Model issuing method and device, model obtaining method and device, UE and network side network element

Abstract

The invention provides a model issuing method and device, a model obtaining method and device, UE, a network side network element, a communication system and a storage medium. The model issuing method comprises the steps that a model request sent by the UE through control plane signaling is received; in response to the model request, sending first response information to the UE through the user plane network element, or sending second response information to the UE through the control plane network element; wherein the first response information comprises target model information matched with the model request; the second response information comprises model request rejected information. The security of interaction can be ensured, and the method is suitable for transmission of a large amount of data; cross-domain model transmission can be realized, the security is high, supervision on the model can be enhanced, and resources such as storage and computing power can be fully utilized.

Description

Technical Field

[0001] The present invention relates to the field of communication technologies, and in particular, to a method for model distribution, model acquisition, an apparatus, a UE, a network-side network element, a communication system, and a storage medium. Background Art

[0002] In the related art, models such as machine learning (ML) models can be applied in various application scenarios such as information feedback, positioning enhancement, and beam management, which can bring a better application experience. In domains such as the 5GC (5th Generation Core) domain and the RAN (Radio Access Network) domain of a communication system, operations such as training, transmission, and interaction of models such as ML models are all completed within each domain, and cross-domain model transmission, interaction, etc. are not possible. However, within the RAN domain, due to factors such as storage, computing power, complexity, and energy consumption, it is difficult for both the UE (User Equipment) side and the base station side to independently complete the training of models such as ML models, and the UE side has a need to obtain models such as ML models stored in network elements in other domains. Summary of the Invention

[0003] In view of this, a technical problem to be solved by the present invention is to provide a method for model distribution, model acquisition, an apparatus, a UE, a network-side network element, a communication system, and a storage medium.

[0004] According to a first aspect of the present disclosure, there is provided a model distribution method applied to a network-side network element, including: receiving a model request sent by a user equipment (UE) through control plane signaling; in response to the model request, sending the first response information to the UE through a user plane network element, or sending second response information to the UE through the control plane network element; wherein, the first response information includes: target model information matching the model request; and the second response information includes: information indicating that the model request is rejected.

[0005] Optionally, the model request includes: the UE information and / or the model request information; wherein, the UE information includes: UE identification information and / or UE location information; and the model request information includes: model ID and / or model-related information.

[0006] Optionally, the model-related information includes at least one of: model filter information, location range information applicable to the model, time range information for using the model, use case information, use case context information, model interoperability information, manufacturer identification information, and accuracy level information.

[0007] Optionally, the model request sent by the receiving UE via control plane signaling includes: receiving the model request sent by a control plane network element; wherein, the UE sends the control plane signaling to the control plane network element, and the control plane network element sends the model request to the network side network element based on the control plane signaling.

[0008] Optionally, the control plane network element sending the model request to the network side network element based on the control plane signaling includes: the control plane network element selects the network side network element based on the UE information in the control plane signaling, or based on the UE information and the model request information in the control plane signaling, and generates the model request to send to the network side network element.

[0009] Optionally, the control plane signaling includes: NAS signaling; wherein, the model request information is encapsulated in the NAS signaling in a container encapsulation manner or a protocol encapsulation manner.

[0010] Optionally, sending the first response information to the UE via a user plane network element, or sending a second response information to the UE via the control plane network element includes: generating the first response information and sending it to the UE via the user plane network element, or generating the second response information and sending it to the UE via the control plane network element.

[0011] Optionally, generating the first response information includes: obtaining the model subscription information of the UE; determining whether the UE is allowed to obtain the target model according to the model request and the model subscription information; generating the first response information when it is determined that the UE is allowed to obtain the target model and it is determined that the target model information can be obtained.

[0012] Optionally, sending the first response information to the UE via the user plane network element includes: sending the first response information to the user plane network element, so that the user plane network element sends the first response information to the UE.

[0013] Optionally, sending the first response information to the UE via the user plane network element includes: sending the first response information to the user plane network element through the AF, so that the user plane network element sends the first response information to the UE.

[0014] Optionally, sending the first response information to the user plane network element through the AF includes: when it is determined that the AF is a trusted AF, sending the first response information to the AF, so that the AF sends the first response information to the user plane network element.

[0015] Optionally, the step of sending the first response information to the user plane network element through the AF includes: when it is determined that the AF is an untrusted AF, selecting the NEF according to the model request; sending the first response information to the NEF, and the NEF sends the first response information to the user plane network element through the AF.

[0016] Optionally, the step of generating the second response information and sending it to the UE through the control plane network element includes: when it is determined according to the model request and the model subscription information that the UE is not allowed to obtain the target model or cannot obtain the target model information, determining the information that the model request is rejected and generating the second response information; sending the second response information to the control plane network element, so that the control plane network element sends the second response information to the UE.

[0017] Optionally, according to the model request, it is judged whether the target model information is stored locally; if so, the target model information stored locally is obtained, and if not, the target model information is obtained through the storage network element.

[0018] Optionally, the step of obtaining the target model information through the storage network element includes: generating retrieval information based on the model request and sending the retrieval information to the storage network element; obtaining the detection result returned by the storage network element; wherein, the retrieval result includes: the target model information, or the reason information for not retrieving the target model information.

[0019] Optionally, the target model information includes: target model object information and / or the download address of the target model; the information that the model request is rejected includes: the indication information that the model request is rejected and / or the reason information.

[0020] Optionally, the network side network element includes: NWDAF, ADRF, LMF; the control plane network element includes: AMF; the user plane network element includes: UPF; the model includes: machine learning ML model; the target model includes: ML target model.

[0021] According to a second aspect of the present disclosure, there is provided a model acquisition method applied to a user equipment UE, including: sending a model request to a network side network element through a control plane signaling; receiving the first response information sent by the network side network element through the user plane network element, or receiving the second response information sent by the network side network element through the control plane network element; wherein, the first response information includes: target model information matching the model request; the second response information includes: the information that the model request is rejected.

[0022] Optionally, the model request includes: the UE information and / or the model request information; wherein, the UE information includes: UE identification information and / or UE location information; the model request information includes: model ID and / or model-related information.

[0023] Optionally, the model-related information includes at least one of: model filter information, location range information applicable to the model, time range information used by the model, use case information, use case context information, model interoperability information, manufacturer identification information, accuracy level information.

[0024] Optionally, sending the model request to the network-side network element via control plane signaling includes: sending the control plane signaling to the control plane network element; wherein, the control plane network element sends the model request to the network-side network element based on the control plane signaling.

[0025] Optionally, the control plane network element sending the model request to the network-side network element based on the control plane signaling includes: the control plane network element selecting the network-side network element based on the UE information in the control plane signaling, or based on the UE information and the model request information in the control plane signaling, and generating the model request to send to the network-side network element.

[0026] Optionally, the control plane signaling includes: NAS signaling; wherein, the model request information is encapsulated in the NAS signaling in a container encapsulation manner or a protocol encapsulation manner.

[0027] Optionally, the user plane network element includes: UPF; receiving the first response information sent by the network-side network element through the user plane network element includes: receiving the first response information sent by the network-side network element through the UPF.

[0028] Optionally, the control plane network element includes: AMF; receiving the second response information sent by the network-side network element through the control plane network element includes: receiving the second response information sent by the network-side network element through the AMF.

[0029] Optionally, the target model information includes: target model object information and / or the download address of the target model; the model request rejected information includes: the indication information that the model request is rejected and / or the reason information.

[0030] Optionally, the network-side network element includes: NWDAF, ADRF, LMF; the model includes: a machine learning ML model; the target model includes: an ML target model.

[0031] According to a third aspect of the present disclosure, there is provided a model distribution device, which is applied to a network-side network element and includes: a receiving module, configured to receive a model request sent by a user equipment (UE) through control plane signaling; a response module, configured to, in response to the model request, send first response information to the UE through a user plane network element, or send second response information to the UE through the control plane network element; wherein, the first response information includes: target model information matching the model request; and the second response information includes: information indicating that the model request is rejected.

[0032] According to a fourth aspect of the present disclosure, there is provided a model distribution device, which is applied to a network-side network element and includes: a memory; and a processor coupled to the memory, where the processor is configured to execute the model distribution method as described above based on instructions stored in the memory.

[0033] According to a fifth aspect of the present disclosure, there is provided a network-side network element, which includes: the model distribution device as described above.

[0034] According to a sixth aspect of the present disclosure, there is provided a model acquisition device, which is applied to a user equipment (UE) and includes: a request sending module, configured to send a model request to a network-side network element through control plane signaling; a response receiving module, configured to receive the first response information sent by the network-side network element through a user plane network element, or receive the second response information sent by the network-side network element through the control plane network element; wherein, the first response information includes: machine learning target model information matching the model request; and the second response information includes: information indicating that the model request is rejected.

[0035] According to a seventh aspect of the present disclosure, there is provided a model acquisition device, which is applied to a UE and includes: a memory; and a processor coupled to the memory, where the processor is configured to execute the model acquisition method as described above based on instructions stored in the memory.

[0036] According to an eighth aspect of the present disclosure, there is provided a UE, which includes: the model acquisition device as described above.

[0037] According to a ninth aspect of the present disclosure, there is provided a communication system, which includes: the network-side network element as described above and the UE as described above.

[0038] According to a tenth aspect of the present disclosure, there is provided a computer-readable storage medium, where the computer-readable storage medium stores computer instructions, and the instructions are executed by a processor to perform the method as described above.

[0039] The method and device for model distribution, model acquisition, UE, network-side network element, communication system, and storage medium according to the present disclosure. The UE sends a model request to the network-side network element through the control plane, and the network-side network element sends target model information to the UE through the user plane, which can ensure the security of the interaction and is suitable for the transmission of large amounts of data; it can realize cross-domain model transfer, with strong security and can strengthen the supervision of the model, and can make full use of resources such as storage and computing power. Description of the Drawings

[0040] By describing the embodiments of the present disclosure in more detail in conjunction with the accompanying drawings, the above and other objects, features, and advantages of the present disclosure will become more obvious. The accompanying drawings are used to provide a further understanding of the embodiments of the present disclosure, and constitute a part of the specification. Together with the embodiments of the present disclosure, they are used to explain the present disclosure and do not constitute a limitation to the present disclosure. The following describes the above and other objects and advantages of the present disclosure in more detail with reference to specific embodiments and the accompanying drawings. In the drawings, the same or corresponding technical features or components will be represented by the same or corresponding reference numerals.

[0041] Figure 1 It is a schematic flowchart of an embodiment of the model distribution method according to the present disclosure;

[0042] Figure 2 It is a schematic flowchart of generating the first response information in an embodiment of the model distribution method according to the present disclosure;

[0043] Figure 3 It is a schematic flowchart of sending the first response information in an embodiment of the model distribution method according to the present disclosure;

[0044] Figure 4 It is a schematic flowchart of obtaining target model information through a storage network element in an embodiment of the model distribution method according to the present disclosure;

[0045] Figure 5 It is an information interaction schematic diagram of another embodiment of the model distribution method according to the present disclosure;

[0046] Figure 6 It is an information interaction schematic diagram of obtaining target model information through a storage network element in another embodiment of the model distribution method according to the present disclosure;

[0047] Figure 7 It is a schematic flowchart of an embodiment of the model acquisition method according to the present disclosure;

[0048] Figure 8A It is a module schematic diagram of an embodiment of the model distribution device according to the present disclosure;

[0049] Figure 8BSchematic diagram of modules of another embodiment of the model distribution device according to the present disclosure;

[0050] Figure 8C Schematic diagram of modules of yet another embodiment of the model distribution device according to the present disclosure;

[0051] Figure 9A Schematic diagram of modules of an embodiment of the model acquisition device according to the present disclosure;

[0052] Figure 9B Schematic diagram of modules of another embodiment of the model acquisition device according to the present disclosure. Detailed implementation manners

[0053] In the following, exemplary embodiments of the present disclosure will be described in conjunction with the accompanying drawings. For clarity and conciseness, not all features of the embodiments are described in the specification. However, it should be understood that many implementation-specific settings must be made during the implementation of the embodiments to achieve the specific goals of the developers, for example, to comply with those restrictions related to the devices and services, and these restrictions may vary with different implementations. In addition, it should also be understood that although the development work may be very complex and time-consuming, for those skilled in the art who benefit from the present disclosure, such development work is only a routine task.

[0054] It should be noted that: Unless otherwise specifically stated, the relative arrangements, numerical expressions, and numerical values of the components and steps set forth in these embodiments do not limit the scope of the present disclosure.

[0055] Those skilled in the art can understand that the terms "first", "second", etc. in the embodiments of the present disclosure are only used to distinguish different steps, devices, or modules, etc., and neither represent any specific technical meaning nor indicate an inevitable logical order between them.

[0056] It should also be understood that in the embodiments of the present disclosure, "a plurality of" may refer to two or more, and "at least one" may refer to one, two, or more.

[0057] It should also be understood that for any component, data, or structure mentioned in the embodiments of the present disclosure, unless otherwise clearly defined or given a contrary indication in the context, it can generally be understood as one or more.

[0058] In addition, the term "and / or" in the present disclosure is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in the present disclosure generally represents an "or" relationship between the associated objects before and after.

[0059] It should also be understood that the descriptions of the various embodiments in the present disclosure emphasize the differences between the various embodiments, and the same or similar aspects can be referred to each other. For the sake of brevity, they will not be elaborated one by one.

[0060] At the same time, it should be understood that, for the sake of description convenience, the dimensions of the various parts shown in the drawings are not drawn according to the actual proportional relationship.

[0061] The following description of at least one exemplary embodiment is actually merely illustrative and in no way limits the present disclosure or its application or use.

[0062] Technologies, methods, and devices known to those of ordinary skill in the relevant art may not be discussed in detail, but where appropriate, the said technologies, methods, and devices should be regarded as part of the specification.

[0063] It should be noted that like reference numerals and letters denote like items in the following drawings, and thus, once an item is defined in one drawing, it does not need to be further discussed in subsequent drawings.

[0064] In addition, to avoid obscuring the present disclosure with unnecessary details, only the processing steps and / or device structures closely related to at least the solutions of the present disclosure are shown in the drawings, while other details less related to the present disclosure are omitted. It should also be noted that like reference numerals and letters in the drawings indicate like items, and thus once an item is defined in one drawing, it does not need to be further discussed for subsequent drawings.

[0065] Figure 1 As a flowchart of an embodiment of the model distribution method according to the present disclosure, the model distribution method of the present disclosure is applied to a network-side network element, such as Figure 1 shown:

[0066] Step 101, receive a model request sent by the UE through control plane signaling.

[0067] In one embodiment, the network-side network element can be a network element within a domain such as the 5GC domain. For example, the network-side network element can be NWDAF (Network Data Analytics Function), ADRF (Analytic Data Repository Function), LMF (Location Management Function), etc. The model includes various models such as a machine learning ML model. The model request includes information such as UE information and / or model request information.

[0068] Step 102: In response to the model request, send the first response message to the UE through the user plane network element, or send the second response message to the UE through the control plane network element.

[0069] The user plane network element can be multiple network elements, such as UPF (User Plane Function). The first response message includes information such as target model information matching the model request. The target model information includes information such as target model object information and / or the download address of the target model. The target model includes an ML target model, etc. The second response message includes information such as the model request rejection information. The model request rejection information includes indication information and / or reason information for the rejection of the model request, etc.

[0070] In one embodiment, the network side network element receives the model request sent by the control plane network element. The control plane signaling can be multiple types of signaling, such as NAS (Non-Access Stratum) signaling, etc. The control plane network element can be multiple network elements, such as an AMF (Access and Mobility Management Function) network element, etc. The UE sends control plane signaling such as NAS signaling to the control plane network element such as the AMF, and the control plane network element such as the AMF sends the model request to the network side network element based on the control plane signaling.

[0071] For example, the UE sends a request message to the AMF through NAS signaling. This request message contains UE information and corresponding model request information. The UE information includes at least one of information such as UE identification information and UE location information; the UE identification information is used to indicate the user who issues the model request, and the UE identification information includes at least one of information such as the UE IP address, GPSI (Generic Public Subscription Identifier), and SUPI (Subscription Permanent Identifier). The UE location information is used to indicate the location where the UE is located. For example, the UE location information is the longitude and latitude information of the location where the UE is located, etc.

[0072] The model request information includes at least one of the information such as the model ID, model-related information, etc. The model ID can be an ML model ID, etc., and is used to indicate the model ID requested by the UE. The model-related information is used to describe the information related to the model, and is used to assist the network-side network element in the authorizability and model selection of the model, and determine the model required by the UE. The model request information can be sent in the form of a corresponding container, or directly encapsulated in the NAS information together with the UE information. In the NAS signaling, various existing container encapsulation methods can be used to encapsulate the model request information. The model request information in the container is invisible to the AMF, and the AMF only performs a forwarding function on this container. The model request information can be encapsulated in the NAS signaling in a protocol encapsulation method. The protocol encapsulation method is to encapsulate the model request information in the NAS signaling using the existing NAS signaling format, and the model request information is visible to the AMF.

[0073] The model-related information includes at least one of the information such as model filter information, location range information applicable to the model, time range information for model use, use case information, use case context information, model interoperability information, manufacturer identification information, accuracy level information, etc.

[0074] The model filter information is the model filter information, which is used to provide the corresponding information for screening the model according to different user use cases. For example, it is S-NSSAI (Single Network Slice Selection Assistance Information), DNN (Data Network Name), UEID(s), etc. The location range information applicable to the model is AoI (Area of Interest), which is used to indicate the location range applicable to the requested model. Its value can be TAI (Tracking Area Identity), cell ID(s), etc., or the representation form of the regional area, or other forms.

[0075] The time range information for model use is the ML Model Target Period, which is used to indicate the time range for using the requested model. The use case information is use case (Analytics ID), which is used to indicate the use case applicable to the requested model, or the analysis type used by the model, including three use cases within the RAN domain (AI-based CSI feedback, AI-based positioning enhancement, and AI-based beam management), etc. It can be in the form of identification, character, numerical value, or other representation forms.

[0076] The use case context information is "use case context", which is used to indicate the refined information of the use case to which the requested model applies, assisting the network-side network element to further select a more relevant model, and can be in the form of an identifier, a character form, a numerical form, or other forms. The model interoperability information is "ML Model Interoperability Information", including the model file format, the model execution environment, the encoding method, etc. The vendor identification information is "vendor ID", which is used to indicate the vendor that provides the model. The accuracy level information is "preferred level of accuracy", which is used to indicate the accuracy that the requested model needs to meet, and can be in the form of "low", "medium", "high", or other forms.

[0077] In one embodiment, after receiving the control plane signaling sent by the UE, the control plane network element selects a network-side network element based on the UE information in the control plane signaling, or based on the UE information and the model request information in the control plane signaling, and generates a model request to send to the network-side network element.

[0078] For example, the AMF can retrieve in the NRF (Network Repository Function), etc., to determine network-side network elements such as NWDAF that can provide services such as model transfer for the UE. When the model request information is set in a container, the model request information is in an unresolvable state for the AMF, and the AMF can only determine network-side network elements such as NWDAF based on the UE information. When the model request information and the UE information are directly encapsulated in the NAS signaling, the AMF can resolve the model request information and the UE information, and can determine network-side network elements such as NWDAF based on the UE information and the model request information.

[0079] When the model request information is set in a container, the AMF generates a model request based on the container (i.e., the corresponding model request information) and sends it to network-side network elements such as NWDAF. When the model request information and the UE information are directly encapsulated in the NAS signaling, the AMF generates a model request with the UE information and / or the model request information and sends it to network-side network elements such as NWDAF.

[0080] In one embodiment, the network-side network element generates first response information and sends it to the UE through the user plane network element, or generates second response information and sends it to the UE through the control plane network element.

[0081] Figure 2 It is a schematic flow diagram of generating the first response information in an embodiment of the model distribution method according to the present disclosure, as Figure 2 shown:

[0082] Step 201, obtain the model subscription information of the UE.

[0083] Step 202, determine whether the UE is allowed to obtain the target model according to the model request and the model subscription information.

[0084] In one embodiment, the network-side network element is a network element such as NWDAF. The network-side network element obtains the model subscription information of the UE from a network element with a storage function such as UDM (Unified Data Management) to determine whether the UE is allowed to obtain the target model requested by the UE. The network-side network element can compare and process the UE information and / or the model request information in the model request with the model subscription information of the UE in the UDM, and determine whether the UE is allowed to obtain the target model requested by the UE according to the comparison result. The comparison processing results for determining that the UE is allowed to obtain the target model requested by the UE include at least one of the following:

[0085] Determine that the model subscription information of the UE corresponding to the UE identification information in the UE information includes the ability of the UE to obtain the model;

[0086] Determine that the UE location information in the UE information matches the location range where the model can be used in the model subscription information of the UE;

[0087] Determine that the model ID in the model request information matches the available model ID in the model subscription information of the UE;

[0088] Determine that the use case (Analytics ID) in the model request information matches the available use case (Analytics ID) in the model subscription information corresponding to the UE;

[0089] Determine that the vendor ID in the model request information matches the available vendor ID in the model subscription information of the UE.

[0090] If the network-side network element can determine that the UE has been authorized to obtain the target model requested by the UE, or it is verified by other network elements that the UE is allowed to obtain the target model, the network-side network element does not need to execute Step 201 and Step 202.

[0091] Step 203, generate the first response information when it is determined that the UE is allowed to obtain the target model and it is determined that the target model information can be obtained.

[0092] In one embodiment, the network-side network element may determine whether the target model information is stored locally according to the model request. The target model information may be information such as the model file of the target model and / or the download address of the target model. If the target model information is stored locally, the locally stored target model information is obtained. If the target model information is not stored locally, the target model information is obtained through the storage network element.

[0093] The network-side network element may determine whether the model file of the target model that meets the requirements and / or information such as the download address of the target model is stored locally according to the model request information in the model request. If the model information stored by the network-side network element meets at least one of the following judgment results, it is determined that the target model information is stored locally, and the locally stored target model information can be obtained; wherein, the above judgment results are as follows:

[0094] Determine that the model filter information in the model request information matches the filter information corresponding to a certain model;

[0095] Determine that the AoI information in the model request information matches the AoI information defined by a certain model;

[0096] Determine that the ML Model Target Period in the model request information matches the ML Model Target Period defined by a certain model;

[0097] Determine that the use case (Analytics ID) in the model request information matches the use case (Analytics ID) applicable to a certain model;

[0098] Determine that the use case context in the model request information matches the use case context corresponding to a certain model;

[0099] Determine that the ML Model Interoperability Information in the model request information matches the corresponding attributes of a certain model (i.e., model file format, model execution environment, encoding method, etc.);

[0100] Determine that the vendor ID in the model request information matches the vendor ID of the manufacturer that provides a certain model;

[0101] Determine that the preferred level of accuracy in the model request information matches or is higher than the accuracy level in the request for a certain model.

[0102] In one embodiment, the network-side network element may use multiple methods to send the first response information to the user-plane network element, so that the user-plane network element sends the first response information to the UE.

[0103] Figure 3 FIG. is a schematic flowchart of sending the first response information in an embodiment of the model distribution method according to the present disclosure, as Figure 3 shown:

[0104] Step 301, determine whether the AF (Application Function) is a trusted AF; if so, go to step 302, if not, go to step 303.

[0105] Step 302, send the first response information to the AF, so that the AF sends the first response information to the user-plane network element.

[0106] Step 303, select the NEF according to the model request, send the first response information to the NEF, and the NEF sends the first response information to the user-plane network element through the AF.

[0107] In one embodiment, when it is determined that the AF is a trusted AF, the network-side network element sends the first response information to the AF. When it is determined that the AF is a non-trusted AF, the first response information is sent to the NEF, and the NEF sends the first response information to the user-plane network element through the AF.

[0108] The network-side network element may query in the NRF or the like based on one or more items of information such as UE information in the model request to obtain an AF that can provide services such as model transfer for the UE. If the AF is a non-trusted AF, the network-side network element may query in the NRF or the like based on one or more items of information such as UE information to obtain a NEF (Network Exposure Function) that can provide services such as model transfer for the UE. The NEF may determine according to the operator's configuration that the information in the first response information can be opened to the AF.

[0109] The first response information includes target model information, and the target model information includes target model object information and / or information such as the download address of the target model. The target model object information includes the model file of the target model, and the download address of the target model is an address indicating the download of the target model, which may be a URL address, a FQDN (Fully Qualified Domain Name) address, or other forms of addresses.

[0110] The first response information may also include one or more of the following information: one or several items of model-related information, validity period, spatial validity, subscription correlation ID, etc. The validity period is the "validity period", which is used to indicate the time period for which the provided model is applicable; the spatial validity is the "Spatial validity", which is used to indicate the scope to which the provided model is applicable; the subscription correlation ID is the "Subscription Correlation ID", which is used to identify the corresponding subscription to the model and can be used for subsequent subscription modification and cancellation, etc.

[0111] In one embodiment, when the network-side network element determines that the UE is not allowed to obtain the target model or cannot obtain the target model information according to the model request and the model subscription information, it determines the information that the model request is rejected and generates the second response information. The network-side network element sends the second response information to a control-plane network element such as the AMF, so that the control-plane network element sends the second response information to the UE, and the second response information represents the rejection of the UE's model request.

[0112] The model request rejection information includes information such as the indication information that the model request is rejected and / or the reason information. The indication information that the model request is rejected is used to indicate that the model request is rejected. The reason information is used to indicate the reason for the rejection of the model request. For example, the reason information may be that the UE is not allowed to use the model, no compliant model is found, the requested model is not within the usage range, the requested vendor ID does not exist, the preferred level of accuracy is not met, the usecase (Analytics ID) does not match, etc.

[0113] Figure 4 FIG. is a schematic flow diagram of obtaining target model information through a storage network element in an embodiment of the model distribution method according to the present disclosure, as Figure 4 shown:

[0114] Step 401, generate retrieval information based on the model request and send the retrieval information to the storage network element.

[0115] Step 402, obtain the detection result returned by the storage network element; wherein, the retrieval result includes information such as the target model information or the reason information for not retrieving the target model information.

[0116] In one embodiment, when the network-side network element does not store the target model information, the network-side network element generates retrieval information based on the model request, and the retrieval information includes one or more items of information such as the network-side network element ID and the model request information. The network-side network element sends the retrieval information to the storage network element, and the storage network element may be an ADRF or other network elements with storage functions.

[0117] The method for the storage network element to obtain the target model information is the same as that for the network - side network element to obtain the target model information. The storage network element determines that there is target model information that meets the UE's requirements, and sends the retrieval result to the network - side network element. The retrieval result includes the target model information, etc. The storage network element determines that there is no target model information that meets the UE's requirements, and sends the retrieval result to the network - side network element. The retrieval result includes the reason information for not retrieving the target model information; the reason information for not retrieving the target model information is used to indicate the reason for the model request being rejected. For example, the reason information can be that the requested model is not within the usage range, the requested vendor ID does not exist, the preferred level of accuracy is not met, the use case (AnalyticsID) does not match, etc.

[0118] Figure 5 It is a schematic diagram of information interaction for another embodiment of the model distribution method according to the present disclosure. Among them, the network - side network element is NWDAF, the control - plane network element is AMF, and the user - plane network element is UPF, as Figure 5 shown:

[0119] Step 500, the UE sends a NAS signaling to the UPF; the information carried in the NAS signaling includes UE information, model request information, etc.

[0120] Step 501, the UPF forwards the NAS signaling to the AMF.

[0121] Step 502, the AMF searches for an NWDAF that can provide services such as model transfer for the UE.

[0122] Step 503, the AMF sends a model request to the NWDAF. The model request includes UE information and / or model request information.

[0123] Step 504, the NWDAF confirms whether the UE has the authorization for the model. The NWDAF obtains the UE's model subscription information from the UDM and determines whether the UE is allowed to obtain the model; if the UE is allowed to obtain the model, then step 505 is executed; if the UE is not allowed to obtain the model, then step 511 is executed.

[0124] Step 505, the NWDAF searches whether it stores a model object or a model object storage address that meets the requirements. If not stored, then step 511 is executed.

[0125] Step 506a, if the AF is a non - trusted AF, the NWDAF searches for a NEF that can provide services such as model transfer for the UE.

[0126] Step 506b, if the AF is a trusted AF, the NWDAF searches for an AF that can provide services such as model transfer for the UE.

[0127] Step 507a-1, if the AF is a non-trusted AF, the NWDAF sends a first response message to the NEF. The first response message includes target model information that matches the model request, and the target model information includes target model object information and / or the download address of the target model.

[0128] Step 507a-2, if the AF is a non-trusted AF, the NEF forwards the first response message to the AF.

[0129] Step 507b, if the AF is a trusted AF, the NWDAF sends the first response message to the AF.

[0130] Step 508, the AF sends the first response message to the UPF through the user plane.

[0131] Step 509, the UPF sends the first response message to the UE.

[0132] Step 510, if the first response message contains a model download address, the UE downloads the target model according to the model download address.

[0133] Step 511, the NWDAF rejects the UE's model request. The NWDAF indirectly sends a second response message to the UE from the control plane through the AMF. The second response message includes information that the model request is rejected, and the information that the model request is rejected includes indication information and / or reason information that the model request is rejected.

[0134] Figure 6 It is a schematic diagram of information interaction for obtaining target model information through a storage network element in another embodiment of the model distribution method according to the present disclosure. Here, the storage network element is an ADRF, and the ADRF can store information such as the target model and / or the model address, etc.; in the case where the target model is not stored in the NWDAF, the NWDAF executes steps 605-1 to 605-3, as Figure 6 shown:

[0135] Step 605-1, the NWDAF sends retrieval information to the ADRF to find a model object or the storage address of the model object that matches the UE's model request, etc.

[0136] Step 605-2, the ADRF performs a model search operation.

[0137] Step 605-3, the ADRF returns a detection result to the NWDAF. The retrieval result includes target model information or reason information for not retrieving the target model information.

[0138] Figure 7FIG. 0 is a schematic flowchart of an embodiment of a model acquisition method according to the present disclosure. The model acquisition method of the present disclosure is applied to a UE, as Figure 7 shown:

[0139] Step 701, send a model request to a network-side network element through control plane signaling.

[0140] Step 702, receive first response information sent by the network-side network element through the user plane network element, or receive second response information sent by the network-side network element through the control plane network element.

[0141] For example, the UE sends control plane signaling to the control plane network element, and the control plane network element sends a model request to the network-side network element based on the control plane signaling. The control plane network element selects the network-side network element based on the UE information in the control plane signaling, or based on the UE information and the model request information in the control plane signaling, and generates a model request to send to the network-side network element.

[0142] In the model distribution and model acquisition methods in the above embodiments, the UE sends a model request to the network-side network element through the control plane, without expanding the interaction mode between the UE and the network-side network element on the user plane; the network-side network element sends the target model information to the UE through the user plane, only expanding functions such as data transfer on the user plane; it can ensure the security of the interaction and is suitable for large data volume transmission, can realize cross-domain model transfer, can distribute the model in domains such as the 5GC domain to the UE side, train, store, manage, etc. the model through domains such as the 5GC domain, has strong security and can strengthen the operator's supervision of the model, and can make full use of resources such as storage and computing power in domains such as the 5GC domain; it can improve the user experience.

[0143] In one embodiment, as Figure 8A shown, the present disclosure provides a model distribution device 80, which is applied to a network-side network element. The model distribution device 80 includes a receiving module 81 and a response module 82. The receiving module 81 receives a model request sent by a user equipment UE through control plane signaling. The response module 82, in response to the model request, sends first response information to the UE through the user plane network element, or sends second response information to the UE through the control plane network element.

[0144] In one embodiment, the receiving module 81 receives a model request sent by the control plane network element, where the UE sends control plane signaling to the control plane network element, and the control plane network element sends a model request to the network-side network element based on the control plane signaling. The response module 82 generates first response information and sends it to the UE through the user plane network element, or generates second response information and sends it to the UE through the control plane network element.

[0145] As Figure 8BAs shown in the figure, the model distribution device 80 further includes a decision module 83. The decision module 83 obtains the model subscription information of the UE, and determines whether the UE is allowed to obtain the target model according to the model request and the model subscription information. The response module 82 generates a first response message when it is determined that the UE is allowed to obtain the target model and it is determined that the target model information can be obtained.

[0146] The response module 82 may send the first response message to the user plane network element so that the user plane network element sends the first response message to the UE. The response module 82 may also send the first response message to the user plane network element through the AF so that the user plane network element sends the first response message to the UE.

[0147] For example, when the response module 82 determines that the AF is a trusted AF, it sends the first response message to the AF so that the AF sends the first response message to the user plane network element. When the response module 82 determines that the AF is a non-trusted AF, it selects the NEF according to the model request; the response module 82 sends the first response message to the NEF, and the NEF sends the first response message to the user plane network element through the AF.

[0148] When the response module 82 determines that the UE is not allowed to obtain the target model according to the model request and the model subscription information, or the target model information cannot be obtained, it determines the information that the model request is rejected and generates a second response message. The response module 82 sends the second response message to the control plane network element so that the control plane network element sends the second response message to the UE.

[0149] In one embodiment, as Figure 8B shown, the model distribution device 80 further includes an acquisition module 84. The acquisition module 84 determines whether the target model information is stored locally according to the model request; if so, the acquisition module 84 obtains the target model information stored locally, and if not, the acquisition module 84 obtains the target model information through the storage network element.

[0150] The acquisition module 84 may generate retrieval information based on the model request and send the retrieval information to the storage network element; the acquisition module 84 obtains the detection result returned by the storage network element, and the retrieval result includes the target model information or the reason information for not retrieving the target model information.

[0151] In one embodiment, as Figure 8C shown, the model distribution device may include a memory 85, a processor 86, a communication interface 87, and a bus 88. The memory 85 is used to store instructions, and the processor 86 is coupled to the memory 85. The processor 86 is configured to execute the above-mentioned model distribution method based on the instructions stored in the memory 85.

[0152] The memory 85 can be a high-speed RAM memory, a non-volatile memory, etc. The memory 85 can also be a memory array. The memory 85 may also be partitioned, and the partitions can be combined into virtual volumes according to certain rules. The processor 86 can be a central processing unit CPU, or an application specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the model distribution method of the present disclosure.

[0153] In one embodiment, the present disclosure provides a network-side network element, including the model distribution device in any of the above embodiments.

[0154] In one embodiment, as Figure 9A shown, the present disclosure provides a model acquisition device 90, which is applied to a UE. The model acquisition device 90 includes a request sending module 91 and a response receiving module 92. The request sending module 91 sends a model request to the network-side network element through control plane signaling. The response receiving module 92 receives the first response information sent by the network-side network element through the user plane network element, or receives the second response information sent by the network-side network element through the control plane network element.

[0155] For example, the request sending module 91 sends control plane signaling to the control plane network element, and the control plane network element sends a model request to the network-side network element based on the control plane signaling. The response receiving module 92 receives the first response information sent by the network-side network element through the UPF; the response receiving module 92 receives the second response information sent by the network-side network element through the AMF.

[0156] In one embodiment, as Figure 9B shown, the model distribution device may include a memory 93, a processor 94, a communication interface 95, and a bus 96. The memory 93 is used to store instructions. The processor 94 is coupled to the memory 93, and the processor 94 is configured to execute to implement the above model acquisition method based on the instructions stored in the memory 93.

[0157] The memory 93 can be a high-speed RAM memory, a non-volatile memory, etc. The memory 93 can also be a memory array. The memory 93 may also be partitioned, and the partitions can be combined into virtual volumes according to certain rules. The processor 94 can be a central processing unit CPU, or an application specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the model acquisition method of the present disclosure.

[0158] In one embodiment, the present disclosure provides a UE including a model acquisition device in any of the above embodiments.

[0159] In one embodiment, the present disclosure provides a communication system including: a network-side network element in any of the above embodiments and a UE in any of the above embodiments. The communication system may be a 5G system, a 6G system, etc.

[0160] In one embodiment, the present disclosure provides a computer-readable storage medium storing computer instructions that, when executed by a processor, implement the method in any of the above embodiments.

[0161] The computer-readable storage medium may employ any combination of one or more readable media. The readable media may be a readable signal medium or a readable storage medium. The readable storage medium may, for example, include but is not limited to an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (non-exhaustive list) of the readable storage medium may include: an electrical connection having one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0162] An embodiment of the present disclosure may also be a computer program product including computer program instructions that, when run by a processor, cause the processor to execute the steps in the methods according to various embodiments of the present disclosure described in the "Exemplary Methods" section above of this specification.

[0163] In the above embodiments, for the model distribution, model acquisition methods and devices, UE, network-side network element, communication system, and storage medium, the UE sends a model request to the network-side network element through the control plane, without expanding the interaction mode between the UE and the network-side network element in the user plane, which can ensure the security of the interaction; the network-side network element sends target model information to the UE through the user plane, only needs to expand functions such as data transfer in the user plane, and is suitable for large data volume transmission; it can realize cross-domain model transfer, can distribute the model in domains such as 5GC to the UE side, train, store, manage, etc. the model through domains such as 5GC, has strong security and can strengthen the operator's supervision of the model, and can make full use of resources such as storage and computing power in domains such as 5GC; it can improve the user experience.

[0164] The basic principles of the present disclosure have been described above in connection with specific embodiments. However, it should be noted that the advantages, benefits, effects, etc. mentioned in the present disclosure are only examples and not limitations, and it cannot be considered that these advantages, benefits, effects, etc. are essential for each embodiment of the present disclosure. Additionally, the specific details disclosed above are only for illustrative and easy-to-understand purposes and not limitations, and the above details do not limit the present disclosure to necessarily implementing with the above specific details.

[0165] Each embodiment in this specification is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. For the same or similar parts among the embodiments, reference can be made to each other. For system embodiments, since they basically correspond to method embodiments, the description is relatively simple, and reference can be made to the corresponding parts of the method embodiments for relevant content.

[0166] The block diagrams of the devices, apparatuses, equipment, and systems involved in the present disclosure are only illustrative examples and do not intend to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open-ended terms, meaning "including but not limited to", and can be used interchangeably with each other. The word "or" and "and" used herein refer to the word "and / or", and can be used interchangeably with each other, unless the context clearly indicates otherwise. The word "such as" used herein refers to the phrase "such as but not limited to", and can be used interchangeably with each other.

[0167] It should also be noted that in the devices, equipment, and methods of the present disclosure, each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent solutions of the present disclosure.

[0168] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these aspects are very obvious to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of the present disclosure. Therefore, the present disclosure is not intended to be limited to the aspects shown herein, but rather to the broadest scope consistent with the principles and novel features disclosed herein.

[0169] The above description has been given for purposes of illustration and description. In addition, this description does not intend to limit the embodiments of the present disclosure to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art should understand that the above embodiments are merely illustrative and do not limit the scope of the present disclosure. Those skilled in the art should understand that the above embodiments can be combined, modified, or replaced without departing from the scope and essence of the present disclosure.

Claims

1. A model distribution method, applied to a network-side network element, comprising: receiving a model request sent by a user equipment (UE) through control plane signaling; in response to the model request, sending the first response information to the UE through a user plane network element, or sending second response information to the UE through the control plane network element; wherein, the first response information includes: target model information matching the model request; the second response information includes: information indicating that the model request is rejected.

2. The method according to claim 1, wherein, the model request includes: the UE information and / or the model request information; wherein, the UE information includes: UE identification information and / or UE location information; the model request information includes: model ID and / or model-related information.

3. The method according to claim 2, wherein, the model-related information includes at least one of: model filter information, location range information applicable to the model, time range information for using the model, use case information, use case context information, model interoperability information, manufacturer identification information, accuracy level information.

4. The method according to claim 2, wherein, the receiving the model request sent by the UE through control plane signaling includes receiving the model request sent by the control plane network element; wherein, the UE sends the control plane signaling to the control plane network element, and the control plane network element sends the model request to the network-side network element based on the control plane signaling.

5. The method according to claim 4, wherein, the control plane network element sending the model request to the network-side network element based on the control plane signaling includes: the control plane network element selects the network-side network element based on the UE information in the control plane signaling, or based on the UE information and the model request information in the control plane signaling, and generates the model request to send to the network-side network element.

6. The method according to claim 5, wherein, the control plane signaling includes: NAS signaling; wherein, the model request information is encapsulated in the NAS signaling in a container encapsulation manner or a protocol encapsulation manner.

7. The method according to any one of claims 1 to 6, the sending the first response information to the UE through a user plane network element, or sending second response information to the UE through the control plane network element comprising: generating the first response information and sending it to the UE through a user plane network element, or generating the second response information and sending it to the UE through the control plane network element.

8. The method according to claim 7, the generating the first response information comprising: obtaining the model subscription information of the UE; judging whether the UE is allowed to obtain the target model according to the model request and the model subscription information; when it is determined that the UE is allowed to obtain the target model and it is determined that the target model information can be obtained, generating the first response information.

9. The method according to claim 8, wherein, the sending the first response information to the UE through a user plane network element includes: Send the first response information to the user plane network element so that the user plane network element sends the first response information to the UE.

10. The method according to claim 8, wherein, sending the first response information to the UE through the user plane network element includes: Sending the first response information to the user plane network element through the Application Function (AF) so that the user plane network element sends the first response information to the UE.

11. The method according to claim 10, wherein, sending the first response information to the user plane network element through the AF includes: When it is determined that the AF is a trusted AF, send the first response information to the AF so that the AF sends the first response information to the user plane network element.

12. The method according to claim 10, wherein, sending the first response information to the user plane network element through the AF includes: When it is determined that the AF is an untrusted AF, select the Network Exposure Function (NEF) according to the model request; Send the first response information to the NEF, and the NEF sends the first response information to the user plane network element through the AF.

13. The method according to claim 8, wherein, generating the second response information and sending it to the UE through the control plane network element includes: When it is determined according to the model request and the model subscription information that the UE is not allowed to obtain the target model or cannot obtain the target model information, determine the information that the model request is rejected and generate the second response information; Send the second response information to the control plane network element so that the control plane network element sends the second response information to the UE.

14. The method according to claim 8, further including: According to the model request, determine whether the target model information is stored locally; If so, obtain the target model information stored locally, if not, obtain the target model information through the storage network element.

15. The method according to claim 14, wherein, obtaining the target model information through the storage network element includes: Generate retrieval information based on the model request and send the retrieval information to the storage network element; Obtain the detection result returned by the storage network element; wherein, the retrieval result includes: the target model information, or the reason information for not retrieving the target model information.

16. The method according to claim 1, wherein, the target model information includes: target model object information and / or the download address of the target model; the information that the model request is rejected includes: the indication information that the model request is rejected and / or the reason information.

17. The method according to claim 1, wherein, the network side network elements include: Network Data Analytics Function (NWDAF), Analytics Database Function (ADRF), Location Management Function (LMF); the control plane network element includes: Access and Mobility Management Function (AMF); the user plane network element includes: User Plane Function (UPF); The model includes: a machine learning ML model; the target model includes: an ML target model.

18. A method for obtaining a model, applied to a user equipment UE, including: sending a model request to a network element on the network side through control plane signaling; receiving first response information sent by the network element on the network side through a user plane element, or receiving second response information sent by the network element on the network side through a control plane element; wherein, the first response information includes: target model information matching the model request; the second response information includes: information indicating that the model request is rejected.

19. The method according to claim 18, wherein, the model request includes: the UE information and / or the model request information; wherein, the UE information includes: UE identification information and / or UE location information; the model request information includes: model ID and / or model-related information.

20. The method according to claim 19, wherein, the model-related information includes at least one of: model filter information, location range information applicable to the model, time range information for using the model, use case information, use case context information, model interoperability information, manufacturer identification information, accuracy level information.

21. The method according to claim 19, wherein, sending the model request to the network element on the network side through control plane signaling includes: sending the control plane signaling to the control plane element; wherein, the control plane element sends the model request to the network element on the network side based on the control plane signaling.

22. The method according to claim 21, wherein, the control plane element sending the model request to the network element on the network side based on the control plane signaling includes: the control plane element selects the network element on the network side based on the UE information in the control plane signaling, or based on the UE information and the model request information in the control plane signaling, and generates the model request to send to the network element on the network side.

23. The method according to claim 22, wherein, the control plane signaling includes: NAS signaling; wherein, the model request information is encapsulated in the NAS signaling in a container encapsulation manner or a protocol encapsulation manner.

24. The method according to claim 18, wherein, the user plane element includes: UPF; receiving the first response information sent by the network element on the network side through the user plane element includes: receiving the first response information sent by the network element on the network side through the UPF.

25. The method according to claim 18, wherein, the control plane element includes: AMF; receiving the second response information sent by the network element on the network side through the control plane element includes: receiving the second response information sent by the network element on the network side through the AMF.

26. The method according to claim 18, wherein, the target model information includes: target model object information and / or the download address of the target model; the information indicating that the model request is rejected includes: indication information and / or reason information that the model request is rejected.

27. The method according to claim 18, wherein, The network-side network element includes: NWDAF, ADRF, and LMF; The model includes: a machine learning ML model; the target model includes: an ML target model.

28. A model distribution device, applied to a network-side network element, comprising: a receiving module, configured to receive a model request sent by a user equipment UE through control plane signaling; a response module, configured to, in response to the model request, send first response information to the UE through a user plane network element, or send second response information to the UE through a control plane network element; wherein, the first response information includes: target model information matching the model request; the second response information includes: information indicating that the model request is rejected.

29. A model distribution device, applied to a network-side network element, comprising: a memory; and a processor coupled to the memory, the processor being configured to execute the method according to any one of claims 1 to 17 based on instructions stored in the memory.

30. A network-side network element, comprising: the model distribution device according to claim 28 or 29.

31. A model acquisition device, applied to a user equipment UE, comprising: a request sending module, configured to send a model request to a network-side network element through control plane signaling; a response receiving module, configured to receive the first response information sent by the network-side network element through a user plane network element, or receive the second response information sent by the network-side network element through a control plane network element; wherein, the first response information includes: machine learning target model information matching the model request; the second response information includes: information indicating that the model request is rejected.

32. A model acquisition device, applied to a UE, comprising: a memory; and a processor coupled to the memory, the processor being configured to execute the method according to any one of claims 18 to 27 based on instructions stored in the memory.

33. A UE, comprising: the model acquisition device according to claim 31 or 32.

34. A communication system, comprising: the network-side network element according to claim 30, and the UE according to claim 33.

35. A computer-readable storage medium, storing computer instructions, the instructions being executed by a processor to perform the method according to any one of claims 1 to 27.

Images