Data collection method and apparatus, and device and readable storage medium

Abstract

The present application relates to a data collection method and apparatus, and a device and a readable storage medium. The method comprises: receiving a data collection configuration sent by means of a network device; on the basis of the data collection configuration, collecting and / or recording first training data; and sending the first training data to the network device, wherein the first training data is used for training and / or updating a target model, the target model is used for implementing a target use case, and the target use case comprises at least one of the following use cases: an RRM prediction use case, an event prediction use case and an RLF prediction use case. The present application provides a data collection method for an RRM prediction use case, an event prediction use case and an RLF prediction use case, such that training data is provided on the basis of the method so as to train and / or update a model for implementing the use cases, that is, a target model.

Description

Data collection methods, apparatus, devices and readable storage media

[0001] Related applications

[0002] This application claims priority to Chinese patent application filed on January 26, 2025, with application number 2025101245682, entitled "Data Collection Method, Apparatus, Device and Readable Storage Medium", the entire contents of which are incorporated herein by reference. Technical Field

[0003] This application relates to the field of wireless communication technology, and in particular to a data collection method, apparatus, device, and readable storage medium. Background Technology

[0004] Currently, models such as machine learning models (ML models) and artificial intelligence models (AI models) can be deployed in communication systems, which helps improve the overall performance of the communication system. Models deployed in communication systems can implement different use cases, among which RRM (Radio Resource Management) prediction use cases, event prediction use cases, and RLF (Radio Link Failure) prediction use cases are some typical examples.

[0005] Currently, there is no consideration on how to collect data to provide training data for training and / or updating the models that implement these use cases for the above-mentioned use cases. Summary of the Invention

[0006] This application provides a data collection method, apparatus, device, and readable storage medium.

[0007] In a first aspect, this application provides a data collection method for a terminal, comprising:

[0008] The system receives a data collection configuration sent by a network device; collects and / or records first training data according to the data collection configuration; and sends the first training data to the network device. The first training data is used to train and / or update a target model, and the target model is used to implement a target use case, which includes at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case.

[0009] In some embodiments, the method further includes: sending a configuration request to the network device, the configuration request being used to request the network device to send the data collection configuration.

[0010] In some embodiments, where the target use case includes the RRM prediction use case, the configuration request information includes at least one of the following:

[0011] First model type information, which indicates whether the target model is a single-cell model or a cell cluster model;

[0012] Basic request information;

[0013] The basic request information includes at least one of the following: request indication information, which indicates a request to the network device to configure the collection and / or recording of the first training data; purpose indication information, which indicates the purpose of training and / or updating the target model; reason indication information, which indicates the reason for training and / or updating the target model; model input and / or model output; model window length, which includes observation window length and / or prediction window length; additional conditions on the network side; additional conditions on the terminal side; and information about a third-party server, which is used to train and / or update the target model.

[0014] When the target use case includes the event prediction use case, the configuration request includes at least one of the following: the basic request information; the first model type information; the second model type information, which indicates that the target model is a model that directly performs event prediction, or indicates that the target model is a model that indirectly performs event prediction; the requested event type information; and the requested event configuration parameters.

[0015] When the target use case includes the RLF prediction use case, the configuration request includes at least one of the following: the basic request information; third model type information, which indicates that the target model is a model that directly performs RLF prediction, or indicates that the target model is a model that indirectly performs RLF prediction; and RLM configuration information.

[0016] In some embodiments, the data collection configuration includes at least one of the following configurations: a measurement configuration; a recording configuration; and a reporting configuration.

[0017] In some embodiments, where the target use case includes the RRM prediction use case, the measurement configuration includes: a basic measurement configuration;

[0018] The basic measurement configuration includes at least one of the following configurations: a cell list, comprising at least one cell, used to instruct the terminal to measure and / or record cells in the cell list; a beam list, comprising at least one beam, used to instruct the terminal to measure and / or record beams in the beam list; a frequency list, comprising at least one frequency, used to instruct the terminal to measure and / or record frequencies in the frequency list; first frequency indication information, used to indicate whether to measure and / or record cells on this frequency; first measurement object indication information, used to indicate whether to measure and / or record all measurement objects; second measurement object indication information, used to indicate whether to measure and / or record this measurement object; model-related parameters, used to implicitly indicate the measurement objects that need to be measured and / or recorded; a destination configuration; and a reference signal configuration.

[0019] When the target use case includes the event prediction use case, the measurement configuration includes at least one of the following configurations: the basic measurement configuration; the event configuration; the reporting configuration indication information, which indicates whether to measure and / or record the event and / or measurement object corresponding to this reporting configuration; and the third measurement object indication information, which indicates whether to measure and / or record all measurement objects that satisfy the event.

[0020] When the target use case includes the RLF prediction use case, the measurement configuration includes at least one of the following configurations: reference signal configuration; RLM configuration; destination configuration; cell identifier, which instructs the terminal to perform measurement and / or recording for the cell corresponding to the cell identifier; beam identifier, which instructs the terminal to perform measurement and / or recording for the beam corresponding to the beam identifier; serving beam indication information, which indicates whether to use the beam with the best signal quality as the serving beam for measurement and / or recording; and fourth measurement object indication information, which indicates whether to perform measurement and / or recording for this measurement object.

[0021] In some embodiments, the recording configuration includes at least one of the following configurations: start recording conditions; stop recording conditions.

[0022] In some embodiments, where the target use case includes the RRM prediction use case, the start recording conditions include: basic start conditions;

[0023] The basic start conditions include at least one of the following: signal quality is higher or lower than a signal quality threshold; reconfiguration or access to a new cell occurs; additional conditions on the network side change; additional conditions on the terminal side change.

[0024] When the target use case includes the event prediction use case, the start recording condition includes at least one of the following: the basic start condition; the event entry condition is met; the first trigger timer (TTT) corresponding to the event entry condition is started; the event is met; the first TTT times out; the difference between the event entry condition and the event entry condition reaches a preset threshold; a first recording period, wherein the first recording period is a preset duration before the event entry condition is met and / or a preset duration after the event entry condition is met; and a second recording period, wherein the second recording period is a preset duration before the event is met and / or a preset duration after the event is met.

[0025] When the target use case includes the RLF prediction use case, the start recording condition includes at least one of the following: the basic start condition; the target timer is turned on; the target value related to out-of-step reaches its maximum number; out-of-step is detected; RLF is detected; the number of out-of-step detections reaches a preset threshold; a third recording period, wherein the third recording period is a preset duration before the target condition is met and / or a preset duration after the target condition is met, and the target condition includes at least one of the following: the target timer is turned on, out-of-step is detected, RLF is detected; beam failure is detected; event trigger reporting is detected.

[0026] In some embodiments, where the target use case includes the RRM prediction use case, the stop recording condition includes: a basic stop condition;

[0027] The basic stopping conditions include at least one of the following: the duration of a single recording reaches a preset duration; the total recording duration reaches a preset duration; the number of recordings reaches a preset number; the signal quality is higher or lower than a signal quality threshold; the additional conditions on the network side change; the additional conditions on the terminal side change; the data volume reaches a preset data volume; a reconfiguration occurs or a new cell is accessed; the user enters an idle state or an inactive state; the battery is depleted; or the memory is insufficient.

[0028] When the target use case includes the event prediction use case, the stop recording condition includes at least one of the following: the basic stop condition; the event is met; the first TTT timeout corresponding to the event entry condition; the event exit condition is met; the second TTT is started corresponding to the event exit condition; the second TTT timeout corresponding to the event exit condition; the difference between the event exit condition and the event exit condition reaches a preset threshold.

[0029] When the target use case includes the RLF prediction use case, the stop recording condition includes at least one of the following: the basic stop condition; synchronization is detected; the number of times synchronization is detected reaches a preset number; the target timer stops; the target value related to synchronization reaches a maximum number; RLF is detected.

[0030] In some embodiments, where the target use case includes the RRM prediction use case, the reporting configuration includes at least one of the following configurations: indication information for reporting data of the top K cells with high signal quality, where K is a value greater than or equal to 1; indication information for reporting data of the top K beams with high signal quality, where K is a value greater than or equal to 1; indication information for reporting frequency point information; indication information for whether to record data by frequency point; basic reporting configuration; the basic reporting configuration includes at least one of the following configurations: information of a third-party server; information of an AI data processing entity in the core network; information of an AI data processing entity in OAM; measurement quantity reporting configuration; indication information for recording cell-level or beam-level measurement results; one or more network-side additional conditions; one or more terminal-side additional conditions; observation window length and / or prediction window length; N future instances and / or N future time slots and / or N future samples, where N is a value greater than or equal to 1; reporting method indication information, which is used to indicate that the model input data and model output data in the first training data are recorded separately, or to indicate that the model input data and model output data are recorded uniformly;

[0031] When the target use case includes the event prediction use case, the reporting configuration includes at least one of the following configurations: the basic reporting configuration; a first window length, wherein the first window length is the window length of the probability of the event occurring; and an event type configuration.

[0032] When the target use case includes the RLF prediction use case, the reporting configuration includes at least one of the following configurations: the basic reporting configuration; a second window length, wherein the second window length is the window length for the probability of RLF occurrence.

[0033] In some embodiments, where the target use case includes the RRM prediction use case, the first training data includes at least one of the following: identification information and / or measurement result information of the top K cells with high signal quality, where K is a value greater than or equal to 1; identification information and / or measurement result information of the top K beams with high signal quality, where K is a value greater than or equal to 1; frequency point information; basic data; the basic data includes at least one of the following: information of a third-party server; information of the AI ​​data processing entity in the core network; information of the AI ​​data processing entity in OAM; cell identification information and / or beam identification information; cell measurement result information and / or beam measurement result information; observation window length and / or prediction window length; N future instances and / or N future time slots and / or N future samples, where N is a value greater than or equal to 1; additional condition information on the network side; additional condition information on the terminal side.

[0034] When the target use case includes the event prediction use case, the first training data includes at least one of the following: the basic data; event configuration information; event occurrence time; event occurrence indication information, which indicates whether the corresponding event has occurred; the occurrence time when the event departure condition is met; the occurrence time when the event entry condition is met; indication information that the target model is a model that directly performs event prediction or a model that indirectly performs event prediction; unfiltered data obtained from each measurement during the first TTT and / or second TTT activation process, where the first TTT corresponds to the event entry condition and the second TTT corresponds to the event departure condition.

[0035] When the target use case includes the RLF prediction use case, the first training data includes at least one of the following: the basic data; the occurrence time of RLF; RLF occurrence indication information, which is used to indicate whether RLF has occurred; the time information of detecting out-of-step; the time information of the number of out-of-step detections reaching a preset number; the time information of the time information of the number of synchronization detections reaching a preset number; RLM configuration; and indication information that the target model is a model that directly performs RLF prediction or a model that indirectly performs RLF prediction.

[0036] In some embodiments, sending the first training data to the network device includes: receiving a data acquisition request sent by the network device; and in response to the data acquisition request, sending the first training data to the network device.

[0037] In some embodiments, the method further includes sending an availability indication to the network device, the availability indication being used to indicate that the terminal stores first training data.

[0038] Secondly, this application provides a data collection method for a network device, comprising:

[0039] Send data collection configuration to the terminal; receive first training data sent by the terminal; wherein the first training data is collected and / or recorded by the terminal based on the data collection configuration, the first training data is used to train and / or update the target model, the target model is used to implement the target use case, the target use case includes at least one of the following use cases: RRM prediction use case, event prediction use case, RLF prediction use case.

[0040] In some embodiments, the method further includes: receiving a configuration request sent by the terminal, the configuration request being used to request the network device to send the data collection configuration.

[0041] In some embodiments, the method further includes: sending a data acquisition request to the terminal, wherein the data acquisition request is used to request the terminal to send the first training data.

[0042] In some embodiments, the method further includes: receiving available indication information sent by the terminal, wherein the available indication information is used to indicate that the terminal stores the first training data.

[0043] Thirdly, this application also provides a data collection device for a terminal, the device comprising:

[0044] The receiving unit is used to receive data collection configurations sent by network devices;

[0045] A collection unit is configured to collect and / or record first training data according to the data collection configuration.

[0046] A sending unit is configured to send the first training data to the network device;

[0047] The first training data is used to train and / or update the target model, and the target model is used to implement the target use case, which includes at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case.

[0048] Fourthly, this application also provides a data collection apparatus for network devices, the apparatus comprising:

[0049] The sending unit is used to send data collection configuration to the terminal;

[0050] A receiving unit is configured to receive the first training data sent by the terminal;

[0051] Wherein, the first training data is collected and / or recorded by the terminal based on the data collection configuration, and the first training data is used to train and / or update the target model, the target model is used to implement the target use case, and the target use case includes at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case.

[0052] Fifthly, this application also provides a terminal, the terminal including a memory, a transceiver, and a processor:

[0053] A memory for storing computer programs; a transceiver for sending and receiving data under the control of the processor; and a processor for reading the computer programs from the memory and performing the following operations:

[0054] The transceiver is controlled to receive data collection configuration sent by the network device; first training data is collected and / or recorded according to the data collection configuration; the transceiver is controlled to send the first training data to the network device; wherein the first training data is used to train and / or update the target model, the target model is used to implement the target use case, and the target use case includes at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case.

[0055] Sixthly, this application also provides a network device, the network device including a memory, a transceiver, and a processor:

[0056] A memory for storing computer programs; a transceiver for sending and receiving data under the control of the processor; and a processor for reading the computer programs from the memory and performing the following operations:

[0057] Control the transceiver to send data collection configuration to the terminal;

[0058] Control the transceiver to receive the first training data sent by the terminal;

[0059] Wherein, the first training data is collected and / or recorded by the terminal based on the data collection configuration, and the first training data is used to train and / or update the target model, the target model is used to implement the target use case, and the target use case includes at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case.

[0060] In a seventh aspect, this application also provides a processor-readable storage medium, wherein the processor-readable storage medium stores a program for causing a processor to perform the method described in any of the first aspects above.

[0061] Eighthly, this application also provides a processor-readable storage medium, wherein the processor-readable storage medium stores a program for causing a processor to perform the method described in any of the second aspects above.

[0062] The aforementioned data collection method, apparatus, device, and readable storage medium involve receiving a data collection configuration sent by a network device from a terminal, then collecting and / or recording first training data according to the data collection configuration, and then sending the first training data to the network device. The first training data is used to train and / or update a target model, which is used to implement a target use case. The target use case includes at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case. Thus, a data collection method is provided for RRM prediction use case, event prediction use case, and RLF prediction use case, thereby providing training data based on this method to train and / or update the model implementing the aforementioned use case, i.e., the target model. Attached Figure Description

[0063] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the published drawings without creative effort.

[0064] Figure 1 is a schematic diagram of sub-use case 2 in some embodiments of RRM prediction use cases;

[0065] Figure 2 is a schematic diagram of sub-use case 3 in some embodiments of RRM prediction use cases;

[0066] Figure 3 is a schematic diagram of the LCM process of the model in some embodiments;

[0067] Figure 4 is a flowchart illustrating the data collection method in some embodiments;

[0068] Figure 5 is a flowchart illustrating another data collection method in some embodiments;

[0069] Figure 6 is a flowchart illustrating another data collection method in some embodiments;

[0070] Figure 7 is a structural block diagram of the data collection device in some embodiments;

[0071] Figure 8 is a structural block diagram of another data collection device in some embodiments;

[0072] Figure 9 is an internal structure diagram of the communication device in some embodiments. Detailed Implementation

[0073] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0074] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0075] When used herein, the singular forms of “a,” “an,” and “the” may also include the plural forms unless the context clearly indicates otherwise. It should also be understood that the terms “comprising,” “including,” or “having,” etc., specify the presence of the stated feature, whole, step, operation, component, part, or combination thereof, but do not preclude the possibility of the presence or addition of one or more other features, wholes, steps, operations, components, parts, or combinations thereof.

[0076] Currently, models such as machine learning (ML) models and artificial intelligence (AI) models can be deployed in communication systems, which helps improve the overall performance of the communication system. Models deployed in communication systems can implement different use cases, among which RRM (Radio Resource Management) prediction use cases, event prediction use cases, and RLF (Radio Link Failure) prediction use cases are some typical examples.

[0077] To make the technical solutions provided in the embodiments of this application easy to understand, the RRM prediction use cases, event prediction use cases, and RLF prediction use cases will be described below.

[0078] I. RRM prediction use cases.

[0079] RRM prediction use cases are used to predict unmeasured time-domain measurements based on measured time-domain measurements, or to predict unmeasured frequency-domain measurements based on measured frequency-domain measurements. The measured data may include, for example, RSRP (Reference Signal Received Power), RSRQ (Reference Signal Received Quality), SINR (Signal to Noise or Interference Ratio), etc.

[0080] When implementing RRM prediction use cases, you can input the measured data into the model and the model will output the predicted measurement data.

[0081] Please refer to Tables 1 and 2, which show several exemplary sub-use cases included in the RRM prediction use case. Table 2 is a Chinese translation of Table 1.

[0082] Table 1

[0083] Table 2

[0084] For example, sub-use case 2 can predict measurement data in the time domain with the aim of reducing measurement; sub-use case 4 can predict measurement data in the time domain with the aim of improving switching performance; and sub-use case 3 can predict measurement data in the frequency domain with the aim of reducing measurement. Please refer to Figures 1 and 2, which illustrate sub-use case 2 and sub-use case 3, respectively.

[0085] In practical applications, the prediction results obtained from RRM prediction use cases can be used to implement event prediction use cases and / or RLF prediction use cases.

[0086] II. Event prediction use cases.

[0087] Event prediction use cases are used to predict whether an event will be satisfied within a future period of time based on measured data. Events may include, for example, events such as A3, A4, and A5. Event prediction use cases can improve handover performance. For example, in some cases, the terminal can report the predicted event satisfaction result to the network side in advance before the event is actually satisfied, so that the network side can perform corresponding actions to reduce failures caused by too late handover.

[0088] Currently, event prediction use cases include two prediction methods: direct prediction and indirect prediction.

[0089] Direct prediction involves inputting measured data, event configuration information, etc., into the model, and the model outputs probability information, which indicates the likelihood of an event occurring within a time window.

[0090] Indirect prediction involves first obtaining predicted measurement data through RRM prediction use cases, and then determining the timing of events based on the predicted measurement data. The process of determining the timing of events based on the predicted measurement data does not require the use of a model.

[0091] Generally, the performance of direct prediction can be evaluated using the F1 score, while the performance of indirect prediction can be evaluated using the F1 score, RSRP difference, and the time interval between actual event reporting trigger and predicted event reporting trigger.

[0092] It should be noted that, in general, event prediction use cases can be implemented using models deployed in the terminal.

[0093] III. RLF Prediction Use Cases.

[0094] RLF prediction use cases are used to predict whether an RLF will occur within a certain period of time based on measured data. RLF prediction use cases can improve handover performance. For example, in some cases, the terminal can report the prediction result of the RLF occurrence to the network side in advance before the RLF actually occurs, so that the network side can perform corresponding actions (such as the network side can hand over the terminal to another cell before the terminal experiences an RLF) to reduce the occurrence of RLF.

[0095] Currently, RLF prediction use cases include two prediction methods: direct prediction and indirect prediction.

[0096] Direct prediction involves inputting measured data and RLF monitoring configuration information into the model, and the model outputs probability information, which indicates the likelihood of an RLF occurring within a time window.

[0097] Indirect prediction involves first obtaining predicted measurement data through RRM prediction use cases, and then determining the timing of RLF occurrence based on the predicted measurement data. The process of determining the timing of RLF occurrence based on the predicted measurement data does not require the use of a model.

[0098] Generally, the performance of direct prediction can be evaluated using the F1 score, while the performance of indirect prediction can be evaluated using the F1 score, SINR difference, and the interval between the actual occurrence of RLF and the predicted occurrence of RLF.

[0099] It should be noted that, in general, RLF prediction use cases can be implemented using models deployed in the terminal.

[0100] After describing the RRM prediction use cases, event prediction use cases, and RLF prediction use cases respectively, in order to further help understand the technical solutions provided by the embodiments of this application, the LCM (Life Cycle Management) process of the model (including AI model, ML model, etc.) will be briefly described below.

[0101] As can be seen from Figure 3, based on the current standard discussion, the LCM of the model mainly includes the following processes: 1. Data collection, which includes training data collection, inference data collection, and performance monitoring data collection; 2. Model training; 3. Model identification; 4. Model transfer; 5. Model inference; 6. Model performance monitoring; 7. Model management.

[0102] The model training process refers to training the model using the training data collected during the data collection process to obtain a trained model that can be used for inference. In addition, it may also include the model update process, which refers to retraining the already trained model using the training data collected during the data collection process to update the model.

[0103] Currently, there is no consideration on how to collect data to provide training data for training and / or updating models that implement these use cases, such as RRM prediction use cases, event prediction use cases, and RLF prediction use cases.

[0104] In view of the above, embodiments of this application provide a data collection method, apparatus, device, and readable storage medium. In this data collection method, a terminal receives a data collection configuration sent by a network device, and then collects and / or records first training data according to the data collection configuration. Next, the terminal sends the first training data to the network device. The first training data is used to train and / or update a target model, which is used to implement a target use case. The target use case includes at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case. Thus, a data collection method is provided for RRM prediction use case, event prediction use case, and RLF prediction use case, thereby providing training data based on this method to train and / or update the model implementing the above use cases, i.e., the target model.

[0105] It should be noted that the beneficial effects or technical problems solved by the embodiments of this application are not limited to this one, but may also be other implicit or related problems. For details, please refer to the description of the embodiments below.

[0106] In an exemplary embodiment, as shown in FIG4, a data collection method is provided. Taking the application of this method to a terminal as an example, the method includes the following steps 401-403:

[0107] Step 401: Receive data collection configuration sent by the network device.

[0108] In optional embodiments of this application, the network device is a device deployed on the network side. For example, the network device can be an access network device (e.g., a base station) deployed in the access network, or a core network device deployed in the core network (CN). The network device can also be OAM (Operations, Administration, and Maintenance).

[0109] When the network device is an access network device, the data collection configuration can be carried in RRC (Radio Resource Control) signaling and sent to the terminal via RRC signaling. When the network device is a core network device, the data collection configuration can be carried in NAS (Non-Access Stratum) signaling and sent to the terminal via NAS signaling; alternatively, the data collection configuration can be sent to the terminal via the User Plane (UP) channel. When the network device is an OAM (Operational Access Management) device, the data collection configuration can be sent to the terminal by the OAM through the access network; alternatively, the data collection configuration can be sent to the terminal by the OAM through the core network via the UP channel.

[0110] In optional embodiments of this application, the data collection configuration may be triggered autonomously by the network device or sent by the network device in response to a request from the terminal. This application does not impose specific limitations on this.

[0111] In an optional embodiment of this application, the data collection configuration is used to configure data collection, recording, and / or reporting, etc., and this embodiment of the application does not specifically limit this.

[0112] Step 402: Collect and / or record the first training data according to the data collection configuration.

[0113] In an optional embodiment of this application, the first training data is used to train and / or update the target model, wherein the target model is used to implement the target use case, which includes at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case.

[0114] In an optional embodiment of this application, the target model may be deployed on the network side and / or the terminal side.

[0115] Step 403: Send the first training data to the network device.

[0116] In an optional embodiment of this application, the terminal may send first training data to the network device according to the data collection configuration.

[0117] In optional embodiments of this application, the first training data may be sent autonomously by the terminal (for example, after the amount of the first training data reaches a certain value, the terminal may autonomously trigger the sending of the first training data to the network device), or it may be sent by the terminal in response to a request from the network device. This application does not specifically limit this.

[0118] In optional embodiments of this application, when the network device is an access network device, the terminal can send the first training data to the access network device via RRC signaling. When the network device is a core network device, the terminal can send the first training data to the core network device via NAS signaling, or via the UP channel. When the network device is an OAM, the terminal can send the first training data to the OAM via the access network, or via the UP channel from the core network.

[0119] In an optional embodiment of this application, the first training data includes at least model input data and / or model output data, wherein the model input data refers to the data input to the model, and the model output data can be understood as the label corresponding to the model input data, or as the ideal model output corresponding to the model input data.

[0120] In an optional embodiment of this application, after receiving the first training data, the network device can send the first training data to a third-party server, so that the third-party server can train and / or update the target model based on the first training data. Alternatively, the network device can directly train and / or update the target model based on the first training data. This embodiment of the application does not specifically limit this approach. The third-party server may, for example, be an OTT Server (Over-The-Top Server).

[0121] The data collection method provided in this embodiment involves a terminal receiving a data collection configuration sent by a network device, then collecting and / or recording first training data according to the configuration, and then sending the first training data to the network device. This first training data is used to train and / or update a target model, which is used to implement a target use case. The target use case includes at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case. Thus, a data collection method is provided for RRM prediction use case, event prediction use case, and RLF prediction use case, thereby providing training data based on this method to train and / or update the model implementing the above use cases, i.e., the target model.

[0122] In an optional embodiment of this application, the terminal may further send a configuration request to the network device, wherein the configuration request is used to request the network device to send data collection configuration. In other words, before step 401 above, the terminal may send a configuration request to the network device to request the network device to send data collection configuration to the terminal.

[0123] In an optional embodiment of this application, the configuration request may be triggered by a third-party server and sent by the terminal, or the configuration request may be sent autonomously by the terminal. This application embodiment does not specifically limit this. As mentioned above, the third-party server is used to train and / or update the target model.

[0124] In an alternative embodiment of this application, the configuration request may be sent when the target model is missing, or when the target model needs to be updated.

[0125] In an alternative embodiment of this application, the network device may or may not respond to the configuration request, depending on the network implementation.

[0126] In optional embodiments of this application, when the network device is an access network device, the terminal can send the configuration request to the access network device via RRC signaling. When the network device is a core network device, the terminal can send the configuration request to the core network device via NAS signaling, or via a UP channel. When the network device is an OAM, the terminal can send the configuration request to the OAM via the access network, or via a UP channel from the core network to the OAM.

[0127] In an optional embodiment of this application, in the event of handover and / or reconstruction, the configuration request can be transmitted in the Xn interface so that the newly accessed cell of the terminal can obtain the configuration request and / or respond to the configuration request.

[0128] In an optional embodiment of this application, before the terminal sends a configuration request to the network device, the terminal may receive an allow request indication information from the network device. The allow request indication information is used to indicate that the terminal is allowed to send the configuration request to the network device. In other words, the terminal can only send the configuration request to the network device if it receives the allow request indication information; otherwise, the terminal is not allowed to send the configuration request to the network device. Of course, in some optional embodiments, the terminal can send the configuration request to the network device without obtaining permission from the network device. In this case, the network device may not send the allow request indication information.

[0129] In an optional embodiment of this application, the terminal may receive a data acquisition request sent by the network device, and then, in response to the data acquisition request, send the first training data to the network device. In other words, in an optional embodiment, the terminal may send the first training data to the network device in response to a request from the network device.

[0130] In an optional embodiment of this application, before the network device sends a data acquisition request to the terminal, the terminal may send an availability indication message to the network device, wherein the availability indication message indicates that the terminal has stored the first training data. After determining that the first training data can be acquired from the terminal based on the availability indication message, the network device may send the data acquisition request to the terminal based on its own needs. It should be noted that in some cases, the network device may send a data acquisition request to the terminal without receiving the availability indication message sent by the terminal; that is, in these cases, the terminal may not need to send the availability indication message.

[0131] The following embodiments of this application will describe the configuration request, data collection configuration, and the content of the first training data mentioned above:

[0132] I. Configuration Request.

[0133] In optional embodiments of this application, the request content of the configuration request can be generated by the terminal, generated by a third-party server, or predefined. This application does not impose specific limitations on this.

[0134] A. If the target use case includes an RRM prediction use case, the configuration request includes at least one of the following:

[0135] 1. First model type information, which indicates whether the target model is a per-cell model or a cluster model.

[0136] In other words, the first model type information is used to indicate whether the target model is making predictions for a single cell or for a cell cluster. It should be noted that the cell cluster includes multiple cells.

[0137] 2. Request instruction information, which is used to instruct the requesting network device to configure the collection and / or recording of the first training data.

[0138] In optional embodiments of this application, the request indication information may also indicate the use case number and / or sub-use case number, use case name and / or sub-use case name, function number, function name, purpose of request configuration (obtaining training data), etc., and this application embodiment does not specifically limit this.

[0139] 3. Purpose indication information: Purpose indication information is used to indicate the purpose of training and / or updating the target model.

[0140] In an optional embodiment of this application, the purpose of training and / or updating the target model may be, for example, to improve switching performance or reduce measurements.

[0141] 4. Cause indication information: The cause indication information is used to indicate the reason for training and / or updating the target model.

[0142] In an optional embodiment of this application, the reason for training and / or updating the target model may be, for example, the absence of a target model (i.e., the need to train a new target model) or the need to update the target model.

[0143] 5. Model input and / or model output.

[0144] In an optional embodiment of this application, the model input and / or model output can be understood as: the requested model input and / or the requested model output.

[0145] In an optional embodiment of this application, the model input and / or model output may include, for example, cell identifiers and / or cell lists to request measurement and / or recording of measurement data for one or more neighboring cells included in the cell identifiers and / or cell lists.

[0146] In an optional embodiment of this application, the model input and / or model output may include, for example, a beam identifier and / or a beam list to request measurement and / or recording of measurement data for one or more beams included in the beam identifier and / or beam list.

[0147] In optional embodiments of this application, the model input and / or model output may, for example, request configuration of the reporting format (pattern) of the model input data and / or model output data included in the first training data. For instance, the model input data and model output data can be reported together or separately. Reporting the model input data and model output data together means that the model input data and model output data do not need to be distinguished within the same signaling message, and are reported together. Reporting the model input data and model output data separately means that the model input data and model output data need to be distinguished within the same signaling message, and are reported separately. Of course, other reporting formats can also be requested, and this application embodiment does not specifically limit this.

[0148] In an optional embodiment of this application, the model input and / or model output may, for example, request the recording of information such as the terminal's position and speed.

[0149] In optional embodiments of this application, particularly in the RRM frequency domain prediction sub-example included in the RRM prediction example, the frequency points of the model input measurement and / or recorded measurement data may be different from the frequency points of the model output measurement and / or recorded measurement data, the frequency points of the cell into which the model input measurement and / or recorded measurement data is located may be different from the frequency points of the cell into which the model output measurement and / or recorded measurement data is located, and the frequency points of the beam into which the model input measurement and / or recorded measurement data is located may be different from the frequency points of the beam into which the model output measurement and / or recorded measurement data is located.

[0150] 6. Model window length, which includes the observation window length and / or prediction window length.

[0151] In an optional embodiment of this application, the observation window (OW) refers to the time range within which data is collected, and the data collected within this window serves as the input to the model. The prediction window (PW) refers to the time range within which the model outputs its prediction results; it defines how far into the future the model needs to predict the data.

[0152] In optional embodiments of this application, the observation window length and prediction window length can be characterized by the number of instances, time slots, or samples. For example, in one possible case, the observation window length can be 5 instances and the prediction window length can be 5 instances.

[0153] 7. Additional conditions on the network side.

[0154] In an optional embodiment of this application, the additional conditions on the network side may include, for example, the network environment in which the terminal is located. For example, it can be represented by an associated ID. Different associated ID values ​​represent different network environments, such as different beam patterns, base station antenna downlink angles, base station heights, etc. The same associated ID value represents the same network environment in the same cell, and the same or different network environments in different cells.

[0155] 8. Additional conditions on the terminal side.

[0156] In an optional embodiment of this application, the additional conditions on the terminal side may include, for example, the state of the terminal, such as the speed of the terminal and the location information of the terminal.

[0157] 9. Information about third-party servers.

[0158] In optional embodiments of this application, the information of the third-party server may include, for example, the third-party server's number, tracking number, identifier, tracking reference, etc., and this application does not specifically limit this.

[0159] In an optional embodiment of this application, for the convenience of subsequent description, basic request information is defined, wherein the basic request information includes at least one of the information in 2-9 above.

[0160] B. If the target use case includes an event prediction use case, the configuration request includes at least one of the following:

[0161] 1. First model type information.

[0162] 2. Basic request information, as described above, includes at least one of the following: request indication information; purpose indication information; reason indication information; model input and / or model output; model window length; additional conditions on the network side; additional conditions on the terminal side; information of the third-party server.

[0163] The contents of the basic request information have been described above, and will not be repeated here in this embodiment.

[0164] It should be noted that, when the target use case includes an event prediction use case, the model input and / or model output section in the basic request information, in addition to what has been described above, can also request the recording of configuration information for events that need to be recorded (when the prediction method of the event prediction use case is direct prediction).

[0165] 3. Second model type information, which indicates whether the target model is a model that directly predicts events or a model that indirectly predicts events.

[0166] As mentioned above, event prediction use cases include two prediction methods: direct prediction and indirect prediction. Based on this premise, the configuration request may include second model type information, which can indicate whether the target model is used to predict events directly or indirectly.

[0167] 4. Event type information requested.

[0168] For example, the requested event type may include event A3, event A4, event A5, etc., and this application embodiment does not specifically limit this.

[0169] 5. Requested event configuration parameters.

[0170] For example, the event configuration parameters requested may include the TTT (Time To Trigger) and offset corresponding to the event, etc., which are not specifically limited in this embodiment of the application.

[0171] It should be noted that if the event prediction use case uses an indirect prediction method, the content included in this configuration request can be the same as that included when the target use case is an RRM prediction use case. If the event prediction use case uses a direct prediction method, the configuration request may also include at least one of the requested event type information and the requested event configuration parameters.

[0172] C. If the target use case includes an RLF predicted use case, the configuration request includes at least one of the following:

[0173] 1. Basic request information, as described above, includes at least one of the following: request indication information; purpose indication information; reason indication information; model input and / or model output; model window length; additional conditions on the network side; additional conditions on the terminal side; information of the third-party server.

[0174] The contents of the basic request information have been described above, and will not be repeated here in this embodiment.

[0175] It should be noted that, when the target use case includes RLF prediction use cases, the model input and / or model output section in this basic request information can be:

[0176] a. Request the measurement and / or recording of the terminal's service beam.

[0177] b. Includes serving cell identifiers and / or a list of serving cells, to request measurement and / or recording of one or more serving cells included in the serving cell identifiers and / or the list of serving cells.

[0178] c. Includes a serving beam identifier and / or a serving beam list to request measurement and / or recording of one or more serving beams included in the serving beam identifier and / or the serving beam list, wherein, in the case of multiple serving beams, the multiple serving beams may be across cells or not.

[0179] 2. Third model type information: The third model type information is used to indicate whether the target model is a model that directly performs RLF prediction or a model that indirectly performs RLF prediction.

[0180] As mentioned above, RLF prediction use cases include two prediction methods: direct prediction and indirect prediction. Based on this premise, the configuration request may include third model type information, which can indicate whether the target model is used for direct or indirect prediction of RLF.

[0181] 3. RLM (Radio Link Monitoring) configuration information.

[0182] For example, the configuration information of the RLM may include the configuration information of T310, N310, N311, etc., and this application embodiment does not specifically limit this.

[0183] N310 is a counter that represents the maximum number of consecutive "out-of-synchronization" indications received by the terminal. When the terminal receives N310 consecutive "out-of-synchronization" indications, the T310 timer is started, indicating that the terminal has detected a decline in the quality of the wireless link and may be about to lose its connection with the base station.

[0184] N311 is a counter that represents the maximum number of consecutive "synchronization" instructions received by the terminal. When the terminal receives N311 consecutive "synchronization" instructions, the T310 timer stops, indicating that the wireless link quality has been restored and the terminal has re-established a stable connection with the base station.

[0185] T310 is a timer. When T310 times out, the terminal determines that an RLF has occurred and can initiate connection reconstruction.

[0186] It should be noted that if the RLF prediction test case uses an indirect prediction method, the configuration request may only include basic request information. If the RLF prediction test case uses a direct prediction method, the configuration request may also include RLM configuration information.

[0187] II. Data Collection Configuration.

[0188] In optional embodiments of this application, the data collection configuration may include at least one of the following configurations: S1, measurement configuration; S2, recording configuration; S3, reporting configuration. In optional embodiments of this application, the measurement configuration is used to instruct the terminal to perform measurements and / or recordings according to the measurement configuration. In optional embodiments of this application, the recording configuration is used to instruct the terminal to record measurement data according to the recording configuration. In optional embodiments of this application, the reporting configuration is used to instruct the terminal to report the recorded data (i.e., the first training data) to the network device according to the reporting configuration. The following embodiments of this application will describe the contents of the measurement configuration, recording configuration, and reporting configuration respectively:

[0189] S1, Measurement Configuration.

[0190] A. If the target use case includes an RRM prediction use case, the measurement configuration includes at least one of the following:

[0191] 1. List of residential communities.

[0192] The cell list includes at least one cell and is used to instruct the terminal to perform measurements and / or recordings on the cells in the cell list. Of course, if only one cell is included, the cell list can also be understood as a cell identifier. In an optional embodiment of this application, the cells included in the cell list can be cells at different frequency points.

[0193] 2. Beam list.

[0194] The beam list includes at least one beam and is used to instruct the terminal to perform measurements and / or recordings on the beams in the beam list. Of course, if only one beam is included, the beam list can also be understood as a beam identifier. In an optional embodiment of this application, the beams included in the beam list can be beams at different frequencies.

[0195] In optional embodiments of this application, the beams in the beam list may be configured on a cell-by-cell basis, or the beams in the beam list may be configured on a measurement object basis, or the beams in the beam list may not be configured on a cell-by-cell or measurement object basis.

[0196] 3. Frequency list.

[0197] The frequency list includes at least one frequency point, and is used to instruct the terminal to perform measurements and / or recordings on the frequencies listed. Of course, if only one frequency point is included, the frequency list can also be understood as a frequency point identifier.

[0198] In an optional embodiment of this application, the terminal measuring and / or recording the frequency points in the frequency point list includes: the terminal measuring and / or recording the cell corresponding to the frequency point in the frequency point list, and / or the terminal measuring and / or recording the beam corresponding to the frequency point in the frequency point list.

[0199] 4. First frequency indication information.

[0200] The first frequency point indication information corresponds to a frequency point and is used to indicate whether to measure and / or record the cell on the corresponding frequency point. In an optional embodiment of this application, the measurement and / or recording of the cell on the corresponding frequency point may include measuring and / or recording the cell, or it may include measuring and / or recording the beams in the cell.

[0201] In an optional embodiment of this application, the measurement configuration may include one or more first frequency point indication information, wherein the multiple first frequency point indication information may correspond to different frequency points.

[0202] It should be noted that this first frequency indication information can be configured on a per-measurement basis.

[0203] It should also be noted that either the first frequency indication information or the frequency list can be configured.

[0204] 5. First measurement object indication information.

[0205] The first measurement object indication information is used to indicate whether all measurement objects are measured and / or recorded.

[0206] 6. Second measurement object indication information.

[0207] The second measurement object indication information corresponds to the measurement object, and the second measurement object indication information is used to indicate whether to measure and / or record the corresponding measurement object.

[0208] In an optional embodiment of this application, the measurement configuration may include one or more second measurement object indication information, wherein the multiple second measurement object indication information may correspond to different measurement objects.

[0209] 7. Model-related parameters.

[0210] Among them, the model-related parameters are used to implicitly indicate the measurement objects that need to be measured and / or recorded.

[0211] For example, in a traditional RRM measurement configuration, a new SMTC (SS / PBCH block measurement timing configuration; SSB measurement timing configuration) configuration may be configured, or an SMTC configuration for AI / ML may be configured, or an SMTC configuration for RRM time-domain prediction may be configured, or other time-domain related configurations may be configured. These contents can implicitly indicate the measurement objects that need to be measured and / or recorded.

[0212] 8. Target configuration.

[0213] The purpose configuration is used to configure the purpose of measurement and / or recording. For example, the purpose can be RRM time-domain prediction use case data collection, or the purpose can be RRM frequency-domain prediction use case data collection.

[0214] 9. Reference signal configuration.

[0215] The reference signal configuration is used to configure the reference signal that needs to be measured and / or recorded. For example, the reference signal may include SSB (Synchronization Signal Block) or CSI-RS (Channel State Information Reference Signal), etc. This application embodiment will not describe it in detail.

[0216] In an optional embodiment of this application, for the convenience of subsequent description, a basic measurement configuration is defined, wherein the basic measurement configuration includes at least one of 1-9 above.

[0217] B. When the target use case includes an event prediction use case, the measurement configuration includes at least one of the following configurations:

[0218] 1. Basic measurement configuration, as described above, includes at least one of the following: cell list; beam list; frequency list; first frequency indication information; first measurement object indication information; second measurement object indication information; model-related parameters; purpose configuration, wherein the purpose configuration is used to configure the purpose of measurement and / or recording, for example, the purpose may be event prediction use case data collection; reference signal configuration.

[0219] 2. Event configuration.

[0220] For example, configuration information for parameters such as event TTT and offset.

[0221] 3. Report configuration instructions.

[0222] In an optional embodiment of this application, the reporting configuration indication information may correspond to an event and / or a measurement object, and the reporting configuration indication information is used to indicate whether to measure and / or record the event and / or measurement object corresponding to this reporting configuration.

[0223] In an optional embodiment of this application, the measurement configuration may include one or more reporting configuration indication messages, wherein the multiple reporting configuration indication messages may correspond to different events and / or measurement objects.

[0224] 4. Third measurement object indication information.

[0225] The third measurement object indication information is used to indicate whether all measurement objects that satisfy the event should be measured and / or recorded. In an optional embodiment of this application, the event can be a specific event, or an event specified by the network side. The event can include: event A3, event A4, event A5, etc., and this embodiment of the application does not specifically limit it.

[0226] C. When the target use case includes RLF prediction use cases, the measurement configuration includes at least one of the following configurations:

[0227] 1. Reference signal configuration.

[0228] The reference signal configuration is used to configure the reference signal that needs to be measured and / or recorded. For example, the reference signal may include SSB and / or CSI-RS, etc., which will not be specifically described in the embodiments of this application.

[0229] 2. RLM configuration.

[0230] For example, the RLM configuration may include T310 configuration, N310 configuration, N311 configuration, etc., and this application embodiment does not specifically limit it.

[0231] 3. Target configuration.

[0232] The purpose configuration is used to configure the purpose of measurement and / or recording, for example, the purpose can be RLF prediction use case data collection.

[0233] 4. Community signage.

[0234] The cell identifier is used to instruct the terminal to perform measurements and / or recordings for the cell corresponding to the cell identifier. In an optional embodiment of this application, the cell corresponding to the cell identifier may, for example, include the terminal's serving cell.

[0235] 5. Beam marking.

[0236] The beam identifier is used to instruct the terminal to perform measurements and / or recordings on the beam corresponding to the beam identifier. In an optional embodiment of this application, the beam corresponding to the beam identifier may, for example, include the terminal's serving beam.

[0237] 6. Service beam indication information.

[0238] The service beam indication information is used to indicate whether to use the beam with the best signal quality as the service beam for measurement and / or recording.

[0239] 7. Fourth measurement object indication information.

[0240] In an optional embodiment of this application, the fourth measurement object indication information corresponds to the measurement object, and the fourth measurement object indication information is used to indicate whether to measure and / or record the corresponding measurement object.

[0241] In an optional embodiment of this application, the measurement configuration may include one or more fourth measurement object indication information, wherein the multiple fourth measurement object indication information correspond to different measurement objects.

[0242] S2, Record Configuration.

[0243] In an optional embodiment of this application, the recording configuration may include at least one of the following configurations: S21, start recording condition; S22, stop recording condition.

[0244] S21, Start recording conditions.

[0245] A. When the target use case includes RRM prediction use cases, the conditions for starting recording include at least one of the following:

[0246] 1. The signal quality is higher or lower than the signal quality threshold.

[0247] In an optional embodiment of this application, recording may begin when the signal quality is higher or lower than a signal quality threshold, or measurement data may be recorded when the signal quality is higher or lower than a signal quality threshold.

[0248] In optional embodiments of this application, the signal quality can be the signal quality of the serving cell, the signal quality of the neighboring cell, the signal quality of the neighboring cell beam, the signal quality of the serving cell beam, etc., and this application does not specifically limit this.

[0249] In optional embodiments of this application, the signal quality may include RSRP, RSRQ, SINR, etc., and this application does not specifically limit this.

[0250] In an optional embodiment of this application, the signal quality threshold can be configured by the network side.

[0251] It should be noted that multiple signal quality thresholds appear in the embodiments of this application. These multiple signal quality thresholds are only used to indicate one threshold of signal quality and do not indicate or imply that the multiple signal quality thresholds are the same threshold, or that the multiple signal quality thresholds are equal to each other.

[0252] 2. A reconfiguration or connection to a new cell has occurred.

[0253] In optional embodiments of this application, reconfiguration may refer to the reconfiguration of the data collection configuration. For example, the network device sends a new data collection configuration to the terminal, or the network device sends a configuration change instruction to the terminal to change some of the configurations in the data collection configuration. This application does not specifically limit the reconfiguration process of the data collection configuration.

[0254] In optional embodiments of this application, accessing a new cell may include accessing a new cell through mechanisms such as cell handover, cell selection, or cell reselection.

[0255] 3. Changes to additional conditions on the network side.

[0256] As described above, in an optional embodiment of this application, the additional conditions on the network side may include, for example, the network environment in which the terminal is located, which may be represented by an associated ID.

[0257] Therefore, in an optional embodiment of this application, changes to additional conditions on the network side may include, for example, changes to the associated ID.

[0258] 4. Changes to additional conditions on the terminal side.

[0259] As described above, in an optional embodiment of this application, the additional conditions on the terminal side may include, for example, the state of the terminal, such as the speed of the terminal and the location information of the terminal.

[0260] Based on this, in an optional embodiment of this application, changes to additional conditions on the terminal side may include, for example, changes in the terminal's speed or changes in the terminal's geographical location.

[0261] In an optional embodiment of this application, for the convenience of subsequent description, basic start conditions are defined, wherein the basic start conditions include at least one of 1-4 above.

[0262] B. When the target use case includes an event prediction use case, the conditions for starting recording include at least one of the following:

[0263] 1. Basic start conditions, as described above, include at least one of the following: signal quality is higher or lower than the signal quality threshold; reconfiguration or access to a new cell occurs; additional conditions on the network side change; additional conditions on the terminal side change.

[0264] 2. The conditions for the event to proceed have been met.

[0265] In an optional embodiment of this application, the event entry condition can be, for example, the entry condition of event A3, the entry condition of event A4, or the entry condition of event A5, etc.

[0266] 3. Start the first TTT, where the first TTT corresponds to the event entry condition.

[0267] Under normal circumstances, the first TTT will be activated when the event entry condition is met. In other words, the fulfillment of the event entry condition and the activation of the first TTT can be considered equivalent in some respects.

[0268] 4. The event is satisfied.

[0269] In an optional embodiment of this application, event satisfaction can also be understood as event triggering and reporting satisfaction.

[0270] 5. First TTT timeout.

[0271] In general, the first TTT timeout can be understood as the event being satisfied. In other words, event satisfaction and the first TTT timeout can be considered equivalent in some sense.

[0272] 6. The difference between the value and the event entry condition reaches a preset threshold.

[0273] In an optional embodiment of this application, the event entry condition may correspond to a signal quality threshold. For example, the signal quality threshold may include RSRP threshold, RSRQ threshold, SINR threshold, etc. The difference between the current signal quality (RSRP, RSRQ, SINR) and the signal quality threshold (RSRP threshold, RSRQ threshold, SINR threshold) corresponding to the event entry condition can be understood as the difference between the current signal quality (RSRP, RSRQ, SINR) and the signal quality threshold (RSRP threshold, RSRQ threshold, SINR threshold) corresponding to the event entry condition.

[0274] Understandably, this difference may include RSRP difference, RSRQ difference, and SINR difference.

[0275] In an optional embodiment of this application, recording can begin when the difference between the event entry condition and the event entry condition reaches a preset threshold, or measurement data with a difference between the event entry condition and the preset threshold can be recorded.

[0276] In an optional embodiment of this application, for example, recording is triggered when the difference between the current signal quality measurement value and the event entry condition is less than a preset threshold.

[0277] It should be noted that multiple preset thresholds appear in the embodiments of this application. These multiple preset thresholds are only used to indicate that the threshold is preset, and do not indicate or imply that the multiple preset thresholds are the same threshold, or that the multiple preset thresholds are equal to each other.

[0278] 7. First recording cycle.

[0279] The first recording period is a preset duration before the event entry condition is met and / or a preset duration after the event entry condition is met.

[0280] In other words, data within a first recording period can be recorded, wherein, in an optional embodiment of this application, the first recording period can be an overall time window.

[0281] It should be noted that although the definition of the first recording period includes the concepts of "preset duration before the event entry condition is met" and "preset duration after the event entry condition is met", it should be understood that the two preset durations mentioned here only indicate that the duration is preset, and do not indicate or imply that the two preset durations are equal. In addition, the preset durations mentioned below also only indicate that the duration is preset, and do not indicate or imply that the preset durations are equal to each other.

[0282] 8. Second recording cycle.

[0283] The second recording period is a preset duration before the event is satisfied and / or a preset duration after the event is satisfied.

[0284] In other words, data within a second recording period can be recorded, wherein, in an optional embodiment of this application, the second recording period can be an overall time window.

[0285] C. When the target use case includes an RLF-predicted use case, the conditions for starting recording include at least one of the following:

[0286] 1. Basic start conditions, as described above, include at least one of the following: signal quality is higher or lower than the signal quality threshold; reconfiguration or access to a new cell occurs; additional conditions on the network side change; additional conditions on the terminal side change.

[0287] 2. The target timer is activated, which may include the T310 timer.

[0288] 3. The target value related to the loss of synchronization reaches its maximum value, where the target value related to the loss of synchronization may include N310.

[0289] 4. A loss of synchronization was detected.

[0290] 5. RLF detected.

[0291] 6. The number of times a step loss is detected reaches the preset threshold.

[0292] 7. Third recording cycle.

[0293] The third recording period is a preset duration before and / or after the target condition is met. The target condition includes at least one of the following: the target timer is turned on, which may include the T310 timer, a step loss is detected, and an RLF is detected.

[0294] In other words, data can be recorded within a third recording period, which can be a whole time window.

[0295] 8. Beam failure detected.

[0296] In an optional embodiment of this application, the beam failure refers to the failure of the serving beam, wherein the serving beam can be the beam with the highest signal quality, and the signal quality can include RSRP, RSRQ, SINR, etc., which are not specifically limited in this embodiment of the application.

[0297] 9. An event was detected and a report was triggered.

[0298] In an optional embodiment of this application, the event-triggered reporting refers to a specific event triggering measurement reporting. For example, the specific event may be configured on the network side and may include events such as A3 and A5.

[0299] S22, Stop recording condition.

[0300] A. When the target use case includes an RRM prediction use case, the stop recording condition includes at least one of the following:

[0301] 1. The duration of a single recording reaches the preset duration.

[0302] In an optional embodiment of this application, in response to the data collection configuration, the terminal can make multiple records, wherein the duration of a single record refers to the duration of one of the records, and the duration of a single record can be calculated from the time the record start condition was met most recently.

[0303] 2. The total recording time has reached the preset time.

[0304] As mentioned above, in response to the data collection configuration, the terminal can record multiple times, and the total recording duration can refer to the total duration up to the current recording.

[0305] 3. The number of records has reached the preset number.

[0306] 4. The signal quality is higher or lower than the signal quality threshold.

[0307] In an optional embodiment of this application, recording may be stopped when the signal quality is higher or lower than the signal quality threshold, or measurement data when the signal quality is higher or lower than the signal quality threshold may not be recorded.

[0308] In optional embodiments of this application, the signal quality can be the signal quality of the serving cell, the signal quality of the neighboring cell, the signal quality of the neighboring cell beam, the signal quality of the serving cell beam, etc., and this application does not specifically limit this.

[0309] In optional embodiments of this application, the signal quality may include RSRP, RSRQ, SINR, etc., and this application does not specifically limit this.

[0310] In an optional embodiment of this application, the signal quality threshold can be configured by the network side.

[0311] 5. Changes to additional conditions on the network side.

[0312] As described above, in an optional embodiment of this application, the additional conditions on the network side may include, for example, the network environment in which the terminal is located, which may be represented by an associated ID.

[0313] Therefore, in an optional embodiment of this application, changes to additional conditions on the network side may include, for example, changes to the associated ID.

[0314] 6. Changes to additional conditions on the terminal side.

[0315] As described above, in an optional embodiment of this application, the additional conditions on the terminal side may include, for example, the state of the terminal, such as the speed of the terminal and the location information of the terminal.

[0316] Based on this, in an optional embodiment of this application, changes to additional conditions on the terminal side may include, for example, changes in the terminal's speed or changes in the terminal's geographical location.

[0317] 7. The data volume has reached the preset data volume.

[0318] In an optional embodiment of this application, recording may be stopped when the amount of data collected and / or recorded reaches a preset amount.

[0319] 8. A reconfiguration or connection to a new cell occurs.

[0320] In optional embodiments of this application, reconfiguration may refer to the reconfiguration of the data collection configuration. For example, the network device sends a new data collection configuration to the terminal, or the network device sends a configuration change instruction to the terminal to change some of the configurations in the data collection configuration. This application does not specifically limit the reconfiguration process of the data collection configuration.

[0321] In optional embodiments of this application, accessing a new cell may include accessing a new cell through mechanisms such as cell handover, cell selection, or cell reselection.

[0322] 9. Enter the idle state or the inactive state.

[0323] 10. Battery depleted.

[0324] 11. Insufficient memory.

[0325] In an optional embodiment of this application, for the convenience of subsequent explanation, a basic stopping condition is defined, wherein the basic stopping condition includes at least one of 1-11 above.

[0326] B. When the target use case includes an event prediction use case, the stop recording condition includes at least one of the following:

[0327] 1. Basic stopping conditions, as described above, include at least one of the following: the duration of a single recording reaches a preset duration; the total recording duration reaches a preset duration; the number of recordings reaches a preset number; the signal quality is higher or lower than the signal quality threshold; the additional conditions on the network side change; the additional conditions on the terminal side change; the data volume reaches a preset data volume; reconfiguration occurs or a new cell is accessed; entering an idle state or inactive state; the battery is depleted; or the memory is insufficient.

[0328] 2. The event is satisfied.

[0329] 3. First TTT timeout.

[0330] 4. The event exit condition is met.

[0331] 5. The second TTT is started, where the second TTT corresponds to the event exit condition.

[0332] Under normal circumstances, the second TTT will be activated when the event departure condition is met. In other words, the meeting of the event departure condition and the activation of the second TTT can be considered equivalent in some sense.

[0333] 6. Second TTT timeout.

[0334] 7. The difference between the event departure condition and the preset threshold is reached.

[0335] In an optional embodiment of this application, the event departure condition may correspond to a signal quality threshold. For example, the signal quality threshold may include RSRP threshold, RSRQ threshold, SINR threshold, etc. The difference between the current signal quality (RSRP, RSRQ, SINR) and the signal quality threshold (RSRP threshold, RSRQ threshold, SINR threshold) corresponding to the event departure condition can be understood as the difference between the current signal quality (RSRP, RSRQ, SINR) and the signal quality threshold (RSRP threshold, RSRQ threshold, SINR threshold) corresponding to the event departure condition.

[0336] Understandably, this difference may include RSRP difference, RSRQ difference, and SINR difference.

[0337] In an optional embodiment of this application, recording may stop when the difference between the event departure condition and the event departure condition reaches a preset threshold, or measurement data where the difference between the event departure condition and the event departure condition is a preset threshold may not be recorded.

[0338] In an optional embodiment of this application, for example, recording may be stopped if the difference between the current signal quality and the event departure condition is less than a preset threshold.

[0339] C. When the target use case includes an RLF predicted use case, the stop recording condition includes at least one of the following:

[0340] 1. Basic stopping conditions, as described above, include at least one of the following: the duration of a single recording reaches a preset duration; the total recording duration reaches a preset duration; the number of recordings reaches a preset number; the signal quality is higher or lower than the signal quality threshold; the additional conditions on the network side change; the additional conditions on the terminal side change; the data volume reaches a preset data volume; reconfiguration occurs or a new cell is accessed; entering an idle state or inactive state; the battery is depleted; or the memory is insufficient.

[0341] 2. Synchronization detected.

[0342] 3. The number of synchronization attempts detected has reached the preset number.

[0343] 4. The target timer stops. In an optional embodiment of this application, the target timer may include a T310 timer.

[0344] 5. The target value related to synchronization reaches its maximum value. In an optional embodiment of this application, the target value related to synchronization may include N311.

[0345] 6. RLF detected.

[0346] S3, Report Configuration.

[0347] A. When the target use case includes RRM prediction use cases, the reported configuration includes at least one of the following configurations:

[0348] 1. Information about third-party servers.

[0349] In an optional embodiment of this application, the information of the third-party server may include, for example, the third-party server's number, tracking number, identifier, and / or tracking reference.

[0350] 2. Information about AI data processing entities in the core network.

[0351] In an optional embodiment of this application, the AI ​​data processing entity in the core network may include NWDAF (Network Data Analytics Function).

[0352] In an optional embodiment of this application, the information of the AI ​​data processing entity in the core network may include the number, tracking number, identifier and / or tracking reference of the AI ​​data processing entity in the core network.

[0353] 3. Information about entities in OAM AI data processing.

[0354] In an optional embodiment of this application, the information of the AI ​​data processing entity in OAM may include the entity's number, tracking number, identifier, and / or tracking reference.

[0355] 4. Report the configuration of measurement quantities.

[0356] In an optional embodiment of this application, the measurement reporting configuration is used to configure what type of measurement to report. For example, the measurement reporting configuration can be configured to report measurements such as RSRP, RSRQ, and SINR.

[0357] 5. Indicates information for recording cell-level or beam-level measurement results.

[0358] 6. One or more network-side additional conditions.

[0359] In an optional embodiment of this application, the additional conditions on the network side may include, for example, the network environment in which the terminal is located, such as, the associated ID.

[0360] 7. One or more additional conditions on the terminal side.

[0361] In an optional embodiment of this application, the additional conditions on the terminal side may include, for example, the state of the terminal, such as the speed of the terminal and the location information of the terminal.

[0362] 8. Observation window length and / or prediction window length.

[0363] In an optional embodiment of this application, the observation window length and prediction window length can be characterized by the number of instances, time slots, or samples.

[0364] In an optional embodiment of this application, the terminal can divide the model input data and / or model output data from the recorded measurement data according to the observation window length and / or prediction window length, and report the division results to the network device.

[0365] 9. N future instances and / or N future slots and / or N future samples, where N is a value greater than or equal to 1.

[0366] In an optional embodiment of this application, N future instances and / or N future time slots and / or N future samples are used to indicate the time range of the model output prediction results.

[0367] In an optional embodiment of this application, the terminal can divide the model input data and / or model output data from the recorded measurement data based on N future instances and / or N future time slots and / or N future samples, and report the division results to the network device.

[0368] It should be noted that N, K, etc. appearing in this article refer to numerical values. The appearance of the same numerical labels N and K in different contents does not indicate or imply that the numerical values ​​in different contents are the same or equal.

[0369] 10. Instructions for reporting methods.

[0370] The reporting method instruction information is used to instruct that the model input data and model output data in the first training data be recorded separately, or to instruct that the model input data and model output data be recorded uniformly.

[0371] It should be noted that recording model input and output data uniformly means that when recording and / or reporting data, model input and output data are not distinguished, but rather recorded and / or reported together. Recording model input and output data separately means that when recording and / or reporting data, model input and output data need to be distinguished. For example, model input and output data can be reported using different signaling, or they can be reported using the same signaling. Indication information can be used to indicate which part of the reported data is model input data and which part is model output data; this indication information can be explicit or implicit.

[0372] 11. Report the indication information of the top K cells with high signal quality, where K is a value greater than or equal to 1.

[0373] 12. Report the indication information of the first K beams with high signal quality, where K is a value greater than or equal to 1.

[0374] 13. Instructions for reporting frequency point information.

[0375] 14. Indication information regarding whether data is recorded according to frequency points.

[0376] In an optional embodiment of this application, for the convenience of subsequent description, a basic reporting configuration is defined, wherein the basic reporting configuration includes at least one of the information 1-10 above.

[0377] B. When the target use case includes an event prediction use case, the reporting configuration includes at least one of the following configurations:

[0378] 1. Basic Reporting Configuration: As described above, the basic reporting configuration includes at least one of the following: information of a third-party server; information of the AI ​​data processing entity in the core network; information of the AI ​​data processing entity in OAM; configuration for reporting measurement quantities; indication information for recording cell-level or beam-level measurement results; one or more network-side additional conditions; one or more terminal-side additional conditions; observation window length and / or prediction window length; N future instances and / or N future time slots and / or N future samples; and reporting method indication information.

[0379] 2. First window length.

[0380] The first window length is the window length for determining the probability of an event occurring, which can be expressed in English as "an window for determining the event probability within the window". In other words, in an optional embodiment of this application, the terminal can detect whether an event has occurred or the probability of its occurrence within the first window length, and report data (such as model output data) to the network device based on the detection results.

[0381] 3. Event type configuration.

[0382] In an optional embodiment of this application, the event type configuration is used to configure the type of event that needs to be reported. For example, the type of event that needs to be reported may include: A3 event, A4 event, A5 event, etc.

[0383] C. When the target use case includes RLF predicted use cases, the reporting configuration includes at least one of the following configurations:

[0384] 1. Basic Reporting Configuration: As described above, the basic reporting configuration includes at least one of the following: information of a third-party server; information of the AI ​​data processing entity in the core network; information of the AI ​​data processing entity in OAM; configuration for reporting measurement quantities; indication information for recording cell-level or beam-level measurement results; one or more network-side additional conditions; one or more terminal-side additional conditions; observation window length and / or prediction window length; N future instances and / or N future time slots and / or N future samples; and reporting method indication information.

[0385] 2. Second window length.

[0386] The second window length is the window length for determining the probability of an RLF occurring, which can be expressed in English as "an window for determining the RLF probability within the window". In other words, in an optional embodiment of this application, the terminal can detect whether an RLF has occurred or its probability within the second window length, and report data (such as model output data) to the network device based on the detection results.

[0387] III. First training data.

[0388] A. When the target use case includes RRM prediction use cases, the first training data includes at least one of the following:

[0389] 1. Information about third-party servers.

[0390] 2. Information about AI data processing entities in the core network.

[0391] 3. Information about entities in OAM AI data processing.

[0392] 4. Cell identification information and / or beam identification information, cell measurement results information and / or beam measurement results information.

[0393] In an optional embodiment of this application, the measurement result information of the cell may include, for example, the RSRP measurement result, RSRQ measurement result, SINR measurement result of the cell, and the measurement result information of the beam may include, for example, the RSRP measurement result, RSRQ measurement result, SINR measurement result of the beam.

[0394] In an optional embodiment of this application, the measurement result information of the cell and / or the measurement result information of the beam may include, for example, model input data and / or model output data.

[0395] As described above, in the optional embodiments of this application, model input data and model output data can be reported separately or reported uniformly. It should be noted that uniformly reporting model input data and model output data means that when reporting data, model input data and model output data are not distinguished, but rather they are reported together. Recording model input data and model output data separately means that when recording and / or reporting data, model input data and model output data need to be distinguished. For example, model input data and model output data can be reported using different signaling, or they can be reported using the same signaling. Indication information can be used to indicate which part of the reported data is model input data and which part is model output data; this indication information can be explicit or implicit.

[0396] 5. Observation window length and / or prediction window length.

[0397] 6. N future instances and / or N future time slots and / or N future samples, where N is a value greater than or equal to 1.

[0398] 7. Additional conditions information on the network side.

[0399] 8. Additional conditions information on the terminal side.

[0400] 9. Identification information and / or measurement results of the top K cells with high signal quality, where K is a value greater than or equal to 1.

[0401] In optional embodiments of this application, the measurement result information may include, for example, RSRP measurement results, RSRQ measurement results, SINR measurement results, etc.

[0402] 10. Identification information and / or measurement result information of the first K beams with high signal quality, where K is a value greater than or equal to 1.

[0403] In optional embodiments of this application, the measurement result information may include, for example, RSRP measurement results, RSRQ measurement results, SINR measurement results, etc.

[0404] 11. Frequency information.

[0405] In an optional embodiment of this application, basic data is defined for the convenience of subsequent description, wherein the basic data includes at least one of the information in 1-8 above.

[0406] B. When the target use case includes event prediction use cases, the first training data includes at least one of the following:

[0407] 1. Basic data, as described above, includes at least one of the following: information of a third-party server; information of the AI ​​data processing entity in the core network; information of the AI ​​data processing entity in OAM; cell identification information and / or beam identification information, cell measurement result information and / or beam measurement result information; observation window length and / or prediction window length; N future instances and / or N future time slots and / or N future samples; additional condition information on the network side; additional condition information on the terminal side.

[0408] In cases where the target use case includes an event prediction use case, the measurement result information in the basic data may include L1 filtered measurement results or L3 filtered measurement results.

[0409] 2. Event configuration information.

[0410] In optional embodiments of this application, the event configuration information may include, for example, the configuration information of event A3, the configuration information of event A4, the configuration information of event A5, etc., and this application embodiment does not specifically limit it.

[0411] 3. Time of the event.

[0412] In an optional embodiment of this application, the event occurrence time may include the event trigger reporting time, or the first TTT timeout occurrence time, which may be an absolute time or a relative time.

[0413] 4. Event occurrence indication information.

[0414] In an optional embodiment of this application, different events can correspond to different event occurrence indication information. The event occurrence indication information is used to indicate whether the corresponding event has occurred. In an optional embodiment of this application, the event occurrence indication information can also indicate the number of times the corresponding event has occurred, etc.

[0415] 5. The time when the departure condition of the event occurs.

[0416] In an optional embodiment of this application, the time at which the event departure condition is met can refer to the time when the second TTT is activated. In an optional embodiment of this application, this time information can be an absolute time or a relative time.

[0417] 6. The time when the event entry conditions are met.

[0418] In an optional embodiment of this application, the time at which the event entry condition is met can refer to the time when the first TTT is activated. In an optional embodiment of this application, this time information can be an absolute time or a relative time.

[0419] 7. The target model is an indication of whether it is a model that directly predicts events or a model that indirectly predicts events.

[0420] 8. Unfiltered data obtained from each measurement taken during the first TTT and / or second TTT activation process.

[0421] In an optional embodiment of this application, unfiltered data may refer to the underlying measurement results that have not been L1 filtered.

[0422] 9. First window length.

[0423] C. When the target use case includes RLF prediction use cases, the first training data includes at least one of the following:

[0424] 1. Basic data, as described above, includes at least one of the following: information of a third-party server; information of the AI ​​data processing entity in the core network; information of the AI ​​data processing entity in OAM; cell identification information and / or beam identification information; cell measurement result information and / or beam measurement result information; observation window length and / or prediction window length; N future instances and / or N future time slots and / or N future samples; additional condition information on the network side; additional condition information on the terminal side.

[0425] In cases where the target use case includes an RLF prediction use case, the measurement result information in the basic data may include L1 filtered measurement results, or L3 filtered measurement results, or unfiltered underlying measurement results.

[0426] 2. The occurrence time of RLF.

[0427] In an optional embodiment of this application, the time information can be an absolute time or a relative time.

[0428] 3. RLF (Remote Flow Response) indication message.

[0429] The RLF occurrence indication information is used to indicate whether an RLF has occurred. In an optional embodiment of this application, the RLF occurrence indication information may indicate whether an RLF has occurred within the second window length.

[0430] 4. Time information of out-of-step was detected.

[0431] In an optional embodiment of this application, the time information for detecting a loss of synchronization can refer to the time when the first loss of synchronization is detected, and this time information can be an absolute time or a relative time.

[0432] 5. Time information when the number of timed steps out of step reaches the preset number.

[0433] In an optional embodiment of this application, the time information for detecting the number of time points of step loss to reach a preset number can refer to the time when T310 is turned on. This time information can be an absolute time or a relative time.

[0434] It should be noted that the multiple preset number of times in the embodiments of this application only indicate that the number of times is preset. The embodiments of this application do not explicitly or implicitly suggest that the preset number of times that occur multiple times is the same or equal.

[0435] 6. Synchronized time information was detected.

[0436] In an optional embodiment of this application, the synchronization detection time information can refer to the time when synchronization was first detected. Specifically, the first detection of synchronization refers to the first detection of synchronization after a loss of synchronization occurs, or the first detection of synchronization after T310 has been enabled. In an optional embodiment of this application, this time information can be an absolute time or a relative time.

[0437] 7. Time information indicating when the number of synchronization attempts reaches the preset number.

[0438] In an optional embodiment of this application, the time information indicating when the number of synchronization attempts reaches a preset number can refer to the time when T310 stops. In an optional embodiment of this application, this time information can be an absolute time or a relative time.

[0439] 8. RLM configuration.

[0440] In optional embodiments of this application, the RLM configuration may include, for example, T310 configuration, N310 configuration, N311 configuration, etc., and this application does not specifically limit this.

[0441] 9. Indication information for whether the target model is a model that directly performs RLF prediction or a model that indirectly performs RLF prediction.

[0442] 10. Second window length.

[0443] In an exemplary embodiment, as shown in FIG5, a data collection method is provided. Taking the application of this method to a network device as an example, the method includes the following steps:

[0444] Step 501: Send data collection configuration to the terminal.

[0445] Step 502: Receive the first training data sent by the receiving terminal.

[0446] The first training data is collected and / or recorded by the terminal based on the data collection configuration. The first training data is used to train and / or update the target model. The target model is used to implement the target use case. The target use case includes at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case.

[0447] In an optional embodiment of this application, the network device may also receive a configuration request sent by the terminal, wherein the configuration request is used to request the network device to send data collection configuration.

[0448] In an optional embodiment of this application, the network device may also send an allow request indication message to the terminal, wherein the allow request indication message is used to instruct the terminal to send the configuration request to the network device.

[0449] In an optional embodiment of this application, the network device may also send a data acquisition request to the terminal, wherein the data acquisition request is used to request the terminal to send the first training data.

[0450] In an optional embodiment of this application, the network device may also receive an availability indication message sent by the terminal, wherein the availability indication message is used to indicate that the terminal has stored the first training data.

[0451] It should be noted that the configuration request, data collection configuration, and first training data have already been described above, and the same applies here. For the sake of brevity, the embodiments of this application will not be repeated.

[0452] In an exemplary embodiment, as shown in FIG6, a data collection method is provided, including the following steps:

[0453] Step 601: The network device sends an allow request indication message to the terminal, wherein the allow request indication message is used to instruct the terminal to send a configuration request to the network device.

[0454] It should be noted that this step is optional.

[0455] Step 602: In response to the permission request indication information, the terminal sends a configuration request to the network device.

[0456] It should be noted that this step is optional. In addition, if step 601 is not executed, the terminal can independently trigger the sending of a configuration request to the network device.

[0457] Step 603: In response to the configuration request, the network device sends a data collection configuration to the terminal.

[0458] Without performing step 602, the network device can autonomously trigger the sending of data collection configuration to the terminal.

[0459] Step 604: The terminal collects and / or records the first training data according to the data collection configuration.

[0460] Step 605: The terminal sends an availability indication message to the network device, wherein the availability indication message is used to indicate that the terminal has stored the first training data.

[0461] It should be noted that this step is optional.

[0462] Step 606: In response to the available indication information, the network device sends a data acquisition request to the terminal.

[0463] It should be noted that this step is optional. In addition, if step 605 is not executed, the network device can automatically trigger the sending of a data acquisition request to the terminal.

[0464] Step 607: In response to the data acquisition request, the terminal sends the first training data to the network device according to the data collection configuration.

[0465] Without executing step 606, the terminal can autonomously trigger the sending of the first training data to the network device.

[0466] Step 608: The network device sends the first training data to a third-party server so that the third-party server can train and / or update the target model based on the first training data.

[0467] It should be noted that this step is optional. Without executing step 608, the network device can train and / or update the target model based on the first training data.

[0468] For ease of understanding, the technical solutions provided in the embodiments of this application will be further described below with reference to several embodiments.

[0469] Example 1:

[0470] In Example 1, the target use case is the RRM time-domain prediction use case.

[0471] Step 1: Optionally, the terminal sends a configuration request to the network device, requesting the network device to configure data collection for RRM time-domain prediction. This configuration request can be triggered by a third-party server, such as an OTT server triggering the terminal to send a configuration request to the network device, or it can be triggered autonomously by the terminal. The content of the configuration request can be generated by the terminal, a third-party server, or predefined. The reason for the request may be a lack of relevant models, or the model may need to be updated. The content of the configuration request includes one or more of the following:

[0472] - Request instruction information, which is used to instruct the network device to configure the collection and / or recording of the first training data, wherein the first training data is used for the target model, the target model is used to implement the RRM time-domain prediction use case, and the request instruction information may instruct one or more of the sub-use cases in Table 1 or Table 2 (by instructing use case number or use case name or function number or function name, etc.).

[0473] - Purpose indication information, which indicates the purpose of training and / or updating the target model. In optional embodiments of this application, the purpose of training and / or updating the target model may be, for example, to improve switching performance or reduce measurements.

[0474] - Reason indication information, which indicates the reason for training and / or updating the target model. In an optional embodiment of this application, the reason for training and / or updating the target model may be, for example, the absence of a target model (i.e., the need to train a new target model) or the need to update the target model.

[0475] - Model inputs and / or model outputs, such as the ability to request measurements and / or recordings of one or more neighboring cells, the ability to request measurements and / or recordings of one or more beams, and the ability to configure the reporting format (pattern) of model input data and / or model output data.

[0476] - Additional conditions on the network side, such as those represented by associated ID.

[0477] - Model window length, which includes the observation window (OW) length and / or prediction window (PW) length (e.g., OW is 5 instances, PW is 5 instances).

[0478] - Additional conditions on the terminal side, such as the state of the terminal, for example, the speed of the terminal, the location information of the terminal.

[0479] - First model type information, which indicates whether the target model is a per-cell model or a cluster model.

[0480] - Information about third-party servers, which may include, for example, the server's ID, tracking number, identifier, and tracking reference.

[0481] Note 1: Before sending a configuration request to the network device, the terminal may also receive an indication message (allow request indication message) from the network device indicating that the terminal can send this configuration request.

[0482] Note 2: In scenarios such as handover / reconstruction, this configuration request can be transmitted in an interface such as Xn so that the newly accessed cell of the terminal knows whether and / or how to configure the terminal to collect data.

[0483] Note 3: Whether a network device accepts or rejects this configuration request depends on the network implementation.

[0484] Note 4: When the network device is an access network device, the terminal can send the configuration request to the access network device via RRC signaling. When the network device is a core network device, the terminal can send the configuration request to the core network device via NAS signaling, or via the UP channel. When the network device is an OAM, the terminal can send the configuration request to the OAM via the access network, or via the UP channel from the core network to the OAM.

[0485] Step 2: After receiving the configuration request from the terminal, or triggered autonomously by the network device, the network device sends data collection configuration to the terminal. This data collection configuration may include one or more of the following: measurement configuration, recording configuration, and reporting configuration.

[0486] The measurement configuration may include one or more of the following:

[0487] a) SetB / SetA related information; where SetB is the model input data and SetA is the label (i.e., the model output data).

[0488] - Cell list.

[0489] - Beam list, which may be configured by cell, by measurement object, or not by cell or measurement object.

[0490] - Frequency list.

[0491] - First frequency indication information, which is a display indication information used to indicate whether to collect measurement result information of cells on this frequency point. This first frequency indication information can be configured on a unit basis according to the measurement object.

[0492] - First measurement object indication information, which is the displayed indication information used to indicate whether to collect measurement result information of all measurement objects.

[0493] - Second measurement object indication information. The second measurement object indication information corresponds to a measurement object and is used to indicate whether to measure and / or record the corresponding measurement object. In optional embodiments of this application, the measurement configuration may include one or more second measurement object indication information, wherein multiple second measurement object indication information can correspond to different measurement objects.

[0494] - Reference signal configuration, used to configure the reference signal that needs to be measured and / or recorded, for example, the reference signal may include SSB or CSI-RS, etc.

[0495] - Model-related parameters, where model-related parameters are used to implicitly indicate the measurement objects that need to be measured and / or recorded.

[0496] b) Purpose configuration: Purpose configuration is used to configure the purpose of measurement and / or recording, for example, the purpose can be RRM time-domain prediction use case data collection.

[0497] The recording configuration may include: start recording conditions and / or stop recording conditions.

[0498] a) The conditions for starting recording may include one or more of the following:

[0499] - The signal quality may be higher or lower than a signal quality threshold. In optional embodiments of this application, recording may begin when the signal quality is higher or lower than the signal quality threshold, or measurement data may be recorded when the signal quality is higher or lower than the signal quality threshold. In optional embodiments of this application, the signal quality may be the signal quality of the serving cell, the signal quality of neighboring cells, the signal quality of neighboring cell beams, the signal quality of the serving cell beam, etc., and this application does not specifically limit this. In optional embodiments of this application, the signal quality may include RSRP, RSRQ, SINR, etc., and this application does not specifically limit this. In optional embodiments of this application, the signal quality threshold may be configured by the network side.

[0500] - A reconfiguration or access to a new cell occurs. A reconfiguration can refer to a change in data collection configuration, while access to a new cell can include mechanisms such as cell handover, cell selection, or cell reselection.

[0501] - Changes to additional conditions on the network side, such as changes to the associated ID.

[0502] - Changes in additional conditions on the terminal side, such as changes in the terminal's speed or its geographical location.

[0503] b) The conditions for stopping recording may include one or more of the following:

[0504] - The duration of a single recording reaches the preset duration.

[0505] - The total recording time has reached the preset time.

[0506] - The number of records has reached the preset number.

[0507] - The signal quality may be higher or lower than a signal quality threshold. In optional embodiments of this application, recording may stop when the signal quality is higher or lower than the signal quality threshold, or measurement data for signal quality higher or lower than the signal quality threshold may not be recorded. In optional embodiments of this application, the signal quality may be the signal quality of the serving cell, the signal quality of neighboring cells, the signal quality of neighboring cell beams, the signal quality of the serving cell beam, etc., and this application does not specifically limit these. In optional embodiments of this application, the signal quality may include RSRP, RSRQ, SINR, etc., and this application does not specifically limit these. In optional embodiments of this application, the signal quality threshold may be configured by the network side.

[0508] - Changes to additional conditions on the network side, such as changes to the associated ID.

[0509] - Changes in additional conditions on the terminal side, such as changes in the terminal's speed or its geographical location.

[0510] - The data volume has reached the preset data volume.

[0511] - A reconfiguration or access to a new cell occurs. A reconfiguration can refer to a change in data collection configuration, while access to a new cell can include mechanisms such as cell handover, cell selection, or cell reselection.

[0512] - Enter the idle state or the inactive state.

[0513] - Battery depleted.

[0514] - Insufficient memory.

[0515] The reported configuration may include one or more of the following:

[0516] - Information about third-party servers.

[0517] - Information about AI data processing entities in the core network. AI data processing entities in the core network may include NWDAF. The information about AI data processing entities in the core network may include the entity's number, tracking number, identifier, and / or tracking reference.

[0518] - Information about AI data processing entities in OAM, which may include the entity's number, tracking number, identifier, and / or tracking reference.

[0519] - Reporting measurement configuration: This configuration is used to configure what types of measurements to record and / or report. For example, the reporting measurement configuration can be configured to report measurements such as RSRP, RSRQ, and SINR.

[0520] - Observation window length and / or prediction window length.

[0521] -N future instances and / or N future slots and / or N future samples, where N is a value greater than or equal to 1.

[0522] - Instructions for recording cell-level or beam-level measurement results.

[0523] - One or more network-side additional conditions, which may include, for example, the network environment in which the terminal is located, such as an associated ID.

[0524] - One or more terminal-side additional conditions, which may include, for example, the state of the terminal, such as the speed of the terminal, the location information of the terminal, and the terminal-side additional conditions may be, for example, an identifier.

[0525] - Indication information for the top K cells with high signal quality, where K is a value greater than or equal to 1.

[0526] - Indication information for the top K beams with high signal quality, where K is a value greater than or equal to 1.

[0527] - Reporting method indication information. This information indicates whether model input and output data should be recorded separately, or whether they should be recorded uniformly. It should be noted that recording model input and output data uniformly means that model input and output data are recorded and / or reported together without distinguishing between them. Recording model input and output data separately means that model input and output data need to be distinguished during recording and / or reporting. For example, model input and output data can be reported using different signaling, or the information can indicate which part of the reported data is model input and which part is model output.

[0528] Note 5: This data collection configuration can be generated and configured by gNB, CN, or OAM. When the network device is an access network device, the data collection configuration can be carried in RRC signaling and sent to the terminal via RRC signaling. When the network device is a core network device, the data collection configuration can be carried in NAS signaling and sent to the terminal via NAS signaling. Alternatively, the data collection configuration can be sent to the terminal via the UP channel. When the network device is an OAM, the data collection configuration can be sent to the terminal by the OAM through the access network. Alternatively, the data collection configuration can be sent to the terminal by the OAM through the core network via the UP channel.

[0529] Note 6: This data collection configuration can be based on the current RRM measurement configuration framework.

[0530] Step 3: The terminal collects and / or records data based on the data collection configuration; the content recorded by the terminal (i.e., the first training data) may include one or more of the following:

[0531] - Information about third-party servers.

[0532] - Information about entities processed by AI data processing in the core network.

[0533] - Information about entities in OAM AI data processing.

[0534] - Cell identification information and / or beam identification information, cell measurement result information and / or beam measurement result information. In optional embodiments of this application, the cell measurement result information may include, for example, the cell's RSRP measurement result, RSRQ measurement result, SINR measurement result, etc., and the beam measurement result information may include, for example, the beam's RSRP measurement result, RSRQ measurement result, SINR measurement result, etc. In optional embodiments of this application, the cell measurement result information and / or beam measurement result information may, for example, include model input data and / or model output data. As mentioned above, in optional embodiments of this application, model input data and model output data can be reported separately or reported uniformly. It should be noted that uniformly reporting model input data and model output data means that when reporting data, model input data and model output data are not distinguished, but are mixed together for reporting. Recording model input and output data separately means that when recording and / or reporting data, it is necessary to distinguish between model input and output data. For example, model input and output data can be reported using different signaling methods, or the reported data can be used to indicate which part is model input and which part is model output. For instance, reporting model input and output data uniformly means recording all time slots together; recording model input and output data separately means recording the time slots for model input and output data separately.

[0535] - Observation window length and / or prediction window length.

[0536] -N future instances and / or N future time slots and / or N future samples, where N is a value greater than or equal to 1.

[0537] - Additional conditions information on the network side, which may include, for example, the network environment in which the terminal is located, such as the associated ID.

[0538] - Additional condition information on the terminal side, such as the terminal's state, for example, the terminal's speed or location information.

[0539] - Identification information and / or measurement result information of the top K cells with high signal quality, where K is a value greater than or equal to 1. In optional embodiments of this application, the measurement result information may include, for example, RSRP measurement results, RSRQ measurement results, SINR measurement results, etc.

[0540] - Identification information and / or measurement result information of the top K beams with high signal quality, where K is a value greater than or equal to 1. In optional embodiments of this application, the measurement result information may include, for example, RSRP measurement results, RSRQ measurement results, SINR measurement results, etc.

[0541] Step 4: Optionally, the terminal sends an indication message to the network device.

[0542] Step 5: After receiving the available indication information sent by the terminal, the network device sends a data acquisition request to the terminal, requesting the terminal to report the collected data.

[0543] Step 6: After receiving the data acquisition request sent by the network device, the terminal reports the collected data to the network device.

[0544] Note 7: The data acquisition request in step 5 can be based on the network device's own request, that is, no terminal reports the information that can be obtained to the network device.

[0545] Note 8: In step 6, the terminal can send the collected data to the network device autonomously, i.e., it does not receive a data acquisition request from the network device.

[0546] Note 9: When the network device is an access network device, the terminal can send the first training data to the access network device via RRC signaling. When the network device is a core network device, the terminal can send the first training data to the core network device via NAS signaling, or via the UP channel. When the network device is an OAM, the terminal can send the first training data to the OAM via the access network, or via the UP channel from the core network to the OAM.

[0547] Step 7 (optional): The network device sends the collected data to a third-party server for training the RRM time-domain prediction AI / ML model.

[0548] Note 10: Alternatively, step 7 may be omitted, i.e., the AI / ML model for RRM temporal prediction may be trained directly by the network device.

[0549] Note 11: Applicable to both terminal-side model training data collection and network-side model training data collection.

[0550] Example 2:

[0551] In Example 2, the target use case is the RRM frequency domain prediction use case.

[0552] Step 1: Optionally, the terminal sends a configuration request to the network device, requesting the network device to configure the data collection for RRM frequency domain prediction. This configuration request can be triggered by a third-party server, such as an OTT server triggering the terminal to send a configuration request to the network device, or it can be triggered autonomously by the terminal. The content of the configuration request can be generated by the terminal, a third-party server, or predefined. The reason for the request may be a lack of relevant models, or the model may need to be updated. The content of the configuration request includes one or more of the following:

[0553] - Request instruction information, which is used to instruct the network device to configure the collection and / or recording of the first training data, wherein the first training data is used to train the target model, the target model is used to implement the RRM frequency domain prediction use case, and the request instruction information may instruct one or more of the sub-use cases in Table 1 or Table 2 (by instructing use case number or use case name or function number or function name, etc.).

[0554] - Purpose indication information, which indicates the purpose of training and / or updating the target model. In optional embodiments of this application, the purpose of training and / or updating the target model may be, for example, to improve switching performance or reduce measurements.

[0555] - Reason indication information, which indicates the reason for training and / or updating the target model. In an optional embodiment of this application, the reason for training and / or updating the target model may be, for example, the absence of a target model (i.e., the need to train a new target model) or the need to update the target model.

[0556] - Model inputs and / or model outputs, such as the ability to request measurements and / or recordings of one or more neighboring cells, the ability to request measurements and / or recordings of one or more beams, and the ability to configure the reporting format (pattern) of model input data and / or model output data. The frequency points of the model input measurement and / or recorded measurement data may be different from those of the model output measurement and / or recorded measurement data. The frequency points of the cells in the model input measurement and / or recorded measurement data may be different from those of the cells in the model output measurement and / or recorded measurement data. The frequency points of the beams in the model input measurement and / or recorded measurement data may be different from those of the beams in the model output measurement and / or recorded measurement data.

[0557] - Additional conditions on the network side, such as those represented by associated ID.

[0558] - Additional conditions on the terminal side, such as the state of the terminal, for example, the speed of the terminal, the location information of the terminal.

[0559] - Information about third-party servers, which may include, for example, the server's ID, tracking number, identifier, and tracking reference.

[0560] Note 1: Before sending a configuration request to the network device, the terminal may also receive an indication message (allow request indication message) from the network device indicating that the terminal can send this configuration request.

[0561] Note 2: In scenarios such as handover / reconstruction, this configuration request can be transmitted in an interface such as Xn so that the newly accessed cell of the terminal knows whether and / or how to configure the terminal to collect data.

[0562] Note 3: Whether a network device accepts or rejects this configuration request depends on the network implementation.

[0563] Note 4: When the network device is an access network device, the terminal can send the configuration request to the access network device via RRC signaling. When the network device is a core network device, the terminal can send the configuration request to the core network device via NAS signaling, or via the UP channel. When the network device is an OAM, the terminal can send the configuration request to the OAM via the access network, or via the UP channel from the core network to the OAM.

[0564] Step 2: After receiving the configuration request from the terminal, or triggered autonomously by the network device, the network device sends data collection configuration to the terminal. This data collection configuration may include one or more of the following: measurement configuration, recording configuration, and reporting configuration.

[0565] The measurement configuration may include one or more of the following:

[0566] a) SetB / SetA related information; where SetB is the model input data and SetA is the label (i.e., the model output data).

[0567] - Cell list, which includes cells at different frequencies.

[0568] - Beamlist: This list includes beams at different frequencies.

[0569] - Frequency list.

[0570] - First frequency indication information, which is a display indication information used to indicate whether to collect measurement result information of cells on this frequency point. This first frequency indication information can be configured on a unit basis according to the measurement object.

[0571] - First measurement object indication information, which is the displayed indication information used to indicate whether to collect measurement result information of all measurement objects.

[0572] - Second measurement object indication information. The second measurement object indication information corresponds to a measurement object and is used to indicate whether to measure and / or record the corresponding measurement object. In optional embodiments of this application, the measurement configuration may include one or more second measurement object indication information, wherein multiple second measurement object indication information can correspond to different measurement objects.

[0573] - Reference signal configuration, used to configure the reference signal that needs to be measured and / or recorded, for example, the reference signal may include SSB or CSI-RS, etc.

[0574] - Model-related parameters, where model-related parameters are used to implicitly indicate the measurement objects that need to be measured and / or recorded.

[0575] b) Purpose configuration: Purpose configuration is used to configure the purpose of measurement and / or recording, for example, the purpose can be RRM frequency domain prediction use case data collection.

[0576] The recording configuration may include: start recording conditions and / or stop recording conditions.

[0577] a) The conditions for starting recording may include one or more of the following:

[0578] - The signal quality may be higher or lower than a signal quality threshold. In optional embodiments of this application, recording may begin when the signal quality is higher or lower than the signal quality threshold, or measurement data may be recorded when the signal quality is higher or lower than the signal quality threshold. In optional embodiments of this application, the signal quality may be the signal quality of the serving cell, the signal quality of neighboring cells, the signal quality of neighboring cell beams, the signal quality of the serving cell beam, etc., and this application does not specifically limit this. In optional embodiments of this application, the signal quality may include RSRP, RSRQ, SINR, etc., and this application does not specifically limit this. In optional embodiments of this application, the signal quality threshold may be configured by the network side.

[0579] - A reconfiguration or access to a new cell occurs. A reconfiguration can refer to a change in data collection configuration, while access to a new cell can include mechanisms such as cell handover, cell selection, or cell reselection.

[0580] - Changes to additional conditions on the network side, such as changes to the associated ID.

[0581] - Changes in additional conditions on the terminal side, such as changes in the terminal's speed or its geographical location.

[0582] b) The conditions for stopping recording may include one or more of the following:

[0583] - The duration of a single recording reaches the preset duration.

[0584] - The total recording time has reached the preset time.

[0585] - The number of records has reached the preset number.

[0586] - The signal quality may be higher or lower than a signal quality threshold. In optional embodiments of this application, recording may stop when the signal quality is higher or lower than the signal quality threshold, or measurement data for signal quality higher or lower than the signal quality threshold may not be recorded. In optional embodiments of this application, the signal quality may be the signal quality of the serving cell, the signal quality of neighboring cells, the signal quality of neighboring cell beams, the signal quality of the serving cell beam, etc., and this application does not specifically limit these. In optional embodiments of this application, the signal quality may include RSRP, RSRQ, SINR, etc., and this application does not specifically limit these. In optional embodiments of this application, the signal quality threshold may be configured by the network side.

[0587] - Changes to additional conditions on the network side, such as changes to the associated ID.

[0588] - Changes in additional conditions on the terminal side, such as changes in the terminal's speed or its geographical location.

[0589] - The data volume has reached the preset data volume.

[0590] - A reconfiguration or access to a new cell occurs. A reconfiguration can refer to a change in data collection configuration, while access to a new cell can include mechanisms such as cell handover, cell selection, or cell reselection.

[0591] - Enter the idle state or the inactive state.

[0592] - Battery depleted.

[0593] - Insufficient memory.

[0594] The reported configuration may include one or more of the following:

[0595] - Information about third-party servers.

[0596] - Information about entities processed by AI data processing in the core network.

[0597] - Information about entities in OAM AI data processing.

[0598] - Reporting measurement configuration: This configuration is used to configure what types of measurements to record and / or report. For example, the reporting measurement configuration can be configured to report measurements such as RSRP, RSRQ, and SINR.

[0599] - Instructions for recording cell-level or beam-level measurement results.

[0600] - One or more network-side additional conditions, which may include, for example, the network environment in which the terminal is located, such as an associated ID.

[0601] - One or more terminal-side additional conditions, which may include, for example, the state of the terminal, such as the speed of the terminal, the location information of the terminal, and the terminal-side additional conditions may be, for example, an identifier.

[0602] - Reporting method indication information. This information indicates whether model input and output data should be recorded separately, or whether they should be recorded uniformly. It should be noted that recording model input and output data uniformly means that model input and output data are recorded and / or reported together without distinguishing between them. Recording model input and output data separately means that model input and output data need to be distinguished during recording and / or reporting. For example, model input and output data can be reported using different signaling, or the information can indicate which part of the reported data is model input and which part is model output.

[0603] - Instructions for reporting frequency information.

[0604] - Indication information regarding whether data is recorded according to frequency point.

[0605] Note 5: This data collection configuration can be generated and configured by gNB, CN, or OAM. When the network device is an access network device, the data collection configuration can be carried in RRC signaling and sent to the terminal via RRC signaling. When the network device is a core network device, the data collection configuration can be carried in NAS signaling and sent to the terminal via NAS signaling. Alternatively, the data collection configuration can be sent to the terminal via the UP channel. When the network device is an OAM, the data collection configuration can be sent to the terminal by the OAM through the access network. Alternatively, the data collection configuration can be sent to the terminal by the OAM through the core network via the UP channel.

[0606] Note 6: This data collection configuration can be based on the current RRM measurement configuration framework.

[0607] Step 3: The terminal collects and / or records data based on the data collection configuration; the content recorded by the terminal (i.e., the first training data) may include one or more of the following:

[0608] Information about third-party servers.

[0609] - Information about entities processed by AI data processing in the core network.

[0610] - Information about entities in OAM AI data processing.

[0611] - Cell identification information and / or beam identification information, cell measurement result information and / or beam measurement result information. In optional embodiments of this application, the cell measurement result information may include, for example, the cell's RSRP measurement result, RSRQ measurement result, SINR measurement result, etc., and the beam measurement result information may include, for example, the beam's RSRP measurement result, RSRQ measurement result, SINR measurement result, etc. In optional embodiments of this application, the cell measurement result information and / or beam measurement result information may, for example, include model input data and / or model output data. As mentioned above, in optional embodiments of this application, model input data and model output data can be reported separately or reported uniformly. It should be noted that uniformly reporting model input data and model output data means that when reporting data, model input data and model output data are not distinguished, but are mixed together for reporting. Recording model input and output data separately means that when recording and / or reporting data, it is necessary to distinguish between model input and output data. For example, model input and output data can be reported using different signaling methods, or the reported data can be labeled with indications of which part is model input and which part is model output. For instance, reporting model input and output data uniformly means recording all frequency point measurements together; recording model input and output data separately means recording the frequency points of the model input data and the frequency points of the model output data separately.

[0612] -Frequency information.

[0613] - Additional conditions information on the network side, which may include, for example, the network environment in which the terminal is located, such as the associated ID.

[0614] - Additional condition information on the terminal side, such as the terminal's state, for example, the terminal's speed or location information.

[0615] Step 4: Optionally, the terminal sends an indication message to the network device.

[0616] Step 5: After receiving the available indication information sent by the terminal, the network device sends a data acquisition request to the terminal, requesting the terminal to report the collected data.

[0617] Step 6: After receiving the data acquisition request sent by the network device, the terminal reports the collected data to the network device.

[0618] Note 7: The data acquisition request in step 5 can be based on the network device's own request, that is, no terminal reports the information that can be obtained to the network device.

[0619] Note 8: In step 6, the terminal can send the collected data to the network device autonomously, i.e., it does not receive a data acquisition request from the network device.

[0620] Note 9: When the network device is an access network device, the terminal can send the first training data to the access network device via RRC signaling. When the network device is a core network device, the terminal can send the first training data to the core network device via NAS signaling, or via the UP channel. When the network device is an OAM, the terminal can send the first training data to the OAM via the access network, or via the UP channel from the core network to the OAM.

[0621] Step 7 (optional): The network device sends the collected data to a third-party server for RRM frequency domain prediction AI / ML model training.

[0622] Note 10: Alternatively, step 7 may be omitted, i.e., the AI / ML model for RRM frequency domain prediction may be trained directly by the network device.

[0623] Note 11: Applicable to both terminal-side model training data collection and network-side model training data collection.

[0624] Example 3:

[0625] In Example 3, the target use case is the event prediction use case.

[0626] Step 1: Optionally, the terminal sends a configuration request to the network device, requesting the network device to configure the data collection for event prediction. This configuration request can be triggered by a third-party server, such as an OTT server triggering the terminal to send a configuration request to the network device, or it can be triggered autonomously by the terminal. The content of the configuration request can be generated by the terminal, a third-party server, or predefined. The reason for the request may be a lack of relevant models, or the model may need to be updated. The content of the configuration request includes one or more of the following:

[0627] - A request instruction message, which instructs a network device to configure the collection and / or recording of first training data, wherein the first training data is used to train a target model, and the target model is used to implement event prediction use cases. In optional embodiments of this application, the request instruction message may also indicate a use case number and / or sub-use case number, use case name and / or sub-use case name, function number, function name, the purpose of the configuration request (obtaining training data), etc., and this application embodiment does not specifically limit this.

[0628] - Purpose indication information, which indicates the purpose of training and / or updating the target model. In optional embodiments of this application, the purpose of training and / or updating the target model may be, for example, to improve switching performance or reduce measurements.

[0629] - Reason indication information, which indicates the reason for training and / or updating the target model. In an optional embodiment of this application, the reason for training and / or updating the target model may be, for example, the absence of a target model (i.e., the need to train a new target model) or the need to update the target model.

[0630] - Second model type information, which indicates whether the target model is a model that directly predicts events or a model that indirectly predicts events.

[0631] - Model inputs and / or model outputs, such as the ability to request measurement and / or recording of one or more neighboring cells, the ability to request measurement and / or recording of one or more beams, the ability to request recording of terminal location information, and the ability to request recording of configuration information for events that need to be recorded.

[0632] - Additional conditions on the network side, such as those represented by associated ID (associatedid).

[0633] - Requested event type information, for example, the requested event type may include A3 event, A4 event, A5 event, etc.

[0634] - Request event configuration parameters, such as the event's TTT (Time To Trigger) and offset.

[0635] - Model window length, which includes the observation window (OW) length and / or prediction window (PW) length (e.g., OW is 5 instances, PW is 5 instances).

[0636] - Additional conditions on the terminal side, such as the state of the terminal, for example, the speed of the terminal, the location information of the terminal.

[0637] - First model type information, which indicates whether the target model is a per-cell model or a cluster model.

[0638] - Information about third-party servers, which may include, for example, the server's ID, tracking number, identifier, and tracking reference.

[0639] Note 1: Before sending a configuration request to the network device, the terminal may also receive an indication message (allow request indication message) from the network device indicating that the terminal can send this configuration request.

[0640] Note 2: In scenarios such as handover / reconstruction, this configuration request can be transmitted in an interface such as Xn so that the newly accessed cell of the terminal knows whether and / or how to configure the terminal to collect data.

[0641] Note 3: Whether a network device accepts or rejects this configuration request depends on the network implementation.

[0642] Note 4: When the network device is an access network device, the terminal can send the configuration request to the access network device via RRC signaling. When the network device is a core network device, the terminal can send the configuration request to the core network device via NAS signaling, or via the UP channel. When the network device is an OAM, the terminal can send the configuration request to the OAM via the access network, or via the UP channel from the core network to the OAM.

[0643] Step 2: After receiving the configuration request from the terminal, or triggered autonomously by the network device, the network device sends data collection configuration to the terminal. This data collection configuration may include one or more of the following: measurement configuration, recording configuration, and reporting configuration.

[0644] The measurement configuration may include one or more of the following:

[0645] a) SetB / SetA related information; where SetB is the model input data and SetA is the label (i.e., the model output data).

[0646] - Cell list.

[0647] - Beam list, which may be configured by cell, by measurement object, or not by cell or measurement object.

[0648] - Frequency list.

[0649] - First frequency indication information, which is a display indication information used to indicate whether to collect measurement result information of cells on this frequency point. This first frequency indication information can be configured on a unit basis according to the measurement object.

[0650] - First measurement object indication information, which is the displayed indication information used to indicate whether to collect measurement result information of all measurement objects.

[0651] - Second measurement object indication information. The second measurement object indication information corresponds to a measurement object and is used to indicate whether to measure and / or record the corresponding measurement object. In optional embodiments of this application, the measurement configuration may include one or more second measurement object indication information, wherein multiple second measurement object indication information can correspond to different measurement objects.

[0652] - Report configuration indication information, which may correspond to an event and / or a measurement object. This reporting configuration indication information is used to indicate whether to measure and / or record the event and / or measurement object corresponding to this reporting configuration. The measurement configuration may include one or more reporting configuration indication information, wherein multiple reporting configuration indication information may correspond to different events and / or measurement objects.

[0653] - Third measurement object indication information, wherein the third measurement object indication information is used to indicate whether all measurement objects that meet the event should be measured and / or recorded. The event can be a specific event, which can be an event specified by the network side, and can include: A3 event, A4 event, A5 event, etc.

[0654] - Reference signal configuration, used to configure the reference signal that needs to be measured and / or recorded, for example, the reference signal may include SSB or CSI-RS, etc.

[0655] - Model-related parameters, where model-related parameters are used to implicitly indicate the measurement objects that need to be measured and / or recorded.

[0656] b) Purpose configuration: Purpose configuration is used to configure the purpose of measurement and / or recording, for example, the purpose can be event use case data collection.

[0657] c) Event configuration, such as configuration information for parameters like TTT and offset of an event.

[0658] The recording configuration may include: start recording conditions and / or stop recording conditions.

[0659] a) The conditions for starting recording may include one or more of the following:

[0660] - The event entry condition is met. This event entry condition can be, for example, the entry condition of event A3 is met, or the entry condition of event A4 is met, or the entry condition of event A5 is met, etc.

[0661] - The first TTT is activated, which corresponds to the event entry condition. Generally, the first TTT is activated when the event entry condition is met; in other words, meeting the event entry condition and activating the first TTT are, in some sense, equivalent.

[0662] - Event satisfied. In an optional embodiment of this application, event satisfied can also be understood as the event triggering report being satisfied.

[0663] - First TTT timeout. In general, the first TTT timeout can be understood as the event being satisfied. In other words, event satisfaction and the first TTT timeout can be considered equivalent in some sense.

[0664] - The difference between the current signal quality (RSRP, RSRQ, SINR) and the event entry condition reaches a preset threshold. The event entry condition can correspond to a signal quality threshold, such as RSRP, RSRQ, SINR, etc. The difference between the current signal quality (RSRP, RSRQ, SINR) and the corresponding signal quality threshold (RSRP, RSRQ, SINR) can be understood as the difference between the current signal quality and the event entry condition. This difference can include RSRP, RSRQ, and SINR differences. Recording can begin when the difference between the current signal quality and the event entry condition reaches the preset threshold, or it can record measurement data where the difference is equal to the preset threshold. For example, recording is triggered when the difference between the current signal quality measurement and the event entry condition is less than the preset threshold.

[0665] - The signal quality may be higher or lower than a signal quality threshold. In optional embodiments of this application, recording may begin when the signal quality is higher or lower than the signal quality threshold, or measurement data may be recorded when the signal quality is higher or lower than the signal quality threshold. In optional embodiments of this application, the signal quality may be the signal quality of the serving cell, the signal quality of neighboring cells, the signal quality of neighboring cell beams, the signal quality of the serving cell beam, etc., and this application does not specifically limit this. In optional embodiments of this application, the signal quality may include RSRP, RSRQ, SINR, etc., and this application does not specifically limit this. In optional embodiments of this application, the signal quality threshold may be configured by the network side.

[0666] - First recording period. This first recording period is a preset duration before and / or after the event entry condition is met. In other words, data within the first recording period can be recorded. In an optional embodiment of this application, the first recording period can be a single, continuous time window.

[0667] - Second recording period. This second recording period is a preset duration before and / or after the event is satisfied. In other words, data can be recorded within the second recording period. In an optional embodiment of this application, the second recording period can be a single overall time window.

[0668] - A reconfiguration or access to a new cell occurs. A reconfiguration can refer to a change in data collection configuration, while access to a new cell can include mechanisms such as cell handover, cell selection, or cell reselection.

[0669] - Changes to additional conditions on the network side, such as changes to the associated ID.

[0670] - Changes in additional conditions on the terminal side, such as changes in the terminal's speed or its geographical location.

[0671] b) The conditions for stopping recording may include one or more of the following:

[0672] - The event is satisfied.

[0673] - First TTT timeout.

[0674] - The event exit condition is met.

[0675] - The second TTT is activated, which corresponds to the event exit condition. In general, the second TTT is activated when the event exit condition is met; in other words, the fulfillment of the event exit condition and the activation of the second TTT can be considered equivalent in some respects.

[0676] - Second TTT timeout.

[0677] - The difference between the current signal quality and the event departure condition reaches a preset threshold. In an optional embodiment of this application, the event departure condition may correspond to a signal quality threshold, such as an RSRP threshold, an RSRQ threshold, a SINR threshold, etc. The difference between the current signal quality (RSRP, RSRQ, SINR) and the signal quality threshold (RSRP, RSRQ, SINR) corresponding to the event departure condition can be understood as the difference between the current signal quality (RSRP, RSRQ, SINR) and the signal quality threshold (RSRP, RSRQ, SINR) corresponding to the event departure condition. Understandably, this difference may include RSRP difference, RSRQ difference, and SINR difference. In an optional embodiment of this application, recording may stop when the difference between the current signal quality and the event departure condition reaches a preset threshold, or measurement data where the difference between the current signal quality and the event departure condition is a preset threshold may not be recorded. In an optional embodiment of this application, for example, recording may stop when the difference between the current signal quality and the event departure condition is less than a preset threshold.

[0678] - The duration of a single recording reaches the preset duration.

[0679] - The total recording time has reached the preset time.

[0680] - The number of records has reached the preset number.

[0681] - The signal quality may be higher or lower than a signal quality threshold. In optional embodiments of this application, recording may stop when the signal quality is higher or lower than the signal quality threshold, or measurement data for signal quality higher or lower than the signal quality threshold may not be recorded. In optional embodiments of this application, the signal quality may be the signal quality of the serving cell, the signal quality of neighboring cells, the signal quality of neighboring cell beams, the signal quality of the serving cell beam, etc., and this application does not specifically limit these. In optional embodiments of this application, the signal quality may include RSRP, RSRQ, SINR, etc., and this application does not specifically limit these. In optional embodiments of this application, the signal quality threshold may be configured by the network side.

[0682] - Changes to additional conditions on the network side, such as changes to the associated ID.

[0683] - Changes in additional conditions on the terminal side, such as changes in the terminal's speed or its geographical location.

[0684] - The data volume has reached the preset data volume.

[0685] - A reconfiguration or access to a new cell occurs. A reconfiguration can refer to a change in data collection configuration, while access to a new cell can include mechanisms such as cell handover, cell selection, or cell reselection.

[0686] - Enter the idle state or the inactive state.

[0687] - Battery depleted.

[0688] - Insufficient memory.

[0689] The reported configuration may include one or more of the following:

[0690] - Information about third-party servers.

[0691] - Information about entities processed by AI data processing in the core network.

[0692] - Information about entities in OAM AI data processing.

[0693] - Reporting measurement configuration: This configuration is used to configure what types of measurements to record and / or report. For example, the reporting measurement configuration can be configured to report measurements such as RSRP, RSRQ, and SINR.

[0694] - Observation window length and / or prediction window length.

[0695] -N future instances and / or N future slots and / or N future samples, where N is a value greater than or equal to 1.

[0696] - First window length. The first window length is the length of the window within which the event is likely to occur, which can be expressed in English as "an window for determining the event probability within the window". In other words, in an optional embodiment of this application, the terminal can detect whether an event has occurred or the probability of the event occurring within the first window length, and report data (such as model output data) to the network device based on the detection results.

[0697] - Event type configuration. In an optional embodiment of this application, the event type configuration is used to configure the types of events that need to be reported. For example, the types of events that need to be reported may include: A3 event, A4 event, A5 event, etc.

[0698] - Instructions for recording cell-level or beam-level measurement results.

[0699] - One or more network-side additional conditions, which may include, for example, the network environment in which the terminal is located, such as an associated ID.

[0700] - One or more terminal-side additional conditions, which may include, for example, the state of the terminal, such as the speed of the terminal, the location information of the terminal, and the terminal-side additional conditions may be, for example, an identifier.

[0701] - Reporting method indication information. This information indicates whether model input and output data should be recorded separately, or whether they should be recorded uniformly. It should be noted that recording model input and output data uniformly means that model input and output data are recorded and / or reported together without distinguishing between them. Recording model input and output data separately means that model input and output data need to be distinguished during recording and / or reporting. For example, model input and output data can be reported using different signaling, or the information can indicate which part of the reported data is model input and which part is model output.

[0702] Note 5: This data collection configuration can be generated and configured by gNB, CN, or OAM. When the network device is an access network device, the data collection configuration can be carried in RRC signaling and sent to the terminal via RRC signaling. When the network device is a core network device, the data collection configuration can be carried in NAS signaling and sent to the terminal via NAS signaling. Alternatively, the data collection configuration can be sent to the terminal via the UP channel. When the network device is an OAM, the data collection configuration can be sent to the terminal by the OAM through the access network. Alternatively, the data collection configuration can be sent to the terminal by the OAM through the core network via the UP channel.

[0703] Note 6: This data collection configuration can be based on the current RRM measurement configuration framework.

[0704] Step 3: The terminal collects and / or records data based on the data collection configuration; the content recorded by the terminal (i.e., the first training data) may include one or more of the following:

[0705] - Information about third-party servers.

[0706] - Information about entities processed by AI data processing in the core network.

[0707] - Information about entities in OAM AI data processing.

[0708] - Cell identification information and / or beam identification information, cell measurement result information and / or beam measurement result information. In optional embodiments of this application, the cell measurement result information may include, for example, the cell's RSRP measurement result, RSRQ measurement result, SINR measurement result, etc., and the beam measurement result information may include, for example, the beam's RSRP measurement result, RSRQ measurement result, SINR measurement result, etc. The measurement result information may include L1 filtered measurement results, or L3 filtered measurement results. In optional embodiments of this application, the cell measurement result information and / or beam measurement result information may, for example, include model input data and / or model output data. As mentioned above, in optional embodiments of this application, model input data and model output data can be reported separately or reported uniformly. It should be noted that uniformly reporting model input data and model output data means that when reporting data, model input data and model output data are not distinguished, but are mixed together for reporting. Recording model input data and model output data separately means that when recording and / or reporting data, it is necessary to distinguish between model input data and model output data. For example, model input data and model output data can be reported through different signaling, or it can be indicated which part of the reported data is model input data and which part is model output data.

[0709] - The time when the event entry condition is met. In an optional embodiment of this application, the time when the event entry condition is met may refer to the time when the first TTT is turned on. In an optional embodiment of this application, this time information may be an absolute time or a relative time.

[0710] - The time when the event departure condition is met. In an optional embodiment of this application, the time when the event departure condition is met may refer to the time when the second TTT is started. In an optional embodiment of this application, this time information may be an absolute time or a relative time.

[0711] - Unfiltered data obtained from each measurement performed during the first TTT and / or second TTT activation process. In an optional embodiment of this application, unfiltered data may refer to the underlying measurement results without L1 filtering.

[0712] - Event configuration information. In optional embodiments of this application, the event configuration information may include, for example, configuration information for event A3, configuration information for event A4, configuration information for event A5, etc., and this application embodiment does not specifically limit this.

[0713] - Observation window length and / or prediction window length.

[0714] -N future instances and / or N future time slots and / or N future samples, where N is a value greater than or equal to 1.

[0715] - First window length.

[0716] - Event occurrence time. In an optional embodiment of this application, the event occurrence time may include the event trigger reporting time, or the first TTT timeout time, which may be an absolute time or a relative time.

[0717] - Event occurrence indication information. In optional embodiments of this application, different events may correspond to different event occurrence indication information. This event occurrence indication information is used to indicate whether the corresponding event has occurred. In optional embodiments of this application, the event occurrence indication information may also indicate the number of times the corresponding event has occurred, etc.

[0718] - Additional conditions information on the network side, which may include, for example, the network environment in which the terminal is located, such as the associated ID.

[0719] - Additional condition information on the terminal side, such as the terminal's state, for example, the terminal's speed or location information.

[0720] -Indicative information for a target model that directly predicts events, or a target model that indirectly predicts events.

[0721] Step 4: Optionally, the terminal sends an indication message to the network device.

[0722] Step 5: After receiving the available indication information sent by the terminal, the network device sends a data acquisition request to the terminal, requesting the terminal to report the collected data.

[0723] Step 6: After receiving the data acquisition request sent by the network device, the terminal reports the collected data to the network device.

[0724] Note 7: The data acquisition request in step 5 can be based on the network device's own request, that is, no terminal reports the information that can be obtained to the network device.

[0725] Note 8: In step 6, the terminal can send the collected data to the network device autonomously, i.e., it does not receive a data acquisition request from the network device.

[0726] Note 9: When the network device is an access network device, the terminal can send the first training data to the access network device via RRC signaling. When the network device is a core network device, the terminal can send the first training data to the core network device via NAS signaling, or via the UP channel. When the network device is an OAM, the terminal can send the first training data to the OAM via the access network, or via the UP channel from the core network to the OAM.

[0727] Step 7 (optional): The network device sends the collected data to a third-party server for training the event prediction AI / ML model.

[0728] Note 10: Alternatively, step 7 may be omitted, i.e., the AI / ML model for event prediction may be trained directly by the network device.

[0729] Note 11: Applicable to terminal-side model training data collection.

[0730] Example 4:

[0731] In Example 4, the target use case is the RLF prediction use case.

[0732] Step 1: Optionally, the terminal sends a configuration request to the network device, requesting the network device to configure the data collection for RLF prediction. This configuration request can be triggered by a third-party server, such as an OTT server triggering the terminal to send a configuration request to the network device, or it can be triggered autonomously by the terminal. The content of the configuration request can be generated by the terminal, a third-party server, or predefined. The reason for the request may be a lack of relevant models, or the model may need to be updated. The content of the configuration request includes one or more of the following:

[0733] - A request instruction message, which instructs a network device to configure the collection and / or recording of first training data, wherein the first training data is used to train a target model, and the target model is used to implement RLF prediction use cases. In optional embodiments of this application, the request instruction message may also indicate use case number and / or sub-use case number, use case name and / or sub-use case name, function number, function name, the purpose of the requested configuration (obtaining training data), etc., and this application embodiment does not specifically limit this.

[0734] - RLM configuration information. For example, the RLM configuration information may include the configuration information of T310, N310, N311, etc., and this application embodiment does not specifically limit this.

[0735] - Purpose indication information, which indicates the purpose of training and / or updating the target model. In optional embodiments of this application, the purpose of training and / or updating the target model may be, for example, to improve switching performance or reduce measurements.

[0736] - Reason indication information, which indicates the reason for training and / or updating the target model. In an optional embodiment of this application, the reason for training and / or updating the target model may be, for example, the absence of a target model (i.e., the need to train a new target model) or the need to update the target model.

[0737] - Third model type information, which indicates whether the target model is a model that directly performs RLF prediction or a model that indirectly performs RLF prediction.

[0738] - Model inputs and / or model outputs, such as requests to measure and / or record the serving beam of the terminal; may include serving cell identifiers and / or a list of serving cells to request measurement and / or recording of one or more serving cells included in the serving cell identifiers and / or the list of serving cells; may include serving beam identifiers and / or a list of serving beams to request measurement and / or recording of one or more serving beams included in the serving beam identifiers and / or the list of serving beams, wherein, in the case of multiple serving beams, the multiple serving beams may or may not span across cells.

[0739] - Model window length, which includes the observation window (OW) length and / or prediction window (PW) length (e.g., OW is 5 instances, PW is 5 instances).

[0740] - Additional conditions on the terminal side, such as the state of the terminal, for example, the speed of the terminal, the location information of the terminal.

[0741] - Information about third-party servers, which may include, for example, the server's ID, tracking number, identifier, and tracking reference.

[0742] - Additional conditions on the network side, such as those represented by associated ID.

[0743] Note 1: Before sending a configuration request to the network device, the terminal may also receive an indication message (allow request indication message) from the network device indicating that the terminal can send this configuration request.

[0744] Note 2: In scenarios such as handover / reconstruction, this configuration request can be transmitted in an interface such as Xn so that the newly accessed cell of the terminal knows whether and / or how to configure the terminal to collect data.

[0745] Note 3: Whether a network device accepts or rejects this configuration request depends on the network implementation.

[0746] Note 4: When the network device is an access network device, the terminal can send the configuration request to the access network device via RRC signaling. When the network device is a core network device, the terminal can send the configuration request to the core network device via NAS signaling, or via the UP channel. When the network device is an OAM, the terminal can send the configuration request to the OAM via the access network, or via the UP channel from the core network to the OAM.

[0747] Step 2: After receiving the configuration request from the terminal, or triggered autonomously by the network device, the network device sends data collection configuration to the terminal. This data collection configuration may include one or more of the following: measurement configuration, recording configuration, and reporting configuration.

[0748] The measurement configuration may include one or more of the following:

[0749] a) SetB / SetA related information; where SetB is the model input data and SetA is the label (i.e., the model output data).

[0750] - Cell identifier. This cell identifier is used to instruct the terminal to perform measurements and / or recordings for the cell corresponding to the cell identifier. In an optional embodiment of this application, the cell corresponding to the cell identifier may, for example, include the terminal's serving cell.

[0751] - Beam identifier. The beam identifier is used to instruct the terminal to measure and / or record the beam corresponding to the beam identifier. In an optional embodiment of this application, the beam corresponding to the beam identifier may, for example, include the terminal's serving beam.

[0752] - Service beam indication information. This service beam indication information is used to indicate whether the beam with the best signal quality should be used as the service beam for measurement and / or recording.

[0753] - Reference signal configuration. This reference signal configuration is used to configure the reference signal that needs to be measured and / or recorded. For example, the reference signal may include SSB and / or CSI-RS, etc., which will not be specifically described in the embodiments of this application.

[0754] - Fourth Measurement Object Indication Information. In an optional embodiment of this application, the fourth measurement object indication information corresponds to a measurement object, and is used to indicate whether to measure and / or record the corresponding measurement object. In an optional embodiment of this application, the measurement configuration may include one or more fourth measurement object indication information, wherein the multiple fourth measurement object indication information correspond to different measurement objects.

[0755] b) Purpose configuration: Purpose configuration is used to configure the purpose of measurement and / or recording, for example, the purpose can be RLF use case data collection.

[0756] b) RLM configuration. For example, the RLM configuration may include T310 configuration, N310 configuration, N311 configuration, etc., and this application embodiment does not specifically limit this.

[0757] The recording configuration may include: start recording conditions and / or stop recording conditions.

[0758] a) The conditions for starting recording may include one or more of the following:

[0759] - The target timer is enabled, which may include the T310 timer.

[0760] - The target value associated with step loss reaches a maximum number, where the target value associated with step loss may include N310.

[0761] - A loss of synchronization was detected.

[0762] - RLF detected.

[0763] - The signal quality may be higher or lower than a signal quality threshold. In optional embodiments of this application, recording may begin when the signal quality is higher or lower than the signal quality threshold, or measurement data may be recorded when the signal quality is higher or lower than the signal quality threshold. In optional embodiments of this application, the signal quality may be the signal quality of the serving cell, the signal quality of neighboring cells, the signal quality of neighboring cell beams, the signal quality of the serving cell beam, etc., and this application does not specifically limit this. In optional embodiments of this application, the signal quality may include RSRP, RSRQ, SINR, etc., and this application does not specifically limit this. In optional embodiments of this application, the signal quality threshold may be configured by the network side.

[0764] - The number of times a step failure is detected reaches a preset threshold.

[0765] - Third recording period. The third recording period is a preset duration before and / or after the target condition is met. The target condition includes at least one of the following: the target timer is turned on, which may include the T310 timer; a step loss is detected; or an RLF is detected. In other words, data within the third recording period can be recorded. The third recording period can be an overall time window.

[0766] - Beam failure detected. In an optional embodiment of this application, the beam failure refers to the failure of the serving beam, wherein the serving beam can be the beam with the highest signal quality, and the signal quality can include RSRP, RSRQ, SINR, etc., which are not specifically limited in this embodiment of the application.

[0767] - An event was detected that triggered a reporting. In an optional embodiment of this application, the event-triggered reporting refers to a specific event triggering a measurement reporting. For example, the specific event may be configured on the network side and may include events such as A3 and A5.

[0768] - A reconfiguration or access to a new cell occurs. A reconfiguration can refer to a change in data collection configuration, while access to a new cell can include mechanisms such as cell handover, cell selection, or cell reselection.

[0769] - Changes to additional conditions on the network side, such as changes to the associated ID.

[0770] - Changes in additional conditions on the terminal side, such as changes in the terminal's speed or its geographical location.

[0771] b) The conditions for stopping recording may include one or more of the following:

[0772] - Synchronization detected.

[0773] - The number of times synchronization was detected reached the preset number.

[0774] - The target timer stops. In an optional embodiment of this application, the target timer may include a T310 timer.

[0775] - The target value related to synchronization reaches its maximum value. In an optional embodiment of this application, the target value related to synchronization may include N311.

[0776] - RLF detected.

[0777] - The duration of a single recording reaches the preset duration.

[0778] - The total recording time has reached the preset time.

[0779] - The number of records has reached the preset number.

[0780] - The signal quality may be higher or lower than a signal quality threshold. In optional embodiments of this application, recording may stop when the signal quality is higher or lower than the signal quality threshold, or measurement data for signal quality higher or lower than the signal quality threshold may not be recorded. In optional embodiments of this application, the signal quality may be the signal quality of the serving cell, the signal quality of neighboring cells, the signal quality of neighboring cell beams, the signal quality of the serving cell beam, etc., and this application does not specifically limit these. In optional embodiments of this application, the signal quality may include RSRP, RSRQ, SINR, etc., and this application does not specifically limit these. In optional embodiments of this application, the signal quality threshold may be configured by the network side.

[0781] - Changes to additional conditions on the network side, such as changes to the associated ID.

[0782] - Changes in additional conditions on the terminal side, such as changes in the terminal's speed or its geographical location.

[0783] - The data volume has reached the preset data volume.

[0784] - A reconfiguration or access to a new cell occurs. A reconfiguration can refer to a change in data collection configuration, while access to a new cell can include mechanisms such as cell handover, cell selection, or cell reselection.

[0785] - Enter the idle state or the inactive state.

[0786] - Battery depleted.

[0787] - Insufficient memory.

[0788] The reported configuration may include one or more of the following:

[0789] - Information about third-party servers.

[0790] - Information about entities processed by AI data processing in the core network.

[0791] - Information about entities in OAM AI data processing.

[0792] - Reporting measurement configuration: This configuration is used to configure what types of measurements to record and / or report. For example, the reporting measurement configuration can be configured to report measurements such as RSRP, RSRQ, and SINR.

[0793] - Observation window length and / or prediction window length.

[0794] -N future instances and / or N future slots and / or N future samples, where N is a value greater than or equal to 1.

[0795] - Second window length, where the second window length is the window length for determining the probability of RLF occurrence, which can be expressed in English as "an window for determining the RLF probability within the window". The second window length refers to the length of the window for detecting the probability of RLF occurrence. In other words, in an optional embodiment of this application, the terminal can detect whether RLF has occurred or the probability of RLF occurrence within the second window length, and report data (such as model output data) to the network device based on the detection results.

[0796] - Instructions for recording cell-level or beam-level measurement results.

[0797] - One or more network-side additional conditions, which may include, for example, the network environment in which the terminal is located, such as an associated ID.

[0798] - One or more terminal-side additional conditions, which may include, for example, the state of the terminal, such as the speed of the terminal, the location information of the terminal, and the terminal-side additional conditions may be, for example, an identifier.

[0799] - Reporting method indication information. This information indicates whether model input and output data should be recorded separately, or whether they should be recorded uniformly. It should be noted that recording model input and output data uniformly means that model input and output data are recorded and / or reported together without distinguishing between them. Recording model input and output data separately means that model input and output data need to be distinguished during recording and / or reporting. For example, model input and output data can be reported using different signaling, or the information can indicate which part of the reported data is model input and which part is model output.

[0800] Note 5: This data collection configuration can be generated and configured by gNB, CN, or OAM. When the network device is an access network device, the data collection configuration can be carried in RRC signaling and sent to the terminal via RRC signaling. When the network device is a core network device, the data collection configuration can be carried in NAS signaling and sent to the terminal via NAS signaling. Alternatively, the data collection configuration can be sent to the terminal via the UP channel. When the network device is an OAM, the data collection configuration can be sent to the terminal by the OAM through the access network. Alternatively, the data collection configuration can be sent to the terminal by the OAM through the core network via the UP channel.

[0801] Note 6: This data collection configuration can be based on the current RadioLinkMonitoringConfig configuration framework.

[0802] Step 3: The terminal collects and / or records data based on the data collection configuration; the content recorded by the terminal (i.e., the first training data) may include one or more of the following:

[0803] - Information about third-party servers.

[0804] - Information about entities processed by AI data processing in the core network.

[0805] - Information about entities in OAM AI data processing.

[0806] - Cell identification information and / or beam identification information, cell measurement result information and / or beam measurement result information. In optional embodiments of this application, the cell measurement result information may include, for example, the cell's RSRP measurement result, RSRQ measurement result, SINR measurement result, etc., and the beam measurement result information may include, for example, the beam's RSRP measurement result, RSRQ measurement result, SINR measurement result, etc. The measurement result information may include L1 filtered measurement results, or L3 filtered measurement results, or unfiltered underlying measurement results. In optional embodiments of this application, the cell measurement result information and / or beam measurement result information may, for example, include model input data and / or model output data. As mentioned above, in optional embodiments of this application, model input data and model output data may be reported separately or uniformly. It should be noted that uniformly reporting model input data and model output data means that when reporting data, model input data and model output data are not distinguished, but are mixed together for reporting. Recording model input data and model output data separately means that when recording and / or reporting data, it is necessary to distinguish between model input data and model output data. For example, model input data and model output data can be reported through different signaling, or it can be indicated which part of the reported data is model input data and which part is model output data.

[0807] - Observation window length and / or prediction window length.

[0808] -N future instances and / or N future time slots and / or N future samples, where N is a value greater than or equal to 1.

[0809] -Second window length.

[0810] - Additional conditions information on the network side, which may include, for example, the network environment in which the terminal is located, such as the associated ID.

[0811] - Additional condition information on the terminal side, such as the terminal's state, for example, the terminal's speed or location information.

[0812] - The occurrence time of the RLF. In an optional embodiment of this application, this time information can be an absolute time or a relative time.

[0813] -RLF occurrence indication information. This RLF occurrence indication information indicates whether an RLF has occurred. In an optional embodiment of this application, the RLF occurrence indication information may indicate whether an RLF has occurred within the second window length.

[0814] - Time information for detecting a loss of synchronization. In an optional embodiment of this application, the time information for detecting a loss of synchronization may refer to the time when a loss of synchronization is first detected, and this time information may be an absolute time or a relative time.

[0815] - Time information indicating when the number of detected step loss reaches a preset number. In an optional embodiment of this application, the time information indicating when the number of detected step loss reaches the preset number can refer to the time when T310 is turned on. This time information can be an absolute time or a relative time.

[0816] - Synchronization detection time information. In an optional embodiment of this application, the synchronization detection time information may refer to the time when synchronization was first detected, wherein the first detection of synchronization refers to the first detection of synchronization after a loss of synchronization occurs, or the first detection of synchronization after T310 has been enabled. In an optional embodiment of this application, the time information may be an absolute time or a relative time.

[0817] - Time information indicating when the number of synchronizations detected reaches a preset number. In an optional embodiment of this application, the time information indicating when the number of synchronizations detected reaches the preset number can refer to the time when T310 stops. In an optional embodiment of this application, this time information can be an absolute time or a relative time.

[0818] -RLM configuration. In optional embodiments of this application, the RLM configuration may include, for example, T310 configuration, N310 configuration, N311 configuration, etc., and this application embodiment does not specifically limit this.

[0819] -Indication information for the target model, whether it is a model that directly performs RLF prediction or a model that indirectly performs RLF prediction.

[0820] -Second window length.

[0821] Step 4: Optionally, the terminal sends an indication message to the network device.

[0822] Step 5: After receiving the available indication information sent by the terminal, the network device sends a data acquisition request to the terminal, requesting the terminal to report the collected data.

[0823] Step 6: After receiving the data acquisition request sent by the network device, the terminal reports the collected data to the network device.

[0824] Note 7: The data acquisition request in step 5 can be based on the network device's own request, that is, no terminal reports the information that can be obtained to the network device.

[0825] Note 8: In step 6, the terminal can send the collected data to the network device autonomously, i.e., it does not receive a data acquisition request from the network device.

[0826] Note 9: When the network device is an access network device, the terminal can send the first training data to the access network device via RRC signaling. When the network device is a core network device, the terminal can send the first training data to the core network device via NAS signaling, or via the UP channel. When the network device is an OAM, the terminal can send the first training data to the OAM via the access network, or via the UP channel from the core network to the OAM.

[0827] Step 7 (optional): The network device sends the collected data to a third-party server for RLF prediction AI / ML model training.

[0828] Note 10: Alternatively, step 7 may be omitted, i.e., the AI / ML model for RLF prediction may be trained directly by the network device.

[0829] Note 11: Applicable to terminal-side model training data collection.

[0830] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0831] Based on the same inventive concept, this application also provides a data collection apparatus for implementing the data collection method described above. The solution provided by this apparatus is similar to the implementation scheme described in the above method; therefore, the specific limitations in one or more data collection apparatus embodiments provided below can be found in the limitations of the data collection method described above, and will not be repeated here.

[0832] In an exemplary embodiment, as shown in FIG7, a data collection device 700 is provided for a terminal, including: a receiving unit 701, a collecting unit 702 and a sending unit 703.

[0833] The receiving unit 701 is used to receive data collection configuration sent by the network device.

[0834] Collection unit 702 is used to collect and / or record first training data according to data collection configuration.

[0835] The sending unit 703 is used to send first training data to the network device; wherein the first training data is used to train and / or update the target model, the target model is used to implement the target use case, and the target use case includes at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case.

[0836] In an optional embodiment of this application, the sending unit 703 is further configured to send a configuration request to the network device, the configuration request being used to request the network device to send data collection configuration.

[0837] In an optional embodiment of this application, when the target use case includes an RRM prediction use case, the configuration request information includes at least one of the following: first model type information, which indicates whether the target model is a single-cell model or a cell cluster model; basic request information;

[0838] The basic request information includes at least one of the following: request indication information, which instructs the requesting network device to configure the collection and / or recording of the first training data; purpose indication information, which instructs the purpose of training and / or updating the target model; reason indication information, which instructs the reason for training and / or updating the target model; model input and / or model output; model window length, which includes the observation window length and / or prediction window length; additional conditions on the network side; additional conditions on the terminal side; and information about a third-party server, which is used by the third-party server to train and / or update the target model.

[0839] When the target use case includes an event prediction use case, the configuration request includes at least one of the following: basic request information; first model type information; second model type information, which indicates that the target model is a model that directly performs event prediction, or indicates that the target model is a model that indirectly performs event prediction; requested event type information; and requested event configuration parameters.

[0840] When the target use case includes an RLF prediction use case, the configuration request includes at least one of the following: basic request information; third model type information, which indicates that the target model is a model that directly performs RLF prediction, or indicates that the target model is a model that indirectly performs RLF prediction; and RLM configuration information.

[0841] In optional embodiments of this application, the data collection configuration includes at least one of the following configurations: measurement configuration; recording configuration; and reporting configuration.

[0842] In an optional embodiment of this application, when the target use case includes an RRM prediction use case, the measurement configuration includes: a basic measurement configuration; wherein the basic measurement configuration includes at least one of the following configurations: a cell list, which includes at least one cell, used to instruct the terminal to measure and / or record cells in the cell list; a beam list, which includes at least one beam, used to instruct the terminal to measure and / or record beams in the beam list; a frequency list, which includes at least one frequency, used to instruct the terminal to measure and / or record frequencies in the frequency list; a first frequency indication information, used to indicate whether to measure and / or record cells on this frequency; a first measurement object indication information, used to indicate whether to measure and / or record all measurement objects; a second measurement object indication information, used to indicate whether to measure and / or record this measurement object; model-related parameters, used to implicitly indicate the measurement objects that need to be measured and / or recorded; a target configuration; and a reference signal configuration.

[0843] When the target use case includes an event prediction use case, the measurement configuration includes at least one of the following configurations: a basic measurement configuration; an event configuration; a reporting configuration indication information, which indicates whether to measure and / or record the event and / or measurement object corresponding to this reporting configuration; and a third measurement object indication information, which indicates whether to measure and / or record all measurement objects that satisfy the event.

[0844] When the target use case includes an RLF prediction use case, the measurement configuration includes at least one of the following configurations: reference signal configuration; RLM configuration; destination configuration; cell identifier, which instructs the terminal to perform measurement and / or recording for the cell corresponding to the cell identifier; beam identifier, which instructs the terminal to perform measurement and / or recording for the beam corresponding to the beam identifier; serving beam indication information, which indicates whether to use the beam with the best signal quality as the serving beam for measurement and / or recording; and fourth measurement object indication information, which indicates whether to perform measurement and / or recording for this measurement object.

[0845] In an optional embodiment of this application, the recording configuration includes at least one of the following configurations: start recording conditions; stop recording conditions.

[0846] In an optional embodiment of this application, when the target use case includes an RRM prediction use case, the start recording conditions include: basic start conditions; wherein, the basic start conditions include at least one of the following: signal quality is higher or lower than a signal quality threshold; reconfiguration or access to a new cell occurs; additional conditions on the network side change; additional conditions on the terminal side change;

[0847] When the target use case includes an event prediction use case, the start recording condition includes at least one of the following: a basic start condition; an event entry condition is met; a first trigger timer (TTT) corresponding to the event entry condition is started; the event is met; the first TTT times out; the difference between the event entry condition and the event entry condition reaches a preset threshold; a first recording period, which is a preset duration before the event entry condition is met and / or a preset duration after the event entry condition is met; and a second recording period, which is a preset duration before the event is met and / or a preset duration after the event is met.

[0848] When the target use case includes an RLF prediction use case, the start recording conditions include at least one of the following: basic start conditions; target timer is turned on; target value related to out-of-step reaches its maximum number; out-of-step is detected; RLF is detected; the number of out-of-step detections reaches a preset threshold; third recording period, which is a preset duration before and / or after the target conditions are met. The target conditions include at least one of the following: target timer is turned on, out-of-step is detected, RLF is detected; beam failure is detected; event trigger reporting is detected.

[0849] In an optional embodiment of this application, when the target use case includes an RRM prediction use case, the stop recording condition includes: a basic stop condition; wherein the basic stop condition includes at least one of the following: the duration of a single recording reaches a preset duration; the total recording duration reaches a preset duration; the number of recordings reaches a preset number; the signal quality is higher or lower than a signal quality threshold; the additional conditions on the network side change; the additional conditions on the terminal side change; the data volume reaches a preset data volume; a reconfiguration occurs or a new cell is accessed; the user enters an idle state or an inactive state; the battery is depleted; or the memory is insufficient.

[0850] When the target use case includes an event prediction use case, the stop recording condition includes at least one of the following: basic stop condition; event met; first TTT timeout corresponding to the event entry condition; event exit condition met; second TTT started corresponding to the event exit condition; second TTT timeout corresponding to the event exit condition; the difference between the event exit condition and the event exit condition reaches a preset threshold.

[0851] When the target use case includes an RLF prediction use case, the stop recording condition includes at least one of the following: basic stop condition; synchronization is detected; the number of times synchronization is detected reaches a preset number; the target timer stops; the target value related to synchronization reaches a maximum number; RLF is detected.

[0852] In optional embodiments of this application, when the target use case includes an RRM prediction use case, the reporting configuration includes at least one of the following configurations: indication information for reporting data of the top K cells with high signal quality, where K is a value greater than or equal to 1; indication information for reporting data of the top K beams with high signal quality, where K is a value greater than or equal to 1; indication information for reporting frequency point information; indication information for whether to record data by frequency point; basic reporting configuration; the basic reporting configuration includes at least one of the following configurations: information of a third-party server; information of an AI data processing entity in the core network; information of an AI data processing entity in OAM; measurement quantity reporting configuration; indication information for recording cell-level or beam-level measurement results; one or more network-side additional conditions; one or more terminal-side additional conditions; observation window length and / or prediction window length; N future instances and / or N future time slots and / or N future samples, where N is a value greater than or equal to 1; reporting method indication information, which is used to indicate that the model input data and model output data in the first training data are recorded separately, or to indicate that the model input data and model output data are recorded uniformly;

[0853] When the target use case includes an event prediction use case, the reporting configuration includes at least one of the following configurations: basic reporting configuration; first window length, where the first window length is the window length of the probability of the event occurring; event type configuration;

[0854] When the target use case includes an RLF prediction use case, the reporting configuration includes at least one of the following configurations: basic reporting configuration; second window length, which is the window length of the probability of RLF occurrence.

[0855] In an optional embodiment of this application, when the target use case includes an RRM prediction use case, the first training data includes at least one of the following: identification information and / or measurement result information of the top K cells with high signal quality, where K is a value greater than or equal to 1; identification information and / or measurement result information of the top K beams with high signal quality, where K is a value greater than or equal to 1; frequency point information; basic data; the basic data includes at least one of the following: information of a third-party server; information of the AI ​​data processing entity in the core network; information of the AI ​​data processing entity in OAM; cell identification information and / or beam identification information; cell measurement result information and / or beam measurement result information; observation window length and / or prediction window length; N future instances and / or N future time slots and / or N future samples, where N is a value greater than or equal to 1; additional condition information on the network side; additional condition information on the terminal side.

[0856] When the target use case includes an event prediction use case, the first training data includes at least one of the following: basic data; event configuration information; event occurrence time; event occurrence indication information, which indicates whether the corresponding event has occurred; the occurrence time when the event departure condition is met; the occurrence time when the event entry condition is met; indication information that the target model is a model that directly performs event prediction or a model that indirectly performs event prediction; unfiltered data obtained from each measurement during the first TTT and / or second TTT activation process, where the first TTT corresponds to the event entry condition and the second TTT corresponds to the event departure condition.

[0857] When the target use case includes an RLF prediction use case, the first training data includes at least one of the following: basic data; the occurrence time of RLF; RLF occurrence indication information, which indicates whether an RLF has occurred; the time information of detecting out-of-step; the time information of the number of out-of-step detections reaching a preset number; the time information of detecting synchronization; the time information of the number of synchronization detections reaching a preset number; RLM configuration; and indication information that the target model is a model that directly performs RLF prediction or a model that indirectly performs RLF prediction.

[0858] In an optional embodiment of this application, the receiving unit 701 is further configured to receive a data acquisition request sent by the network device; and the sending unit 703 is configured to send first training data to the network device in response to the data acquisition request.

[0859] In an optional embodiment of this application, the sending unit 703 is further configured to send an availability indication message to the network device, the availability indication message being used to indicate that the terminal stores the first training data.

[0860] In an exemplary embodiment, as shown in FIG8, a data collection device 800 is provided for a network device, including: a transmitting unit 801 and a receiving unit 802.

[0861] The sending unit 801 is used to send data collection configuration to the terminal.

[0862] The receiving unit 802 is used to receive first training data sent by the terminal; wherein the first training data is collected and / or recorded by the terminal based on the data collection configuration, and the first training data is used to train and / or update the target model, and the target model is used to implement the target use case, the target use case including at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case.

[0863] In an optional embodiment of this application, the receiving unit 802 is further configured to receive a configuration request sent by the terminal, the configuration request being used to request the network device to send data collection configuration.

[0864] In an optional embodiment of this application, the sending unit 801 is used to send a data acquisition request to the terminal, wherein the data acquisition request is used to request the terminal to send first training data.

[0865] In an optional embodiment of this application, the receiving unit 802 is further configured to receive available indication information sent by the terminal, wherein the available indication information is used to indicate that the terminal stores the first training data.

[0866] Specific limitations regarding the data collection device can be found in the limitations regarding the data collection method described above, and will not be repeated here. Each unit in the aforementioned data collection device can be implemented entirely or partially through software, hardware, or a combination thereof. These units can be embedded in hardware or independently of the processor in the communication device, or stored in software in the memory of the communication device, so that the processor can call and execute the corresponding operations of each unit.

[0867] It should be noted that the division of units in the embodiments of this application is illustrative and only represents one logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units described above can be implemented in hardware or as software functional units.

[0868] If the integrated unit is implemented as 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 solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application.

[0869] It should be noted that the apparatus provided in this application embodiment can implement all the method steps implemented in the above method embodiment and can achieve the same technical effect. Here, the parts that are the same as those in the method embodiment and the beneficial effects will not be described in detail.

[0870] Figure 9 is a schematic diagram of the structure of a communication device provided in an embodiment of this application. This communication device can be a terminal or a network device. The communication device 900 shown in Figure 9 includes at least one processor 901 and a memory 902. The various components in the communication device 900 are coupled together through a bus system 903. It is understood that the bus system 903 is used to realize the connection and communication between these components. In addition to a data bus, the bus system 903 also includes a power bus, a control bus, and a status signal bus. However, for clarity, all buses are labeled as bus system 903 in Figure 9. Furthermore, this embodiment of the application also includes a transceiver 904. The transceiver can be multiple elements, including a transmitter and a receiver, providing a unit for communicating with various other devices over a transmission medium.

[0871] It is understood that the memory 902 in the embodiments of this application can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DRRAM). The memory 902 of the systems and methods described in the embodiments of this application is intended to include, but is not limited to, these and any other suitable types of memory.

[0872] In some implementations, memory 902 stores elements such as executable modules or data structures, or subsets thereof, or extended sets thereof: an operating system. The operating system 9021 includes various system programs, such as a framework layer, a core library layer, and a driver layer, used to implement various basic business functions and handle hardware-based tasks.

[0873] In this embodiment of the application, when the communication device 900 is a terminal, the processor 901 performs the following operations by calling the program or instructions stored in the memory 902:

[0874] The transceiver 904 is controlled to receive data collection configuration sent by the network device; collect and / or record first training data according to the data collection configuration; and send the first training data to the network device. The first training data is used to train and / or update the target model, and the target model is used to implement the target use case. The target use case includes at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case.

[0875] In some embodiments, the processor 901 is further configured to perform the following operations: control the transceiver 904 to send a configuration request to the network device, the configuration request being used to request the network device to send data collection configuration.

[0876] In some embodiments, when the target use case includes an RRM prediction use case, the configuration request information includes at least one of the following: first model type information, which indicates whether the target model is a single-cell model or a cell cluster model; basic request information; wherein the basic request information includes at least one of the following: request indication information, which indicates that the network device is requested to configure the collection and / or recording of the first training data; purpose indication information, which indicates the purpose of training and / or updating the target model; reason indication information, which indicates the reason for training and / or updating the target model; model input and / or model output; model window length, which includes the observation window length and / or prediction window length; additional conditions on the network side; additional conditions on the terminal side; information of a third-party server, which is used to train and / or update the target model;

[0877] When the target use case includes an event prediction use case, the configuration request includes at least one of the following: basic request information; first model type information; second model type information, which indicates that the target model is a model that directly performs event prediction, or indicates that the target model is a model that indirectly performs event prediction; requested event type information; and requested event configuration parameters.

[0878] When the target use case includes an RLF prediction use case, the configuration request includes at least one of the following: basic request information; third model type information, which indicates that the target model is a model that directly performs RLF prediction, or indicates that the target model is a model that indirectly performs RLF prediction; and RLM configuration information.

[0879] In some embodiments, the data collection configuration includes at least one of the following configurations: measurement configuration; recording configuration; reporting configuration.

[0880] In some embodiments, where the target use case includes an RRM prediction use case, the measurement configuration includes: a basic measurement configuration; wherein the basic measurement configuration includes at least one of the following configurations: a cell list, the cell list including at least one cell, the cell list being used to instruct the terminal to measure and / or record cells in the cell list; a beam list, the beam list including at least one beam, the beam list being used to instruct the terminal to measure and / or record beams in the beam list; a frequency list, the frequency list including at least one frequency, the frequency list being used to instruct the terminal to measure and / or record frequencies in the frequency list; first frequency indication information, the first frequency indication information being used to indicate whether to measure and / or record cells on this frequency; first measurement object indication information, the first measurement object indication information being used to indicate whether to measure and / or record all measurement objects; second measurement object indication information, the second measurement object indication information being used to indicate whether to measure and / or record this measurement object; model-related parameters, the model-related parameters being used to implicitly indicate the measurement objects that need to be measured and / or recorded; a target configuration; and a reference signal configuration.

[0881] When the target use case includes an event prediction use case, the measurement configuration includes at least one of the following configurations: a basic measurement configuration; an event configuration; a reporting configuration indication information, which indicates whether to measure and / or record the event and / or measurement object corresponding to this reporting configuration; and a third measurement object indication information, which indicates whether to measure and / or record all measurement objects that satisfy the event.

[0882] When the target use case includes an RLF prediction use case, the measurement configuration includes at least one of the following configurations: reference signal configuration; RLM configuration; destination configuration; cell identifier, which instructs the terminal to perform measurement and / or recording for the cell corresponding to the cell identifier; beam identifier, which instructs the terminal to perform measurement and / or recording for the beam corresponding to the beam identifier; serving beam indication information, which indicates whether to use the beam with the best signal quality as the serving beam for measurement and / or recording; and fourth measurement object indication information, which indicates whether to perform measurement and / or recording for this measurement object.

[0883] In some embodiments, the recording configuration includes at least one of the following configurations: start recording conditions; stop recording conditions.

[0884] In some embodiments, when the target use case includes an RRM prediction use case, the start recording conditions include: basic start conditions; wherein the basic start conditions include at least one of the following: signal quality is higher or lower than a signal quality threshold; reconfiguration or access to a new cell occurs; additional conditions on the network side change; additional conditions on the terminal side change;

[0885] When the target use case includes an event prediction use case, the start recording condition includes at least one of the following: a basic start condition; an event entry condition is met; a first trigger timer (TTT) corresponding to the event entry condition is started; the event is met; the first TTT times out; the difference between the event entry condition and the event entry condition reaches a preset threshold; a first recording period, which is a preset duration before the event entry condition is met and / or a preset duration after the event entry condition is met; and a second recording period, which is a preset duration before the event is met and / or a preset duration after the event is met.

[0886] When the target use case includes an RLF prediction use case, the start recording conditions include at least one of the following: basic start conditions; target timer is turned on; target value related to out-of-step reaches its maximum number; out-of-step is detected; RLF is detected; the number of out-of-step detections reaches a preset threshold; third recording period, which is a preset duration before and / or after the target conditions are met. The target conditions include at least one of the following: target timer is turned on, out-of-step is detected, RLF is detected; beam failure is detected; event trigger reporting is detected.

[0887] In some embodiments, when the target use case includes an RRM prediction use case, the stop recording condition includes: a basic stop condition; wherein the basic stop condition includes at least one of the following: the duration of a single recording reaches a preset duration; the total recording duration reaches a preset duration; the number of recordings reaches a preset number; the signal quality is higher or lower than a signal quality threshold; the additional conditions on the network side change; the additional conditions on the terminal side change; the data volume reaches a preset data volume; a reconfiguration occurs or a new cell is accessed; the user enters an idle state or an inactive state; the battery is depleted; or the memory is insufficient.

[0888] When the target use case includes an event prediction use case, the stop recording condition includes at least one of the following: basic stop condition; event met; first TTT timeout corresponding to the event entry condition; event exit condition met; second TTT started corresponding to the event exit condition; second TTT timeout corresponding to the event exit condition; the difference between the event exit condition and the event exit condition reaches a preset threshold.

[0889] When the target use case includes an RLF prediction use case, the stop recording condition includes at least one of the following: basic stop condition; synchronization is detected; the number of times synchronization is detected reaches a preset number; the target timer stops; the target value related to synchronization reaches a maximum number; RLF is detected.

[0890] In some embodiments, when the target use case includes an RRM prediction use case, the reporting configuration includes at least one of the following configurations: indication information for reporting data of the top K cells with high signal quality, where K is a value greater than or equal to 1; indication information for reporting data of the top K beams with high signal quality, where K is a value greater than or equal to 1; indication information for reporting frequency point information; indication information for whether to record data by frequency point; basic reporting configuration; the basic reporting configuration includes at least one of the following configurations: information of a third-party server; information of an AI data processing entity in the core network; information of an AI data processing entity in OAM; measurement quantity reporting configuration; indication information for recording cell-level or beam-level measurement results; one or more network-side additional conditions; one or more terminal-side additional conditions; observation window length and / or prediction window length; N future instances and / or N future time slots and / or N future samples, where N is a value greater than or equal to 1; reporting method indication information, which is used to indicate that the model input data and model output data in the first training data are recorded separately, or to indicate that the model input data and model output data are recorded uniformly;

[0891] When the target use case includes an event prediction use case, the reporting configuration includes at least one of the following configurations: basic reporting configuration; first window length, where the first window length is the window length of the probability of the event occurring; event type configuration;

[0892] When the target use case includes an RLF prediction use case, the reporting configuration includes at least one of the following configurations: basic reporting configuration; second window length, which is the window length of the probability of RLF occurrence.

[0893] In some embodiments, where the target use case includes an RRM prediction use case, the first training data includes at least one of the following: identification information and / or measurement result information of the top K cells with high signal quality, where K is a value greater than or equal to 1; identification information and / or measurement result information of the top K beams with high signal quality, where K is a value greater than or equal to 1; frequency point information; basic data; the basic data includes at least one of the following: information of a third-party server; information of the AI ​​data processing entity in the core network; information of the AI ​​data processing entity in OAM; cell identification information and / or beam identification information; cell measurement result information and / or beam measurement result information; observation window length and / or prediction window length; N future instances and / or N future time slots and / or N future samples, where N is a value greater than or equal to 1; additional condition information on the network side; additional condition information on the terminal side;

[0894] When the target use case includes an event prediction use case, the first training data includes at least one of the following: basic data; event configuration information; event occurrence time; event occurrence indication information, which indicates whether the corresponding event has occurred; the occurrence time when the event departure condition is met; the occurrence time when the event entry condition is met; indication information that the target model is a model that directly performs event prediction or a model that indirectly performs event prediction; unfiltered data obtained from each measurement during the first TTT and / or second TTT activation process, where the first TTT corresponds to the event entry condition and the second TTT corresponds to the event departure condition.

[0895] When the target use case includes an RLF prediction use case, the first training data includes at least one of the following: basic data; the occurrence time of RLF; RLF occurrence indication information, which indicates whether an RLF has occurred; the time information of detecting out-of-step; the time information of the number of out-of-step detections reaching a preset number; the time information of detecting synchronization; the time information of the number of synchronization detections reaching a preset number; RLM configuration; and indication information that the target model is a model that directly performs RLF prediction or a model that indirectly performs RLF prediction.

[0896] In some embodiments, the processor 901 is further configured to perform the following operations: control the transceiver 904 to receive a data acquisition request sent by the network device; and in response to the data acquisition request, control the transceiver 904 to send first training data to the network device.

[0897] In some embodiments, the processor 901 is further configured to perform the following operations: control the transceiver 904 to send an availability indication message to the network device, the availability indication message being used to indicate that the terminal stores the first training data.

[0898] In this embodiment of the application, when the communication device 900 is a network device, the processor 901 performs the following operations by calling the program or instructions stored in the memory 902:

[0899] The transceiver 904 is controlled to send a data collection configuration to the terminal; the transceiver 904 is controlled to receive the first training data sent by the terminal; wherein the first training data is collected and / or recorded by the terminal based on the data collection configuration, the first training data is used to train and / or update the target model, the target model is used to implement the target use case, and the target use case includes at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case.

[0900] In some embodiments, the processor 901 is further configured to perform the following operations: control the transceiver 904 to receive a configuration request sent by the terminal, the configuration request being used to request the network device to send data collection configuration.

[0901] In some embodiments, the processor 901 is further configured to perform the following operations: control the transceiver 904 to send a data acquisition request to the terminal, wherein the data acquisition request is used to request the terminal to send first training data.

[0902] In some embodiments, the processor 901 is further configured to perform the following operations: control the transceiver 904 to receive available indication information sent by the terminal, wherein the available indication information is used to indicate that the terminal stores first training data.

[0903] As described above, some or all of the methods disclosed in the embodiments of this application can be applied to processor 901, implemented by processor 901, or implemented by processor 901 in conjunction with other components (e.g., transceiver 904). Processor 901 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in processor 901 or by instructions in the form of software. The processor 901 mentioned above may be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied as being executed by a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software module can reside in a mature storage medium in the field, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory 902, and processor 901 reads the information from memory 902 and, in conjunction with its hardware, completes the steps of the above method.

[0904] It is understood that the embodiments described in this application can be implemented using hardware, software, firmware, middleware, microcode, or a combination thereof. For hardware implementation, the processing unit can be implemented in one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions described in this application, or combinations thereof.

[0905] For software implementation, the technology described in the embodiments of this application can be implemented by modules (e.g., procedures, functions, etc.) that perform the functions described in the embodiments of this application. The software code can be stored in memory and executed by processor 901. The memory can be implemented in processor 901 or external to processor 901.

[0906] In some embodiments, a processor-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the steps implemented by the terminal in the above method embodiments.

[0907] In some embodiments, a processor-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps implemented by the network device in the above method embodiments.

[0908] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0909] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0910] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A data collection method, wherein, For a terminal, the method includes: Receive data collection configuration sent by network devices; Collect and / or record the first training data according to the data collection configuration; Send the first training data to the network device; The first training data is used to train and / or update the target model, and the target model is used to implement the target use case, which includes at least one of the following use cases: Radio Resource Management (RRM) prediction use case, event prediction use case, and Radio Link Failure (RLF) prediction use case.

2. The method according to claim 1, wherein, The method further includes: A configuration request is sent to the network device, the configuration request being used to request the network device to send the data collection configuration.

3. The method according to claim 2, wherein, If the target use case includes the RRM prediction use case, the configuration request information includes at least one of the following: First model type information, which indicates whether the target model is a single-cell model or a cell cluster model; Basic request information; The basic request information includes at least one of the following: The request instruction information is used to instruct the network device to configure the collection and / or recording of the first training data; Purpose indication information, which is used to indicate the purpose of training and / or updating the target model; Reason indication information, which is used to indicate the reason for training and / or updating the target model; Model inputs and / or model outputs; Model window length, which includes observation window length and / or prediction window length; Additional conditions on the network side; Additional conditions on the terminal side; Information about a third-party server, which is used to train and / or update the target model; If the target use case includes the event prediction use case, the configuration request includes at least one of the following: The basic request information; The first model type information; Second model type information, which is used to indicate whether the target model is a model that directly predicts events or a model that indirectly predicts events; The event type information requested; The requested event configuration parameters; If the target use case includes the RLF prediction use case, the configuration request includes at least one of the following: The basic request information; The third model type information is used to indicate whether the target model is a model that directly performs RLF prediction or a model that indirectly performs RLF prediction. Configuration information for the Radio Link Monitoring (RLM).

4. The method according to any one of claims 1 to 3, wherein, The data collection configuration includes at least one of the following configurations: Measurement configuration; recording configuration; reporting configuration.

5. The method according to claim 4, wherein, If the target use case includes the RRM prediction use case, the measurement configuration includes: a basic measurement configuration; The basic measurement configuration includes at least one of the following configurations: A cell list, the cell list including at least one cell, the cell list being used to instruct the terminal to perform measurements and / or recordings on the cells in the cell list; A beam list, the beam list including at least one beam, the beam list being used to instruct the terminal to perform measurements and / or recordings for the beams in the beam list; A frequency point list, the frequency point list including at least one frequency point, the frequency point list being used to instruct the terminal to perform measurements and / or recordings for the frequency points in the frequency point list; First frequency point indication information, the first frequency point indication information is used to indicate whether to measure and / or record the cell on this frequency point; First measurement object indication information, which is used to indicate whether all measurement objects are measured and / or recorded; Second measurement object indication information, which is used to indicate whether to measure and / or record this measurement object; Model-related parameters, which implicitly indicate the measurement objects that need to be measured and / or recorded; Target configuration; Reference signal configuration; When the target use case includes the event prediction use case, the measurement configuration includes at least one of the following configurations: The basic measurement configuration; Event configuration; The configuration instruction information is reported, which is used to indicate whether to measure and / or record the event and / or measurement object corresponding to this reported configuration. The third measurement object indication information is used to indicate whether all measurement objects that satisfy the event should be measured and / or recorded. When the target use case includes the RLF prediction use case, the measurement configuration includes at least one of the following configurations: Reference signal configuration; RLM configuration; Target configuration; Cell identifier, the cell identifier being used to instruct the terminal to perform measurements and / or recordings for the cell corresponding to the cell identifier; Beam identifier, the beam identifier being used to instruct the terminal to measure and / or record the beam corresponding to the beam identifier; Service beam indication information, which is used to indicate whether to use the beam with the best signal quality as the service beam for measurement and / or recording; The fourth measurement object indication information is used to indicate whether to measure and / or record this measurement object.

6. The method according to claim 4, wherein, The record configuration includes at least one of the following configurations: Start recording conditions; Stop recording conditions.

7. The method according to claim 6, wherein, When the target use case includes the RRM prediction use case, the start recording conditions include: basic start conditions; The basic starting conditions include at least one of the following: The signal quality is higher or lower than the signal quality threshold; A reconfiguration or connection to a new cell has occurred; Additional conditions on the network side have changed; Additional conditions on the terminal side have changed; When the target use case includes the event prediction use case, the start recording condition includes at least one of the following: The basic start conditions; The event entry conditions are met; Start the first trigger timer (TTT) corresponding to the event entry condition; Event satisfied; The first TTT timed out; The difference between the value and the event entry condition reaches a preset threshold; The first recording period is a preset duration before the event entry condition is met and / or a preset duration after the event entry condition is met. The second recording period is a preset duration before the event is satisfied and / or a preset duration after the event is satisfied. When the target use case includes the RLF prediction use case, the start recording condition includes at least one of the following: The basic start conditions; Target timer started; The target value associated with the loss of synchronization reaches its maximum. A loss of synchronization was detected; RLF detected; The number of times out-of-step was detected reached a preset threshold. The third recording period is a preset duration before and / or after the target condition is met. The target condition includes at least one of the following: the target timer is turned on, a step loss is detected, and an RLF is detected. Beam failure detected; An event was detected and a report was triggered.

8. The method according to claim 6, wherein, When the target use case includes the RRM prediction use case, the stop recording condition includes: basic stop condition; The basic stopping condition includes at least one of the following: The duration of a single recording reaches the preset duration; The total recording time has reached the preset time. The preset number of records has been reached; The signal quality is higher or lower than the signal quality threshold; Additional conditions on the network side have changed; Additional conditions on the terminal side have changed; The data volume has reached the preset data volume; A reconfiguration or connection to a new cell has occurred; Enter the idle or inactive state; Battery depleted; Insufficient memory; When the target use case includes the event prediction use case, the stop recording condition includes at least one of the following: The basic stopping conditions; Event satisfied; The first TTT timeout corresponding to the event entry condition; The event exit condition is met; The second TTT corresponding to the event exit condition is activated; The second TTT timed out; The difference between the value and the event departure condition reaches a preset threshold; When the target use case includes the RLF prediction use case, the stop recording condition includes at least one of the following: The basic stopping conditions; Synchronization detected; The number of synchronizations detected has reached the preset number; The target timer has stopped. The target value related to synchronization reaches its maximum value; RLF detected.

9. The method according to claim 4, wherein, When the target use case includes the RRM prediction use case, the reporting configuration includes at least one of the following configurations: The system reports an indication of the top K cells with the highest signal quality, where K is a value greater than or equal to 1. The system reports the data of the top K beams with the highest signal quality, where K is a value greater than or equal to 1. Instructions for reporting frequency point information; Indication information regarding whether data is recorded according to frequency point; Basically, report the configuration; The basic reporting configuration includes at least one of the following configurations: Information about the third-party server; Information about AI data processing entities in the core network (CN); Manage and maintain information about AI data processing entities in OAM; Report the measurement configuration; Indicator information for recording cell-level or beam-level measurement results; One or more network-side additional conditions; One or more terminal-side additional conditions; Observation window length and / or prediction window length; N future instances and / or N future time slots and / or N future samples, where N is a value greater than or equal to 1; The reporting method indication information is used to indicate that the model input data and model output data in the first training data are recorded separately, or to indicate that the model input data and model output data are recorded uniformly. When the target use case includes the event prediction use case, the reporting configuration includes at least one of the following configurations: The basic reporting configuration; The first window length is the window length at which the event is likely to occur; Event type configuration; When the target use case includes the RLF prediction use case, the reporting configuration includes at least one of the following configurations: The basic reporting configuration; The second window length is the window length at which the probability of RLF occurring.

10. The method according to any one of claims 1 to 3, wherein, When the target use case includes the RRM prediction use case, the first training data includes at least one of the following: Identification information and / or measurement results of the top K cells with high signal quality, where K is a value greater than or equal to 1; The identification information and / or measurement result information of the first K beams with high signal quality, where K is a value greater than or equal to 1; Frequency information; Basic data; The basic data includes at least one of the following: Information about the third-party server; Information about entities processed by AI data in the core network; OAM AI data processing entity information; The identification information of the community and / or the identification information of the beam; Measurement results information for the cell and / or beam measurement results information; Observation window length and / or prediction window length; N future instances and / or N future time slots and / or N future samples, where N is a value greater than or equal to 1; Additional conditional information from the network side; Additional condition information on the terminal side; When the target use case includes the event prediction use case, the first training data includes at least one of the following: The basic data; Event configuration information; Time of the event; Event occurrence indication information, which is used to indicate whether the corresponding event has occurred; The time when the event leaves the condition being met; The time when the event entry condition is met; The target model is an indication of whether it is a model that directly predicts events or a model that indirectly predicts events. Unfiltered data obtained from each measurement during the activation of the first TTT and / or the second TTT, wherein the first TTT corresponds to the event entry condition and the second TTT corresponds to the event exit condition; When the target use case includes the RLF prediction use case, the first training data includes at least one of the following: The basic data; The timing of RLF occurrence; RLF occurrence indication information, which is used to indicate whether an RLF has occurred; Time information indicating a loss of synchronization was detected. Information on the time when the number of timed steps out of step reaches a preset number; Synchronization time information was detected; The time information indicating when the number of synchronizations has reached the preset number; RLM configuration; The target model is an indication of whether it is a model that directly performs RLF prediction or a model that indirectly performs RLF prediction.

11. The method according to any one of claims 1 to 3, wherein, Sending the first training data to the network device includes: Receive the data acquisition request sent by the network device; In response to the data acquisition request, the first training data is sent to the network device.

12. The method according to claim 11, wherein, The method further includes: Send an available indication message to the network device, the available indication message being used to indicate that the terminal stores the first training data.

13. A data collection method, wherein, For use in network devices, the method includes: Send data collection configuration to the terminal; Receive the first training data sent by the terminal; Wherein, the first training data is collected and / or recorded by the terminal based on the data collection configuration, and the first training data is used to train and / or update the target model, the target model is used to implement the target use case, and the target use case includes at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case.

14. The method according to claim 13, wherein, The method further includes: The system receives a configuration request sent by the terminal, the configuration request being used to request the network device to send the data collection configuration.

15. The method according to claim 13, wherein, The method further includes: A data acquisition request is sent to the terminal, wherein the data acquisition request is used to request the terminal to send the first training data.

16. The method according to claim 15, wherein, The method further includes: The terminal receives an available indication message, wherein the available indication message indicates that the terminal stores the first training data.

17. A data collection device, wherein, For a terminal, the device includes: The receiving unit is used to receive data collection configurations sent by network devices; A collection unit is configured to collect and / or record first training data according to the data collection configuration. A sending unit is configured to send the first training data to the network device; The first training data is used to train and / or update the target model, and the target model is used to implement the target use case, which includes at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case.

18. A data collection device, wherein, For a network device, the means includes: The sending unit is used to send data collection configuration to the terminal; The receiving unit is used to receive the first training data sent by the terminal; Wherein, the first training data is collected and / or recorded by the terminal based on the data collection configuration, and the first training data is used to train and / or update the target model, the target model is used to implement the target use case, and the target use case includes at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case.

19. A terminal, wherein, The terminal includes a memory, a transceiver, and a processor. The memory is used to store computer programs; the transceiver is used to send and receive data under the control of the processor. Processor, configured to read the computer program in the memory and perform the following operations: Control the transceiver to collect and configure data collection sent by the network device; Collect and / or record the first training data according to the data collection configuration; Control the transceiver to send the first training data to the network device; The first training data is used to train and / or update the target model, and the target model is used to implement the target use case, which includes at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case.

20. A network device, wherein, The network device includes a memory, a transceiver, and a processor. A memory for storing computer programs; a transceiver for sending and receiving data under the control of the processor; and a processor for reading the computer programs from the memory and performing the following operations: Control the transceiver to send data collection configuration to the terminal; Control the transceiver to receive the first training data sent by the terminal; Wherein, the first training data is collected and / or recorded by the terminal based on the data collection configuration, and the first training data is used to train and / or update the target model, the target model is used to implement the target use case, and the target use case includes at least one of the following use cases: RRM prediction use case, event prediction use case, and RLF prediction use case.

21. A processor-readable storage medium, wherein, The processor-readable storage medium stores a program for causing the processor to perform the method according to any one of claims 1 to 12.

22. A processor-readable storage medium, wherein, The processor-readable storage medium stores a program for causing the processor to perform the method according to any one of claims 13 to 16.

Images