Measurement event prediction method based on artificial intelligence (AI) / ML model

Abstract

A measurement event prediction method based on an artificial intelligence (AI) / ML model. The method comprises: a user equipment reporting an inference quantity of an AI / ML model, wherein the inference quantity at least comprises at least one of the following: one or more predicted trigger events at one or more moments in the future, parameters of a measurement event trigger configuration for inference, adjustment quantities of the parameters of the measurement event trigger configuration for inference, parameters of a measurement gap configuration for inference, and time information; and a base station receiving the inference quantity of the AI / ML model.

Description

Measurement event prediction method based on artificial intelligence AI / ML model Technical Field

[0001] The present application relates to the field of wireless communication technology, and in particular to an artificial intelligence-based measurement event prediction method, user equipment, and base station. Background Art

[0002] Currently, the integration of artificial intelligence (AI) and communication systems has become a key research focus for 5.5G (3GPP Release 18) and 6G. The application of AI machine learning (ML) in wireless communications has been widely discussed, and its application scope includes the physical layer, high layer, and system architecture layer. Applications at the physical layer include AI-based channel coding and decoding, channel estimation, interference cancellation, channel state information (CSI) prediction and estimation, positioning, and beam management. Applications at the high layer include AI-based network energy saving, mobility management, load balancing, and radio resource management (RRM) algorithms. Applications at the system architecture layer include NWADF data analysis and application.

[0003] AI / ML-assisted L3-based mobility operations, such as RRM management and event prediction based on AI / ML, can optimize handover decisions. Technical issues:

[0004] In RRC_CONNECTED, the user equipment (UE) measures multiple beams (at least one) of the cell and averages the measurement results (power values) to obtain the cell quality.

[0005] The network can configure an RRC_CONNECTED UE to perform measurements. The network can provide measurement configuration via dedicated signaling (e.g., RRCReconfiguration or RRCResume). After configuration, the UE reports measurement results based on the measurement configuration or performs conditional reconfiguration evaluation based on conditional reconfiguration. For cell measurements, the network can configure RSRP, RSRQ, SINR, RSCP, or EcN0 as trigger quantities.

[0006] The measurement configuration includes the following parameters:

[0007] 1) Measurement objects (MO): Measurement objects can be represented by a list of objects that the UE needs to measure.

[0008] 2. Reporting configurations

[0009] For intra-frequency measurement and inter-frequency measurement, the measurement object indicates the frequency / time position and subcarrier spacing of the reference signal (RS) to be measured. Associated with this measurement object, the network can configure a cell-specific offset list, a list of 'exclude-listed' cells, and a list of 'allow-listed' cells. Excluded list cells are not applicable for event evaluation or measurement reporting. Allowed list cells are cells that are applicable for event evaluation or measurement reporting. The measObjectId of the MO corresponding to each serving cell is indicated by the servingCellMO in the serving cell configuration.

[0010] The reporting configuration includes a reporting criterion, a reference signal type (RS type), and a reporting format. The reporting criterion indicates the conditions or criteria for triggering the UE to send a measurement report. The reporting criterion can be periodic or a description of a single event.

[0011] 2.1. Reporting Condition: The condition that triggers the UE to send a measurement report. This can be a periodic or single event description.

[0012] 2.2 RS type: RS (SS / PBCH block or CSI-RS) used by the UE for beam and cell measurement results.

[0013] 2.3. Report format: The UE includes in the measurement report the number of each cell and each beam (e.g. RSRP) and other relevant information, such as the maximum number of cells to be reported and the maximum number of beams per cell.

[0014] Detailed measurement configuration information and measurement results are contained in MeasConfig and MeasResult. For the measId that triggers the measurement report procedure, the UE shall set measResults in the MeasurementReport message as follows:

[0015] Table 1: MeasurementReport Information

[0016] Table 2: MeasResults Element

[0017] The reporting configuration includes the reporting configuration information element (IE) ReportConfigNR and the event trigger configuration information element (IE) EventTriggerConfig. The IE ReportConfigNR specifies the trigger criteria for NR measurement report events, CHO, CPA, or CPC events, or L2 U2N relay measurement report events. For events marked with A1, A2, ..., measurement report events and CHO, CPA, or CPC events are based on cell measurement results, which can be derived based on SS / PBCH blocks or CSI-RS.

[0018] The event trigger configuration element EventTriggerConfig is shown in the following table.

[0019] Table 3: EventTriggerConfig Element

[0020] The fields shown are explained in the table below:

[0021] Table 4

[0022] The network sends a measurement configuration (measconfig) to the UE, which contains the following information:

[0023] 1) Report configuration (reportconfig): specifies how the UE reports measurement results;

[0024] 2) Event trigger reporting configuration (EventTriggerConfig): specifies when the UE reports the measurement results.

[0025] The UE performs measurements according to the measurement configuration and determines whether to report the measurement results according to the event-triggered reporting configuration.

[0026] If the measurement result meets the conditions of the event-triggered reporting configuration and lasts for the time indicated by the trigger measurement time (timeToTrigger, TTT), the UE will report the measurement result to the network. The report content includes:

[0027] 1). Measurement ID (meas ID);

[0028] 2) Measurement results (RSRP, RSRQ, SINR based on cell / RS index);

[0029] 3). Maximum number of cells;

[0030] 4). Maximum number of beams.

[0031] In high-frequency scenarios, dense base station deployment can shorten the distance between UEs and base stations, significantly improving the link quality between them and enhancing signal coverage. However, due to the reduced coverage radius, UE movement can lead to more frequent handovers, which incurs increased signaling overhead. Furthermore, changes in UE location can cause rapid fluctuations in signal strength, potentially leading to handover failures or ping-pong handovers, thus degrading network performance and user experience. Furthermore, high-speed UE movement is a major cause of frequent handovers and handover failures. This is because between the time the UE prepares for a handover and the time it receives the handover command, channel quality may have significantly changed, potentially causing handover failures.

[0032] In traditional handover, the base station makes a handover decision (Handover decision) based on the measurement report of the mobile device. Specifically, the base station configures measurements for the UE, including measurement objects (Measurement Object, MO), measurement quantities (for example, RSRP, RSRQ, SINR) and measurement reporting methods (reportType, for example, periodic, event triggered). The UE measures the signal quality of the serving cell and / or adjacent cells, and reports the measurement quantities to the base station based on the measurement configuration. The base station makes subsequent behavioral judgments based on the reported measurement results, such as whether handover is required or whether to add a secondary cell. In traditional handover (Handover, HO), on the one hand, too frequent measurement reports are not conducive to UE energy saving, and on the other hand, a large amount of data needs to be concentrated on the base station side, which greatly increases the complexity of the base station decision. Therefore, 3GPP defines a series of preset measurement reporting mechanisms (i.e., event-triggered reporting mechanisms), including the threshold for triggering measurement reports, the hysteresis factor (Hys) for avoiding ping-pong handovers, and the time to trigger (TTT) for conditional measurement report triggering. By setting these parameters, the goal of reducing handover failures and handover ping-pong is achieved.

[0033] However, improper parameter configuration can also cause handover failures. For example, if TTT is configured too high, the UE may miss the optimal handover time, resulting in handover failure. Alternatively, if TTT and Hys are configured too low, this may result in more frequent event reporting and an increase in the number of ping-pong handovers. Therefore, proper parameter configuration is crucial.

[0034] In FR2 and ultra-high-speed mobility scenarios, current passive mobility management mechanisms cannot achieve optimal network performance. Therefore, there is widespread consensus that AI / ML-based measurement event prediction can be used to assist in optimizing handover decisions and ensure user experience. However, the practical operation methods and standards for AI / ML-based measurement event prediction are still under development.

[0035] Therefore, a measurement event prediction method based on artificial intelligence AI / ML model is needed.

[0036] Summary of the Invention

[0037] One objective of the present disclosure is to propose a measurement event prediction method, user equipment, and base station based on an artificial intelligence (AI) / ML model.

[0038] In a first aspect, the present invention provides a measurement event prediction method based on an artificial intelligence (AI) / ML model, executed in a user device, comprising:

[0039] Report the inference amount of the AI / ML model, where the inference amount includes at least one of the following: one or more triggering events at one or more predicted future moments, parameters of the inferred measurement event trigger configuration, adjustments to the parameters of the inferred measurement event trigger configuration, parameters of the inferred measurement gap configuration, and time information.

[0040] In a second aspect, the present invention provides a measurement event prediction method based on an artificial intelligence (AI) / ML model, executed in a user device, characterized by:

[0041] updating / modifying parameters of a measurement event trigger and / or a measurement gap configuration based on an inference obtained by the AI / ML model, wherein the inference includes at least one of the following: one or more predicted triggering events at one or more future moments, inferred parameters of the measurement event trigger configuration, an adjustment amount of the inferred parameters of the measurement event trigger configuration, inferred parameters of the measurement gap configuration, and time information;

[0042] Report at least one of the following information: an event identifier, parameter modification indication information, and the inference amount, wherein the parameter modification indication information is used to indicate at least one of the following: the update / modification of the parameter configured by the measurement event trigger by the user equipment, or the measurement amount reported according to the updated / modified parameter configured by the measurement event trigger.

[0043] In a third aspect, the present invention proposes a measurement event prediction method based on an artificial intelligence (AI) / ML model, executed in a base station, characterized by:

[0044] Receive inference data from an AI / ML model, where the inference data includes at least one of the following: one or more triggering events at one or more predicted future moments, parameters of the inferred measurement event trigger configuration, adjustments to the parameters of the inferred measurement event trigger configuration, parameters of the inferred measurement gap configuration, and time information.

[0045] In a fourth aspect, the present invention proposes a measurement event prediction method based on an artificial intelligence (AI) / ML model, executed in a base station, characterized by:

[0046] Receive at least one of the following information: an event identifier (event ID), parameter modification indication information, and the inferred amount, wherein the parameter modification indication information is used to indicate at least one of the following: the update / modification of the parameter configured for the measurement event trigger by the user equipment, or the measurement amount reported based on the updated / modified parameter configured for the measurement event trigger, the inferred amount including at least one of the following: one or more triggering events predicted at one or more future moments, the inferred parameter configured for the measurement event trigger, the adjustment amount of the inferred parameter configured for the measurement event trigger, the inferred parameter configured for the measurement gap, and time information.

[0047] In a fifth aspect, an embodiment of the present invention provides a wireless communication device, comprising a processor and a memory, wherein the processor is configured to call and execute a computer program stored in the memory so that the device equipped with the processor performs the disclosed method.

[0048] The disclosed method can be programmed as computer-executable instructions stored in a non-transitory computer-readable medium. The non-transitory computer-readable medium, when loaded into a computer, instructs the processor of the computer to execute the disclosed method.

[0049] The non-transitory computer-readable medium may include at least one of the group consisting of a hard disk, a CD-ROM, an optical storage device, a magnetic storage device, a read-only memory, a programmable read-only memory, an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), and a flash memory.

[0050] The disclosed method can be programmed as a computer program product, which causes a computer to perform the disclosed method.

[0051] The disclosed method may be programmed as a computer program, which causes a computer to perform the disclosed method.

[0052] Technical effect:

[0053] During the handover preparation phase, the UE uses AI / ML-based methods to perform measurement event prediction and infer one or more of the following inference quantities:

[0054] 1) A triggering event that may occur at one or more moments in the future.

[0055] 2). The parameters of the configuration triggered by the predicted measurement event,

[0056] 3). Predict the amount of adjustment to the configured parameters triggered by the measurement event, and

[0057] 4). Time information.

[0058] The UE reports the above inference information to the original base station to help the original base station perform more reasonable measurement event triggering configuration, thereby achieving the following goals: reducing handover ping-pong, increasing the probability of handover success, optimizing handover performance, and ensuring user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] One or more embodiments are exemplarily illustrated by the figures in the corresponding drawings. These exemplifications do not limit the embodiments. Elements with the same reference numerals in the drawings represent similar elements. Unless otherwise specified, the figures in the drawings are not intended to be proportional. The following divisions of the various embodiments are for ease of description and are not intended to limit the specific implementation of the present invention. The various embodiments may be combined and referenced with each other as long as there is no contradiction.

[0060] FIG1 is a schematic diagram showing a communication system.

[0061] FIG2 is a schematic diagram showing an embodiment of a measurement event prediction method based on an artificial intelligence (AI) / ML model according to the present invention.

[0062] FIG3 is a schematic diagram showing another embodiment of a measurement event prediction method based on an artificial intelligence (AI) / ML model according to the present invention.

[0063] FIG4 is a schematic diagram showing parameters of a network-focused configuration event trigger configuration.

[0064] FIG5 is a schematic diagram showing a method of modifying / updating parameters of a measurement event trigger configuration on the UE side.

[0065] FIG6 is a schematic diagram showing a handover decision made by the network side based on measurement values ​​reported under different reporting conditions.

[0066] FIG. 7 is a schematic diagram showing a monitoring process at the UE side.

[0067] FIG8 is a schematic diagram showing the monitoring process at the NW side.

[0068] FIG9 is a schematic diagram showing a process of adjusting measurement gap parameters on the network side.

[0069] FIG10 is a schematic diagram showing static capability reporting of a user equipment.

[0070] FIG11 is a schematic diagram showing dynamic capability reporting of a user equipment.

[0071] FIG12 is a schematic diagram showing a user equipment according to the present invention.

[0072] FIG13 is a schematic diagram showing a network node according to the present invention.

[0073] FIG14 is a schematic diagram showing an integrated circuit (IC) chip of the present invention.

[0074] FIG15 is a schematic diagram showing an integrated circuit (IC) chip of the present invention. DETAILED DESCRIPTION

[0075] In order to make the purpose, technical solutions and advantages of this application more clear, some embodiments of this application are further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain this application and are not used to limit this application.

[0076] The relevant technical terms in this article are explained as follows:

[0077] Table 1

[0078] In the AI / ML-based predicted measurement event use case, the following questions need to be investigated:

[0079] Question 1: The reporting mechanism of the inference amount;

[0080] Question 2: How to determine whether the UE is allowed to modify the parameters of the configured measurement event trigger configuration?

[0081] Problem 3: Performance monitoring mechanism for measuring event prediction;

[0082] Question 4: Measurement gap adjustment;

[0083] Question 5: UE capability reporting.

[0084] Therefore, to better support this use case, embodiments of the present disclosure provide potential solutions to the problem.

[0085] Referring to Figure 1 , one or more core network devices (e.g., core network device 30) that execute core network network functions are connected to multiple base stations, including base stations 20a, 20b, ..., 20m. Base stations 20a and 20b may be referred to as a first base station and a second base station. In some examples, base station 20a may serve as a source base station, and base station 20b may serve as a target base station. The multiple base stations are connected to multiple user equipment (UEs) 10a, 10b, ..., 10n via wireless channels. m and n may be positive integers.

[0086] The network entity device 30 may be a node in a CN. The CN may include an LTE CN or a 5G core (5GC), which includes a user plane function (UPF), a session management function (SMF), an access and mobility management function (AMF), a unified data management (UDM), a policy control function (PCF), a control plane (CP) / user plane (UP) separation (CUPS), an authentication server (AUSF), a network slice selection function (NSSF), and a network exposure function (NEF).

[0087] Examples of UEs described herein may include one of the UEs 10a, UE 10b, ..., or UE 10n. Examples of base stations described herein may include the base stations 20a or 20b. A network may be a network node of the base station or a core network. Examples of base stations may be gNBs, eNBs, or other base stations. Uplink (UL) transmission of control signals or data may be a transmission operation from a UE to a base station. Downlink (DL) transmission of control signals or data may be a transmission operation from a base station to a UE. DL control signals may include medium access control (MAC) control elements (CE), downlink control information (DCI), or radio resource control (RRC) signals from a base station to a UE.

[0088] Model inference is performed in a scenario where the UE side (e.g., at least one of UEs 10a-10n) infers the predicted results (including reporting of candidate / target cells) based on the model and sends them to the source base station (e.g., base station 20a). The source base station makes a handover decision based on the reported prediction results.

[0089] Referring to FIG2 , a measurement event prediction method based on an artificial intelligence (AI) / ML model is described below. The method is described using base stations 20 a - 20 m and user equipment 10 a as examples. However, it should be understood that the method can be executed on one or more of base stations 20 a - 20 m and one or more of user equipment 10 a - 10 n.

[0090] Step S001: The user device 10a reports the inference data of the AI / ML model to the base station 20a. The inference data includes at least one of the following: one or more predicted triggering events at one or more future time points, parameters configured for the inferred measurement event triggering, adjustments to the parameters configured for the inferred measurement event triggering, parameters configured for the inferred measurement gap, and time information. The user device 10a reports the inference data to the base station 20a, helping the base station 20a to configure more reasonable measurement event triggering, thereby achieving the following objectives: reducing handover ping-pong, increasing the probability of handover success, optimizing handover performance, and ensuring user experience.

[0091] In some embodiments of the present invention, the user equipment 10a reports an event ID, and the base station 20a receives the event ID.

[0092] In some embodiments of the present invention, the base station 20a sends a configuration message for AI / ML model reasoning to the user equipment 10a. The user equipment 10a receives the configuration message for AI / ML model reasoning, where the configuration message includes a reporting configuration, where the reporting configuration includes at least one of the following: an inference amount, a measurement time identifier ID, a reporting rule, and a measurement amount.

[0093] In some embodiments of the present invention, the reporting of the inference amount is based on a time-based reporting rule, wherein the time-based reporting rule includes at least one of the following: a start time, a duration, an end time, an offset time, and an interval between two reporting times; or

[0094] The reporting of the inference amount is based on a reporting rule for the amount of inference amount; or

[0095] The reporting of the inference amount is based on a reporting rule of a difference between the inference amount and a traditional configuration parameter.

[0096] In some embodiments of the present invention, the method further includes monitoring the performance of the AI / ML model. The performance monitoring may be performed on the user equipment 10a or the base station 20a.

[0097] The performance monitoring includes comparing one or more measurement events reported at one or more future moments inferred by the AI / ML model or other means with one or more measurement events reported at one or more future moments obtained by a traditional measurement method.

[0098] In some embodiments of the present invention, the method further includes determining an accuracy of performance monitoring of the AI / ML model; wherein the accuracy is based on a threshold, and the threshold is based on at least one of: time and number of times.

[0099] Referring to FIG3 , another embodiment of a measurement event prediction method based on an artificial intelligence (AI) / ML model is described below. The method is described using base stations 20 a - 20 m and user equipment 10 a as examples. However, it should be understood that the method can be executed on one or more of base stations 20 a - 20 m and one or more of user equipment 10 a - 10 n.

[0100] Step S011: The user equipment 10a updates or modifies parameters of the measurement event trigger and / or measurement gap configuration based on the inference amount of the AI / ML model, where the inference amount includes at least one of the following: one or more triggering events predicted at one or more future time moments, inferred measurement event trigger configuration parameters, adjustment amounts of inferred measurement event trigger configuration parameters, inferred measurement gap configuration parameters, and time information.

[0101] Step S012: The user equipment 10a reports at least one of the following information: an event ID, parameter modification indication information, and the inference amount to the base station 20a, wherein the parameter modification indication information is used to indicate at least one of the following: the update / modification of the parameters configured for the measurement event trigger by the user equipment, or the measurement amount reported based on the updated / modified parameters configured for the measurement event trigger. The user equipment 10a reports the inference amount to the base station 20a to help the base station 20a perform more reasonable measurement event triggering configuration, thereby achieving the following objectives: reducing handover ping-pong, improving the probability of handover success, optimizing handover performance, and ensuring user experience.

[0102] In some embodiments of the present invention, the base station 20a sends a configuration message for AI / ML model reasoning to the user equipment 10a. The method further includes receiving a configuration message for AI / ML model reasoning, the configuration message including a reporting configuration, the reporting configuration including at least one of the following: an inference amount, a measurement time identifier ID, a reporting rule, and a measurement amount.

[0103] In some embodiments of the present invention, the base station 20a sends modification permission indication information to the user equipment 10a. The user equipment 10a receives the modification permission indication information, where the indication information is used to indicate that the user equipment is allowed to modify the measurement event triggering and / or measurement gap configuration parameters.

[0104] In some embodiments of the present invention, the indication information is carried in a system message, a paging message, or dedicated information.

[0105] In some embodiments of the present invention, the indication information is carried in the system message or paging message, wherein the indication information is based on at least one of the following: per model level, per functionality level, per cell level, per RAN notification area level, per tracking area level, per TA level, per PLMN level, per AI-specific area level, or per UE level. The functionality indication is based on the functions or features supported by the AIML / AIML model, such as based on AIML RRM prediction, based on AIML event prediction (including measurement event prediction, handover failure rate prediction, link failure prediction), based on AIML candidate or target cell / beam prediction.

[0106] In some embodiments of the present invention, the indication information is carried in the dedicated message, and the indication information is based on at least one of the following: per model level, per functionality level, per cell level, per RAN notification area level, per tracking area level, per PLMN level, per AI-specific area level, or per parameter level.

[0107] In some embodiments of the present invention, the method further includes monitoring the performance of the AI / ML model. The performance monitoring may be performed on the user equipment 10a or the base station 20a.

[0108] The performance monitoring includes comparing one or more measurement events reported at one or more future moments inferred by the AI / ML model or other means with one or more measurement events reported at one or more future moments obtained by a traditional measurement method.

[0109] In some embodiments of the present invention, the method further includes determining an accuracy of performance monitoring of the AI / ML model; wherein the accuracy is based on a threshold, and the threshold is based on at least one of: time and number of times.

[0110] In some embodiments of the present invention, the base station 20a sends a capability request query to query whether the user equipment supports modifying parameters of the measurement event triggering configuration. The base station 20a receives the capability request query.

[0111] The user equipment 10a reports a capability message of the user equipment in response to the capability request query. The base station 20a receives a capability message of the user equipment in response to the capability request query, the capability message being used to indicate whether the user equipment supports modifying parameters of the measurement event triggering configuration.

[0112] In some embodiments of the present invention, the capability request query is a capability query message carrying a "modify configured measurement event trigger configuration parameters - request" information element; and the capability message is a message carrying a "modify configured measurement event trigger configuration parameters" indication.

[0113] In some embodiments of the present invention, the capability request query is an RRC reconfiguration message; and the capability message is an RRC reconfiguration complete message.

[0114] Several embodiments of the present invention are described below:

[0115] Measurement event prediction is performed by a UE-side model (UE sided model). The UE-side model (UE sided model) is an AI / ML model located on the UE side (e.g., user equipment 10a, 10b, ... and / or 10n) that is used to perform model inference. Specifically, the UE (e.g., one or more of user equipment 10a, 10b, ... and 10n) performs model inference based on model inference input data (inference input). The model inference input data may be configuration / data received from the network side.

[0116] Alternatively, the input data for the model inference may be internal data of the UE itself, or historical data. The input data for the model inference may include at least one of the following: configured parameters, such as offset, hysteresis (Hys), threshold, measurement trigger time (TTT), and T312 timer. The measured data may include at least one or more of the following: signal quality (e.g., at least one or a combination of RSRP, RSRQ, SINR, RSCP, and EcN0), UE behavior trajectory, channel condition information, etc. All of the above information may be historical information.

[0117] Model inference output:

[0118] The UE (e.g., one or more of the user equipment 10a, 10b, ... and 10n) runs the AI / ML model to infer one or more inference results / inference quantities, where the inference results include at least one or more of the following: one or more triggering events (predicted events) at one or more predicted future moments, inferred (UE preferred) measurement event triggering configuration parameters, inferred (UE preferred) adjustment amounts of the measurement event triggering configuration parameters, inferred measurement gap configuration parameters, and time information, etc.

[0119] 1) One or more predicted trigger events (predicted measurement events) at one or more future moments indicate that one or more event reports will be triggered at one or more future moments. For example, at time T1, a UE (e.g., one or more of user equipment 10a, 10b, ..., and 10n) meets the conditions of event A3, triggering the UE to report the event. The moment can be understood as a specific point in time or a period of time. Examples of events are as follows:

[0120] Event A1: Serving cell quality exceeds an absolute threshold. Event A1 is about service quality.

[0121] Event A2: Serving cell quality becomes worse than an absolute threshold. Event A2 is about degradation of service quality.

[0122] Event A3: The neighboring cell quality is better than the PCell / PSCell by an offset. Event A3 reflects the relationship between the neighboring cell quality and the PCell / PSCell based on the offset.

[0123] Event A4: Neighboring cell quality exceeds the absolute threshold. Event A4 is about neighboring cell quality.

[0124] Event A5: PCell / PSCell becomes worse than absolute threshold 1, and neighboring cell / SCell becomes better than another absolute threshold 2. Event A5 is about PCell / PSCell quality degradation and neighboring cell / SCell quality improvement

[0125] Event A6: The quality of the neighboring cell is better than the SCell by an offset. Event A6 reflects the relationship between the quality of the neighboring cell and the SCell based on the offset.

[0126] 2) Examples of inferred measurement event trigger configuration parameters include: preferred offset, preferred hysteresis, preferred threshold, preferred TTT, and preferred T312. These parameters can be represented by XXX_AI and are used to represent parameters of the measurement event trigger configuration inferred by the AI / ML model. XXX is used to represent parameters of the measurement event trigger configuration configured in a traditional manner. The inferred measurement event trigger configuration parameters can be used to modify / update the measurement event trigger configuration parameters configured in a traditional manner.

[0127] 3) The inferred adjustment amount for the measurement event trigger configuration parameter is expressed as the increase or decrease in the traditionally configured measurement event trigger parameter. This can be represented by XXX_offset_AI and can be the adjustment amount for offset (per offset), hysteresis (per Hys), threshold (per Threshold), TTT (per TTT), or T312 timer (per T312). XXX is used to represent the traditionally configured measurement event trigger parameter.

[0128] 4) Examples of inferred measurement gap configuration parameters include the following: preferred measurement gap length (preferred mgl), preferred measurement gap offset (preferred gapOffset), preferred measurement gap repetition period (preferred mgrp), and preferred measurement gap timing advance (preferred mgta). XXX__AI represents the gap configuration parameters inferred by the AI / ML model and can be used to modify / update the measurement gap configuration configured using traditional methods. XXX is used to represent the parameters of the measurement event trigger configuration configured using traditional methods.

[0129] 5) Time information, indicating when one or more inferences are valid at a specific moment or moments within a defined timeframe. This time information is associated with the parameters of the one or more measurement event trigger configurations described above. For example, an AI / ML model may predict that the offset value at time T1 is offset1 and the offset value at time T2 is offset2.

[0130] The above-mentioned inference results may be based on inference results at different levels, for example, per event, per cell, per beam, or per UE. Parameters associated with the above-mentioned inference results are parameters at the said levels.

[0131] The above reasoning results can be implemented in the following ways:

[0132] Method 1: The AI / ML model predicts one or more triggering events at one or more future moments. For example, assume that the AI / ML model predicts that at time T1, the measurement values ​​of the UE (e.g., one or more of user equipment 10a, 10b, ..., and 10n) meet the triggering conditions of Event A1. Then, the UE will report its measurement results at time T1.

[0133] In this approach, measurement results can be derived from AI / ML models or measured using traditional methods. Similarly, event trigger configuration parameters can be derived from AI / ML models or configured using traditional methods. Possible conditional expressions are as follows:

[0134] Option 1:

[0135] Inequality A1-1 (entry condition): Ms–Hys>Thresh

[0136] Inequality A1-2 (exit condition): Ms+Hys <Thresh

[0137] Option 2:

[0138] Inequality A1-1 (entry condition): Ms–Hys-AI > Thresh-AI

[0139] Inequality A1-2 (exit condition): Ms+Hys-AI <Thresh-AI

[0140] Option 3:

[0141] Inequality A1-1 (entry condition): Ms–Hys-AI > Thresh-AI

[0142] Inequality A1-2 (exit condition): Ms+Hys-AI <Thresh-AI

[0143] Among them, Ms represents the measurement value;

[0144] Hys represents the hysteresis;

[0145] Hys-AI stands for hysteresis amount;

[0146] Thresh stands for threshold value;

[0147] Thresh-AI stands for threshold value.

[0148] In Option 1, the measurement value (Ms) is predicted by the AI / ML model, and the parameters Hys and Thresh are configured using traditional methods. In Option 2, the measurement value (Ms) is measured using traditional methods, and the parameters Hys-AI and Thresh-AI are inferred using the AI / ML model. In Option 3, the measurement value (Ms) and the parameters Hys-AI and Thresh-AI are inferred using the AI / ML model. The parameters here include at least one of the parameters configured by the measurement event trigger, including but not limited to Hys and TTT.

[0149] Method 2: Use AI / ML models to directly infer the parameters configured for measurement event triggers. These parameters are used to replace / update the parameters configured in the traditional way. This can be represented by XXX_AI. Taking event A1 as an example, possible conditional expressions are as follows:

[0150] Inequality A1-1 (entry condition): Ms–Hys_AI>Thresh-AI

[0151] Inequality A1-2 (exit condition): Ms+Hys_AI <Thresh-AI

[0152] In this scheme, the measurement value Ms is obtained by traditional measurement methods, but one or more parameters of the measurement event (i.e., Hys_AI and Thresh-AI) can be inferred through AI / ML.

[0153] Method 3: Directly infer the adjustment amount of the parameters configured by one or more measurement event triggers through the AI / ML model. This can be expressed as XXX_offset_AI and is used to adjust the conditions of the measurement events configured in the traditional way. Taking event A1 as an example, its conditional expression is as follows:

[0154] Inequality A1-1 (entry condition): Ms–Hys-Hys_offset_AI>Thresh

[0155] Inequality A1-2 (exit condition): Ms+Hys+Hys_offset_AI <Thresh

[0156] Example 1: Network side updates / modifies parameters of measurement event trigger configuration

[0157] As shown in FIG4 , the network (NW) side (e.g., network 220 ) can update or modify parameters of the measurement event trigger configuration (EventTriggerConfig) in the measurement configuration and / or the inferred measurement gap configuration based on the inference value inferred by the AI / ML model. The specific steps of the process are as follows:

[0158] Step S0: A UE (e.g., one or more of user equipment 10a, 10b, ..., and 10n) receives a measurement configuration message, also referred to as a model-inferred configuration message, sent from an original base station (e.g., base station 20a). The model-inferred configuration message includes a measurement event trigger configuration (EventTriggerConfig) for measuring a measurement object of a wireless communication air interface, including different event trigger conditions and / or an inferred measurement gap configuration.

[0159] Step S1: The UE performs a model inference function. At least a portion or all of the parameters configured for the measurement event triggering can serve as input data for the model inference. The input data for the model inference can be configurations / data received from the network. Alternatively, the input data for the model inference can be internal data of the UE itself, or historical data. The input data for the model inference can include at least one of the following: parameters configured for event triggering conditions, such as offset, hysteresis (Hys), threshold, measurement time to trigger (TTT), and T312 timer. Specifically, the UE runs one or more appropriate models based on the input data for the model inference, performs model inference, and generates an inference result / inference output. The inference result / inference output includes at least one of the following: one or more predicted triggering events at one or more future moments, inferred (UE preferred) measurement event triggering configuration parameters, adjustments to inferred (UE preferred) measurement event triggering configuration parameters, inferred (UE preferred) measurement gap configuration parameters, time information, etc.

[0160] Step S2: The UE reports the inference result (inference amount), and the inference result is used to assist the network side in reconfiguring more reasonable parameters for measurement event triggering configuration.

[0161] Optionally, the reported inference amount may also carry an event ID, which is used to indicate which one or more events trigger the need to modify / adjust / reconfigure the configured parameters.

[0162] Step S3: The original base station (eg, base station 20a) updates or modifies the parameters of the measurement event triggering configuration based on the inference result (inference amount), and reconfigures the updated / modified parameters of the measurement event triggering configuration to the UE.

[0163] In the above solution, it can be seen that in order to assist the network in reconfiguring more reasonable measurement event triggering (condition) parameters and sending them to the UE, in step S2, the UE can report the inference amount to the network side. In order to implement the reporting of the inference amount, there are several optional methods:

[0164] Method 1: Introducing reporting rules for inference metrics. The UE (e.g., one or more of user equipment 10a, 10b, ..., and 10n) reports inference metrics according to these reporting rules. Reasonable reporting rules can reduce reporting resource overhead and signaling overhead. These reporting rules can be defined using any of the following schemes.

[0165] Solution 1: The reporting rules for inference volume can be time-based. Specifically, for example, you can define that all inference volume within a period needs to be reported, or you can define that the latest inference volume within a period needs to be reported, or you can define that the average inference volume within a period needs to be reported, etc. In this solution, you can define at least one of the following information: start time, duration, end time, offset time, and the interval between two reporting times. The following is an explanation:

[0166] Start time: indicates the start time of reporting the inference amount;

[0167] Duration: indicates the duration of the inference amount reported;

[0168] End time: indicates the time when the inference data reporting ends;

[0169] Offset time: indicates the offset of the inference amount reporting time, where the reporting time is the start time plus the offset; the interval between two times: indicates the interval between two reports.

[0170] The time unit may be a subframe, a time slot, milliseconds, seconds, minutes, hours, days, or weeks, etc. Solution 1 may also be described as being based on periodic reporting.

[0171] Solution 2: The reporting rule for the inference amount can be a reporting rule based on the number of inference amounts. Specifically, the reporting rule may include a threshold based on the number. When the number of inference amounts inferred by the AI / ML model reaches the threshold, the UE (for example, one or more of the user devices 10a, 10b, ... and 10n) is triggered to report the inference amount. For example, if the threshold is defined as 5, the unit is times / piece, when the number of predicted values ​​inferred by the AI / ML model, such as the preferred hysteresis (preferred Hys) is 5 times / piece, the UE reports the inference amount based on the above threshold and the preferred hysteresis. The reported amount can be all the inference amounts, the latest inference amount, the average inference amount value, or a randomly selected inference amount, etc.

[0172] In this example, the unit "times" can be understood as the number of times an AI / ML model can infer one or more identical or different inference quantities from an inference. Furthermore, in this example, the quantity can be based on measurement parameters, events (per event), models (per model), functionality (per functionality), or user equipment (per UE).

[0173] Solution 3: The reporting rule for the inference amount can be based on the difference between the inference amount and traditional configuration parameters. Specifically, the reporting rule can include one or more thresholds based on the difference between the inference amount and the traditional configuration parameters. When the difference between the inference amount inferred by the AI / ML model and the traditional configuration parameters meets the threshold, the UE (e.g., one or more of user equipment 10a, 10b, ..., and 10n) can be triggered to report the inference amount. For example, if a threshold value is defined, "Threshold 1_Difference", when the difference between the inference amount and the traditional configuration parameters is greater than the defined threshold, it can indicate that the preconfigured parameter values ​​are inappropriate, and the UE can be triggered to report the inference amount. Alternatively, if the difference between the inference amount and the traditional configuration parameters is significantly greater than the defined threshold, it can indicate that the predicted value may be inaccurate, which may be due to a malfunction or inappropriateness of the AI / ML model. Alternatively, if the difference between the inference amount and the traditional configuration parameters is less than the defined threshold, it can indicate that the predicted value is not significantly different from the inference amount, the configured parameters are appropriate, and the UE may not need to report the inference amount. Optionally, the reporting rule may also include one or more threshold ranges based on the difference between the inference amount and the traditional configuration parameter.

[0174] In some examples, the reported inference amount may also carry an event ID to indicate which one or more events trigger the need for modification / adjustment / reconfiguration of the (conditional) parameters.

[0175] The parameters or thresholds involved in the reporting rules can be fixed, pre-defined, configured, or pre-configured. The inference amount can be reported to the network side through one of the following signaling, including UE assistance information (UAI), dedicated RRC, MAC CE, UCI, PUSCH, PUCCH, etc.

[0176] Method 2: The inference amount is reported together with the traditional measurement reporting signaling, such as the RRC reconfiguration completion message (RRCReconfigurationComplete) message. Specifically, when the event triggering condition is met, the UE (for example, one or more of the user equipment 10a, 10b, ... and 10n) triggers the measurement report and transmits the reported amount to the network side through the RRC message. The RRC message carries the inference amount of AI / ML. Optionally, the inference amount of AI / ML carried in the RRC message can be all the inference amounts, the latest inference amount, the average inference amount value, or a randomly selected inference amount.

[0177] In some examples, the UE's updating / modification of the event-triggered configuration parameters may be based on the UE's implementation behavior or on some defined or configured rules. For example, the UE may modify / update the received event-triggered configuration parameters only when these rules are met. The above rules may refer to the three solutions (Solution 1, Solution 2, and Solution 3) in Method 1.

[0178] In some examples, the network 220 sends measurement configuration information or model reasoning configuration information to the UE (e.g., one or more of user equipment 10a, 10b, ... and 10n), which may carry at least one of the following: rules for reporting reasoning amount, such as time information, a threshold for the number of reasoning amounts, a threshold for the difference between the reasoning amount and the traditional configuration parameters, etc., and an event identifier (event ID).

[0179] Example 2: UE side updates / modifies parameters of measurement event trigger configuration:

[0180] The UE (e.g., one or more of the user equipment 10a, 10b, ..., and 10n) may directly update or modify the parameters of the measurement event trigger configuration (EventTriggerConfig) in the measurement configuration and / or the inferred measurement gap configuration on the UE side based on the inference amount inferred by the AI / ML model. The process is shown in FIG5 , and the specific steps are as follows:

[0181] Step B0: A UE (e.g., one or more of user equipment 10a, 10b, ..., and 10n) receives a measurement configuration message, also called a model inference configuration message, sent from an original base station (e.g., base station 20a). The configuration message includes a measurement event trigger configuration (EventTriggerConfig), which includes various event trigger conditions and / or an inferred measurement gap configuration.

[0182] Step B1: The UE performs a model inference function based on the measurement event trigger configuration. Specifically, the UE runs one or more appropriate models based on the input data of the model inference, performs model inference, and generates an inference result / inference output. The inference result / inference output includes at least one of the following: one or more predicted triggering events at one or more future moments, parameters of the inferred (UE preferred) measurement event trigger configuration, adjustment amounts of the parameters of the inferred (UE preferred) measurement event trigger configuration, parameters of the inferred (UE preferred) measurement gap configuration, time information, etc.

[0183] Step B2: The UE side may directly update / modify the parameters of the measurement event triggering configuration configured on the network side based on the inference result of the AI / ML model.

[0184] Step B3: When the UE's measurement value meets the updated / modified measurement event triggering condition, the UE reports the measurement result to the network side. The measurement result includes a measurement indication, such as RSRP, RSRQ, SINR, cell ID, and beam ID. Optionally, the inference result can be reported to the network 220, or only the configured parameter modification indication information can be reported to the network 220, or no information can be reported to the network 220. The configured parameter modification indication information is used to indicate that the configured parameters of the original configured measurement event have been modified / updated by the UE.

[0185] In some examples, the AI / ML model on the UE side infers that at one or more future moments, one or more event triggering conditions will be met and the one or more events will be reported, and the UE reports the measurement value and / or event identifier (event ID). The event triggering can be described as: at the future moment T, the measurement value of the UE will meet the measurement event triggering reporting condition and trigger the UE to report the predicted measurement result or measurement result. Optionally, the measurement result can also report the event identifier (event ID).

[0186] It can be seen from the above scheme that in step B3, the UE reports the measurement value to the network 220, which may carry a parameter modification indication information of the reported configuration. The parameter modification indication information can be used to indicate that the configured parameters have been updated / modified by the UE, or it can also be used to indicate the measurement amount reported by the UE based on the updated / modified parameters (including triggering conditions) of the measurement event trigger configuration. The network 220 can know whether the UE side has modified the parameters of the configured event trigger configuration or whether the parameters of the previously configured event trigger configuration are reasonable after receiving the indication. Optional. Based on the indication information, the network side can independently decide whether the parameters of the event trigger configuration need to be reconfigured. The indication information can be indicated by 1 bit, for example, set to 0 to indicate no modification, and set to 1 to indicate modification. The method 3 can be used in the use case where the UE side updates / modifies the parameters of the event trigger configuration configured on the network side. Optionally, in step B3, the UE reports the measurement value to the network side, which may carry the parameters of the inferred (UE preferred) measurement event trigger configuration. The parameters may be the final modified / adjusted parameters of the UE, or may be parameters used as conditions for event triggering.

[0187] It can be seen from the above scheme that in step B2, the UE can update / modify the parameters of the received configured measurement event. However, generally speaking, the UE cannot change the parameters of the received configuration without authorization. However, considering the signaling saving aspect, for example, the parameters of the measurement event trigger configuration predicted by the UE have changed, but the change is not large, or the duration is not long, then the inference amount does not need to be reported to the network side, and the UE side directly updates / modifies the parameters of the received configured measurement event. Optionally, the UE is triggered to report the measurement amount based on the updated / modified event trigger condition predicted by the AI / ML within the valid time. Therefore, the network side needs to allow the UE to modify the parameters of the measurement event trigger configuration based on the results predicted by the AI / ML model. To support the above function, the network side can configure an authorization modification indication information to indicate that it allows the UE to modify the parameters of the measurement event trigger configuration. For example, the information "indication of instructing the UE to modify the parameters of the measurement event trigger configuration" can be indicated by 1 bit, and there are several possible implementation methods:

[0188] Mode 1: If this information is set to "enable", the UE is allowed to modify the parameters of the measurement event trigger configuration when it receives this information; if this information is set to "disable", the UE is not allowed to modify the parameters of the measurement event trigger configuration when it receives this information.

[0189] Mode 2: If the information is set to "present", the UE is allowed to modify the parameters of the measurement event trigger configuration when it receives the information; if the information is set to "absent", the UE is not allowed to modify the parameters of the measurement event trigger configuration when it receives the information.

[0190] Mode 3: If this information is set to "1", the UE is allowed to modify the parameters of the measurement event trigger configuration when it receives this information; if this information is set to "0", the UE is not allowed to modify the parameters of the measurement event trigger configuration when it receives this information.

[0191] The above-mentioned indication information needs to be sent from the network 220 to the UE. There are several optional solutions:

[0192] In solution 1, the indication information is carried in system messages, such as MIB, SIBx, paging, short paging, etc. The indication information can be based on the model level (per model), based on the function level (per functionality), based on the cell level (per cell), based on the radio access network notification area level (per RAN notification area), based on the tracking area level (per TA), based on the public land mobile network level (per PLMN), based on the AI ​​specific area level (per AI-specific area), or based on the user equipment level (per UE).

[0193] Solution 2: The indication information is carried in a dedicated message, such as a dedicated RRC, MAC CE, DCI, etc. The indication information can be based on the model level (per model), based on the function level (per functionality), based on the cell level (per cell), based on the radio access network notification area level (per RAN notification area), based on the tracking area level (per TA), based on the public land mobile network level (per PLMN), based on the AI ​​specific area level (per AI-specific area), or based on the parameter level (per parameter).

[0194] Here, the parameter level (per parameter) indicates that the parameters of certain configured measurement event trigger configurations can be modified. For example, if the indication information is based on TTT (per TTT), it means that only the UE can modify the TTT parameter configured on the network side, while other parameters (such as threshold and hys) cannot be modified.

[0195] For example, network 220 sends a system message (e.g., SIBx) to the UE side, carrying indication information instructing the UE to modify the parameters of the measurement event trigger configuration. If the indication information is set to enable, after receiving the indication information, the UE can modify the parameters / conditions of the measurement event trigger configuration based on the inference amount inferred by the AI / ML model. If the indication message is based on the AI / ML model, then the SIB message also needs to carry model identification (model ID) information or meta information (meta info) to indicate that the UE can modify the parameters / conditions of the measurement event trigger configuration based on the inference amount inferred by the model. Meta information is used to indicate other information about the AI / ML model, such as the model version and model creation time. If the indication message is based on the functionality level (per functionality), then the SIB message also needs to carry functionality identification (functionality ID) information to indicate that the UE can modify the parameters / conditions of the measurement event trigger configuration based on the inference amount inferred by the model that is mapped to the functionality.

[0196] An example of a short paging is as follows, wherein the short paging includes indication information for indicating that the UE is allowed to modify the parameters of the measurement event trigger configuration.

[0197] Table 5: Short Messages

[0198] In addition, in some examples, if the network 220 is configured with the above-mentioned indication information, it indicates that only when the measurement value of the serving cell or neighboring cell of the UE (e.g., one or more of the user equipment 10a, 10b, ... and 10n) meets the measurement event triggering condition, the UE will be triggered to report the measurement value, and the UE reporting the measurement value is based on the measurement event triggering condition after the AI / ML model inference.

[0199] In addition, in some examples, if the network 220 is configured with the above-mentioned indication information, it can be considered that the triggering condition of the traditional measurement event needs to be suspended, that is, after the UE receives the indication information configured on the network side, the UE reports the measurement value based on the condition of the measurement event triggered after the AI / ML model reasoning.

[0200] In addition, in some examples, if no measurement event triggering configuration is present and the network side has configured the aforementioned indication information, the UE may also autonomously update / modify the parameters of the received configured measurement event. The update / modification may be implemented at the UE or based on predefined protocol rules. The predefined rules may refer to the three solutions (Solution 1, Solution 2, and Solution 3) in Method 1 of Example 1.

[0201] Example 3: Performance Monitoring for Measurement Event Prediction

[0202] To ensure the adaptability of AI / ML models, the 3GPP communication standard Release 18 (R18) introduces performance monitoring in the AI ​​over air interface section. This monitoring function, such as comparing the model's inference output with the ground truth, is used to determine the accuracy of the AI / ML model's results and, therefore, its adaptability. In the AI / ML aided mobility for network triggered L3-based handover use case, the introduction of this new use case introduces new performance monitoring challenges, including monitoring performance indicators, monitoring benchmarks, and the monitoring process.

[0203] In the measurement event prediction use case, as described above, the results of AI / ML inference can include measurement event predictions. The measurement event predictions can have the following possible monitoring scenarios:

[0204] Monitoring scenario 1: After the AI / ML model infers that at one or more moments in the future, one or more event trigger conditions will be met and the one or more events will be reported, the UE (for example, one or more of the user equipment 10a, 10b, ... and 10n) reports the measurement value and / or event identifier (event ID). The event trigger predicted by the measurement event can be described as: at the future moment T, the measurement value of the UE will meet the measurement event trigger reporting condition, triggering the UE to report the measurement result or the predicted measurement result. Optionally, the UE can also report the event identifier (event ID). In this example, the future satisfied trigger event is directly predicted;

[0205] For example, the measurement event prediction includes: reporting the measurement result of event A1 at time T1 in the future, that is, the measurement value and serving cell ID of the serving cell;

[0206] For example, the measurement event prediction includes: at the future time T2, reporting the measurement result of event A3, ie, the measurement value of the neighboring cell, the measurement value of the serving cell, the neighboring cell ID, and the serving cell ID.

[0207] The measurement values ​​described in monitoring scenario 1 can be obtained through AI / ML model reasoning or through traditional measurement methods.

[0208] Monitoring scenario 2: The measurement value is obtained through traditional measurement methods. The measurement value (RRM meas results) and trigger conditions are obtained through AI / ML model reasoning, such as the preferred hysteresis (preferred Hys) and preferred threshold (preferred threshold). The event trigger of the measurement event prediction can be described as: when the measurement value (meas results) of the UE meets the trigger condition of AI / ML model reasoning, the UE is triggered to report the measurement value to assist the network side in executing the handover decision (HO decision).

[0209] For example: Inequality A1-1 (entry condition): Ms–Hys-AI>Thresh-AI

[0210] Monitoring scenario 3: The measurement value is obtained through AI / ML model reasoning, that is, the predicted measurement result (predicted meas results) at a certain moment or multiple moments in the future. The trigger condition is obtained through AI / ML model reasoning, such as the preferred hysteresis (preferred Hys) and the preferred threshold (preferred threshold). The event trigger of the measurement event prediction can be described as: when the predicted measurement result (predicted meas results) of the UE meets the trigger condition of the AI / ML model reasoning, the UE is triggered to report the predicted measurement result, which can be used to assist the network side in executing the handover decision (HO decision).

[0211] For example: Inequality A1-1 (entry condition): Ms–Hys-AI>Thresh-AI

[0212] Monitoring scenario 4: The measurement value is obtained through AI / ML model reasoning, that is, the predicted measurement results (predicted meas results) at a certain time or multiple time points in the future, and the trigger condition is configured through traditional measurement. The event trigger of the measurement event prediction can be described as: when the predicted measurement value (meas results) of the UE meets the trigger condition of the traditional measurement configuration, the UE is triggered to report the predicted measurement result value to assist the network side in making a handover decision (HO decision).

[0213] For example: Inequality A1-1 (entry condition): Ms–Hys-AI>Thresh-AI

[0214] The performance metric(s) / methods for model monitoring should include at least one of the following possible solutions:

[0215] Alternative 1: Key performance indicators (KPIs) related to measuring the accuracy of event predictions, such as measuring the accuracy of event-triggered predictions;

[0216] Alternative 2: Key performance indicators (KPIs) related to link quality, such as throughput, RSRP, RSRQ, SINR, and hypothetical BLER;

[0217] Alternative 3: Performance metrics based on the input / output data distribution of AI / ML.

[0218] A possible benchmark / reference for monitoring comparisons is:

[0219] Alternative Solution 1: Whether the measurement values ​​obtained through traditional measurement methods will trigger an event report at some point in the future; Alternative Solution 2: The network side makes switching decisions based on the measurement values ​​obtained through traditional measurement methods, the switching failure rate during the switching process, or switching ping-pong, etc.

[0220] Possible scenarios for monitoring comparison are as follows:

[0221] Solution 1: The performance monitoring of the model inference includes: comparing the one or more events inferred to be reported at the future time or time periods with the measurement events that are actually triggered to be reported at the future time or time periods.

[0222] For example, if the predicted measurement value / predicted measurement value meets a certain event trigger condition between t0 and t1, the measurement result is reported under that condition. Then, based on traditional measurement methods, observe whether event A1 is actually triggered and reported between t0 and t1. If event A1 is triggered, the prediction result is relatively accurate; otherwise, it is inaccurate. Optionally, multiple comparisons can be performed. For example, if event A1 is triggered in 4 of 5 monitoring comparisons and only 1 is not triggered, this also proves that the monitoring result is relatively accurate.

[0223] Solution 2: Compare the handover failure rate and handover ping-pong. Specifically, the measurement value can be reported under the event triggering conditions of the traditional configuration, and / or the measurement value can be reported under the triggering conditions of the AI / ML model reasoning. As shown in Figure 6, the user equipment 10a performs measurements according to the conditions of the traditional event triggering configuration in step C0a, and performs measurements according to the conditions of the event triggering configuration of the AI / ML reasoning in step C0b. The network 220 makes a handover decision based on the measurement values ​​reported under different triggering conditions in steps C1a and C1b, and compares the handover failure rate and handover ping-pong under different conditions. For example, if the handover decision is made by the network 220 based on the measurement values ​​reported by the AI / ML model reasoning, the handover failure rate is 25%, but the handover failure rate calculated by the handover decision based on the measurement values ​​under the event triggering conditions of the traditional configuration on the network side is 50% (for example, switching too slowly will lead to an increase in handover failures), then it is considered that the result of the AI / ML model reasoning is more accurate.

[0224] In the above-mentioned monitoring comparison scheme, for measurement event prediction, one or more thresholds can be configured through network-side configuration or pre-definition, and the threshold is a threshold for the accuracy of the measurement event prediction. The accuracy threshold is used to measure the accuracy of the model reasoning. Specifically, the threshold for the accuracy of the measurement event prediction includes any of the following possible schemes:

[0225] Solution 1: The accuracy threshold for measurement event prediction can be time-based. For example, a time-based threshold T can be defined or configured. Using traditional measurement methods, the system observes whether an event triggers measurement reporting within T. If so, the measurement event prediction accuracy is considered high; if not, the measurement event prediction accuracy is considered low. Optionally, the time T can be a start time t, a time duration t1, and / or a time offset t'.

[0226] Solution 2: The accuracy threshold for measurement event prediction can be based on a number of occurrences. For example, a threshold N based on the number of event trigger reports can be defined or configured. Using traditional measurement methods, the system observes whether event reporting measurements are triggered within one or more time periods, and compares these results multiple times. If the number of event trigger reports is greater than N, the accuracy of the measurement event prediction is considered high. If the number of event trigger reports is less than N, the accuracy of the measurement event prediction is considered low. The one or more time reference accuracy thresholds are similar to Solution 1.

[0227] Optionally, the network side may configure the threshold via RRC, DCI, MAC CE message, etc.

[0228] Optionally, the prediction accuracy may be expressed as high, medium, or low (high, medium, and low may correspond to a range of percentages) or percentage accuracy;

[0229] The process of performance monitoring is as follows.

[0230] As shown in FIG7 below, the monitoring is located at the UE side, and the specific steps of the monitoring process are as follows:

[0231] Step D1: Network 220 sends a performance monitoring configuration message to the UE. The monitoring activation message includes at least one of the following performance monitoring configuration parameters: performance parameter / method, performance comparison method, measurement event trigger prediction accuracy threshold, monitoring start time, monitoring execution time, monitoring reporting time, monitoring object, etc. The monitoring object may include an AI model (represented by a model ID) and a functionality (represented by a functionality ID). The performance monitoring configuration message may be a DCI, MAC CE, or RRC.

[0232] Step D2: The UE performs monitoring calculation / comparison. The UE may perform monitoring comparison according to the performance monitoring parameters / methods and / or performance comparison methods configured in step D1 to indicate whether the monitoring model or function is adapted.

[0233] Step D3: The UE sends a monitoring performance reporting message to the network 220. The operation performed by the network 220 according to the monitoring performance reporting message includes at least one of the following: performance monitoring parameters / methods, monitoring calculation results, or lifecycle management operations.

[0234] Performance monitoring parameters / methods: as described above in detail.

[0235] Monitoring calculation results: The monitoring calculation / comparison results are used to display whether the indicator model / function is compatible. The monitoring calculation results may vary based on the performance parameters / methods. The performance monitoring calculation results may be expressed as a high or 90% accuracy of the prediction of the measurement event trigger. For example, if the accuracy of the measurement event trigger prediction is low, the monitoring calculation results indicate that the monitored AI model or AI function is incompatible, and an incompatibility indication is reported. The incompatibility indication may be mapped one-to-one or one-to-many to the AI ​​model or AI function.

[0236] Lifecycle management operations: Lifecycle management operations include at least one of the following: model activation, model deactivation, model update, model switch, model selection, and rollback. For example, the report message includes at least one of the above operation instructions or related information, such as a rollback operation.

[0237] The monitoring is located on the NW side. The monitoring process is shown in Figure 8. The specific steps are as follows:

[0238] Step E1: The network 220 collects monitored data, which may include data obtained through traditional measurement methods, data predicted through AI methods, or data calculated by the UE.

[0239] Step E2: The network 220 performs monitoring calculation / comparison. The network 220 may perform monitoring calculation / comparison based on data obtained by a traditional measurement method.

[0240] Step E3: The network 220 sends the performance monitoring result as feedback. The content of the performance monitoring result is the same as step D3 of the performance monitoring process on the UE side.

[0241] In addition, in some examples, the UE or network 220 may also enable an AI / ML model based on the monitoring results. The system (i.e., the UE or network 220) that enables the AI / ML model instruction uses the parameters inferred by the AI / ML model. For example, the UE reports the parameters inferred by the AI / ML model; or the UE updates / modifies the parameters of the received configured measurement event based on the parameters inferred by the AI / ML model; the UE reports the measurement value based on the condition triggered by the measurement event after the AI / ML model inference, etc.

[0242] Example 4: Meas Gap Adjustment

[0243] In NR, when the UE needs to perform inter-frequency measurement, the base station (for example, one of the base stations 20a-20m) needs to re-message the UE through RRC, and configure the measurement gap (meas gap) through the MeasGapConfig IE in MeasConfig, including the measurement gap type, gapOffset, mgl, mgrp, and mgta. However, if the configured gap length (mgl) is too short when the UE performs inter-frequency measurement, the UE will not receive all synchronization signals within the measurement time, resulting in inaccurate measurement results; if the configured mgl is too long, the data interruption time will be too long, reducing the system throughput and wasting time and frequency resources. Therefore, configuring a suitable mgl is extremely important. In traditional measurement configuration, since the NW side configures the measurement object (including the serving cell and the neighboring cell), the network 220 can configure the corresponding mgl according to the measurement object. However, in an AI / ML-based system, the network 220 may not know the cells that the UE can predict using the AI / ML model. Therefore, if the UE performs inter-frequency measurements, the mgl configured in the traditional measurement may not be appropriate, and therefore, the measurement gap needs to be reconfigured. For example, in model monitoring, the measurement value obtained by traditional measurement methods can generally be considered a benchmark / reference for model monitoring. The performance of the AI / ML model is judged by comparing the difference between the measured value and the predicted value through monitoring. Therefore, traditional measurement methods are also needed in model monitoring. The results of model monitoring can also be used to determine whether the system enables an AI / ML-based solution, namely model activation. Model activation is to enable an inactive model as an active model. In this case, the network 220 does not know the AI / ML model prediction results, for example, the network 220 does not know the candidate cells predicted by the AI / ML model. Therefore, during the monitoring process, the mgl previously used for traditional measurement configuration may be inappropriate, and therefore, it needs to be reconfigured.

[0244] In the measurement gap adjustment, the AI / ML model can infer the UE preferred measurement event (UE preferred measurement event), and the UE preferred measurement event (UE preferred measurement event) includes at least one of the following: inferred measurement gap configuration, inferred measurement result, inferred serving cell ID, inferred neighboring cell ID, inferred neighboring cell / serving cell beam, etc. The inferred measurement gap configuration includes preferred measurement gap length (preferred mgl), preferred measurement gap offset (preferred gapOffset), preferred measurement gap repetition period (preferred mgrp), and preferred measurement gap timing advance (preferred mgta). The inferred measurement results include predicted RSRP, predicted RSRQ, and predicted SINR.

[0245] As shown in FIG9 , the network 220 may update or modify the parameters of the measurement gap configuration (meas config) in the measurement configuration based on the inference amount (e.g., the UE preferred measurement event) inferred by the AI / ML model. The specific steps of the process are as follows:

[0246] Step F0. Measurement Configuration and Reporting. The UE (e.g., one or more of user equipment 10a, 10b, ..., and 10n) receives a measurement configuration message sent by network 220. The UE reports measurement quantities based on the content of the measurement configuration message, such as periodically reporting measurement quantities or triggering measurement quantities based on an event.

[0247] Step F1. Optionally, the network 220 sends an inference configuration, which may include a model ID or a functionality ID. For example, if the UE receives a configuration message including a functionality ID, such as a functionality ID for time-frequency beamforming prediction, the UE may independently run a model for the functionality it supports.

[0248] Step F2. The UE side runs the inference function of one or more AI / ML models, as shown in steps F2a, 2b, and 2c in Figure 9, but the inference results do not need to be reported.

[0249] Step F3. Network 220 sends a model monitoring configuration to the UE. The monitoring configuration may include the model monitoring method, the monitored object, the monitoring period, the monitoring comparison method, and the KPIs for the monitoring comparison. The UE monitors the model according to the monitoring configuration. The monitored object can be a model or a functionality, and a functionality can be associated with multiple models.

[0250] Step F4. When the trigger condition is met during the monitoring process (a periodic or event-triggered condition, used to trigger the UE to report a trigger message), the UE side may send a trigger message. The trigger message is used to trigger the network 220 to reconfigure the appropriate gap length (mgl) of the measurement. The trigger message carries UE preferred measurement event (UE preferred measurement event) information. The UE preferred measurement event (UE preferred measurement event) includes at least one of the following: inferred measurement gap configuration, inferred measurement result, inferred service cell ID, inferred neighboring cell ID, inferred neighboring cell beam or service cell beam. The cell indication contained in the UE preferred measurement event information is only used to assist the network 220 in updating the measurement gap configuration. The trigger message may be UAI, or MAC CE, or UCI. In addition, in order to distinguish it from the role of other predicted alternative / target cells, there may be the following ways:

[0251] Method 1: The trigger message in step F4 is a newly introduced dedicated message specifically used to trigger measurement reconfiguration (i.e., reconfigure measurement gap configuration) in model monitoring. This message carries UE preferred measurement event information. The message can be UCI, MAC CE, or RRC.

[0252] Method 2: The trigger message of step F4 reuses a traditional message, and the message carries the inferred cell. In order to distinguish, the trigger message has indication information, which is used to indicate that the inferred cell in the trigger message is used to trigger measurement reconfiguration in model monitoring (i.e., reconfiguration of measurement gap configuration).

[0253] In some cases, the trigger message is reported when the model monitor is enabled, such as when the model monitor is enabled by an event trigger or when the model monitor is enabled periodically. Alternatively, the trigger message can be reported based on a reporting rule. The reporting rule is the same as described in Example 1. The above steps are not in any logical order, and all steps are optional.

[0254] Step F5. The network 220 receives the message of step F4 and performs measurement reconfiguration according to the content of the trigger message of step F4 and sends a measurement reconfiguration message to the UE. The measurement reconfiguration message includes new measurement gap parameters, such as mgl, etc.

[0255] In addition, in some examples, the measurement gap parameter in step F5, such as the mgl, is temporary and can be expressed as "T-mgl", that is, it is only valid in one monitoring comparison or monitoring process.

[0256] In addition, in some examples, the reporting of the UE preferred measurement event information may be reported by the UE itself, or may be reported according to some rules, which may be predefined.

[0257] Example 5: UE capability reporting:

[0258] As discussed above, based on the inference results of the AI / ML model, a UE (e.g., one or more of user equipment 10a, 10b, ..., and 10n) can support modification of configured measurement event trigger parameters, such as TTT, threshold, hysteresis (Hys), offset, etc. To provide the appropriate configuration to the UE, network 220 needs to know whether the UE has such capabilities. Several possible methods are described below to support the UE in reporting its capabilities.

[0259] Method 1: Static capability reporting:

[0260] As shown in Figure 10, the UE has the ability to modify the parameters of the configured measurement event trigger configuration, can actively report the UE capability, or receive a capability request query from the network 220, and report the capability to the network 220 by responding to the capability request query. This capability is a static capability, that is, after accessing the network, it only needs to be reported once.

[0261] Step G1: The network 220 sends a capability query message (UECapabilityEnquiry) to the UE (e.g., one or more of the user equipments 10a, 10b, ..., and 10n) to request the capabilities of the UE. The capability query message carries a "Modify configured measurement event trigger configuration parameters - request" information element (IE) for querying whether the UE supports modifying the configured measurement event trigger configuration parameters.

[0262] Step G2: After receiving the UECapabilityEnquiry message, the UE reports the UE capability information to the network 220. The UE capability message carries an indication of "modifying parameters of the configured measurement event trigger configuration." For example, when the indication of "modifying parameters of the configured measurement event trigger configuration" is set to "support," it indicates that the UE supports modifying parameters of the configured measurement event trigger configuration.

[0263] The UE capability message may be included in the UE-NR-Capability message, carried in a capability query message (UECapabilityEnquiry) and uploaded to the network 220:

[0264] Table 6: UE-NR-Capability information element

[0265] Table 7: Capability query message UECapabilityEnquiry

[0266] Method 2: Dynamic capability reporting:

[0267] The UE has the ability to modify the parameters of the measurement event trigger configuration, and can query the capability through the RRC reconfiguration message (RRCReconfiguration), and / or report the capability to the network 220 through the RRC reconfiguration complete message (RRCReconfigurationComplete). This capability is a dynamic capability and can be reported to the network 220 multiple times. The process is shown in Figure 11:

[0268] Step H1: The network 220 sends an RRC reconfiguration message to the UE, carrying a "Modify parameters of configured measurement event trigger configuration - request" information element, which is used to query whether the UE supports modifying the parameters of the configured measurement event trigger configuration.

[0269] Step H2: After the UE receives the RRC reconfiguration message, if the UE supports modification of the configured measurement event trigger, the UE carries an indication of "modified parameters of the configured measurement event trigger configuration" in an RRC reconfiguration complete message (RRCReconfigurationComplete) and reports it to the network 220. For example, when "modified parameters of the configured measurement event trigger configuration" is set to "support", it indicates that the UE supports modification of the configured measurement event trigger configuration parameters.

[0270] Prediction is an AI / ML model capability, not a UE capability. The UE only needs to support model inference.

[0271] Referring to Figure 12, UE 10 may include a processor 11a, a memory 12a, and a transceiver 13a. The processor 11a is configured to invoke and execute a computer program stored in the memory 12a, causing the UE 10, in which the processor 11 is installed, to perform the disclosed methods, steps, and / or UE functions. UE 10 is an example of a UE described herein (e.g., one of UEs 10a-10n). The transceiver 13a may include baseband circuitry and radio frequency (RF) circuitry.

[0272] Referring to Figure 13 , network node 200 is a network device that may include a processor 21a, a memory 22a, and a transceiver 23a. The processor 21a is configured to invoke and execute a computer program stored in the memory 22a, causing the network node 200, in which the processor 21a is installed, to perform methods, steps, and / or functions of a network node. Network node 200 is an example of a network 220, a CN network entity, a network node, a radio node, a base station, or a gNB described herein. The transceiver 23a may include baseband circuitry and radio frequency (RF) circuitry.

[0273] 14 , the present embodiment further provides a chip 70. This chip 70 may correspond to the user device 10 in the present embodiment, and the chip 70 may implement the corresponding processes implemented by the user device 10 in the various methods in the present embodiment. The chip 70 includes a processor 71, which may call and execute computer programs from memory to implement the methods in the present embodiment.

[0274] Optionally, the chip 70 may further include a memory 72. The processor 71 may call and execute a computer program from the memory 72 to implement the method in the embodiment of the present application.

[0275] The memory 72 may be a separate device independent of the processor 71 , or may be integrated into the processor 71 .

[0276] Optionally, the chip 70 may further include an input interface 73. The processor 71 may control the input interface 73 to communicate with other devices or chips, and specifically, may obtain information or data sent by other devices or chips.

[0277] Optionally, the chip 70 may further include an output interface 74. The processor 71 may control the output interface 74 to communicate with other devices or chips, and specifically, may output information or data to other devices or chips.

[0278] 15 , another embodiment of the present application further provides another chip 80. This chip 80 may correspond to the network node 200 in the embodiment of the present application, and this chip 80 may implement the corresponding processes implemented by the network node 200 in each method in the embodiment of the present application. This chip 80 includes a processor 81, which may call and execute computer programs from a memory 82 to implement the methods in the embodiment of the present application.

[0279] Optionally, the chip 80 may further include a memory 82. The processor 81 may call and execute a computer program from the memory 82 to implement the method in the embodiment of the present application.

[0280] The memory 82 may be a separate device independent of the processor 81 , or may be integrated into the processor 81 .

[0281] Optionally, the chip 80 may further include an input interface 83. The processor 81 may control the input interface 83 to communicate with other devices or chips, and specifically, may obtain information or data sent by other devices or chips.

[0282] Optionally, the chip may further include an output interface 84. The processor 81 may control the output interface 84 to communicate with other devices or chips, and specifically, may output information or data to other devices or chips.

[0283] Those skilled in the art will appreciate that the above-mentioned embodiments are specific embodiments for implementing the present application, and that in actual applications, various changes may be made thereto in form and detail without departing from the spirit and scope of the present application.

Claims

1. A measurement event prediction method based on an artificial intelligence (AI) / ML model, executed in a user device, characterized in that: include: Report the inference amount of the AI / ML model, where the inference amount includes at least one of the following: one or more triggering events at one or more predicted future moments, parameters of the inferred measurement event trigger configuration, adjustments to the parameters of the inferred measurement event trigger configuration, parameters of the inferred measurement gap configuration, and time information.

2. The method according to claim 1, characterized in that ,The method also includes: reporting an event identifier.

3. The method according to claim 1, characterized in that The method also includes receiving a configuration message for AI / ML model reasoning, the configuration message including a reporting configuration, and the reporting configuration including at least one of the following: reasoning amount, measurement time identifier ID, reporting rule, and measurement amount.

4. The method according to claim 1, characterized in that The reporting of the inference amount is based on a time-based reporting rule, wherein the time-based reporting rule includes at least one of the following: a start time, a duration, an end time, an offset time, and an interval between two reporting times; or The reporting of the inference amount is based on a reporting rule for the amount of inference amount; or The reporting of the inference amount is based on a reporting rule of a difference between the inference amount and a traditional configuration parameter.

5. The method according to claim 1, further comprising monitoring the performance of the AI / ML model; The performance monitoring includes: Compare the one or more measurement events reported at one or more future moments inferred by the AI / ML model or other means with the one or more measurement events reported at one or more future moments obtained by a traditional measurement method.

6. The method according to claim 5, characterized in that The method also includes determining an accuracy of the performance monitoring of the AI / ML model; wherein the accuracy is based on a threshold, and the threshold is based on at least one of: time and number of times.

7. A measurement event prediction method based on an artificial intelligence (AI) / ML model, executed in a user device, characterized by: updating / modifying parameters of a measurement event trigger and / or a measurement gap configuration based on an inference obtained by the AI / ML model, wherein the inference includes at least one of the following: one or more predicted triggering events at one or more future moments, inferred parameters of the measurement event trigger configuration, an adjustment amount of the inferred parameters of the measurement event trigger configuration, inferred parameters of the measurement gap configuration, and time information; Report at least one of the following information: an event identifier, parameter modification indication information, and the inference amount, wherein the parameter modification indication information is used to indicate at least one of the following: the update / modification of the parameter configured by the measurement event trigger by the user equipment, or the measurement amount reported according to the updated / modified parameter configured by the measurement event trigger.

8. The method according to claim 7, characterized in that The method also includes receiving a configuration message for AI / ML model reasoning, where the configuration message includes a reporting configuration, and the reporting configuration includes at least one of the following: an inference amount, a measurement time identifier ID, a reporting rule, and a measurement amount.

9. The method according to claim 7, characterized in that Also includes: receiving indication information of allowing modification, where the indication information is used to indicate that the user equipment is allowed to modify parameters of the measurement event triggering and / or measurement gap configuration.

10. The method according to claim 9, characterized in that The indication information is carried in a system message, a paging message, or a dedicated message.

11. The method according to claim 10, characterized in that The indication information is carried in the system message or the paging message, wherein the indication information is based on at least one of the following: per model level, per functionality level, per cell level, per RAN notification area level, per TA level, per PLMN level, per AI-specific area level, and per UE level.

12. The method according to claim 10, characterized in that The dedicated message carries the indication information, where the indication information is based on at least one of the following: per model level, per functionality level, per cell level, per RAN notification area level, per TA level, per PLMN level, per AI-specific area level, and per parameter level.

13. The method according to claim 7, characterized in that The method also includes performance monitoring of the AI / ML model, wherein the performance monitoring includes at least one of the following: comparing one or more measurement events reported at one or more future moments inferred by the AI / ML model or other means with one or more measurement events reported at one or more future moments obtained by traditional measurement methods.

14. The method according to claim 13, characterized in that The method also includes determining an accuracy of the performance monitoring of the AI / ML model; wherein the accuracy is based on a threshold, and the threshold is based on at least one of: time and number of times.

15. The method according to claim 7, characterized in that Also includes: receiving a capability request query for querying whether the user equipment supports modifying parameters of the measurement event triggering configuration; and Reporting a capability message of the user equipment in response to the capability request query, the capability message being used to indicate whether the user equipment supports modifying parameters of the measurement event triggering configuration.

16. The method according to claim 15, characterized in that The capability request query is a capability query message carrying a "modify configured measurement event trigger configuration parameters - request" information element; and The capability message is a message carrying an indication of "modifying parameters of a configured measurement event triggering configuration".

17. The method according to claim 15, characterized in that The capability request query is an RRC reconfiguration message; and The capability message is an RRC reconfiguration complete message.

18. A measurement event prediction method based on artificial intelligence (AI) / ML model, executed in a base station, characterized in that: include: Receive inference data from an AI / ML model, where the inference data includes at least one of the following: one or more triggering events at one or more predicted future moments, parameters of the inferred measurement event trigger configuration, adjustments to the parameters of the inferred measurement event trigger configuration, parameters of the inferred measurement gap configuration, and time information.

19. The method according to claim 18, characterized in that ,The method also includes: receiving a reporting event identifier (event ID).

20. The method according to claim 18, characterized in that The method also includes sending a configuration message for AI / ML model reasoning, the configuration message including a reporting configuration, and the reporting configuration including at least one of the following: reasoning amount, measurement time identifier ID, reporting rule, and measurement amount.

21. The method according to claim 18, characterized in that The reporting of the inference amount is based on a time-based reporting rule, wherein the time-based reporting rule includes at least one of the following: a start time, a duration, an end time, an offset time, and an interval between two reporting times; or The reporting of the inference amount is based on a reporting rule for the amount of inference amount; or The reporting of the inference amount is based on a reporting rule of a difference between the inference amount and a traditional configuration parameter.

22. The method of claim 18, further comprising monitoring the performance of the AI / ML model; The performance monitoring includes: Compare the one or more measurement events reported at one or more future moments inferred by the AI / ML model or other means with the one or more measurement events reported at one or more future moments obtained by a traditional measurement method.

23. The method according to claim 22, characterized in that The method also includes determining an accuracy of the performance monitoring of the AI / ML model; wherein the accuracy is based on a threshold, and the threshold is based on at least one of: time and number of times.

24. A measurement event prediction method based on an artificial intelligence (AI) / ML model, executed in a base station, characterized by: receiving at least one of the following information: an event identifier (event ID), parameter modification indication information, and the inference amount, wherein the parameter modification indication information is used to indicate at least one of the following: the update / modification by the user equipment of the parameter configured for triggering the measurement event, or the measurement amount reported based on the updated / modified parameter configured for triggering the measurement event, the inference amount including at least one of the following: one or more triggering events predicted at one or more future moments; Parameters of the inferred measurement event trigger configuration, adjustment amounts of the inferred measurement event trigger configuration parameters, parameters of the inferred measurement gap configuration, and time information.

25. The method according to claim 24, characterized in that The method also includes sending a configuration message for AI / ML model reasoning, where the configuration message includes a reporting configuration, and the reporting configuration includes at least one of the following: an inference amount, a measurement time identifier ID, a reporting rule, and a measurement amount.

26. The method according to claim 24, characterized in that The method further comprises: Sending indication information of allowing modification, where the indication information is used to indicate that the user equipment is allowed to modify parameters of the measurement event triggering and / or measurement gap configuration.

27. The method according to claim 26, characterized in that The indication information is carried in a system message, a paging message, or a dedicated message.

28. The method according to claim 24, characterized in that The method further includes: receiving a reporting event identifier (event ID).

29. The method according to claim 27, characterized in that The indication information is carried in the system message or the paging message, wherein the indication information is based on at least one of the following: per model level, per functionality level, per cell level, per RAN notification area level, per TA level, per PLMN level, per AI-specific area level, and per UE level.

30. The method according to claim 27, wherein The dedicated message carries the indication information, where the indication information is based on at least one of the following: per model level, per functionality level, per cell level, per RAN notification area level, per TA level, per PLMN level, per AI-specific area level, and per parameter level.

31. The method according to claim 24, wherein The method also includes performance monitoring of the AI / ML model, wherein the performance monitoring includes at least one of the following: comparing one or more measurement events reported at one or more future moments inferred by the AI / ML model or other means with one or more measurement events reported at one or more future moments obtained by traditional measurement methods.

32. The method according to claim 31, characterized in that The method also includes determining an accuracy of the performance monitoring of the AI / ML model; wherein the accuracy is based on a threshold, and the threshold is based on at least one of: time and number of times.

33. The method according to claim 24, characterized in that Also includes: Send a capability request query to query whether the user equipment supports modifying the parameters of the measurement event trigger configuration number; and A capability message of the user equipment is received, where the capability message is used to indicate whether the user equipment supports modifying parameters of the measurement event triggering configuration.

34. The method according to claim 33, characterized in that The capability request query is a capability query message carrying a "modify configured measurement event trigger configuration parameters - request" information element; and The capability message is a message carrying an indication of "modifying parameters of a configured measurement event triggering configuration".

35. The method according to claim 33, wherein The capability request query is an RRC reconfiguration message; and The capability message is an RRC reconfiguration complete message.

36. A wireless communication device, characterized in that: include: A processor configured to call and execute a computer program stored in a memory so that a device equipped with the processor executes the method according to any one of claims 1 to 35.

37. A chip, characterized in that: include: A processor configured to call and execute a computer program stored in a memory so that a device equipped with the processor executes the method according to any one of claims 1 to 35.

38. A computer-readable storage medium, characterized in that A computer program is stored therein, wherein the computer program causes a computer to execute the method of any one of claims 1 to 35.

39. A computer program product, characterized in that comprising a computer program, wherein the computer program causes a computer to execute the method of any one of claims 1 to 35.

Images