: Based on time information or time window.

[0147] The base station configures a time information or time window, also known as Timewindow for aiInfo reporting. The time information or time window is used to instruct the UE to wait for a period of time after completing the configuration before sending the UAI message. Sending the UAI message can also be understood as the UE checking whether its AI-related information meets the reporting conditions after configuring the time information; or the triggering conditions configured for the UAI message are officially started after the time information expires. The time information or time window has the following possible representations: (1) Enumeration method, predefining / configuring a set of discrete data, giving possible options for the time window, for example: Timewindow for aiInfo reporting,ENUMERATED {min1, min5,min10, spare3}; (2) Defining a time range, allowing the time value to change within the defined time interval. For example: Timewindow for aiInfo reporting, ::= INTEGER (1...10). For example, the length of the time window can be 2 to 10 minutes.

[0148] Optionally, the time window information is a parameter dynamically configured by the network, and the length of the time can be dynamically adjusted. The network side configures different values ​​for AI-related information or different AI-related information through RRC signaling.

[0149] For time window information, the start point, duration, bias value, and / or end point of the time window can be further defined; Optionally, the specific length of the time window is expressed in time units, such as milliseconds, microseconds, seconds, minutes, hours, etc., which are not limited in this application.

[0150] Optionally, the start time of the time window is calculated after the UE receives and successfully decodes the message carrying the time window; or after the UE successfully sends the RRC reconfiguration completion message; or for a message that carries the time window but is sent.

[0151] Optionally, the relationship between time-based information or time windows and AI-related information can be established through a single IE configuration. An example configuration is as follows: aIInfo-ReportConfigSetupRelease {AIInfo-ReportConfig:}OPTIONAL, AIInfo-ReportConfig ::= SEQUENCE { aIInfo-ReportProhibitTimerENUMERATED {s0, s1, s2, s5, s10, s20, s30, s60, s90, s120, s300, s600, spare3, spare2, spare1} OPTIONAL, aIInfo-reportTypeENUMERATED {applicable functionality, applicablemodel non-applicable functionality, spare1} OPTIONAL, Timewindow for aiInfo reporting,ENUMERATED {min1, min5, min10,spare3} OPTIONAL, Option C: Based on a timer.

[0152] In this application, a timer “Txx” can also be used to indicate the reporting rule. The timer is used to control the time when the UAI message starts sending updated AI-related information after the UE sends an RRC completion message. The timer can be started when the UE sends an RRC reconfiguration completion message, which carries AI-related information. The timer can be stopped when the RRC reconfiguration message is received again. The reconfiguration message carries configuration information related to AI.

[0153] Timeout behavior: If the timer times out, the UE starts sending UAI messages, or the UE starts evaluating the triggering conditions that would cause the UE to send AI-related information. The value of the timer "Txx" is expressed in time units, such as milliseconds, microseconds, seconds, minutes, hours, etc., such as 100 ms, 200 ms, 300 ms, 400 ms, 600 ms, 1000 ms, etc., without specific limitations here. Furthermore, this value can be fixed in the standard, i.e., predefined, or configured through the network, such as by sending it to the UE via RRC signaling.

[0154] Method D: Network-based control.

[0155] In this embodiment, the UAI configuration can also be activated by sending an activation message from the network side. Specifically, although the UAI configuration can be pre-configured to the network in the RRC reconfiguration message, there are two cases, including: Scenario 1: UAI's aiinfo instruction messages are disabled by default. When the network side receives the RRC reconfiguration completion message, which contains AI-related messages, the network side sends an activation message to activate UAI's aiinfo configuration information. The activation message can be MAC CE or DCI, and can also carry any aiinfo message, or reportconfigID, assicted ID, or measID, etc. Scenario 2: UAI's aiinfo instruction messages are enabled, but no rules are configured to trigger reporting. When the network side receives the RRC reconfiguration completion message, which contains AI-related messages, the network side sends an activation message. This activation message may carry an instruction for triggering aiinfo reporting rules. The activation message can be a MAC CE, DCI, or RRC message.

[0156] Optionally, the protocol may default to requiring the sending of messages via RRC reconfiguration before reporting AI-related information. After the message is sent, UAI configuration will be started by default, and UAI messages will be used to update and report AI-related information.

[0157] It should be noted that the signaling reporting rules can be carried in configuration messages related to AI, in request messages, or in other messages; this application does not specify which ones.

[0158] Therefore, by using the above method, the goal of sending the RRC reconfiguration completion message first and then the UAI message is achieved, that is, the timing relationship between the two messages is designed.

[0159] Furthermore, this application also considers that AI-related information can be reported periodically, in which case applicable functionalities are activated autonomously by the UE. In this embodiment, based on the periodically reported AI-related information, a relevant solution is provided so that the network side can further understand which AI-related information the UE is using, such as applicable functionalities or applicable models; thereby improving the network side's management efficiency of AI-related information, such as switching applicable functionalities and reselecting applicable functionalities. Specifically, this includes the following methods: Method 1: Introduce Flag information to identify the applicable functionalities used on the UE side; In this scheme, a flag is introduced to indicate which active applicable functionalities are performing AI-related operations, such as inference. This flag, along with the corresponding applicable functionalities, is then reported to the network side. Upon receiving this information, the network side can determine which applicable functionalities are currently in use. For example, if there are 10 active applicable functionalities, and the UE or network side selects applicable functionality #1 for inference, then the UE identifies applicable functionality #1 as the applicable functionality currently in use by the UE; the other nine do not need to be identified.

[0160] Optionally, all applicable functionalities that meet the conditions need to be reported; in this scheme, the Flag information can be added by the UE itself, or configured by the NW so that the UE can indicate and report it.

[0161] Method 2: The UE only reports the applicable functionalities used on the UE side; Optionally, the UE can also only report the applicable functionalities used on its UE side. For example, by indexing the relevant information of the applicable functionalities through the applicable functionality index, the signaling overhead of periodically reporting applicable functionalities can be further reduced.

[0162] 205. The network side performs AI-related operations based on AI-related information; 206. The UE reports the inference results to the network side.

[0163] NW can perform AI-related operations (management) based on the reported AI-related information, including activation and deactivation.

[0164] If the UE receives an activation message sent by the network side, the activation message may be to activate AI-related information, such as applicable functionality, or to activate the inference function corresponding to AI-related information, such as inference, or to meet the reporting conditions configured in the inference configuration information. In this case, the UE will report the corresponding inference results based on the activation message.

[0165] It is important to consider that while inference results can be reported using PHY signaling and are sensitive to latency requirements, applicable functionalities generally change slowly and are reported using higher-level signaling (such as L3 signaling). If the inference results deteriorate, for example, due to inaccurate predictions, and applicable functionalities have not yet been reported, or updated applicable functionalities have not been reported to the NW in a timely manner, the NW will be unable to select or switch to new applicable functionalities. This will compromise the continuity of applicable functionalities, leading to interruptions in AI-based operations or a revert to traditional methods. This embodiment provides several solutions to address this issue, ensuring the continuity of AI-related operations based on AI-related information, or improving efficiency by reselecting or switching AI-related information. These solutions include the following: Option A: UE-triggered. Specifically, the following UE-triggered methods can be used: Option 1a: UE actively sends a request message In this scheme, the UE proactively initiates a request message for AI-related information. The request message is used to inform the NW that the currently used AI-related information is inappropriate. The inappropriateness can be that the used AI-related information does not achieve the expected results / effects. Specifically, the request message may carry at least one of the following: 1) AI-related information content, which may be AI-related information currently in use, such as a functionality ID, or the content of all AI-related information that meets configuration conditions / rules, where the conditions may be associated IDs or model information; 2) AI-related information association indicators, such as associated IDs, reported IDs, measIDs, or / and AI IDs, which implicitly represent the content of AI-related information; 3) Indicators requesting to switch AI-related information, used to indicate that the UE needs the network side to instruct it to switch AI-related information or reconfigure AI-related information. These indicators may be indicated by bits, such as 1 bit, or may be associated with the reason for requesting to switch AI-related information; 4) Performance monitoring results, which can be understood as the results obtained after model monitoring or function monitoring. These results may be the difference between predicted and actual values, or the average difference between multiple predicted and actual values, contrast, etc. The performance monitoring results can also implicitly indicate that the UE needs to switch AI-related information.

[0166] It should be noted that the request message can be an RRC message, a UAI message, a MAC CE message, a UCI message, or a dedicated instruction message.

[0167] When the UE actively sends this request message, from the network side perspective, the following behaviors or actions are possible, including but not limited to: 1) Upon receiving the request message, the NW can resend a configuration message containing AI-related information to the UE to filter AI-related information; 2) Upon receiving the request message, the NW can send an instruction message to the UE to switch or reselect AI-related information, instructing the UE to switch or reselect AI-related information; 3) Upon receiving the request message, the NW can send the content of AI-related information to the UE, such as identification information of AI-related information, which is used to inform the UE to switch to the resent AI-related information. The identification information can be one or more; 4) After receiving the request message, the NW can also instruct the UE to fall back to non-AI mode; 5) After receiving the request message, the NW can also send a request message for AI-related messages. When the NW receives the AI-related messages, it can reselect one or more AI-related messages for the UE.

[0168] Corresponding to the network-side behavior or action, the UE's behavior or action may also include: 1) The UE performs AI-related information switching based on the configuration information, indication information, or AI-related information received from the network; 2) The UE can send AI-related information to the NW reconfiguration based on the request message or configuration message received from the network.

[0169] Option 2a: The UE actively reports AI-related information, and the network side determines the AI-related information to be used.

[0170] In this scheme, the UE can also actively send updated AI-related information to the NW. The NW then determines one or more suitable AI-related information and sends the newly determined AI-related information to the UE, instructing the UE to use the AI-related information, which can also be called activating the AI-related information.

[0171] Option 3a: UE reports AI-related information after handover.

[0172] In this solution, the UE can also actively select or switch AI-related information, and then send the switched AI-related information to the network side.

[0173] It should be noted that in some embodiments, the AI-related information after switching is inactive, so the NW can still send corresponding activation information to the UE; or, the AI-related information after switching can be active, that is, AI-related operations, such as inference operations, can be performed directly. In this case, the reported AI-related message can also carry activation status indication information of the AI-related information. The activation status indication information is used to indicate the activation status of the AI-related information, including active or inactive state.

[0174] Option B: Based on NW triggering. Specifically, the following NW triggering methods can be used: Option 1b: Triggered based on the On-demand mechanism.

[0175] In this scheme, the network side proactively initiates a request message, which is a request message for AI-related information, such as an applicable functionality request. Optionally, the request message may also include a reason for initiating the request, which indicates why the network side requires the UE to report AI-related information. It should be noted that the reason for the request may also be included in the first configuration information or inference configuration information, or it may be carried in other messages; this application does not impose any specific limitations on this.

[0176] Optionally, after receiving the request message, the UE can send AI-related information that meets the conditions to the network. The network selects one or more suitable AI-related information for the UE and then performs corresponding AI-related operations, such as activation. Optionally, after receiving the request message, the UE can send AI-related information that meets the conditions to the network side. The network side then sends an indication message, instructing the UE to select one or more suitable AI-related information. Optionally, the NW can proactively send an instruction message to the UE, instructing the UE to independently select one or more suitable AI-related information.

[0177] - Option 2b: NW Select or Reconfigure In this solution, the network side actively reselects AI-related information and then sends the reselected AI-related information to the UE; alternatively, the NW resends the configuration information of the AI-related information, and after the UE evaluates the AI-related information that meets the configuration or conditions, it reports it to the NW, so that the NW can select the new AI-related information from it and then send the new AI-related information to the UE.

[0178] It should be noted that if the inference configuration reporting is also periodic, then inference is also autonomously activated, and its configuration and the periodic reporting of AI-related information can be done in a similar way, without being limited here.

[0179] In this embodiment, for different use cases of AI information, a two-level signaling combination is designed, including the format of configuration signaling, the format of reported signaling, information content, rules, and / or indication information, so that AI-related information is associated with corresponding AI-related operations. This solves the problem of association / pairing failure caused by the network's inability to identify the AI-related operations corresponding to AI-related information, thereby enabling the NW to correctly manage AI-related functions / AI models and improve management efficiency.

[0180] For scenario 2, please refer to Figure 3 The following is a flowchart illustrating another possible communication method provided in this application, including but not limited to at least one of the following steps: 301. The network side sends a second message to the UE; This embodiment provides a scheme where the network side proactively triggers a request for the UE to report AI-related information (such as applicable functionalities). Specifically, when the network side needs the UE to report AI-related information, it proactively initiates a relevant request message / or request indication message. After receiving the message / indication, the UE sends AI-related information that meets the (configuration) conditions. Once the network side obtains the AI-related information, it can better perform AI-related operation management, such as activating a functionality, switching a functionality, reselecting a functionality, or activating inference, thereby improving NW management efficiency. The second message can be sent in various ways, such as direct sending, unicasting configuration information followed by sending the second message, or broadcasting configuration information followed by sending the second message. These will be described in detail below.

[0181] Sending method 1: Directly send the second message to request the UE to report AI-related information.

[0182] The base station sends a second message to the UE, which carries associated ID information. On the one hand, this informs the UE of the network-side condition information, assisting the UE in determining the AI-related information (such as applicable functionality) that can be reported. On the other hand, when the network receives this second message or the associated ID information, it can also trigger the UE to report the AI-related information (such as applicable functionality) that it has finally determined. It should be noted that this second message can be an RRC message, such as a UEInformationRequest message, or a newly defined RRC message; the specifics are not limited here. Optionally, the second message may also carry filter conditions, which include any one of the following: Condition 1: Duration, used to measure the stability of AI-related information (such as applicable functionality). Only when the AI-related information remains unchanged for a configured duration, such as 100ms, can it be reported. This is to prevent the ping-pong effect. The Duration can consist of any of the following information: start time, duration, time hysteresis value, or end time. The time unit of the Duration can be a time slot, frame, ms, s, min, hour, or day, etc.

[0183] Condition 2: Reporting Quantity. The reporting quantity refers to the number of AI-related information (such as applicable functionality) that meets the reporting requirement. For example, if 10 applicable functionality meet the condition and the configured reporting quantity is 5, then the UE can report a maximum of 5. These 5 can be randomly selected by the UE, or they can be selected by the UE based on its internal conditions, such as memory size. For example, if the models supporting applicable functionality #1 are model #2 and model #3, and the model supporting applicable functionality #2 is model #3, and the model supporting applicable functionality #3 is also model #3, then the UE can evaluate and choose to report applicable functionality #2 and applicable functionality #3, but not applicable functionality #1. It should be noted that this reporting quantity can be the maximum or minimum number of reports, and the unit is a positive integer.

[0184] Condition 3: Threshold or condition for status information. The threshold for status information is used to measure the stability of AI-related information (such as applicable functionality). If the configured threshold for status information is met, AI-related information can be reported to prevent the ping-pong effect. The threshold can be a range of values; the condition can be direct indication information, such as "high," which indicates that the stability of AI-related information is higher than a certain set value or level.

[0185] Optionally, in some cases, the message may also carry a request reason, which indicates why the network side needs the UE to report AI-related information. The reason could be that the AI-related information has expired, is invalid, or its performance is unsatisfactory. Upon receiving the request reason, the UE can either resend the required AI-related information to the network side, or reactivate the AI-related information, unload the corresponding model, or update the model.

[0186] A configuration example is shown below. The second message is a UEInformationRequest, which carries aIInfo-ReportReq to indicate that the UE needs / should report AI-related information. The aIInfo-ReportReq information may not contain any information, or it may contain at least one of the following, such as associated ID or filter conditions. Optionally, the associated ID is directly included in the UEInformationRequest. UEInformationRequest ::= SEQUENCE { ra-ReportReq-r16ENUMERATED {true}OPTIONAL, -- Need N rlf-ReportReq-r16ENUMERATED {true}OPTIONAL, -- Need N aIInfo-ReportReq ENUMERATED {true}OPTIONAL, -- Need N ...}, aIInfo-ReportReq ::= SEQUENCE { noneNull, associated IDOPTIONAL, DurationOPTIONAL MaxnumberOPTIONAL, MinnumberOPTIONAL ... } For example, please refer to Figure 3a This is a flowchart of a possible communication method provided in an embodiment of this application, used by a base station to trigger a UE to report AI-related information. The specific steps are described below: Step 1: The UE receives a second message sent by the base station. The second message can be a UE information request message or other RRC messages.

[0187] The second message is used to request AI-related information (such as applicable functionality) reported by the UE. The second message carries at least any of the following information: AI-related information such as AIInfo-ReportRe, associated ID, and / or filter conditions.

[0188] When the UE receives a second message containing AI-related information sent by the base station, it can determine the AI-related information it needs to report, such as applicable functionalities, based on the second message. The applicable functionalities can be one or more.

[0189] Step 2: The UE reports AI-related information to the base station via RRC signaling (such as UE information response).

[0190] Sending method 2: The base station unicasts the configuration message and then sends the second message.

[0191] In some scenarios, network-side condition information is fixed or infrequently changing. It may be unique based on signaling, base station, region, cell, or PLMN. Therefore, the associated ID, the identifier for network-side condition information, can be pre-sent to the UE via RRC configuration message unicast by the network side. This associated ID is then reconfigured after the network changes. This associated ID informs the UE of the network-side condition information, assisting the UE in determining the AI-related information (such as applicable functionality) that can ultimately be reported. Subsequently, when the base station sends a second message to the UE, this message carries an indication for AI-related information. This indication indicates that the request is primarily for AI-related information, not other information. Furthermore, the indication message may also include or indicate the specific type of AI-related information being requested, such as whether it requests applicable functionality or an applicable model.

[0192] Optionally, the second message may also carry filter conditions, and the message may also carry a request reason. The filter conditions and / or request reason are as described above and will not be repeated here. The second message may be an RRC message, such as a UE information request message, a MAC CE, or a DCI message; specific details are not limited here.

[0193] For example, please refer to Figure 3b This is a flowchart of a possible communication method provided in an embodiment of this application, used by a base station to trigger a UE to report AI-related information. The specific steps are described below: Step 1: The UE receives an RRC reconfiguration message or unicast message sent by the base station. The RRC reconfiguration message or unicast message includes the associated ID.

[0194] Step 2: The UE receives a second message sent by the base station. The second message can be a UE information request message or other RRC messages.

[0195] The second message is used to request AI-related information (such as applicable functionality) reported by the UE. The second message carries at least any of the following information: AI-related information indication for AIInfo or filter conditions, etc.

[0196] When the UE receives a second message containing AI-related information sent by the base station, it can determine the AI-related information it needs to report, such as applicable functionalities, based on the second message. The applicable functionalities can be one or more.

[0197] Step 3: The UE reports AI-related information to the base station via RRC signaling (such as UE information response message). The UE information response message includes AI-related information, such as at least one applicable functionality.

[0198] Sending method 3: The base station broadcasts / multicasts configuration messages, and then sends a second message.

[0199] In some scenarios, network-side condition information is fixed or infrequently changing. It may be unique based on signaling, base station, region, cell, or PLMN. Therefore, the associated ID, the identifier for network-side condition information, can be pre-sent to the UE via RRC configuration message broadcast or multicast by the network side. This associated ID is then reconfigured after the network sends a new message. This associated ID informs the UE of the network-side condition information, assisting the UE in determining the AI-related information (such as applicable functionality) that can ultimately be reported. Subsequently, when the base station sends a second message to the UE, this second message carries a request indication for AI-related information (indication for AIInfo). In some cases, the request message may also carry filter conditions; optionally, the message may also carry a request reason. The filter conditions and / or request reason are as described above and will not be repeated here. The second message can be an RRC message, such as a UE information request message, a MAC CE, or a DCI message; the specific type is not limited here.

[0200] For example, please refer to Figure 3c This is a flowchart of a possible communication method provided in an embodiment of this application, used by a base station to trigger a UE to report AI-related information. The specific steps are described below: Step 0: The base station receives the associated ID; Step 1: The UE receives a multicast message or broadcast message sent by the base station. The multicast message or broadcast message includes the associated ID and cell / area related information.

[0201] Step 2: The UE receives a second message sent by the base station. The second message can be a UE information request message or other RRC messages.

[0202] The second message is used to request AI-related information (such as applicable functionality) reported by the UE. The second message carries at least any of the following information: AI-related information indication for AIInfo, reason for request, and / or filter conditions, etc.

[0203] When the UE receives a second message containing AI-related information sent by the base station, it can determine the AI-related information it needs to report, such as applicable functionalities, based on the second message. The applicable functionalities can be one or more.

[0204] Step 3: The UE reports AI-related information to the base station via RRC signaling (such as UE information response message). The UE information response message includes AI-related information, such as at least one applicable functionality.

[0205] It should be noted that when the UE receives the Step 1 message, it can be in an RRC idle, RRC inactive, or RRC connected state; specifically, when the UE receives the Step 2 message, it is usually in an RRC connected state. Optionally, if a network element in the core network needs to collect AI-related information from the UE, the UE can also be in an inactive state. The second message can also be LPP signaling, etc.

[0206] 302. UE determines AI-related information; 303. The UE sends a reporting message to the network side; 304. The network side performs AI-related operations based on AI-related information; 305. The UE reports the inference results to the network side.

[0207] In this embodiment, steps 302 to 305 are... Figure 2 Steps 203 to 206 in the illustrated embodiment are similar and will not be repeated here.

[0208] This embodiment considers an on-demand reporting mechanism for AI-related information. This mechanism involves the NW (Network Controller) proactively sending requests, and further, the content format of the request message is designed to trigger the UE (User Equipment) to send AI-related information. This mechanism allows the NW to proactively and on-demand acquire AI-related information, adapting to various AI use cases, reducing the management burden on the network and UE, and minimizing signaling overhead.

[0209] It should be noted that the associated identifier, or associated ID, in this application is used to represent identification information of network-side information, which may include, but is not limited to, at least one of the following: Information related to beam measurement includes: number / mode of beams, time window, transmit beam codebook (or beam pointing order / index / shape), receive beam assumptions (e.g., beam shape and beam pointing order), and antenna configuration.

[0210] Cell measurement-related information: such as cell frequency information, number of cells, measurement time window, and antenna configuration.

[0211] Network-related information, specific information about the site / cell, gNB deployment, transmit power, and receive power; User Equipment (UE) related information: UE distribution or deployment, UE speed.

[0212] Model-related information, (training / input / output) dataset identifier, model structure, model parameters, model version data distribution, accuracy information, model complexity information, and model size information; The model's input and / or output information; Inference-related configuration information or parameters: resource information, measurement object information, auxiliary information / signal information, prediction accuracy information, prediction result information, prediction timing information, etc. For example, quantization-related information for CSI compression. Another example is beam shape-related information used for beam prediction (e.g., relative power information for each beam at each angle).

[0213] Scene information, configuration information, coordinate information; vendor information; AI function-related information; AI business-related information; AI service-related information; AIQoS-related information, such as inference accuracy and inference latency; AI service accuracy: The ability to perform high-precision processing on a given AI inference / training task within a given latency range, measured by the degree to which the output of the AI ​​service is identical to the true value of the given input.

[0214] AI service density: refers to the number of AI services that can be executed within a unit coverage area under constraints of unit time, unit computing power (TFLOPS), and unit bandwidth (MHz), meeting the accuracy and latency requirements of AI services. AI service efficiency (resource utilization efficiency): Within a unit coverage area, given latency range and the number and accuracy of AI inference / training tasks, the comprehensive consumption of network communication resources, computing resources, and data resources.

[0215] AI service latency: The end-to-end latency of a network performing a given AI inference / learning task under a given accuracy requirement. This end-to-end latency consists of three parts: 1) deployment latency; 2) processing latency; and 3) transmission latency.

[0216] Furthermore, the AI-related information in this application refers to information related to AI technology or information generated during the execution of AI-related operations, such as AI functionality, AI model, AI feature, AI feature group, or AI operations. AI functionality and / or AI model may include one or any combination of multiple terms, such as: applicable functionalities / AI models, functionalities / AI models, supported functionalities / AI models, activated applicable functionalities / AI models, deactivated applicable functionalities / AI models, activated functionalities / AI models, and / or deactivated functionalities / AI models, etc.

[0217] AI functionality should include at least the following different types of functionality, such as supported functionality. Applicable functionality, Activated functionality, non-applicable functionality, or deactivated functionality, etc.

[0218] The AI ​​functionality includes or relates to at least one of the following: Identification information for functional configuration: Used to identify a specific functional configuration, which can be a base station configuration, such as a specific configuration used in a CSI report or a specific configuration used in an RRM, or a core network configuration, such as a specific configuration used in an LPP report; the functional configuration can be AI-based or non-AI-based. Resource set identification information: can be used to identify a resource set, or to represent different combinations of different resource sets, and can distinguish different functional conditions, such as in the SETA resource set and the SETB resource set; Dataset identifier: can be used to identify the input and / or output of a dataset, which may be a training dataset, an inference dataset, a performance monitoring dataset, or a combination of the above datasets; Measurement information includes Channel Report Indicator (CRI), Reference Signal Received Power (RSRP), Reference Signal Received Quality (RSRQ), Signal to Interference plus Noise Ratio (SINR), Beam Index, Cell ID, Measurement event, and other measurement information. Probability information: used to represent the probability of certain events occurring, such as using AI technology to predict the probability of certain things happening at a certain time or over a certain period of time; Time point: A specific point in time used for measurement or prediction; Time interval: The time interval used for measurement or prediction; Output content: Predicted information, such as predicted CRI (Channel Report Indicator), RSRP (Reference Signal Received Power), RSRQ, SINR, Beam Index, Cell ID, Measurement event, and other measurement information; Output type: Indicates different model output types, such as classification outputs, which are usually class labels; and regression outputs, which are continuous numerical values. Probability Distribution: The output is the probability distribution of each category; Vector Outputs: The output is a vector; Structured Outputs: The output is an object with a complex structure, such as predicting a sequence or graph structure; Confidence or Probability Information: The output contains confidence or probability information about the prediction result; Multi-task Learning Outputs: The output is the joint result for multiple tasks; Anomaly Detection Outputs: The output indicates whether the input data is within the normal range. Output range: The range or threshold of the model output, or the difference from the threshold; Performance metrics: such as throughput, latency, the difference or average difference between the actual and predicted values, variance, error of variance (MSE), cross-entropy; Monitoring types: such as real-time monitoring, lenient monitoring, periodic monitoring, etc. The size of the resource set, the size of the dataset, the settings for repeating windows, etc.; model identification information; The structure and / or components of the model; such as input layer, hidden layer, output layer, activation function, loss function, optimizer, regularization, interpretability tools; The type of model, such as the Transformer model; AI-related operations; the AI ​​operations are used to represent operations performed using AI technology. AI-related functions, which may refer to executing a specific AI use case, a sub-use case, or a specific AI operation; AI operations: These represent operations that utilize AI technology, such as training, inference, performance monitoring, rollback, activation, deactivation, switching, or selection. AI feature: to represent the characteristics of AI, which may include at least one of the following: data transmission mechanism, configuration, activation, capability reporting resource set, time window, or content, etc. AI feature group: Used to represent groups of AI features, such as AI-based beam management, AI-based mobility management, AI-based CSI prediction, and AI lifecycle management.

[0219] In some cases, AI features can be similar to or the same as AI feature group content.

[0220] In addition, AI-related information can be presented in several forms, including but not limited to the following: Format 1: AI-related information is identified by an index value.

[0221] The index value is typically an integer or an enumeration value, used to uniquely identify a specific AI-related information or multiple related AI-related information, or the content of AI-related information, as described above. An example is shown in Table 1 below: Table 1

[0222] In some cases, applicable functionality can be associated with supported functionality. For example, if supported functionality has an index value, applicable functionality can also be represented by a local index. This local index can be locally allocated and used to distinguish different applicable functionalities corresponding to a single supported functionality. The specific content of the applicable functionality does not need to be specified here; the identifier of the supported functionality can be predefined or dynamically allocated. An example is as follows: Index=1, supported functionality#1, local index=1 (applicable functionality#1); Index=1, supported functionality#1, local index=2 (applicable functionality#1).

[0223] Format 2: Table format AI-related information can be represented in tabular form. The table can define the value range, mapping relationship, conditional logic, etc. of AI-related information. The table content can contain any one or more AI-related information contents, and can also contain AI-related rule information, conditional information, etc. The table can be predefined or dynamic.

[0224] Similar to Form 1 in some descriptive forms, an example of a static table is shown in Table 2 below: Table 2

[0225] In some descriptive formats, an example of a dynamic table is shown in Table 3 below. Table 3

[0226] Alternatively, the table in question can be of the type of per AI information, as shown in Table 4 below: Table 4

[0227] In some cases, applicable functionality can be associated with supported functionality, similar to form 1, which will not be elaborated here.

[0228] Form 3: Encoding form or parameter list form The message is defined by the content containing AI-related information, and the message contains at least any of the messages mentioned above.

[0229] The message can be structured using ASN.1 or MAC PDU, and the content of the AI-related information is as described above. In some cases, message identifiers, such as RRC IDs, can be used to represent AI-related information.

[0230] The above figures illustrate in detail the communication method provided in the embodiments of this application. Please refer to... Figure 4 , Figure 4 This is a schematic diagram of the storage of a wireless communication device according to an embodiment of this application. The wireless communication device includes a processor and a memory. The memory stores computer programs, and the processor calls and runs the computer programs stored in the memory to execute the methods provided by any embodiment of the channel determination method or channel switching method of this application, as well as any non-conflicting combination thereof. The storage medium 20 of the wireless communication device in this embodiment stores instruction / program data 21. When the instruction / program data 21 is executed, it implements the methods provided by any embodiment of the communication method of this application, as well as any non-conflicting combination thereof. The instruction / program data 21 can be formed into a program file and stored in the storage medium 20 in the form of a software product, so that a computer device (which may be a personal computer, server, or network device, etc.) or processor executes all or part of the steps of the methods of various embodiments of this application. The aforementioned storage medium 20 includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, or a terminal device such as a computer, server, mobile phone, or tablet.

[0231] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between apparatuses or units, and may be electrical, mechanical, or other forms.

[0232] Furthermore, 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. The integrated unit can be implemented in hardware or as a software functional unit.

[0233] The above are merely embodiments of this application and do not limit the scope of this patent application. Any equivalent structural or procedural changes made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of this application.

[0234] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or 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 website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0235] It should be understood that the term "and / or" in this article 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. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0236] In this application, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0237] It should be understood that 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.

[0238] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0239] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0240] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0241] 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.

[0242] 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.

[0243] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0244] 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 applied to the user equipment (UE) side, characterized in that, include: The UE receives a first message and / or a second message sent by the network side, wherein the first message includes first configuration information related to AI, and the second message is used to request the UE to report the AI-related information. A reporting message is sent to the network side, the reporting message including AI-related information and / or UE assistance information.

2. The method according to claim 1, characterized in that, The first message also includes inference configuration information based on AI algorithms; or, before sending the AI-related information and / or UE assistance information to the network side, the method further includes: Receive a third message sent by the network side, the third message including the inference configuration information based on the AI ​​algorithm.

3. The method according to claim 2, characterized in that, The method further includes: The inference results corresponding to the inference configuration information are reported to the network side.

4. The method according to any one of claims 1 to 3, characterized in that, The first configuration information includes, but is not limited to, at least one of the following: measurement resource information dedicated to AI-related information, and / or its identification information, AI-related reporting configuration identification information, AI-related reporting type, inference configuration resource identification information, first identification information of inference reporting configuration, association identification, reporting content, inference configuration result information, inference configuration bandwidth information, AI-related measurement object information, measurement identification, identification information of inference measurement object, first identification information of inference reporting configuration, AI-related configuration association information, AI-related configuration first association information, AI-related configuration second association information, reporting priority rules, and signaling reporting rules.

5. The method according to any one of claims 1 to 4, characterized in that, The inference configuration information includes an association identifier.

6. The method according to any one of claims 2 to 5, characterized in that, The first configuration information is included in the RRC reconfiguration message, and the inference configuration information is included in the CSI reporting configuration CSI-ReportConfig configured on the network side.

7. The method according to claim 6, characterized in that, The inference configuration information is included in the RRC reconfiguration message, and the inference configuration information includes association information.

8. The method according to any one of claims 2 to 5, characterized in that, The first configuration information is contained in OtherConfig configured on the network side, and the inference configuration information is contained in CSI-ReportConfig configured on the network side.

9. The method according to claim 8, characterized in that, The OtherConfig carries an AI reporting instruction, which is used to indicate whether the UE should report the AI-related information.

10. The method according to claim 9, characterized in that, The AI ​​reporting instructions can be implemented in ways including but not limited to at least one of the following: by enabling / disabling, by parameter indication, or by bit value indication.

11. The method according to any one of claims 2 to 5, characterized in that, The first configuration information is contained in the measurement MeasConfig configured on the network side, and the inference configuration information is contained in the CSI-ReportConfig configured on the network side.

12. The method according to any one of claims 2 to 5, characterized in that, Both the first configuration information and the inference configuration information are included in the CSI-ReportConfig configured on the network side. The CSI-ReportConfig includes at least a first part and a second part. The first part includes the first configuration information, and the second part includes the inference configuration information.

13. The method according to any one of claims 1 to 12, characterized in that, When the first configuration information is reported periodically, the method further includes: Receive the periodic reporting configuration sent by the network side.

14. The method according to claim 13, characterized in that, The periodic reporting configuration is associated with resource information used to report the inference results, or the periodic reporting configuration is associated with resource information for inference or measurement corresponding to the inference configuration information.

15. The method according to any one of claims 3 to 14, characterized in that, The inference results reported to the network side corresponding to the inference configuration information include: The inference results are reported to the network side via L1 signaling or L3 signaling.

16. The method according to any one of claims 2 to 5, characterized in that, Both the first configuration information and the inference configuration information are included in the MeasConfig configured on the network side. The MeasConfig includes at least a third part and a fourth part. The third part includes the first configuration information, and the fourth part includes the inference configuration information.

17. The method according to any one of claims 2 to 5, characterized in that, The first configuration information is contained in OtherConfig configured on the network side, and the inference configuration information is contained in MeasConfig configured on the network side.

18. The method according to claim 17, characterized in that, The OtherConfig carries an AI reporting rule indication, which is used to indicate whether the UE should report the AI-related information. The AI ​​reporting indication can be implemented in ways including but not limited to at least one of the following: by enabling / disabling, by parameter indication, or by bit value indication.

19. The method according to any one of claims 1 to 18, characterized in that, The sending of the reporting message to the network side includes: The AI-related information is reported to the network side via L1 or L3 signaling.

20. The method according to claim 1, characterized in that, The second message includes, but is not limited to, at least one of the following: request indication information, associated identifier, and / or, filtering condition information. The request indication information is used to request the reporting of AI-related information. The filtering condition information is used to indicate the conditions that must be met before reporting the AI-related information.

21. The method according to claim 20, characterized in that, The filtering conditions include time conditions and / or the reporting quantity of AI-related information.

22. The method according to claim 1, characterized in that, Before sending the reporting message to the network side, the method further includes: The network side receives a fourth message, which carries an association identifier.

23. The method according to claim 22, characterized in that, The second message carries request indication information or filtering condition information, the request indication information being used to request the reporting of AI-related information.

24. The method according to claim 22 or 23, characterized in that, The fourth message can be sent via unicast, multicast, or broadcast.

25. The method according to any one of claims 1 to 24, characterized in that, The sending of the reporting message to the network side includes: The AI-related information is carried by at least one of the following signaling methods: L3 signaling, L1 signaling, or MAC CE signaling, to be reported to the network side.

26. The method according to claim 25, characterized in that, The L3 signaling includes, but is not limited to, one of the following signaling types: RRC reconfiguration complete message, UAI complete message, and UE-side information response complete message.

27. The method according to any one of claims 1 to 26, characterized in that, The sending of the reporting message to the network side includes: Based on the reporting priority rules, the AI-related information is reported to the network side.

28. The method according to claim 27, characterized in that, The reporting rules are configured directly by the network side, or reported based on a predefined reporting priority, or reported based on preset rules.

29. The method according to claim 28, characterized in that, The predefined reporting priority is related to the real-time nature of the business or the size of the data volume; The preset rules include reporting based on the message type of the second message, or reporting based on the reporting type of the AI-related information.

30. The method according to any one of claims 4 to 29, characterized in that, The signaling reporting rules for the first configuration information include: reporting the initial AI-related information based on the first type of message, and reporting the updated AI-related information based on the second type of message.

31. The method according to claim 30, characterized in that, The signaling reporting rules can be indicated in the following ways: by displaying instructions through fields, by indicating instructions through time windows, by indicating instructions through timers, or by controlling them on the network side.

32. The method according to claim 31, characterized in that, The time window includes a set time value or time range, used to indicate the time interval between the UE sending the first type of message and the second type of message. The timer is used to control the time when the UE starts sending the AI-related information via the second type of message after sending the first type of message.

33. The method according to claims 1 to 32, characterized in that, Before sending the reporting message to the network side, the method further includes: Determine the AI-related information to be reported.

34. The method according to claim 33, characterized in that, When the AI-related information includes multiple AI information, the reporting message includes at least one of the following: the status information of each AI information, the number of models corresponding to each AI information, the generalization ability corresponding to each AI information, and the device capability for running each AI information.

35. The method according to claim 33 or 34, characterized in that, After sending the reporting message to the network side, the method further includes: The system receives a first indication message sent by the network side, the first indication message being used to indicate target AI information from the plurality of AI information.

36. The method according to claim 33, characterized in that, The AI-related information to be reported includes: The AI-related information is determined based on the determination rules for the AI ​​information included in the first configuration information, and then reported.

37. The method according to claim 36, characterized in that, The rules for determining the AI ​​information are related to at least one of the following factors: the state information of the AI ​​information, the number of models corresponding to the AI ​​information, the generalization ability corresponding to the AI ​​information, and the device capability for running the AI ​​information.

38. The method according to claim 37, characterized in that, The AI-related information to be reported includes: Based on the UE's capability information, the AI-related information is determined from the AI ​​information that satisfies the first configuration information.

39. The method according to any one of claims 4 to 38, characterized in that, When the first configuration information includes multiple associated IDs, the step of determining the AI-related information to be reported includes: The first AI information set is determined based on the associated ID corresponding to the first setting. and / or The second AI information set is determined based on the associated ID corresponding to the second setting. The AI-related information is determined based on the overlap information between the first AI information set and / or the second AI information set.

40. The method according to claim 39, characterized in that, The AI-related information is any one of the following: the first AI information set, the second AI information set, the first AI information set and the second AI information set, or overlapping information of the first AI information set and the second AI information set.

41. The method according to claim 33, characterized in that, The AI-related information to be reported includes: The AI-related information is determined based on the association identifier.

42. The method according to claim 40 or 41, characterized in that, The first configuration information includes the associated ID and at least one usage instruction information, wherein the usage instruction information is used to indicate the AI ​​information associated with each associated ID.

43. The method according to claim 42, characterized in that, The different indication states of the usage indication information are used to indicate whether the corresponding associated ID is available.

44. The method according to claim 42, characterized in that, The associated ID is either pre-configured or determined by the UE.

45. The method according to claim 4, characterized in that, The first configuration information includes the AI-related information reporting type, which is used to indicate periodic reporting.

46. ​​The method according to any one of claims 1 to 45, characterized in that, The reported message includes UE-side tagging information, which is used to tag the AI ​​information being used by the UE.

47. The method according to any one of claims 1 to 45, wherein the reported information includes an AI information index, the AI ​​information index being used to indicate the AI ​​information being used by the UE, and the AI-related information includes the AI ​​information being used by the UE.

48. The method according to any one of claims 1 to 47, characterized in that, The method further includes: A first request message is sent to the network side, the first request message being used to indicate that the third AI-related information used by the UE is lower than a preset value.

49. The method according to claim 48, characterized in that, The first request message includes, but is not limited to, at least one of the following: content information of the third AI-related information, association indicator of the third AI-related information, indicator of requesting to switch the third AI-related information, and / or performance monitoring results.

50. The method according to claim 49, characterized in that, The first request message carries at least one of the following signaling types: RRC information, UAI information, MAC CE, UCI information, or proprietary information.

51. The method according to any one of claims 48 to 50, characterized in that, After sending the first request message to the network side, the method further includes: Receive a fifth message sent by the network side, the fifth message including but not limited to one of the following: second configuration information, second indication information, new AI-related information content information, and rollback indication information; The second configuration information is used to redetermine new AI-related information, the second indication information is used to indicate switching or reselecting the new AI-related information, and the rollback indication information is used to indicate switching to non-AI mode.

52. The method according to claim 51, characterized in that, After receiving the fifth message sent by the network side, the method further includes: Based on the fifth message, perform the corresponding operation, which includes, but is not limited to, one of the following actions: reselect the new AI-related information, switch to the new AI-related information, or switch to non-AI mode.

53. The method according to any one of claims 1 to 52, characterized in that, The second message includes the reason for the request.

54. The method according to any one of claims 1 to 53, characterized in that, The second message is either an existing RRC message or a newly defined RRC message.

55. The method according to any one of claims 1 to 54, characterized in that, The method further includes: The network receives a sixth message, which is used to activate the AI-related information or the reasoning function corresponding to the AI-related information.

56. A communication method applied to the network side, characterized in that, include: Send a first message and / or a second message to the UE, wherein the first message includes first configuration information about AI-related information, and the second message is used to request the UE to report the AI-related information; Receive a reporting message sent by the UE, the reporting message including AI-related information and / or UE assistance information.

57. The method according to claim 56, characterized in that, The first message also includes inference configuration information based on AI algorithms; or, Before receiving the reporting message sent by the UE, the method further includes: A third message is sent to the UE, the third message including the inference configuration information.

58. The method according to claim 57, characterized in that, The method further includes: Receive the inference result sent by the UE corresponding to the inference configuration information.

59. The method according to claim 56, characterized in that, When the first configuration information is reported periodically, the method further includes: Send the periodic reporting configuration to the UE.

60. The method according to claim 56, characterized in that, Before receiving the reporting message sent by the UE, the method further includes: A fourth message is sent to the UE, the fourth message carrying an association identifier, the association identifier being used to represent the identification information of the network side information.

61. The method according to claim 60, characterized in that, The fourth message can be sent in the following ways: unicast, multicast, or broadcast.

62. The method according to any one of claims 56 to 61, characterized in that, After receiving the reporting message sent by the UE, the method further includes: Send a first indication information to the UE, the first indication information being used to indicate target AI information from the plurality of AI information.

63. The method according to any one of claims 1 to 62, characterized in that, The method further includes: The system receives a first request message sent by the UE, the first request message being used to indicate that the third AI-related information used by the UE is lower than a preset value.

64. The method according to claim 63, characterized in that, After receiving the first request message sent by the UE, the method further includes: Receive a fifth message sent by the network side, the fifth message including but not limited to one of the following: second configuration information, second indication information, new AI-related information content information, and rollback indication information; The second configuration information is used to redetermine new AI-related information, the second indication information is used to indicate switching or reselecting the new AI-related information, and the rollback indication information is used to indicate switching to non-AI mode.

65. The method according to any one of claims 1 to 64, characterized in that, The method further includes: Based on the AI-related information, perform AI-related operations.

66. The method according to claim 65, characterized in that, The method further includes: A sixth message is sent to the UE, the sixth message being used to activate the AI-related information or the inference function corresponding to the AI-related information.

67. A wireless communication device, comprising: A processor and a memory, the memory being used to store a computer program, the processor being used to invoke and run the computer program stored in the memory to perform the method as described in any one of claims 1 to 66.

68. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes instructions that, when executed, cause the method according to any one of claims 1 to 66 to be implemented.

Images