Model update method, device and apparatus, and storage medium

Abstract

Provided in the present disclosure are a model update method, device and apparatus, and a storage medium. The method comprises: a data analytics function entity of a terminal sending a first data collection request to a data collection function entity of the terminal; the data analytics function entity of the terminal receiving first data provided by the data collection function entity of the terminal on the basis of the first data collection request; and the data analytics function entity of the terminal performing inference on the basis of the first data, and updating a local model on the basis of an inference result.

Description

Model updating method, device, apparatus, and storage medium

[0001] Cross-reference to Related Applications

[0002] This application claims priority to the Chinese Patent Application No. 202410945294.9, filed on July 15, 2024, and entitled “Model updating method, device, apparatus, and storage medium”, which is incorporated by reference in its entirety. TECHNICAL FIELD

[0003] The present disclosure relates to the technical field of wireless communication, and particularly relates to a model updating method, device, apparatus, and storage medium. BACKGROUND

[0004] The 3rd Generation Partnership Project (3GPP) defines a Network Data Analytics Function (NWDAF), which can perform mobility prediction of a terminal (or User Equipment, UE) based on an Artificial Intelligence (AI) model (hereinafter referred to as a model). The data required for model training mainly comes from an Access and Mobility Management Function (AMF) (in the 5th Generation mobile communication (5G) system). When the UE changes location, some communication events of the base station are triggered, and these events are notified as data (mainly information about the base station transition of the terminal) sent by the AMF to the NWDAF. However, if the terminal moves within the area of the Tracking Area (TA) list after entering the idle mode, there will be no base station transition event, which leads to the fact that the AMF cannot provide sufficient model training data for the NWDAF, and directly affects the inference performance of the model. SUMMARY

[0005] The present disclosure provides a model updating method, device, apparatus, and storage medium to solve the problem that the AMF cannot provide sufficient model training data for the NWDAF, thereby affecting the inference performance of the model.

[0006] In a first aspect, the present disclosure provides a model updating method applied to a terminal, comprising:

[0007] The data analytics function entity of the terminal sends a first data collection request to the data collection function entity of the terminal;

[0008] The data analysis function entity of the terminal receives first data provided by the data collection function entity of the terminal based on the first data collection request;

[0009] The data analysis function entity of the terminal performs inference based on the first data, and updates the local model based on an inference result.

[0010] In some embodiments, the method further comprises:

[0011] The data analysis function entity of the terminal sends model parameters of the updated local model to the network node;

[0012] The data analysis function entity of the terminal receives model parameters of the global model sent by the network node, and updates the local model based on the model parameters of the global model;

[0013] The network node comprises a centralized network node or a distributed network node.

[0014] In some embodiments, the first data collection request contains one or more of the following information:

[0015] Cell identity;

[0016] Base station identity;

[0017] Timestamp;

[0018] Time range information.

[0019] In some embodiments, the method further comprises:

[0020] The data analysis function entity of the terminal sends a first model download request to the centralized network node or the distributed network node, and the first model download request contains one or more of the following information:

[0021] User hidden identifier of the terminal;

[0022] Identification code of the data analysis function entity of the terminal;

[0023] Model type information.

[0024] In some embodiments, the local model comprises a terminal mobility prediction model.

[0025] In a second aspect, the disclosure also provides a model updating method, applied to an access network device, comprising:

[0026] The data analysis function entity of the access network device sends a second data collection request to the data collection function entity of the terminal;

[0027] The data analysis function entity of the access network device receives second data provided by the data collection function entity of the terminal based on the second data collection request;

[0028] The data analysis function entity of the access network device performs inference based on the second data, and updates the local model based on the inference result.

[0029] In some embodiments, the method further comprises:

[0030] The data analysis function entity of the access network device sends model parameters of the updated local model to the network node;

[0031] The data analysis function entity of the access network device receives model parameters of the global model sent by the network node, and updates the local model based on the model parameters of the global model;

[0032] The network node comprises a centralized network node or a distributed network node.

[0033] In some embodiments, the second data collection request contains one or more of the following information:

[0034] Cell identity;

[0035] Base station identity;

[0036] Timestamp;

[0037] Time range information.

[0038] In some embodiments, the method further comprises:

[0039] The data analysis function entity of the access network device sends a second model download request to the centralized network node or the distributed network node, and the second model download request contains one or more of the following information:

[0040] User hidden identifier of one or more terminals;

[0041] Identification code of the data analysis function entity of the access network device;

[0042] Model type information.

[0043] In some embodiments, the local model comprises a terminal mobility prediction model.

[0044] In a third aspect, the disclosure also provides a terminal comprising a memory, a transceiver, and a processor.

[0045] The memory is configured to store a computer program; the transceiver is configured to transceive data under the control of the processor; and the processor is configured to read the computer program in the memory and perform the following operations:

[0046] The data analysis function entity of the terminal sends a first data collection request to the data collection function entity of the terminal;

[0047] The data analysis function entity of the terminal receives first data provided by the data collection function entity of the terminal based on the first data collection request;

[0048] The data analysis function entity of the terminal performs inference based on the first data, and updates the local model based on an inference result.

[0049] In some embodiments, the operation further includes:

[0050] The data analysis function entity of the terminal sends model parameters of the updated local model to the network node;

[0051] The data analysis function entity of the terminal receives model parameters of the global model sent by the network node, and updates the local model based on the model parameters of the global model;

[0052] The network node includes a centralized network node or a distributed network node.

[0053] In some embodiments, the first data collection request contains one or more of the following information:

[0054] Cell identity;

[0055] Base station identity;

[0056] Timestamp;

[0057] Time range information.

[0058] In some embodiments, the operation further includes:

[0059] The data analysis function entity of the terminal sends a first model download request to the centralized network node or the distributed network node, and the first model download request contains one or more of the following information:

[0060] User hidden identifier of the terminal;

[0061] Identification code of the data analysis function entity of the terminal;

[0062] Model type information.

[0063] In some embodiments, the local model includes a terminal mobility prediction model.

[0064] In a fourth aspect, the disclosure also provides an access network device, including a memory, a transceiver, and a processor;

[0065] The memory is configured to store a computer program; the transceiver is configured to transceive data under the control of the processor; and the processor is configured to read the computer program in the memory and perform the following operations:

[0066] The data analysis function entity of the access network device sends a second data collection request to the data collection function entity of the terminal;

[0067] The data analysis function entity of the access network device receives second data provided by the data collection function entity of the terminal based on the second data collection request;

[0068] The data analysis function entity of the access network device performs inference based on the second data, and updates the local model based on the inference result.

[0069] In some embodiments, the operation further includes:

[0070] The data analysis function entity of the access network device sends model parameters of the updated local model to the network node;

[0071] The data analysis function entity of the access network device receives model parameters of the global model sent by the network node, and updates the local model based on the model parameters of the global model;

[0072] The network node includes a centralized network node or a distributed network node.

[0073] In some embodiments, the second data collection request contains one or more of the following information:

[0074] Cell identity;

[0075] Base station identity;

[0076] Timestamp;

[0077] Time range information.

[0078] In some embodiments, the operation further includes:

[0079] The data analysis function entity of the access network device sends a second model download request to the centralized network node or the distributed network node, and the second model download request contains one or more of the following information:

[0080] User hidden identifier of one or more terminals;

[0081] Identification code of the data analysis function entity of the access network device;

[0082] Model type information.

[0083] In some embodiments, the local model includes a terminal mobility prediction model.

[0084] In a fifth aspect, the disclosure also provides a model updating apparatus, comprising:

[0085] The first sending unit is configured to send, by a data analysis function entity of a terminal, a first data collection request to a data collection function entity of the terminal.

[0086] The first receiving unit is configured to receive, by the data analysis function entity of the terminal, first data provided by the data collection function entity of the terminal based on the first data collection request.

[0087] The first updating unit is configured to update, by the data analysis function entity of the terminal, a local model based on a result of inference based on the first data.

[0088] In a sixth aspect, the present disclosure further provides a model updating apparatus, comprising:

[0089] The second sending unit is configured to send, by a data analysis function entity of an access network device, a second data collection request to a data collection function entity of a terminal.

[0090] The second receiving unit is configured to receive, by the data analysis function entity of the access network device, second data provided by the data collection function entity of the terminal based on the second data collection request.

[0091] The second updating unit is configured to update, by the data analysis function entity of the access network device, a local model based on a result of inference based on the second data.

[0092] In a seventh aspect, the present disclosure further provides a non-transitory readable storage medium, which stores a program for causing a processor to execute the model updating method according to the first aspect or the model updating method according to the second aspect.

[0093] In an eighth aspect, the present disclosure further provides a communication device, which stores a program for causing the communication device to execute the model updating method according to the first aspect or the model updating method according to the second aspect.

[0094] In a ninth aspect, the present disclosure further provides a processor-readable storage medium, which stores a program for causing a processor to execute the model updating method according to the first aspect or the model updating method according to the second aspect.

[0095] In a tenth aspect, the present disclosure further provides a chip product, which stores a program for causing the chip product to execute the model updating method according to the first aspect or the model updating method according to the second aspect.

[0096] The model update method, device, apparatus, and storage medium disclosed herein request data from the terminal's data collection function entity through the terminal's data analysis function entity, and perform inference and update the local model based on the data provided by the terminal's data collection function entity. This ensures that the terminal's local model has sufficient training data to support it, thereby avoiding the problem of insufficient training data affecting model inference performance and effectively improving the model's inference performance. Furthermore, performing local analysis and inference on the terminal side can make the terminal more intelligent, reduce dependence on the network core, and also help enhance user privacy protection and data security. Attached Figure Description

[0097] To more clearly illustrate the technical solutions in the embodiments or related technologies of this disclosure, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0098] Figure 1 is a schematic flowchart of one of the model update methods provided in the embodiments of this disclosure.

[0099] Figure 2 is a second schematic flowchart of the model update method provided in the embodiments of this disclosure.

[0100] Figure 3 is an example diagram of a network architecture applicable to 6G distributed networks provided in the embodiments of this disclosure.

[0101] Figure 4 is a flowchart of Example 1 provided in the embodiments of this disclosure.

[0102] Figure 5 is a flowchart of Example 2 provided in the embodiments of this disclosure.

[0103] Figure 6 is a flowchart of Example 3 provided in the embodiments of this disclosure.

[0104] Figure 7 is a flowchart of Example 4 provided in the embodiments of this disclosure.

[0105] Figure 8 is a schematic diagram of the structure of the terminal provided in the embodiment of this disclosure.

[0106] Figure 9 is a schematic diagram of the structure of the access network device provided in an embodiment of this disclosure.

[0107] Figure 10 is one of the structural schematic diagrams of the model update device provided in the embodiments of this disclosure.

[0108] Figure 11 is a second schematic diagram of the structure of the model update device provided in the embodiments of this disclosure. Detailed Implementation

[0109] The term "and / or" in the embodiments of the present disclosure describes the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B can represent the following three cases: A exists alone, A and B exist together, and B exists alone. The character " / " generally represents an "or" relationship between the associated objects before and after it.

[0110] The term "a plurality of" in the embodiments of the present disclosure means two or more, and other quantifiers are similar.

[0111] The terms "first", "second", and the like in the embodiments of the present disclosure are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the terms used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present disclosure can be implemented in an order other than those illustrated or described here, and the objects distinguished by "first", "second" are usually of the same type and do not limit the number of objects, for example, the first object can be one or more.

[0112] The technical solutions in the embodiments of the present disclosure will be described clearly and completely in combination with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, not all. Based on the embodiments in the present disclosure, all other embodiments obtained by those of ordinary skill in the art without creative labor are within the scope of protection of the present disclosure.

[0113] In order to more clearly understand the technical solutions of the embodiments of the present disclosure, first, some technical contents related to the embodiments of the present disclosure are introduced.

[0114] 1、5G network data collection

[0115] 3GPP defines network data analysis function (NWDAF), data collection coordination function (DCCF), messaging framework adaptor function (MFAF), management data analytics management function (MDAMF), etc., mainly focusing on collecting data to provide data analysis reports to management data analysis consumers.

[0116] The data service framework defined by the Zero-touch network and Service Management (ZSM) working group of the European Telecommunications Standards Institute (ETSI) includes data persistence and data sharing between different domains, supports different storage and database technologies and can automatically select and adapt to the corresponding technology, and can provide management data such as performance, alarm, tracking, configuration, log, network topology, directory data, etc.

[0117] 2. UE trajectory prediction (or mobility prediction)

[0118] In the 5G system, the data required for model training mainly comes from the AMF, and the demand for trajectory prediction generally comes from the Operation Administration and Maintenance (OAM) node. The OAM sends a request to the NWDAF to subscribe to the trajectory prediction service. In addition, multiple entities such as Service Function (SF) can also subscribe to the trajectory prediction service from the NWDAF. When the UE changes its location, some communication events of the base station are triggered. These events are notified as data (mainly information about the base station transition of the UE, including timestamp, UE unique identifier, new Radio Base Station (RBS) identifier) sent by the AMF to the NWDAF.

[0119] 3. 5G mobility management

[0120] 5G before the UE enters idle mode, the signaling interaction between UE and network remains minimal, when there is incoming downlink data towards the UE, the network still needs to be able to reach the UE, needs to rely on the base station to paging, the process is that the AMF finds the last base station connected to the UE broadcasts paging messages, if the UE does not reply, the AMF instructs the adjacent base station to broadcast the message, and the range is expanded until the UE is found. The paging process in this way is inefficient, so later there is the concept of tracking area, the base station or even the cell in each base station periodically sends the tracking area identifier (TAI) composed of the tracking area code (TAC) and the network identifier of the communication service provider (CSP). Each UE will have a TAI list, when the UE moves within the area of its TAI list, it does not need to contact the network, and when it exceeds the area of the TAI list, it needs to contact the AMF to update the TAI list, which is called periodic registration area update (RAU) in 5G.

[0121] The training of the UE mobility prediction model needs sufficient base station transition events to provide data support, which is contrary to the design of keeping the mobility processing signaling in the 5G system to a minimum. The solution proposed in the disclosure takes the lack of accuracy of UE mobility prediction as the breakthrough point, improves the model inference performance through data collection and local analysis of the UE at the terminal side, and can support the scenario of massive terminals under the large-scale ubiquitous connection of the 6th Generation mobile communication (6G) network.

[0122] FIG. 1 is a flowchart of a model updating method provided by an embodiment of the disclosure, which is applied to a terminal, as shown in FIG. 1, the method comprises the following steps 101, 102 and 103.

[0123] Step 101, the data analysis function entity of the terminal sends a first data collection request to the data collection function entity of the terminal.

[0124] Step 102, the data analysis function entity of the terminal receives first data provided by the data collection function entity of the terminal based on the first data collection request.

[0125] Step 103, the data analysis function entity of the terminal performs inference based on the first data, and updates the local model based on the inference result.

[0126] Specifically, in the present disclosure, the terminal has two logical function entities, a data analysis function entity and a data collection function entity, the data collection function entity is used to implement data collection related functions, and the data analysis function entity is used to perform analysis and inference using a model.

[0127] In some embodiments, the data collection function entity can be mutually replaced with a Network Data Collection Function (NDCF) or a similar term.

[0128] In some embodiments, the data analysis function entity can be mutually replaced with a NWDAF or a similar term.

[0129] In some embodiments, the data analysis function entity of the terminal can be a simplified data analysis function entity (referred to as S-data analysis function) with partial functions of the data analysis function entity, and the S-data analysis function has partial analysis and inference functions of the data analysis function entity, which can be used to adapt to the terminal side or the edge side with limited resources.

[0130] In some embodiments, the data collection function entity of the terminal can be a simplified data collection function entity (referred to as S-data collection function or S-NDCF) with partial functions of the data collection function entity, and the S-data collection function has partial data collection functions of the data collection function entity, which can be used to adapt to the terminal side or the edge side with limited resources.

[0131] The data analysis function entity of the terminal can send a first data collection request to the data collection function entity of the terminal according to the data requirements of model training, and the data collection function entity of the terminal can collect data according to the requested data type and time information after receiving the first data collection request, and then send the collected first data to the data analysis function entity of the terminal. The data analysis function entity of the terminal performs model inference using the first data sent by the data collection function entity of the terminal, and updates the local model by comparing the inference result with the actual information.

[0132] In some embodiments, the local model includes a terminal mobility prediction model. For example, the data analysis function entity of the terminal can send a first data collection request to the data collection function entity of the terminal to request to collect terminal mobility related data, the data collection function entity of the terminal sends the terminal mobility related data to the data analysis function entity of the terminal after performing data collection, and the data analysis function entity of the terminal performs model inference through the local terminal mobility prediction model, and updates the local terminal mobility prediction model by comparing the inference result with the actual location information of the terminal.

[0133] It should be noted that the model of each embodiment of the present disclosure is not limited to a terminal mobility prediction model, but can also be other models that require terminal assistance, and the present disclosure does not limit.

[0134] The model updating method provided by the embodiments of the present disclosure requests data from the data collection function entity of the terminal by the data analysis function entity of the terminal, and updates the local model based on the data provided by the data collection function entity of the terminal, so that the model of the terminal can be supported by sufficient training data, thereby avoiding the problem of affecting the inference performance of the model due to insufficient training data, and effectively improving the inference performance of the model. Moreover, local analysis and inference on the terminal side can make the terminal more intelligent, reduce the dependence on the network core, and also be conducive to enhancing user privacy protection and data security.

[0135] In some embodiments, the method further includes:

[0136] The data analysis function entity of the terminal sends the model parameters of the updated local model to the network node;

[0137] The data analysis function entity of the terminal receives the model parameters of the global model sent by the network node, and updates the local model based on the model parameters of the global model;

[0138] The network node includes a centralized network node or a distributed network node.

[0139] Specifically, in some embodiments, the model training can be performed using horizontal federated learning. After updating the local model, the terminal sends the model parameters of the updated local model to the centralized network node or the distributed network node. The centralized network node or the distributed network node collects the model parameters of the updated local model sent by multiple terminals, aggregates the model parameters through federated averaging, stochastic gradient descent, etc. to update the model parameters of the global model, and then sends the updated model parameters of the global model to each terminal. Each terminal can update the local model using the model parameters of the global model sent by the network node.

[0140] In the present disclosure, the centralized network node refers to a centralized core network, and the distributed network node refers to a distributed core network. For example, the distributed network architecture of the 6G system includes a centralized network node and a distributed network node.

[0141] For the case that the data analysis function entity of the terminal sends the model parameters of the updated local model to the distributed network node, the distributed network node can serve as a node for model parameter aggregation, for example, the distributed network node maintains a global model for a plurality of terminals served thereby, and after receiving the model parameters of the updated local model sent by each terminal, the distributed network node aggregates the model parameters and updates the model parameters of the global model, and then sends the updated model parameters of the global model to each terminal; or the distributed network node can serve as an intermediate node to assist in collecting and aggregating model parameter updates, that is, after receiving the model parameters of the updated local model sent by the terminal, the distributed network node sends these model parameters to the centralized network node, and the centralized network node aggregates the model parameters.

[0142] In some embodiments, the first data collection request comprises one or more of the following information:

[0143] (1) Cell identity. The cell identity is used to indicate the cell range of data collection.

[0144] (2) Base station identity. The base station identity is used to indicate the base station range of data collection.

[0145] (3) Time stamp. The time stamp is used to identify the time information when the first data collection request is sent.

[0146] (4) Time range information. The time range information is used to indicate the time range of data collection.

[0147] In some embodiments, the method further comprises:

[0148] The data analysis function entity of the terminal sends a first model download request to the centralized network node or the distributed network node, and the first model download request comprises one or more of the following information:

[0149] (1) User Concealed Identifier (SUCI) of the terminal.

[0150] (2) Identification code of the data analysis function entity of the terminal.

[0151] (3) Model type information.

[0152] In some embodiments, the distributed network node can assist in model request and model distribution as an intermediate node. For example, after receiving the first model download request sent by the data analysis function entity of the terminal, the distributed network node can request a model from the centralized network node, for example, send the first model download request to the centralized network node, after receiving the model provided by the centralized network node according to the model download request, send the model to the terminal, and the terminal performs local configuration after receiving the model.

[0153] Figure 2 is a flowchart of a model updating method provided by an embodiment of the present disclosure, which is applied to an access network device (for example, a base station). As shown in Figure 2, the method comprises the following steps 201, 202 and 203.

[0154] In step 201, the data analysis function entity of the access network device sends a second data collection request to the data collection function entity of the terminal.

[0155] In step 202, the data analysis function entity of the access network device receives second data provided by the data collection function entity of the terminal based on the second data collection request.

[0156] In step 203, the data analysis function entity of the access network device performs inference based on the second data, and updates the local model based on the inference result.

[0157] Specifically, in the present disclosure, the access network device has a data analysis function entity, and the terminal has a data collection function entity. The data collection function entity is used to implement functions related to data collection, and the data analysis function entity is used to perform analysis and inference by using a model.

[0158] In some embodiments, the data collection function entity can be replaced by NDCF or a similar term with the same meaning.

[0159] In some embodiments, the data analysis function entity can be replaced by NWDAF or a similar term with the same meaning.

[0160] In some embodiments, the data analysis function entity of the access network device can be a simplified data analysis function entity (referred to as S-data analysis function) with part of the functions of the data analysis function entity.

[0161] In some embodiments, the data collection function entity of the terminal can be a simplified data collection function entity (referred to as S-data collection function or S-NDCF) with part of the functions of the data collection function entity.

[0162] The data analysis function entity of the access network device can send a second data collection request to the data collection function entity of the terminal according to the data requirements of model training. After receiving the second data collection request, the data collection function entity of the terminal can collect data according to the requested data type and time, and then send the collected second data to the data analysis function entity of the access network device. The data analysis function entity of the access network device performs model inference using the second data sent by the data collection function entity of the terminal, and updates the local model by comparing the inference result with the actual information.

[0163] In some embodiments, the local model comprises a terminal mobility prediction model. For example, the data analysis function entity of the access network device can send a second data collection request to the data collection function entity of the terminal, requesting to collect mobility-related data of the terminal, the data collection function entity of the terminal sends the mobility-related data of the terminal to the data analysis function entity of the access network device after performing data collection, and the data analysis function entity of the access network device performs model inference through the local terminal mobility prediction model, and updates the local terminal mobility prediction model by comparing the inference result with the actual location information of the terminal.

[0164] It should be noted that the model of each embodiment of the present disclosure is not limited to a terminal mobility prediction model, but can also be other models that require terminal assistance, and the present disclosure does not limit.

[0165] The model updating method provided by the embodiments of the present disclosure requests data from the data collection function entity of the terminal by the data analysis function entity of the access network device, and updates the local model based on the data provided by the data collection function entity of the terminal, so that the model of the access network device can be supported by sufficient training data, thereby avoiding the problem of affecting the inference performance of the model due to insufficient training data, and effectively improving the inference performance of the model. Moreover, local analysis and inference on the access network device side can make the access network device more intelligent and reduce the dependence on the network core.

[0166] In some embodiments, the method further comprises:

[0167] The data analysis function entity of the access network device sends the model parameters of the updated local model to the network node;

[0168] The data analysis function entity of the access network device receives the model parameters of the global model sent by the network node, and updates the local model based on the model parameters of the global model;

[0169] The network node comprises a centralized network node or a distributed network node.

[0170] Specifically, in some embodiments, the model training can be performed using horizontal federated learning, the access network device sends the model parameters of the updated local model to the centralized network node or the distributed network node after updating the local model, the centralized network node or the distributed network node collects the model parameters of the updated local model sent by multiple access network devices, and aggregates the model parameters through federated averaging, stochastic gradient descent, etc., thereby updating the model parameters of the global model, and then sending the updated model parameters of the global model to each access network device. Each access network device can update the local model using the model parameters of the global model sent by the network node.

[0171] In the present disclosure, the centralized network node refers to a centralized core network, and the distributed network node refers to a distributed core network. For example, the distributed network architecture of the 6G system includes a centralized network node and a distributed network node.

[0172] For the case that the data analysis function entity of the access network device sends the model parameters of the updated local model to the distributed network node, the distributed network node can serve as a node for aggregating the model parameters, for example, the distributed network node maintains a global model for a plurality of access network devices served thereby, and after receiving the model parameters of the updated local model sent by each access network device, the distributed network node aggregates the model parameters and updates the model parameters of the global model, and then sends the updated model parameters of the global model to each access network device; or the distributed network node can serve as an intermediate node to assist in collecting and aggregating the model parameter updates, that is, after receiving the model parameters of the updated local model sent by the access network device, the distributed network node sends these model parameters to the centralized network node, and the centralized network node aggregates the model parameters.

[0173] In some embodiments, the second data collection request contains one or more of the following information:

[0174] (1) Cell identification. The cell identification is used to indicate the cell range of data collection.

[0175] (2) Base station identification. The base station identification is used to indicate the base station range of data collection.

[0176] (3) Time stamp. The time stamp is used to identify the time information when the second data collection request is sent.

[0177] (4) Time range information. The time range information is used to indicate the time range of data collection.

[0178] In some embodiments, the method further comprises:

[0179] The data analysis function entity of the access network device sends a second model download request to the centralized network node or the distributed network node, and the second model download request contains one or more of the following information:

[0180] (1) One or more user concealed identifiers (SUCI) of terminals.

[0181] (2) Identification code of the data analysis function entity of the access network device.

[0182] (3) Model type information.

[0183] In some embodiments, the distributed network node can assist the model request and model distribution as an intermediate node. For example, after receiving the second model download request sent by the data analysis function entity of the access network device, the distributed network node can request the model from the centralized network node, for example, send the second model download request to the centralized network node, after receiving the model provided by the centralized network node according to the model download request, send the model to the access network device, and after the access network device receives the model, perform local configuration.

[0184] The methods provided by the embodiments of the present disclosure are based on the same technical concept, and the implementation of each method can be referred to each other, and the repeated parts will not be described.

[0185] The methods provided by the embodiments of the present disclosure are based on the same technical concept, and the implementation of each method can be referred to each other, and the repeated parts will not be described.

[0186] FIG. 3 is an example of a network architecture applicable to a 6G distributed network according to an embodiment of the present disclosure. As shown in FIG. 3, the centralized network node, the distributed network node, the radio access network (RAN) node and the UE all have two logical functions of NDCF and data analysis function for data collection and analysis inference.

[0187] In FIG. 3, PCF refers to a policy control function (PCF), UDM refers to a unified data management (UDM), SMF refers to a session management function (SMF), MANO refers to management and orchestration (MANO), and UPF refers to a user plane function (UPF). It should be noted that the network elements of the core network in FIG. 3 are only examples and are not all core network network elements.

[0188] It should be noted that the following examples take the UE mobility prediction model as an example for introduction, but the model of each example can be replaced by other models that need terminal assistance.

[0189] Example 1: UE mobility prediction model distribution.

[0190] FIG. 4 is a flowchart of example 1 according to an embodiment of the present disclosure. As shown in FIG. 4, the main process includes:

[0191] Step 1-1, the S-data analytics function of the UE sends a first model download request to the centralized network node through the RAN, and the request contains parameters: UE SUCI, identification code of the S-data analytics function, and model type information (in this example, UE mobility prediction).

[0192] Step 1-2, the data analytics function of the centralized network node completes request verification and performs model transmission to the S-data analytics function of the UE.

[0193] Step 1-3, the S-data analytics function of the UE receives the model and performs configuration of the UE mobility prediction model.

[0194] Example 2: UE mobility prediction model update.

[0195] FIG. 5 is a flowchart of Example 2 provided by the embodiment of the present disclosure, as shown in FIG. 5, and the main process includes:

[0196] Step 2-1, the S-data analytics function of the UE sends a first data collection request to the S-NDCF of the UE, and the request contains cell identification or base station identification, a timestamp, and time range information (such as time slot).

[0197] Step 2-2, the S-NDCF of the UE performs data collection and collects data within a specified time range according to the requested data type, and in this example, the collected data is related to UE mobility.

[0198] Step 2-3, the S-NDCF of the UE sends the collected data to the S-data analytics function of the UE.

[0199] Step 2-4, the S-data analytics function of the UE performs model inference through the UE mobility prediction model.

[0200] Step 2-5, comparison is performed based on the inference result and actual location information of the UE.

[0201] Step 2-6, local model update is completed.

[0202] Example 3: centralized network node model update.

[0203] FIG. 6 is a flowchart of Example 3 provided by the embodiment of the present disclosure, as shown in FIG. 6, and the main process includes:

[0204] Step 3-1, the S-data analytics function of the UE sends a first data collection request to the S-NDCF of the UE, and the request contains cell identification or base station identification, a timestamp, and time range information.

[0205] Step 3-2, the S-NDCF of the UE performs data collection according to the requested data type within a specified time range, and in this example, the collection of UE mobility related data.

[0206] Step 3-3, the S-NDCF of the UE sends the collected data to the S-data analysis function of the UE.

[0207] Step 3-4, the S-data analysis function of the UE performs model inference through the UE mobility prediction model.

[0208] Step 3-5, the inference result is compared with the actual location information of the UE and the local model is updated.

[0209] Step 3-6, the model parameters (such as weights, etc.) of the updated local model are directly uploaded to the data analysis function of the centralized network node through the RAN (this method can be applied to the case where the data analysis function of the centralized network node uniformly performs a universal model, such as a UE mobility prediction model without special requirements).

[0210] Step 3-7, the data analysis function of the centralized network node waits for a unit of time, collects the model parameters directly uploaded by each UE, and then aggregates the model parameters through federated averaging, stochastic gradient descent, etc. to update the model parameters of the global model.

[0211] Step 3-8, the data analysis function of the centralized network node distributes the updated model parameters of the global model to the S-data analysis function of each UE.

[0212] Step 3-9, the S-data analysis function of the UE performs model configuration.

[0213] In this example, the model parameters of the global model need to consider the overall effect, and for the model configured by the data analysis function of the centralized network node, scalable model training and execution is a key challenge because the network function (NF) of the control plane is responsible for providing real-time services for millions of UEs. In this case, it is more practical to maintain a global model (rather than a single model) that covers the prediction of millions of UEs.

[0214] In this example, the UE directly interacts with the data analysis function of the centralized network node, and the centralized network node can also be replaced by a distributed node A or a distributed node B to meet different requirements of different groups of UE mobility prediction.

[0215] Example 4: Distributed network node assisted collection and aggregation of model parameter updates.

[0216] FIG. 7 is a flowchart of an example 4 provided by the embodiments of the present disclosure, as shown in FIG. 7, the main processes include:

[0217] Step 4-1, the S-data analysis function of the UE sends a first data collection request to the S-NDCF of the UE, and the request contains a cell identifier or a base station identifier, a timestamp, and time range information.

[0218] Step 4-2, the S-NDCF of the UE performs data collection and collects data within a specified time range according to the requested data type. In this example, the collected data is UE mobility related data.

[0219] Step 4-3, the S-NDCF of the UE sends the collected data to the S-data analysis function of the UE.

[0220] Step 4-4, the S-data analysis function of the UE performs model inference through a UE mobility prediction model.

[0221] Step 4-5, compare the inference result with the actual position information of the UE and update the local model.

[0222] Step 4-6, upload the model parameters (such as weights) of the updated local model to the data analysis function of the distributed network node through the RAN.

[0223] Step 4-7, the distributed network node collects the model parameters uploaded by each UE.

[0224] Step 4-8, the distributed network node sends the collected model parameters uploaded by each UE to the data analysis function of the centralized network node.

[0225] Step 4-9, the data analysis function of the centralized network node waits for a unit of time, collects the model parameters uploaded by each distributed network node, and then aggregates the model parameters through federated averaging, stochastic gradient descent, etc. Update the model parameters of the global model.

[0226] Step 4-10, the data analysis function of the centralized network node updates the model parameters of the global model and distributes them to the S-data analysis function of each UE through the distributed network node.

[0227] Step 4-11, the S-data analysis function of the UE performs model configuration.

[0228] It should be noted that in examples 1-4, the S-data analysis function of the UE can be replaced by the S-data analysis function of the RAN node, for example: the S-data analysis function of the RAN node requests the model to the centralized network node or the distributed network node, the S-data analysis function of the RAN node requests the S-NDCF of the UE to collect data and update the local model of the RAN node, and the S-data analysis function of the RAN node uploads the model parameters of the updated local model to the centralized network node or the distributed network node.

[0229] The technical solution of the present disclosure can be used by multiple use cases, and the application scope is based on mobility prediction. Other use cases include paging and UPF reselection, for example. In the paging scenario, if the system can predict the location and mobility of the UE, the paging signal can be sent more accurately, thereby reducing the broadcast area and improving the efficiency of paging. For UPF reselection, if the network node to which the UE will be connected can be predicted, resource configuration can be prepared in advance, the risk of handover failure can be reduced, and the reliability and user experience of the network can be improved.

[0230] FIG. 8 is a structural schematic diagram of a terminal provided by an embodiment of the present disclosure. As shown in FIG. 8, the terminal includes a memory 820, a transceiver 810 and a processor 800; wherein the processor 800 and the memory 820 can also be arranged physically separately.

[0231] The memory 820 is configured to store a computer program; and the transceiver 810 is configured to transceive data under the control of the processor 800.

[0232] In FIG. 8, the bus architecture can include any number of interconnected buses and bridges, which are linked together by various circuits of the processor 800 represented by one or more processors and the memory 820 represented by the memory. The bus architecture can also link various other circuits such as peripheral devices, voltage stabilizers and power management circuits, which are well known in the art, and thus the present disclosure will not be described in detail. The bus interface provides an interface. The transceiver 810 can be a plurality of elements, i.e., including a transmitter and a receiver, providing a unit for communicating with various other devices on a transmission medium, including wireless channels, wired channels, optical cables and other transmission media. For different user equipment, the user interface 830 can also be an interface that can be connected to the required equipment, including but not limited to a keypad, a display, a speaker, a microphone, a joystick, etc.

[0233] The processor 800 is responsible for managing the bus architecture and general processing, and the memory 820 can store data used by the processor 800 when performing operations.

[0234] The processor 800 can be a central processing unit (CPU), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or a complex programmable logic device (CPLD). The processor can also adopt a multi-core architecture.

[0235] The processor 800 invokes a computer program stored in the memory 820 to execute any of the methods provided by the embodiments of the present disclosure according to the obtained executable instructions, including:

[0236] The data analysis function entity of the terminal sends a first data collection request to the data collection function entity of the terminal.

[0237] The data analysis function entity of the terminal receives first data provided by the data collection function entity of the terminal based on the first data collection request.

[0238] The data analysis function entity of the terminal performs inference based on the first data, and updates the local model based on the inference result.

[0239] In some embodiments, the method further includes:

[0240] The data analysis function entity of the terminal sends the model parameters of the updated local model to the network node.

[0241] The data analysis function entity of the terminal receives the model parameters of the global model sent by the network node, and updates the local model based on the model parameters of the global model.

[0242] The network node includes a centralized network node or a distributed network node.

[0243] In some embodiments, the first data collection request contains one or more of the following information:

[0244] Cell identity;

[0245] Base station identity;

[0246] Timestamp;

[0247] Time range information.

[0248] In some embodiments, the method further includes:

[0249] The data analysis function entity of the terminal sends a first model download request to the centralized network node or the distributed network node, and the first model download request contains one or more of the following information:

[0250] A user hidden identifier of the terminal;

[0251] An identification code of the data analysis function entity of the terminal;

[0252] Model type information.

[0253] In some embodiments, the local model includes a terminal mobility prediction model.

[0254] FIG. 9 is a structural schematic diagram of an access network device provided by an embodiment of the present disclosure, as shown in FIG. 9, the access network device includes a memory 920, a transceiver 910 and a processor 900; wherein the processor 900 and the memory 920 can also be arranged physically separately.

[0255] The memory 920 is configured to store a computer program; and the transceiver 910 is configured to transceive data under the control of the processor 900.

[0256] In FIG. 9, the bus architecture can include any number of interconnected buses and bridges, which are linked together by various circuits of the processor 900 represented by one or more processors and the memory 920 represented by the memory. The bus architecture can also link various other circuits such as peripheral devices, voltage stabilizers and power management circuits, which are well known in the art, and thus the present disclosure will not repeat them. The bus interface provides an interface. The transceiver 910 can be a plurality of elements, that is, it includes a transmitter and a receiver, which provides a unit for communicating with various other devices on a transmission medium, including wireless channels, wired channels, optical cables and other transmission media.

[0257] The processor 900 is responsible for managing the bus architecture and general processing, and the memory 920 can store data used by the processor 900 when performing operations.

[0258] The processor 900 can be a CPU, ASIC, FPGA or CPLD, and the processor can also adopt a multi-core architecture.

[0259] The processor 900 invokes the computer program stored in the memory 920 to execute any of the methods provided by the access network device side of the embodiments of the present disclosure according to the obtained executable instructions, including:

[0260] The data analysis function entity of the access network device sends a second data collection request to the data collection function entity of the terminal;

[0261] The data analysis function entity of the access network device receives second data provided by the data collection function entity of the terminal based on the second data collection request;

[0262] The data analysis function entity of the access network device performs inference based on the second data, and updates the local model based on an inference result.

[0263] In some embodiments, the method further comprises:

[0264] The data analysis function entity of the access network device sends model parameters of the updated local model to the network node;

[0265] The data analysis function entity of the access network device receives model parameters of the global model sent by the network node, and updates the local model based on the model parameters of the global model;

[0266] The network node comprises a centralized network node or a distributed network node.

[0267] In some embodiments, the second data collection request comprises one or more of the following information:

[0268] A cell identifier;

[0269] A base station identifier;

[0270] A timestamp;

[0271] Time range information.

[0272] In some embodiments, the method further comprises:

[0273] The data analysis function entity of the access network device sends a second model download request to the centralized network node or the distributed network node, and the second model download request comprises one or more of the following information:

[0274] User concealed identifiers of one or more terminals;

[0275] An identification code of the data analysis function entity of the access network device;

[0276] Model type information.

[0277] In some embodiments, the local model comprises a terminal mobility prediction model.

[0278] It should be noted that the terminal and the access network device provided by the embodiments of the present disclosure can realize all the method steps achieved by the method embodiments and achieve the same technical effects, and the same parts and beneficial effects of the method embodiments will not be described in detail.

[0279] The model updating apparatus provided by the embodiments of the present disclosure is described below. The model updating apparatus described below can be referred to the model updating method described above.

[0280] FIG. 10 is a structural schematic diagram of a model updating apparatus provided by an embodiment of the present disclosure. As shown in FIG. 10, the apparatus includes:

[0281] A first sending unit 1010, configured to send, by a data analysis function entity of a terminal, a first data collection request to a data collection function entity of the terminal.

[0282] A first receiving unit 1020, configured to receive, by the data analysis function entity of the terminal, first data provided by the data collection function entity of the terminal based on the first data collection request.

[0283] A first updating unit 1030, configured to perform, by the data analysis function entity of the terminal, inference based on the first data, and update a local model based on a result of the inference.

[0284] In some embodiments, the first updating unit 1030 is further configured to:

[0285] send, by the data analysis function entity of the terminal, model parameters of the updated local model to a network node;

[0286] receive, by the data analysis function entity of the terminal, model parameters of a global model sent by the network node, and update the local model based on the model parameters of the global model;

[0287] The network node includes a centralized network node or a distributed network node.

[0288] In some embodiments, the first data collection request contains one or more of the following information:

[0289] a cell identifier;

[0290] a base station identifier;

[0291] a timestamp;

[0292] time range information.

[0293] In some embodiments, the first sending unit 1010 is further configured to:

[0294] send, by the data analysis function entity of the terminal, a first model download request to the centralized network node or the distributed network node, and the first model download request contains one or more of the following information:

[0295] a user hidden identifier of the terminal;

[0296] an identification code of the data analysis function entity of the terminal;

[0297] model type information.

[0298] In some embodiments, the local model comprises a terminal mobility prediction model.

[0299] FIG. 11 is a structural schematic diagram of a model updating apparatus provided by an embodiment of the present disclosure, as shown in FIG. 11, the apparatus comprises:

[0300] The second sending unit 1110 is configured to send, by the data analysis function entity of the access network device, a second data collection request to the data collection function entity of the terminal;

[0301] The second receiving unit 1120 is configured to receive, by the data analysis function entity of the access network device, second data provided by the data collection function entity of the terminal based on the second data collection request;

[0302] The second updating unit 1130 is configured to perform, by the data analysis function entity of the access network device, inference based on the second data, and update the local model based on an inference result.

[0303] In some embodiments, the second updating unit 1130 is further configured to:

[0304] The data analysis function entity of the access network device sends model parameters of the updated local model to a network node;

[0305] The data analysis function entity of the access network device receives model parameters of a global model sent by the network node, and updates the local model based on the model parameters of the global model;

[0306] The network node comprises a centralized network node or a distributed network node.

[0307] In some embodiments, the second data collection request comprises one or more of the following information:

[0308] A cell identifier;

[0309] A base station identifier;

[0310] A timestamp;

[0311] Time range information.

[0312] In some embodiments, the second sending unit 1110 is further configured to:

[0313] The data analysis function entity of the access network device sends a second model download request to the centralized network node or the distributed network node, and the second model download request comprises one or more of the following information:

[0314] One or more user hidden identifiers of the terminal;

[0315] An identification code of the data analysis function entity of the access network device;

[0316] model type information.

[0317] In some embodiments, the local model comprises a terminal mobility prediction model.

[0318] It should be noted that the division of the units in the embodiments of the present disclosure is illustrative, and is merely a logical functional division. In actual implementation, another division manner can be used. In addition, each functional unit in each embodiment of the present disclosure can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0319] When the integrated unit is realized in the form of a software functional unit and sold or used as an independent product, it can be stored in a processor-readable storage medium. Based on this understanding, the technical solutions of the present disclosure, essentially or the part that contributes to the related art, or all or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to perform all or part of the steps of the methods described in the various embodiments of the present disclosure. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program code storage media.

[0320] It should be noted that the model updating apparatus provided by the embodiments of the present disclosure can implement all the method steps achieved by the method embodiments, and can achieve the same technical effects. Here, the same parts and beneficial effects of the method embodiments in the embodiments will not be described in detail.

[0321] On the other hand, the embodiments of the present disclosure also provide a processor-readable storage medium, which stores a program for causing a processor to execute the model updating method provided by each of the embodiments.

[0322] It should be noted that the processor-readable storage medium provided by the embodiments of the present disclosure can implement all the method steps achieved by the method embodiments, and can achieve the same technical effects. Here, the same parts and beneficial effects of the method embodiments in the embodiments will not be described in detail.

[0323] The processor-readable storage medium can be any available medium or data storage device that a processor can access, including but not limited to a magnetic storage device (e.g., floppy diskette, hard disk drive, magnetic tape, MO, etc.), an optical storage device (e.g., CD, DVD, BD, HVD, etc.), and a semiconductor memory device (e.g., ROM, EPROM, EEPROM, NAND FLASH, SSD, etc.), etc.

[0324] The technical solutions provided by the embodiments of the present disclosure can be applied to various systems. For example, the applicable systems can be a long term evolution (LTE) system, an LTE frequency division duplex (FDD) system, an LTE time division duplex (TDD) system, a long term evolution advanced (LTE-A) system, a universal mobile system (UMTS), a worldwide interoperability for microwave access (WiMAX) system, a 5G new radio (NR) system and its evolution communication system, a 6G (sixth generation mobile communication technology) system, etc. The various systems can include terminal devices and network devices. The system can also include a core network part, such as an evolved packet system (EPC), a 5G core network (5GC), etc.

[0325] The terminal involved in the embodiments of the present disclosure can refer to a device that provides voice and / or data connectivity to a user, a handheld device with wireless connection function, or other processing devices connected to a wireless modem, etc. In different systems, the name of the terminal can also be different, for example, in the 5G system, the terminal can be called user equipment (UE). The wireless terminal device can communicate with one or more core networks (CN) through a radio access network (RAN), and the wireless terminal device can be a mobile terminal device, such as a mobile phone (or called "cellular" phone) and a computer with a mobile terminal device, for example, it can be a portable, pocket-sized, handheld, computer-built-in or vehicle-mounted mobile device, which exchanges language and / or data with the radio access network. For example, personal communication service (PCS) phones, cordless phones, session initiation protocol (SIP) phones, wireless local loop (WLL) stations, personal digital assistants (PDAs), etc. The wireless terminal device can also be called a system, a subscriber unit, a subscriber station, a mobile station, a mobile, a remote station, an access point, a remote terminal, an access terminal, a user terminal, a user agent, a user device, which is not limited in the embodiments of the present disclosure.

[0326] The access network device according to the embodiments of the present disclosure can be a base station, which can include a plurality of cells serving terminals. According to different application scenarios, the base station can also be referred to as an access point, or can be a device in an access network that communicates with wireless terminal devices through one or more sectors over an air interface, or other names. The access network device can be used to exchange received air frames and Internet Protocol (IP) packets as a router between the wireless terminal device and the rest of the access network, which can include an Internet Protocol (IP) communication network. The access network device can also coordinate the management of the properties of the air interface. For example, the access network device according to the embodiments of the present disclosure can be an access network device (Base Transceiver Station, BTS) in the Global System for Mobile Communications (GSM) or Code Division Multiple Access (CDMA), or an access network device (NodeB) in Wide-band Code Division Multiple Access (WCDMA), or an evolved access network device (evolutional Node B, eNB or e-NodeB) in a long term evolution (LTE) system, or a 5G base station (gNB) in a next generation system, or a Home evolved Node B (HeNB), a relay node, a femto, a pico, etc., which are not limited in the embodiments of the present disclosure. In some network structures, the access network device can include a centralized unit (CU) node and a distributed unit (DU) node, and the centralized unit and the distributed unit can also be arranged geographically apart.

[0327] Those skilled in the art should understand that the embodiments of the present disclosure can be provided as a method, a system, or a computer program product. Therefore, the present disclosure can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present disclosure can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, magnetic disk storage and optical storage, etc.) containing computer-usable program code.

[0328] The computer executable instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operations steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart and / or block diagram block or blocks.

[0329] These processor executable instructions can also be stored in a processor readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the processor readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart and / or block diagram block or blocks.

[0330] These processor executable instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operations steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart and / or block diagram block or blocks.

[0331] Obviously, numerous modifications and variations of the present disclosure are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims and their equivalents, the present disclosure can be practiced otherwise than as specifically described.

Claims

1. A method for model updating, applied to a terminal, comprising: sending, by a data analysis function entity of the terminal, a first data collection request to a data collection function entity of the terminal; receiving, by the data analysis function entity of the terminal, first data provided by the data collection function entity of the terminal based on the first data collection request; performing, by the data analysis function entity of the terminal, inference based on the first data, and updating a local model based on an inference result.

2. The model updating method of claim 1, wherein, The method further comprises: sending, by the data analysis function entity of the terminal, model parameters of the updated local model to a network node; receiving, by the data analysis function entity of the terminal, model parameters of a global model sent by the network node, and updating the local model based on the model parameters of the global model; The network node comprises a centralized network node or a distributed network node.

3. The model updating method according to claim 1 or 2, wherein, The first data collection request comprises one or more of the following information: a cell identifier; a base station identifier; a timestamp; time range information.

4. The model updating method of claim 1, wherein, The method further comprises: sending, by the data analysis function entity of the terminal, a first model download request to a centralized network node or a distributed network node, the first model download request comprising one or more of the following information: a user hidden identifier of the terminal; an identification code of the data analysis function entity of the terminal; model type information.

5. The model updating method according to claim 1 or 2 or 4, wherein, The local model comprises a terminal mobility prediction model. 6.A method for model updating, applied to an access network device, comprising: sending, by a data analysis function entity of the access network device, a second data collection request to a data collection function entity of a terminal; receiving, by the data analysis function entity of the access network device, second data provided by the data collection function entity of the terminal based on the second data collection request; performing, by the data analysis function entity of the access network device, inference based on the second data, and updating a local model based on an inference result.

7. The model updating method of claim 6, wherein, The method further comprises: sending, by the data analysis function entity of the access network device, model parameters of the updated local model to a network node; receiving, by the data analysis function entity of the access network device, model parameters of a global model sent by the network node, and updating the local model based on the model parameters of the global model; The network node comprises a centralized network node or a distributed network node.

8. The model updating method according to claim 6 or 7, wherein The second data collection request comprises one or more of the following information: a cell identifier; a base station identifier; a timestamp; time range information.

9. The model updating method of claim 6, wherein, The method further comprises: sending, by the data analysis function entity of the access network device, a second model download request to a centralized network node or a distributed network node, the second model download request comprising one or more of the following information: user hidden identifiers of one or more terminals; an identification code of the data analysis function entity of the access network device; model type information.

10. The model updating method according to claim 6 or 7 or 9, wherein, The local model comprises a terminal mobility prediction model. 11.A terminal, comprising a memory, a transceiver, and a processor; The memory is configured to store a computer program; The transceiver is configured to transceive data under the control of the processor; The processor is configured to read the computer program in the memory and perform the following operations: The data analysis function entity of the terminal sends a first data collection request to the data collection function entity of the terminal; The data analysis function entity of the terminal receives first data provided by the data collection function entity of the terminal based on the first data collection request; The data analysis function entity of the terminal performs inference based on the first data, and updates the local model based on the inference result.

12. The terminal according to claim 11, wherein The operations further include: The data analysis function entity of the terminal sends model parameters of the updated local model to a network node; The data analysis function entity of the terminal receives model parameters of a global model sent by the network node, and updates the local model based on the model parameters of the global model; The network node includes a centralized network node or a distributed network node.

13. The terminal according to claim 11 or 12, wherein, The first data collection request contains one or more of the following information: Cell identity; Base station identity; Timestamp; Time range information.

14. The terminal of claim 11, wherein, The operations further include: The data analysis function entity of the terminal sends a first model download request to a centralized network node or a distributed network node, and the first model download request contains one or more of the following information: User hidden identifier of the terminal; Identification code of the data analysis function entity of the terminal; Model type information.

15. The terminal according to claim 11 or 12 or 14, wherein, The local model includes a terminal mobility prediction model.

16. An access network device, comprising a memory, a transceiver, and a processor; The memory is used to store a computer program; The transceiver is used to transceive data under the control of the processor; The processor is used to read the computer program in the memory and perform the following operations: The data analysis function entity of the access network device sends a second data collection request to the data collection function entity of the terminal; The data analysis function entity of the access network device receives second data provided by the data collection function entity of the terminal based on the second data collection request; The data analysis function entity of the access network device performs inference based on the second data, and updates the local model based on the inference result.

17. The access network device of claim 16, wherein, The operations further include: The data analysis function entity of the access network device sends model parameters of the updated local model to a network node; The data analysis function entity of the access network device receives model parameters of a global model sent by the network node, and updates the local model based on the model parameters of the global model; The network node includes a centralized network node or a distributed network node.

18. The access network device of claim 16 or 17, wherein, The second data collection request contains one or more of the following information: Cell identity; Base station identity; Timestamp; Time range information.

19. The access network device of claim 16, wherein, The operations further include: The data analysis function entity of the access network device sends a second model download request to a centralized network node or a distributed network node, and the second model download request contains one or more of the following information: User hidden identifier of one or more terminals; Identification code of the data analysis function entity of the access network device; Model type information.

20. The access network device of claim 16 or 17 or 19, wherein, The local model includes a terminal mobility prediction model.

21. A processor-readable storage medium having stored a program for causing a processor to execute the method of any one of claims 1 to 5, or to execute the method of any one of claims 6 to 10.

Images