WO2023050106A1

WO2023050106A1 - Terminal selection method and apparatus, device, and medium

Info

Publication number: WO2023050106A1
Application number: PCT/CN2021/121495
Authority: WO
Inventors: 陈景然; 许阳
Original assignee: Oppo广东移动通信有限公司
Priority date: 2021-09-28
Filing date: 2021-09-28
Publication date: 2023-04-06
Also published as: CN117693968A

Abstract

The present application relates to the technical field of transfer learning, and relates to a terminal selection method and apparatus, a device, and a medium. The method comprises: receiving a transfer learning request sent by a third-party device, the transfer learning request carrying an identifier of a source domain UE and screening condition information for selecting a candidate UE; determining a plurality of candidate UEs according to the screening condition information and the identifier of the source domain UE; and sending UE indication information to the third-party device, the UE indication information comprising identifiers of the plurality of candidate UEs, and the UE indication information being used by the third-party device to determine a target domain UE from the plurality of candidate UEs. The technical solution provided by embodiments of the present application can realize selection of the target domain UE.

Description

Terminal selection method, device, equipment and medium

technical field

The present application relates to the technical field of transfer learning, and in particular to a terminal selection method, device, equipment and medium.

Background technique

As an important branch of machine learning, transfer learning can use the similarity between data, tasks or models to apply the model learned in the old field to a new field with certain relevance. Source domain and target domain are very important concepts in transfer learning. The source domain refers to the old domain in transfer learning, and the target domain refers to the new domain in transfer learning. Generally speaking, the data set and target of the source domain Domain datasets have a certain similarity to ensure the ideal results of transfer learning. In the OTT (English full name: Over The Top) business field, the application layer data of UE (English full name: User Equipment; Chinese: User Equipment) can be used for model training to obtain a model that assists the OTT business. For a good model to perform transfer learning, it is necessary to ensure that the UE in the source domain has a certain similarity with the UE in the target domain.

At present, how to select the target domain UE has become an urgent problem to be solved in transfer learning in OTT services.

Contents of the invention

Based on this, the present application provides a terminal selection method, device, equipment and medium.

In a first aspect, a terminal selection method is provided, which is used in a first network element of a core network, and the method includes:

Receive a migration learning request sent by a third-party device, the migration learning request carries the identity of the UE in the source domain and screening condition information for selecting candidate UEs; determine multiple candidate UEs according to the filtering condition information and the identity of the UE in the source domain; The indication information is sent to the third-party device, the UE indication information includes identities of the plurality of candidate UEs, and the UE indication information is used for the third-party device to determine a target domain UE from the plurality of candidate UEs.

In the second aspect, a terminal selection method is provided, which is used in a third-party device, and the method includes:

Sending a migration learning request to the first network element of the core network, where the migration learning request carries the identity of the UE in the source domain and screening condition information for selecting candidate UEs;

receiving UE indication information sent by the first network element, where the UE indication information includes identifiers of multiple candidate UEs, and the multiple candidate UEs are determined according to the screening condition information and the identifier of the source domain UE;

Determine the UE in the target domain from the plurality of candidate UEs according to the UE indication information.

In a third aspect, a device for selecting a terminal is provided, and the device includes:

The receiving module is configured to receive a migration learning request sent by a third-party device, where the migration learning request carries the identity of the UE in the source domain and screening condition information for selecting candidate UEs;

A determining module, configured to determine a plurality of candidate UEs according to the screening condition information and the identity of the UE in the source domain;

A sending module, configured to send UE indication information to the third-party device, where the UE indication information includes identities of the plurality of candidate UEs, and the UE indication information is used for the third-party device to determine a target domain from the plurality of candidate UEs UE.

In a fourth aspect, a device for selecting a terminal is provided, and the device includes:

A sending module, configured to send a migration learning request to a first network element of the core network, where the migration learning request carries the identity of the UE in the source domain and screening condition information for selecting a candidate UE;

A receiving module, configured to receive UE indication information sent by the first network element, where the UE indication information includes identifiers of multiple candidate UEs, and the multiple candidate UEs are determined according to the screening condition information and the identifier of the source domain UE;

The determining module is configured to determine the UE in the target domain from the plurality of candidate UEs according to the UE indication information.

In a fifth aspect, a communication device is provided, including: a processor, a memory, and a transceiver, the processor, the memory, and the transceiver communicate with each other through an internal connection path, the memory is used to store program codes; the processor , for calling the program code stored in the memory, to cooperate with the transceiver to realize the steps of the method described in the first aspect above, or to cooperate with the transceiver to realize the steps of the method described in the second aspect above.

A sixth aspect provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method described in the first aspect above are implemented, or, the computer program is executed by the processor When implementing the steps of the method described in the second aspect above.

The above terminal selection method, device, device, and medium receive a transfer learning request sent by a third-party device, wherein the transfer learning request carries the identity of the UE in the source domain and screening condition information for selecting candidate UEs, and then, according to the screening condition information and source domain UE identities, determine a plurality of candidate UEs, and then send UE indication information to a third-party device, where the UE indication information includes identifications of a plurality of candidate UEs, and the UE indication information is used for the third-party device to select from multiple Determine the target domain UE among the candidate UEs, so that the third-party device can obtain multiple candidate UEs determined to be similar to the source domain UE according to the screening condition information after providing the source domain UE identification and the screening condition information for selecting the candidate UE , and then, in a subsequent step, the target domain UE is determined from the candidate UEs, thereby realizing the selection of the candidate UEs, and then in the subsequent process, transfer learning can be performed using the application layer data of the candidate UEs.

Description of drawings

Fig. 1 is the structural representation of neural network model;

Fig. 2 is the result schematic diagram of convolutional neural network model;

Fig. 3 is the schematic diagram of transfer learning process;

FIG. 4 is a schematic diagram of an implementation environment involved in an embodiment of the present application;

FIG. 5 is a schematic diagram of another implementation environment involved in the embodiment of the present application;

FIG. 6 is a flow chart of a terminal selection method provided by an embodiment;

FIG. 7 is a flow chart of a terminal selection method provided by an embodiment;

FIG. 8 is a flow chart of a terminal selection method provided by an embodiment;

FIG. 9 is a flow chart of a terminal selection method provided by an embodiment;

FIG. 10 is a flowchart of a terminal selection method provided by an embodiment;

Fig. 11 is a block diagram of a terminal selection device provided by an embodiment;

Fig. 12 is a block diagram of a terminal selection device provided by an embodiment;

Fig. 13 is a block diagram of a terminal selection device provided by an embodiment;

Fig. 14 is a block diagram of a communication device provided by an embodiment.

Detailed ways

In order to make the purpose, technical solution and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present application, and are not intended to limit the present application.

In order to make it easy for readers to understand the technical solutions provided by the embodiments of the present application, the concepts involved in the embodiments of the present application will be briefly introduced below:

1. Neural network model.

Generally speaking, the basic structure of a simple neural network model includes: input layer, hidden layer and output layer, as shown in Figure 1. Data is input from the input layer, processed by the hidden layer, and the final result is generated in the output layer. Among them, each node represents a processing unit, which can be regarded as simulating a neuron. Multiple neurons form a layer of neural network model, and multi-layer information transmission and processing constructs an overall neural network model.

With the continuous development of research, in recent years, the deep learning algorithm of the neural network model has been proposed, and more hidden layers have been introduced, which greatly improves the processing ability of the neural network model, and in pattern recognition, signal processing, optimization combination and Anomaly detection and other aspects are widely used.

With the development of deep learning algorithms, convolutional neural network models have also been further studied. In a convolutional neural network model, its basic structure includes: an input layer, multiple convolutional layers, multiple pooling layers, a fully connected layer, and an output layer, as shown in Figure 2. The introduction of the convolutional layer and the pooling layer effectively controls the sharp increase of network parameters in the neural network model, limits the number of network parameters and taps the characteristics of the local structure, improving the robustness of the algorithm.

Generally speaking, the values of the network parameters in the neural network model need to be obtained through model training. In a typical model training method, the training sample data and the real labels corresponding to the training sample data can be obtained, and then the training sample data Input through the input layer of the neural network model to obtain the training label generated by the output layer, then determine the difference between the training label and the real label, and adjust the value of the network parameter in the neural network model based on the difference, after multiple rounds of adjustment, The model training can be ended until the difference between the training label generated by the output layer of the neural network model and the real label is small, so as to obtain a trained neural network model.

2. Transfer Learning

As an important branch of machine learning, transfer learning can use the similarity between data, tasks or neural network models to apply the neural network model trained in the old field to a new field with certain relevance. For example, assuming that a neural network model has been trained to identify cat breeds, and there is currently a task of identifying tiger breeds, since cats and tigers are both cats and have similarities, transfer learning can be used The way to apply the neural network model for identifying cat breeds to the field of identifying tiger breeds is that the cost is much lower than separately training a neural network model for identifying tiger breeds.

Please refer to Figure 3, which shows a simple transfer learning process, where the left figure in Figure 3 shows the traditional machine learning method, in this traditional machine learning method, the training corresponding to the cat can be used The sample data is trained to obtain the neural network model A used to identify the cat breed. At the same time, the training sample data corresponding to the tiger is used to train the neural network model B used to identify the tiger breed. The right figure in Figure 3 shows the migration The learning method, in transfer learning, can use the corresponding training sample data of cats to train to obtain the neural network model A used to identify the cat breed, and then perform migration training on the neural network model A to obtain the tiger breed. Neural Network Model B.

Generally speaking, transfer learning can be divided into transfer based on sample weight, transfer based on feature transformation and transfer based on model pre-training according to the learning method. Sample-based transfer accomplishes transfer by weighted usage of training sample data in the source domain. Feature-based transfer accomplishes the transfer by mapping the source and target domains to another space and minimizing the distance between the source and target domains. Model-based pre-training is to fine-tune the neural network model trained in the source domain in the target domain using the training sample data of the target domain.

Among them, the source domain and target domain mentioned above are very important concepts in transfer learning. The source domain refers to the old domain in transfer learning (such as identifying cat breeds in the above example), and the target domain refers to transfer learning. In new domains (such as identifying the species of tigers in the above example), generally speaking, the datasets in the source domain and the datasets in the target domain have a certain similarity to ensure that the results of transfer learning are ideal. In the above example, it is Since both cats and tigers are cats, they are relatively similar. Therefore, the neural network model for identifying tiger species obtained after migration training can be guaranteed to have a high recognition rate. If the similarity is not enough, the migration results will not be accurate. Ideally, negative transfer may even occur. For example, if the neural network model for recognizing tiger breeds is obtained through migration training based on the neural network model for recognizing dog breeds, the similarity between dogs and tigers will be insufficient, resulting in The recognition rate of the neural network model is low.

3. OTT business

OTT business is the abbreviation of Over The Top in English. This word comes from basketball and other sports, and it means "over the top pass". In the field of wireless communication technology, OTT business means that Internet companies surpass operators and develop various video and data service businesses based on the open Internet.

In the OTT business, the application layer data of the UE (Chinese: User Equipment, English: User Equipment) can be used for model training to obtain a neural network model that assists the OTT business. For example, in the OTT business, the UE can be used to conduct video The application uses behavioral data as training sample data to train the neural network model to obtain a neural network model that predicts user behavior.

In the OTT business, if you want to carry out transfer learning for the trained neural network model, you must ensure that the source domain UE has a certain similarity with the candidate UE, otherwise, the transfer result will be unsatisfactory, and even negative transfer will occur. .

However, in practical applications, due to the sensitivity of UE information, the OTT application server cannot know UE information, and thus cannot determine whether there is similarity between UEs. Therefore, migration learning for OTT services faces the difficulty of candidate UE selection.

In view of this, the embodiment of the present application provides a terminal selection method, device, device and medium, which can use the knowledge of the operator network (such as 5GS) on the UE information to select the UE in the target domain, thereby assisting OTT services in migration learning.

Please refer to FIG. 4, which shows a schematic diagram of the implementation environment involved in the terminal selection method provided by the embodiment of the present application. As shown in FIG. 4, the implementation environment includes a third-party device 101 and a first network element in the core network 102 and the second network element 103 in the core network. Optionally, the first network element 102 is responsible for directly connecting with the third-party device 101, for example, receiving a transfer learning request sent by the third-party device 101, and based on the transfer learning Request to return the identity of the candidate UE to the third-party device 101. In addition, the first network element 102 can also obtain UE information through interaction with the second network element 103, so as to determine the candidate UE through the obtained UE information, for The third party device 101 determines the target domain UE from the candidate UEs.

Please refer to FIG. 5. In an optional embodiment, the implementation environment involved in the terminal selection method provided by the embodiment of the present application includes a third-party device 101, a first network element 102 in the core network, and a second network element 102 in the core network. In addition to the network element 103, it may also include a third network element 104 in the core network. In addition to obtaining UE information through interaction with the second network element 103, the first network element 102 may also obtain UE information through interaction with the third network element 104. Obtain UE information interactively.

It should be pointed out that the core network mentioned above can be the core network in GSM (Chinese: Global System for Mobile Communication, English: global system for mobile communication), CDMA (Chinese: Code Division Multiple Access, English: code division multiple access), core network in WCDMA (Chinese: wideband code division multiple access, English: wideband code division multiple access), core network in LTE (Chinese: long term evolution, English: long term evolution), CRAN ( Chinese: cloud wireless access network, English: cloud radio access network) in the core network or the core network in the 5G communication system.

Optionally, in the 5G communication system, the first network element 102 can reuse the existing NF (Chinese: Network Function, English: Network Function) network element in the core network, or a new network element can also be set in the core network The NF network element is used as the first network element 102, the second network element 103 can include the NWDAF (Chinese: Network Data Analysis Function, English: Network Data Analytics Function) network element in the core network, and the third network element 104 can include the NF network element. The NF network element included in the third network element 104 may be an AMF (Chinese: Access and Mobility Management Function; English: Access and Mobility Management Function) network element and/or an SMF (Chinese: Session Management Function; English : Session Management Function) network element.

It should also be pointed out that, in this embodiment of the present application, the third-party device 101 may be an OTT application server or a terminal, where the terminal may communicate with the first network element 102 through an access network device.

Among them, the terminals mentioned above may include PDA (Chinese: personal digital processing, English: personal digital assistant), handheld devices with wireless communication functions, computing devices or other processing devices connected to wireless modems, vehicle-mounted devices, wearable Equipment, user equipment in the 5G network or user equipment in the future evolved PLMN (Chinese: public land mobile network, English: public land mobile network) network, etc.

The OTT application server mentioned above can be a single server or a server cluster composed of multiple servers.

The above-mentioned access network equipment can be transmission reception point (transmission reception point, TRP), base station, relay station or access point, etc., and the access network equipment can be the access network equipment in the 5G communication system or the future evolution network The access network device in the network can also be a wearable device or a vehicle-mounted device, etc. In addition, it can also be: a BTS (Chinese base transceiver station, English: base transceiver station) in a GSM or CDMA network, or a BTS in a WCDMA network. NB (NodeB) can also be eNB or eNodeB (English full name: evolutional NodeB) in LTE, and the access network device can also be a wireless controller in the CRAN scenario.

Please refer to FIG. 6, which shows a flowchart of a terminal selection method provided by an embodiment of the present application. The terminal selection method is applied to the first network element 102 in FIG. 4 or FIG. 5, as shown in FIG. 6, The terminal selection method includes the following steps:

Step 601, the first network element receives a migration learning request sent by a third-party device.

Wherein, the migration learning request carries the identity of the UE in the source domain and the screening condition information for selecting candidate UEs.

In an optional embodiment of the present application, the filter condition information is used to indicate one or more filter conditions, and the filter condition may be, for example, at least one of the following:

It is expected to arrive in the area where the UE in the source domain is currently or historically located in the future;

The expected movement trajectory in the future is similar to the current or historical movement trajectory of the UE in the source domain;

The expected communication quality within the future target time period meets the communication quality threshold;

The type of service expected to be requested in the future is similar to the type of service currently or historically requested by the UE in the source domain;

Currently located in the same area as the UE in the source domain;

The current movement trajectory is similar to the current movement trajectory of the UE in the source domain;

The current communication quality meets the communication quality threshold;

The type of the service currently requested is similar to the type of service currently requested by the UE in the source domain;

The number k of candidate UEs.

It should be noted that in practical applications, whether the moving trajectories are similar can be characterized by whether the curves corresponding to the moving trajectories are similar, or whether the distance between vectors obtained after vectorizing the moving trajectories is less than a certain threshold can be used to represent , or, it can be characterized by whether the number of intersections between the moving trajectories is greater than a certain threshold, or it can be characterized by whether the proportion of overlapping areas between the moving trajectories is greater than a certain threshold, which is not specifically limited in this embodiment of the present application.

It should also be noted that whether the service types are similar can be characterized by whether the number of the same service types is greater than a certain threshold, or whether the proportion of the same service types is greater than a certain threshold. Not specifically limited.

In addition, the "current" in the above should be interpreted in an expanded way. The so-called current can refer to the current moment, or a period of time before the current moment (including the current moment) that is less than a certain threshold, or it can refer to the current moment before ( A time period whose duration is less than a certain threshold (excluding the current moment). The history above refers to the time period before the "current" in the time dimension, and the future refers to the time period after the "current" in the time dimension.

Optionally, the third-party device and the first network element may pre-agree on a filter condition set, each filter condition included in the filter condition set corresponds to a condition identifier, and the third-party device may select from the filter condition before sending the migration learning request. The filter condition to be used is selected from the condition set, and the filter condition information is generated based on the condition identifier of the selected filter condition, so that the first network element selects multiple candidate UEs according to the filter condition indicated by the filter condition information.

Optionally, the set of filtering conditions and the condition identifiers corresponding to the filtering conditions included in the set of filtering conditions may not only be agreed upon by the third-party device and the first network element, but also may be determined by other network elements in the core network based on High-level signaling is respectively configured for the third-party device and the first network element.

Step 602, the first network element determines a plurality of candidate UEs according to the screening condition information and the identifier of the UE in the source domain.

Optionally, the first network element may select UEs meeting at least n screening conditions as the candidate UEs according to the screening condition information and the identity of the UE in the source domain, where n is a positive integer. In an optional embodiment of the present application , n can be a value pre-configured by the network, for example, n can be configured to the first network element by other network elements in the core network through high-level signaling, and n can also be indicated by a third-party device, for example, the third-party device sends The value of n can be carried in the migration learning request.

Optionally, in this embodiment of the application, the above-mentioned migration learning request may also carry a confidence threshold. In this case, the first network element may, according to the screening condition information and the identity of the UE in the source domain, set the Meet at least n screening conditions, and among the satisfied screening conditions, there is a UE whose corresponding confidence level is greater than the confidence level threshold as the candidate UE.

For example, for a certain UE, it meets m (m is greater than or equal to n) screening conditions, and there is one screening condition corresponding to the confidence degree greater than the confidence degree threshold among the m screening conditions, then the first network element can set This UE is used as a candidate UE.

Optionally, the first network element may obtain UE information through communication interaction with the second network element, or through communication interaction with the second network element and the third network element, and implement the UE information based on the obtained UE information. Determination of multiple candidate UEs.

Step 603, the first network element sends UE indication information to a third-party device.

Wherein, the UE indication information includes identifiers of multiple candidate UEs, and the UE indication information is used for the third-party device to determine the target domain UE from the multiple candidate UEs. After determining the target domain UE, the third-party device can obtain the application layer data of the target domain UE based on the identity of the target domain UE, for example, the browsing data of the target domain UE, the search data of the target domain UE, etc., and then, based on the target domain The application layer data of the UE is transferred to the existing neural network model for learning.

As mentioned above, the screening condition may include the number k of candidate UEs. In this case, after the first network element determines a plurality of candidate UEs, it may select k candidate UEs from the candidate UEs, and based on the k Each candidate UE generates UE indication information, and then sends the UE indication information to a third-party device.

Optionally, the first network element may select k candidate UEs with the highest similarity from multiple candidate UEs according to the similarity between the candidate UE and the UE in the source domain. Wherein, the similarity between the candidate UE and the source domain UE can be characterized by the number of screening conditions satisfied by the candidate UE, or can be represented by the number of screening conditions satisfied by the candidate UE and the confidence corresponding to each satisfied screening condition Generally speaking, the more the filter conditions are met, the higher the confidence and the greater the similarity.

Optionally, the UE indication information may also include an identifier of the UE in the source domain.

Optionally, the UE indication information may also include screening conditions that the candidate UEs meet and confidence levels corresponding to the screening conditions.

In this embodiment, by receiving a migration learning request sent by a third-party device, wherein the migration learning request carries the identity of the UE in the source domain and screening condition information for selecting candidate UEs, and then according to the screening condition information and the UE in the source domain identify multiple candidate UEs, and then send UE indication information to a third-party device, where the UE indication information includes identifications of multiple candidate UEs, and the UE indication information is used for the third-party device to determine targets from the multiple candidate UEs In this way, the third-party device can obtain multiple candidate UEs determined to be similar to the source domain UE according to the screening condition information after providing the identity of the source domain UE and the screening condition information for selecting candidate UEs, and then, in the subsequent In the step, the UE in the target domain is determined from the candidate UEs, so as to realize the selection of the candidate UEs, and then in the subsequent process, the application layer data of the candidate UEs can be used for migration learning.

Further, in the embodiment of the present application, since the third-party device cannot obtain the UE information of the candidate UE, that is, sensitive UE information will not be exposed to the third-party device, therefore, the security of the UE information can be guaranteed.

Please refer to FIG. 7, which shows an optional implementation method for determining candidate UEs provided by the embodiment of the present application. As shown in FIG. 7, the implementation method includes the following steps:

Step 701, the first network element sends a first UE information acquisition request to the second network element in the core network.

As mentioned above, the second network element may include the NWDAF network element in the core network. The NWDAF network element is a data-aware analysis network element, which can automatically perceive and analyze the network based on network data. The NWDAF network element can be obtained from NF network elements, AF (Chinese: Application Function, English: Application Function) network elements and OAM (Chinese: Operation Management and Maintenance, English: Operation Administration and Maintenance) network elements collect original network data, and perform Intelligent analysis, output analysis result information. For example, the analysis result information output by the NWDAF network element can be the current or historical movement trajectory of the UE, the future expected movement trajectory of the UE, the current or historical area of the UE, the expected future area of the UE, and the future expected communication quality of the UE. , the types of services currently or historically requested by the UE, and the types of services expected to be requested by the UE in the future.

Wherein, the process of the first network element sending the first UE information acquisition request to the NWDAF network element may include steps A1 to A2:

A1. The first network element determines multiple analysis tasks according to the screening condition information.

Exemplarily, the analysis task may be a "Network Performance (Chinese: Network Performance)" task, which is used to provide statistics or forecast gNB resource utilization in the AF demand area, the number of UEs, PDU (Chinese: Protocol Data Unit , English: Protocol Data Unit) session creation success rate, handover success rate, etc. This analysis task can also be the "UE mobility (Chinese: UE mobility)" task, which is used to provide UE location statistics or Location prediction information.

A2. The first network element sends the first UE information acquisition request to the NWDAF network element.

Wherein, the first UE information acquisition request carries the identification of the UE in the source domain and the identification of each analysis task (English: Analytics ID), and the first UE information acquisition request is used to instruct the NWDAF network element to separately analyze the UE in the source domain and the analysis tasks based on the source domain. Multiple UEs selected by the UE execute each analysis task.

Wherein, the identifier of the analysis task may be, for example, the identifier of the "Network Performance (Chinese: Network Performance)" task, or may be the identifier of the "UE mobility (Chinese: UE Mobility)" task.

Step 702, the first network element receives the first feedback information sent by the second network element.

Wherein, the first feedback information includes the first UE information of the source domain UE and the first UE information of multiple UEs selected based on the source domain UE.

In the case that the second network element is a NWDAF network element, the first UE information is analysis result information obtained after the NWDAF network element performs various analysis tasks.

It should be pointed out that, while sending the analysis result information, the NWDAF network element can also send the confidence (English: confidence) corresponding to each analysis result information to the first network element, and the confidence degree is used to indicate the corresponding analysis result information. Credibility.

Step 703, the first network element determines multiple candidate UEs according to the first feedback information and screening condition information.

In an optional embodiment of the present application, the first feedback information may also include the identifiers of multiple UEs selected based on the UE in the source domain, and the first network element may also send a second UE information acquisition request to a third network element of the core network. request, as described above, the third network element may be an NF network element (including an AMF network element and/or an SMF network element), and the second UE information acquisition request carries the source domain UE identifier and the UE selected based on the source domain UE Then, the first network element can receive the second feedback information sent by the third network element. The second feedback information includes the second UE information of the source domain UE and multiple UEs selected based on the source domain UE. Second UE information of the UE, where the second UE information includes at least one of mobility information and session information, and then, the first network element may, according to the first UE information of the source domain UE and based on the multiple UE information selected by the source domain UE The first UE information, the second feedback information and the screening condition information of a UE to determine a plurality of candidate UEs.

In an optional embodiment of the present application, the first network element may determine multiple candidate UEs from multiple candidate UEs that are expected to meet the filtering condition information in the future according to the UE information of the UE in the source domain and the UE information of multiple candidate UEs and/or, according to the UE information of the UE in the source domain and the UE information of multiple candidate UEs, determine multiple candidate UEs that currently meet the screening condition information from multiple candidate UEs.

Wherein, corresponding to the filtering condition indicated by the filtering condition information, the first network element determines from the plurality of candidate UEs which are expected to meet the filtering condition information in the future according to the UE information of the UE in the source domain and the UE information of the plurality of candidate UEs. Multiple candidate UEs, including at least one of the following:

According to the UE information of the source domain UE and the UE information of multiple candidate UEs, determine the candidate UE that is expected to arrive in the area where the source domain UE is currently or historically located in the future from the multiple candidate UEs;

According to the UE information of the UE in the source domain and the UE information of a plurality of candidate UEs, determine a candidate UE whose future expected movement trajectory is similar to the current or historical movement trajectory of the source domain UE from the plurality of candidate UEs;

According to the UE information of the UE in the source domain and the UE information of multiple candidate UEs, determine the candidate UEs whose communication quality is expected to be better than the communication quality threshold in the future target time period from the multiple candidate UEs;

According to the UE information of the UE in the source domain and the UE information of multiple candidate UEs, determine from multiple candidate UEs the candidate UE whose type of service expected to be requested in the future is similar to the type of service currently or historically requested by the source domain UE.

In addition, corresponding to the filtering conditions indicated by the filtering condition information, the first network element determines from multiple candidate UEs that currently meet the filtering condition information according to the UE information of the UE in the source domain and the UE information of multiple candidate UEs. Multiple candidate UEs, including at least one of the following:

According to the UE information of the source domain UE and the UE information of multiple candidate UEs, determine the candidate UE currently located in the same area as the source domain UE from the multiple candidate UEs;

According to the UE information of the UE in the source domain and the UE information of multiple candidate UEs, determine a candidate UE whose current trajectory is similar to the current trajectory of the source domain UE from multiple candidate UEs;

According to the UE information of the UE in the source domain and the UE information of multiple candidate UEs, determine the candidate UE whose current communication quality is better than the communication quality threshold from the multiple candidate UEs;

According to the UE information of the UE in the source domain and the UE information of multiple candidate UEs, a candidate UE whose type of service currently requested is similar to the type of service currently requested by the UE of the source domain is determined from the multiple candidate UEs.

In an optional embodiment of the present application, before sending the UE indication information to the third-party device, the UE indication information needs to be generated first, please refer to FIG. 8, the sending process includes the following steps:

Step 801, the first network element sorts the plurality of candidate UEs according to the number of screening conditions that each candidate UE meets and the confidence that each candidate UE meets the screening condition, or, according to the number of screening conditions that each candidate UE meets .

In a possible sorting manner, the first network element may first sort the plurality of candidate UEs in descending order according to the number of candidate UEs meeting the filtering conditions. If there are multiple candidate UEs meeting the filtering conditions If the numbers are equal, the plurality of candidate UEs are sorted according to the descending order of the confidence of the candidate UEs meeting the screening conditions.

Please refer to Table 1, which shows an exemplary result of ranking candidate UEs.

Table 1

In Table 1, the filter conditions that are met are recorded as 1, and the filter conditions that are not met are recorded as 0. The percentage in Table 1 is the confidence level. Please refer to Table 1. Target UE-1 meets the filter condition (1), and, meets The confidence level of screening condition (1) is 90%, Target UE-2 does not meet the screening condition of (2), and the confidence level of not meeting the screening condition of (2) is 75%.

Step 802, the first network element generates UE indication information according to the sorting result.

Please refer to FIG. 9, which shows a flow chart of a terminal selection method provided by an embodiment of the present application. The terminal selection method is applied to the third-party device 101 in FIG. 4 or FIG. 5. As shown in FIG. 9, the The terminal selection method includes the following steps:

Step 901, the third-party device sends a migration learning request to the first network element of the core network.

Currently located in the same area as the UE in the source domain;

The current communication quality meets the communication quality threshold;

The number k of candidate UEs.

Step 902, the third-party device receives UE indication information sent by the first network element.

Wherein, the UE indication information includes identifiers of multiple candidate UEs, and the multiple candidate UEs are determined according to the screening condition information and the identifiers of the UEs in the source domain.

Step 903, the third-party device determines the UE in the target domain from multiple candidate UEs according to the UE indication information.

Please refer to FIG. 10, which shows a flow chart of a terminal selection method provided by an embodiment of the present application. The terminal selection method is applied to the implementation environment shown in FIG. 5. As shown in FIG. 10, the terminal selection method includes the following step:

Step 1001, a third-party device sends a migration learning request to a first network element.

Wherein, the transfer learning request carries the identity of the UE in the source domain, screening condition information for selecting candidate UEs, and a confidence threshold, where the screening condition information is used to indicate multiple screening conditions, and the multiple screening conditions include the number k of candidate UEs.

Step 1002, the first network element determines multiple analysis tasks according to the filtering condition information.

Step 1003, the first network element sends a first UE information acquisition request to the NWDAF network element.

Wherein, the first UE information acquisition request carries the identifier of the source domain UE and the identifiers of each analysis task, and the first UE information acquisition request is used to instruct the NWDAF network element to separately perform the source domain UE and the multiple UEs selected based on the source domain UE. The UE performs various analysis tasks.

Step 1004, the NWDAF network element sends the analysis result information of the source domain UE, the identifiers of multiple UEs selected based on the source domain UE, the analysis result information of multiple UEs selected based on the source domain UE, and each analysis result information to the first network element. The confidence level of the resulting information.

Step 1005, the first network element sends a second UE information acquisition request to an NF network element (AMF network element and/or SMF network element).

Wherein, the second UE information acquisition request carries the identifier of the UE in the source domain and the identifiers of multiple UEs selected based on the UE in the source domain.

Step 1006, the NF network element sends the second UE information of the source domain UE and the second UE information of multiple UEs selected based on the source domain UE to the first network element.

Wherein, the second UE information includes at least one of mobility information and session information.

Step 1007, the first network element based on the analysis result information of the source domain UE, the analysis result information of multiple UEs selected based on the source domain UE, the second UE information of the source domain UE, and the multiple UEs selected based on the source domain UE The second UE information and the screening condition information are determined to meet at least n screening conditions, and there are k candidate UEs whose corresponding confidences are greater than the confidence threshold in the matching screening conditions.

Step 1008, the first network element sorts the plurality of candidate UEs according to the number of screening conditions that each candidate UE meets and the confidence of each candidate UE meeting the screening conditions, or according to the number of screening conditions that each candidate UE meets .

Step 1009, the first network element generates UE indication information according to the sorting result.

Step 1010, the first network element sends the generated UE indication information to a third-party device.

It should be understood that although the various steps in the flow charts in FIGS. 6-10 are displayed sequentially as indicated by the arrows, these steps are not necessarily executed sequentially in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least some of the steps in Figures 6-10 may include a plurality of sub-steps or stages, these sub-steps or stages are not necessarily executed at the same time, but may be executed at different times, these sub-steps or stages The order of execution is not necessarily performed sequentially, but may be performed alternately or alternately with at least a part of other steps or sub-steps or stages of other steps.

Please refer to FIG. 11 , which shows a schematic diagram of a terminal selection device 1100 provided by an embodiment of the present application. As shown in FIG. 11 , the terminal selection device 1100 includes a receiving module 1101 , a determining module 1102 and a sending module 1103 .

Wherein, the receiving module 1101 is configured to receive a transfer learning request sent by a third-party device, where the transfer learning request carries the identity of the UE in the source domain and screening condition information for selecting candidate UEs.

The determining module 1102 is configured to determine a plurality of candidate UEs according to the screening condition information and the identity of the UE in the source domain.

The sending module 1103 is configured to send UE indication information to the third-party device, the UE indication information includes identities of the plurality of candidate UEs, and the UE indication information is used for the third-party device to determine from the plurality of candidate UEs Target domain UE.

In an optional embodiment of the present application, the determining module 1102 is specifically configured to: send a first UE information acquisition request to a second network element in the core network; receive first feedback information sent by the second network element, the The first feedback information includes the first UE information of the source domain UE and the first UE information of multiple UEs selected based on the source domain UE; according to the first feedback information and the screening condition information, determine the multiple candidate UEs .

In an optional embodiment of the present application, the determination module 1102 is specifically configured to: determine a plurality of analysis tasks according to the screening condition information; send the first UE information acquisition request to the NWDAF network element, and the first UE information The acquisition request carries the identity of the source domain UE and the identity of each analysis task, and the first UE information acquisition request is used to instruct the NWDAF network element to perform the analysis on the source domain UE and multiple UEs selected based on the source domain UE. each analysis task.

In an optional embodiment of the present application, the first UE information is analysis result information obtained after the NWDAF network element performs each of the analysis tasks.

In an optional embodiment of the present application, the first feedback information further includes identities of multiple UEs selected based on the source domain UE, and the determining module 1102 is specifically configured to: send the second Two UE information acquisition requests, the second UE information acquisition request carries the identity of the UE in the source domain and the identity of multiple UEs selected based on the UE in the source domain; receiving the second feedback information sent by the third network element, the first The second feedback information includes second UE information of the source domain UE and second UE information of multiple UEs selected based on the source domain UE, where the second UE information includes at least one of mobility information and session information; according to The multiple candidate UEs are determined based on the first UE information of the source domain UE, the first UE information of multiple UEs selected based on the source domain UE, the second feedback information, and the screening condition information.

In an optional embodiment of the present application, the filtering condition information is used to indicate multiple filtering conditions; the multiple filtering conditions include at least one of the following:

Currently located in the same area as the UE in the source domain;

The current communication quality meets the communication quality threshold;

The number of candidate UEs.

In an optional embodiment of the present application, the determination module 1102 is specifically configured to: determine at least n candidate UEs that meet the screening conditions according to the screening condition information and the identity of the UE in the source domain, where n is a positive integer.

In an optional embodiment of the present application, the transfer learning request also carries a confidence threshold, and the determining module 1102 is specifically configured to: determine that at least n of the filtering conditions are met according to the filtering condition information and the identity of the UE in the source domain , and, among the matching screening conditions, there is the candidate UE whose corresponding confidence is greater than the screening condition of the confidence threshold.

In an optional embodiment of the present application, the UE indication information also includes the screening conditions met by the candidate UE and the confidence level corresponding to the screening conditions met; or,

The UE indication information also includes the filtering condition that the candidate UE meets, the confidence level corresponding to the matching filtering condition, and the identity of the UE in the source domain.

In an optional embodiment of the present application, the first network element is an NF network element used to determine the candidate UE.

In an optional embodiment of the present application, the third-party device is an OTT application server or terminal.

Please refer to FIG. 12, as shown in FIG. 12, the embodiment of the present application also provides another terminal selection device 1200. In addition to the modules included in the terminal selection device 1100, the terminal selection device optionally also includes a sorting module 1104 and generating module 1105.

The sorting module 1104 is configured to, according to the number of filtering conditions that each candidate UE meets and the confidence corresponding to the filtering conditions that each of the candidate UEs meet, or, according to the number of filtering conditions that each of the candidate UEs meet, sort the Multiple candidate UEs are sorted.

The generating module 1105 is configured to generate the UE indication information according to the sorting result.

The implementation principle and technical effect of the device for selecting a terminal provided in the foregoing embodiments are similar to those of the foregoing method embodiments, and will not be repeated here.

For specific limitations on the terminal selection device, refer to the above-mentioned limitations on the terminal selection method, which will not be repeated here. Each module in the above-mentioned device for selecting a terminal may be fully or partially realized by software, hardware or a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, and can also be stored in the memory of the computer device in the form of software, so that the processor can invoke and execute the corresponding operations of the above-mentioned modules.

Please refer to FIG. 13 , which shows a schematic diagram of a terminal selection device 1300 provided by an embodiment of the present application. As shown in FIG. 13 , the terminal selection device 1300 includes a sending module 1301 , a receiving module 1302 and a determining module 1303 .

The sending module 1301 is configured to send a transfer learning request to a first network element of the core network, where the transfer learning request carries an identifier of a UE in a source domain and screening condition information for selecting a candidate UE.

The receiving module 1302 is configured to receive UE indication information sent by the first network element, where the UE indication information includes identifiers of a plurality of candidate UEs, and the plurality of candidate UEs are determined according to the screening condition information and the identifier of the UE in the source domain of.

The determining module 1303 is configured to determine a target domain UE from the plurality of candidate UEs according to the UE indication information.

Currently located in the same area as the UE in the source domain;

The current communication quality meets the communication quality threshold;

The number of candidate UEs.

In an optional embodiment of the present application, the UE indication information further includes a screening condition that the candidate UE meets and a confidence level corresponding to the screening condition.

In an optional embodiment of the present application, the transfer learning request also carries a confidence threshold.

In one embodiment, a communication device is provided. The communication device may be a network element device or a terminal in a core network, and its internal structure diagram may be shown in FIG. 14 . FIG. 14 is an internal structure of the communication device in an embodiment structure diagram. The communications device includes a processor, memory, and transceivers connected by an internal link pathway, such as a system bus. Wherein, the processor of the communication device is used to provide calculation and control capabilities. The memory of the communication device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The transceiver of the communication device is used to communicate with external devices in a wired or wireless manner, and the wireless manner can be realized by an operator's network or other technologies. When the computer program is executed by the processor, a terminal selection method is realized.

Those skilled in the art can understand that the structure shown in Figure 14 is only a block diagram of a partial structure related to the solution of this application, and does not constitute a limitation on the communication equipment to which the solution of this application is applied. The specific communication equipment can be More or fewer components than shown in the figures may be included, or some components may be combined, or have a different arrangement of components.

In one embodiment, a communication device is provided, including a memory and a processor, a computer program is stored in the memory, and the processor implements the following steps when executing the computer program:

Receive a migration learning request sent by a third-party device, the migration learning request carries the identity of the UE in the source domain and screening condition information for selecting candidate UEs; determine multiple candidate UEs according to the filtering condition information and the identification of the UE in the source domain; UE indication information is sent to the third-party device, where the UE indication information includes identities of the plurality of candidate UEs, and the UE indication information is used for the third-party device to determine a target domain UE from the plurality of candidate UEs.

In one embodiment, when the processor executes the computer program, the following steps are further implemented: sending a first UE information acquisition request to a second network element in the core network; receiving first feedback information sent by the second network element, the The first feedback information includes the first UE information of the source domain UE and the first UE information of multiple UEs selected based on the source domain UE; according to the first feedback information and the screening condition information, determine the multiple candidate UEs .

In one of the embodiments, the second network element includes a NWDAF network element, and when the processor executes the computer program, the following steps are further implemented: determining a plurality of analysis tasks according to the screening condition information; sending the first analysis task to the NWDAF network element A UE information acquisition request, the first UE information acquisition request carries the identity of the source domain UE and the identity of each analysis task, and the first UE information acquisition request is used to instruct the NWDAF network element to separately perform the source domain UE and based on the Multiple UEs selected by the UE in the source domain perform each analysis task.

In one of the embodiments, the first UE information is analysis result information obtained after the NWDAF network element performs each analysis task.

In one of the embodiments, the first feedback information further includes identifiers of multiple UEs selected based on the source domain UE, and the processor further implements the following steps when executing the computer program: sending the second Two UE information acquisition requests, the second UE information acquisition request carries the identity of the UE in the source domain and the identity of multiple UEs selected based on the UE in the source domain; receiving the second feedback information sent by the third network element, the first The second feedback information includes second UE information of the source domain UE and second UE information of multiple UEs selected based on the source domain UE, where the second UE information includes at least one of mobility information and session information; according to The multiple candidate UEs are determined based on the first UE information of the source domain UE, the first UE information of multiple UEs selected based on the source domain UE, the second feedback information, and the screening condition information.

In one of the embodiments, the filter condition information is used to indicate multiple filter conditions; the multiple filter conditions include at least one of the following:

Currently located in the same area as the UE in the source domain;

The current communication quality meets the communication quality threshold;

The number of candidate UEs.

In one embodiment, when the processor executes the computer program, the following steps are further implemented: according to the screening condition information and the identity of the UE in the source domain, determine the candidate UEs meeting at least n screening conditions, where n is a positive integer.

In one of the embodiments, the migration learning request also carries a confidence threshold, and the processor further implements the following steps when executing the computer program: according to the screening condition information and the identity of the UE in the source domain, determine that at least n of the screening conditions are met , and, among the matching screening conditions, there is the candidate UE whose corresponding confidence is greater than the screening condition of the confidence threshold.

In one of the embodiments, the UE indication information also includes the filter condition that the candidate UE meets and the confidence level corresponding to the filter condition; or,

In one of the embodiments, when the processor executes the computer program, the following steps are further implemented: according to the number of screening conditions that each candidate UE meets and the confidence level corresponding to the screening conditions that each candidate UE meets, or, according to each of the candidate UEs The multiple candidate UEs are sorted according to the number of screening conditions that the candidate UEs meet; the UE indication information is generated according to the sorting results.

In one of the embodiments, the first network element is an NF network element used to determine the candidate UE.

In one of the embodiments, the third-party device is an OTT application server or terminal.

The implementation principle and technical effect of the communication device provided in the foregoing embodiments are similar to those of the foregoing method embodiments, and will not be repeated here.

Sending a migration learning request to the first network element of the core network, the migration learning request carrying the identity of the UE in the source domain and screening condition information for selecting candidate UEs; receiving UE indication information sent by the first network element, the UE indication information includes multiple An identifier of a candidate UE, the plurality of candidate UEs is determined according to the screening condition information and the identifier of the source domain UE; and a target domain UE is determined from the plurality of candidate UEs according to the UE indication information.

In one of the embodiments, the filtering condition information is used to indicate multiple filtering conditions; the multiple filtering conditions include at least one of the following:

Currently located in the same area as the UE in the source domain;

The current communication quality meets the communication quality threshold;

The number of candidate UEs.

In one of the embodiments, the UE indication information also includes the screening condition that the candidate UE meets and the confidence level corresponding to the matching screening condition.

In one of the embodiments, the transfer learning request also carries a confidence threshold.

In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

In one of the embodiments, when the computer program is executed by the processor, the following steps are further implemented: sending a first UE information acquisition request to a second network element in the core network; receiving first feedback information sent by the second network element, the The first feedback information includes the first UE information of the source domain UE and the first UE information of multiple UEs selected based on the source domain UE; according to the first feedback information and the screening condition information, determine the multiple candidate UEs .

In one embodiment, the second network element includes a NWDAF network element, and when the computer program is executed by the processor, the following steps are further implemented: determining a plurality of analysis tasks according to the filtering condition information; sending the first analysis task to the NWDAF network element A UE information acquisition request, the first UE information acquisition request carries the identity of the source domain UE and the identity of each analysis task, and the first UE information acquisition request is used to instruct the NWDAF network element to separately perform the source domain UE and based on the Multiple UEs selected by the UE in the source domain perform each analysis task.

In one of the embodiments, the first feedback information further includes identifiers of multiple UEs selected based on the source domain UE, and when the computer program is executed by the processor, the following steps are further implemented: sending the first feedback information to the third network element of the core network Two UE information acquisition requests, the second UE information acquisition request carries the identity of the UE in the source domain and the identity of multiple UEs selected based on the UE in the source domain; receiving the second feedback information sent by the third network element, the first The second feedback information includes second UE information of the source domain UE and second UE information of multiple UEs selected based on the source domain UE, where the second UE information includes at least one of mobility information and session information; according to The multiple candidate UEs are determined based on the first UE information of the source domain UE, the first UE information of multiple UEs selected based on the source domain UE, the second feedback information, and the screening condition information.

Currently located in the same area as the UE in the source domain;

The current communication quality meets the communication quality threshold;

The number of candidate UEs.

In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: according to the screening condition information and the identity of the UE in the source domain, determine the candidate UEs meeting at least n screening conditions, where n is a positive integer.

In one embodiment, the transfer learning request also carries a confidence threshold, and the computer program is executed by the processor to further implement the following steps: according to the screening condition information and the identity of the UE in the source domain, determine that at least n of the screening conditions are met , and, among the matching screening conditions, there is the candidate UE whose corresponding confidence is greater than the screening condition of the confidence threshold.

In one of the embodiments, when the computer program is executed by the processor, the following steps are further implemented: according to the number of screening conditions that each candidate UE meets and the confidence level corresponding to the screening conditions that each candidate UE meets, or, according to the number of screening conditions that each candidate UE meets, The multiple candidate UEs are sorted according to the number of screening conditions that the candidate UEs meet; the UE indication information is generated according to the sorting results.

The implementation principle and technical effect of the computer-readable storage medium provided by the foregoing embodiments are similar to those of the foregoing method embodiments, and details are not repeated here.

Currently located in the same area as the UE in the source domain;

The current communication quality meets the communication quality threshold;

The number of candidate UEs.

In one of the embodiments, the UE indication information further includes the filtering condition that the candidate UE meets and the confidence level corresponding to the matching filtering condition.

Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be implemented through computer programs to instruct related hardware, and the computer programs can be stored in a non-volatile computer-readable memory In the medium, when the computer program is executed, it may include the processes of the embodiments of the above-mentioned methods. Wherein, any references to memory, storage, database or other media used in the various embodiments provided in the present application may include non-volatile and/or volatile memory. Nonvolatile memory can include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Chain Synchlink DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

The technical features of the above embodiments can be combined arbitrarily. To make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be It is considered to be within the range described in this specification. The above-mentioned embodiments only represent several implementation modes of the present application, and the description thereof is relatively specific and detailed, but it should not be construed as limiting the scope of the patent for the invention. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the scope of protection of the patent application should be based on the appended claims.

Claims

A terminal selection method, characterized in that it is used in a first network element of a core network, and the method includes:

receiving a migration learning request sent by a third-party device, where the migration learning request carries the identity of the UE in the source domain and screening condition information for selecting candidate UEs;

Determine a plurality of candidate UEs according to the screening condition information and the identity of the UE in the source domain;

sending UE indication information to the third-party device, where the UE indication information includes identities of the plurality of candidate UEs, and the UE indication information is used for the third-party device to determine targets from the plurality of candidate UEs Domain UE.
The method according to claim 1, wherein the determining a plurality of candidate UEs according to the screening condition information and the identity of the UE in the source domain comprises:

sending a first UE information acquisition request to a second network element in the core network;

Receive first feedback information sent by the second network element, where the first feedback information includes first UE information of the source domain UE and first UE information of multiple UEs selected based on the source domain UE;

Determine the plurality of candidate UEs according to the first feedback information and the filtering condition information.
The method according to claim 2, wherein the second network element comprises a NWDAF network element, and before sending the first UE information acquisition request to the second network element in the core network, the method further comprises:

Determine multiple analysis tasks according to the screening condition information;

The sending the first UE information acquisition request to the second network element in the core network includes:

Sending the first UE information acquisition request to the NWDAF network element, the first UE information acquisition request carrying the identifier of the UE in the source domain and the identifiers of each analysis task, and the first UE information acquisition request using Instructing the NWDAF network element to perform each of the analysis tasks on the source domain UE and multiple UEs selected based on the source domain UE.
The method according to claim 3, wherein the first UE information is analysis result information obtained after the NWDAF network element performs each of the analysis tasks.
The method according to claim 3, wherein the first feedback information further includes identities of multiple UEs selected based on the source domain UE, and the first feedback information and the filter condition information to determine the plurality of candidate UEs, including:

Sending a second UE information acquisition request to a third network element of the core network, where the second UE information acquisition request carries the identifier of the source domain UE and the identifiers of multiple UEs selected based on the source domain UE;

receiving second feedback information sent by the third network element, where the second feedback information includes second UE information of the source domain UE and second UE information of multiple UEs selected based on the source domain UE, The second UE information includes at least one of mobility information and session information;

Determine the multiple candidate UEs according to the first UE information of the source domain UE, the first UE information of multiple UEs selected based on the source domain UE, the second feedback information, and the screening condition information .
The method according to any one of claims 1 to 5, wherein the filter condition information is used to indicate multiple filter conditions; the multiple filter conditions include at least one of the following:

It is expected to arrive in the area where the UE in the source domain is currently or historically located in the future;

The expected movement trajectory in the future is similar to the current or historical movement trajectory of the UE in the source domain;

The expected communication quality within the future target time period meets the communication quality threshold;

The type of service expected to be requested in the future is similar to the type of service currently or historically requested by the UE in the source domain;

Currently located in the same area as the UE in the source domain;

The current movement trajectory is similar to the current movement trajectory of the UE in the source domain;

The current communication quality meets the communication quality threshold;

The type of the service currently requested is similar to the type of service currently requested by the UE in the source domain;

The number of candidate UEs.
The method according to claim 6, wherein the determining a plurality of candidate UEs according to the screening condition information and the identity of the UE in the source domain comprises:

Determine, according to the screening condition information and the identifier of the UE in the source domain, the candidate UEs meeting at least n screening conditions, where n is a positive integer.
The method according to claim 6, wherein the migration learning request also carries a confidence threshold, and determining a plurality of candidate UEs according to the screening condition information and the identity of the UE in the source domain includes:

According to the screening condition information and the identity of the UE in the source domain, determine that at least n screening conditions are met, and there are candidates for the screening conditions whose corresponding confidence is greater than the confidence threshold among the matching screening conditions. UE.
The method according to claim 6, wherein the UE indication information further includes the filter conditions met by the candidate UE and the confidence corresponding to the met filter conditions; or,

The UE indication information also includes the filtering conditions that the candidate UEs meet, the confidence level corresponding to the matching filtering conditions, and the identity of the UE in the source domain.
The method according to claim 9, wherein before the UE indication information is sent to the third-party device, the method further comprises:

According to the number of screening conditions met by each of the candidate UEs and the confidence corresponding to the screening conditions met by each of the candidate UEs, or according to the number of screening conditions met by each of the candidate UEs, the plurality of candidates UE sorts;

The UE indication information is generated according to the sorting result.
The method according to any one of claims 1 to 5, wherein the first network element is an NF network element for determining the candidate UE.
The method according to any one of claims 1 to 5, wherein the third-party device is an OTT application server or a terminal.
A terminal selection method, characterized in that it is used in a third-party device, and the method includes:

Sending a migration learning request to the first network element of the core network, where the migration learning request carries the identity of the UE in the source domain and screening condition information for selecting candidate UEs;

Receive UE indication information sent by the first network element, where the UE indication information includes identifiers of multiple candidate UEs, and the multiple candidate UEs are determined according to the screening condition information and the identifier of the UE in the source domain;

Determine a target domain UE from the plurality of candidate UEs according to the UE indication information.
The method according to claim 13, wherein the filtering condition information is used to indicate multiple filtering conditions; the multiple filtering conditions include at least one of the following:

It is expected to arrive in the area where the UE in the source domain is currently or historically located in the future;

The expected movement trajectory in the future is similar to the current or historical movement trajectory of the UE in the source domain;

The expected communication quality within the future target time period meets the communication quality threshold;

The type of service expected to be requested in the future is similar to the type of service currently or historically requested by the UE in the source domain;

Currently located in the same area as the UE in the source domain;

The current movement trajectory is similar to the current movement trajectory of the UE in the source domain;

The current communication quality meets the communication quality threshold;

The type of the service currently requested is similar to the type of service currently requested by the UE in the source domain;

The number of candidate UEs.
The method according to claim 13 or 14, wherein the UE indication information further includes screening conditions met by the candidate UE and confidence levels corresponding to the screening conditions met.
The method according to claim 13 or 14, wherein the transfer learning request also carries a confidence threshold.
The method according to claim 13 or 14, wherein the first network element is an NF network element for determining the candidate UE.
The method according to claim 13 or 14, wherein the third-party device is an OTT application server or a terminal.
A terminal selection device, characterized in that the device includes:

The receiving module is configured to receive a migration learning request sent by a third-party device, where the migration learning request carries the identity of the UE in the source domain and screening condition information for selecting candidate UEs;

A determining module, configured to determine a plurality of candidate UEs according to the screening condition information and the identity of the UE in the source domain;

A sending module, configured to send UE indication information to the third-party device, where the UE indication information includes identities of the plurality of candidate UEs, and the UE indication information is used for the third-party device to obtain from the plurality of candidate UEs. The target domain UE is determined among the candidate UEs.
A terminal selection device, characterized in that the device includes:

A sending module, configured to send a migration learning request to a first network element of the core network, where the migration learning request carries the identity of the UE in the source domain and screening condition information for selecting candidate UEs;

A receiving module, configured to receive UE indication information sent by the first network element, where the UE indication information includes identities of a plurality of candidate UEs, and the plurality of candidate UEs are selected according to the filtering condition information and the source Determined by the identity of the domain UE;

A determining module, configured to determine a target domain UE from the plurality of candidate UEs according to the UE indication information.
A communication device, comprising: a processor, a memory, and a transceiver, the processor, the memory, and the transceiver communicate with each other through an internal connection path, wherein,

The memory is used to store program codes;

The processor is configured to call the program code stored in the memory, so as to cooperate with the transceiver to realize the steps of the method in any one of claims 1 to 12, or to cooperate with the transceiver to realize the claims The step of any one of 13 to 18.
A computer-readable storage medium, on which a computer program is stored, characterized in that, when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 12 are implemented, or, the computer The steps of the method described in any one of claims 13 to 18 are implemented when the program is executed by the processor.