Reporting for aperiodic ai / ML model performance monitoring

Abstract

Method, comprising triggering a terminal by transmitting a first trigger message to the terminal at a first trigger time, wherein the first trigger message triggers the terminal to measure a set of one or more monitoring reference signals at each of one or more time instances to obtain a monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances, a monitoring metric based on the monitoring measurement instances for the one or more time instances without triggering the terminal to report the monitoring metric; triggering the terminal by transmitting a second trigger message to the terminal at a second trigger time being later than the first trigger time, wherein the second trigger message triggers the terminal to report the monitoring metric for the one or more time instances in a monitoring report.

Description

[0001] REPORTING FOR APERIODIC AI / ML MODEL PERFORMANCE MONITORING

[0002] Technical Field

[0003] The present disclosure relates to reporting for aperiodic AI / ML model performance monitoring.

[0004] Abbreviations

[0005] 3GPP 3rdGeneration Partnership Project

[0006] 4G / 5G / 6G 4th / 5th / 6thGeneration

[0007] Al Artificial Intelligence

[0008] AP Aperiodic

[0009] BM Beam Management

[0010] CSI Channel State Information

[0011] DCI Downlink Control Information

[0012] DMRS Demodulation Reference Signal

[0013] ID Identifier

[0014] MAC-CE Medium Access Control-Control Element

[0015] ML Machine Learning

[0016] NW Network

[0017] PDCCH Physical Downlink Control Channel

[0018] PTRS Phase Tracking Reference Signal

[0019] PUSCH Physical Uplink Shared Channel

[0020] RRC Radio Resource Control

[0021] RS Reference Signal

[0022] RSRP Reference Signal Received Power

[0023] SRS Sounding Reference Signal

[0024] TRS Time Tracking Reference Signal

[0025] UE User Equipment

[0026] UL UplinkBackground

[0027] In UE-sided assisted AI / ML model monitoring, a AI / ML model is employed on the UE. The UE performs continuously (typically periodically) measurements of some RSs for inference by the AI / ML model and reports the outcome of the AI / ML model to the network (e.g. a base station such as a gNB or eNB). In addition, the UE may perform AP CSI measurements. These measurements are triggered by a DCI message. Here, the UE measures RSs (“monitoring RSs”) which are typically different from those measured for input the measurement results into AI / ML model for inference. Typically, AP CSI measurements are performed less frequent than measurements of RSs for inference.

[0028] The AI / ML model employed on the UE outputs predictions for the monitoring RSs. For monitoring the performance of the inference by the AI / ML model, the predictions of the monitoring RSs from the AI / ML model are compared with the measurement results of the monitoring RSs. These comparisons (sub-metrics), or a summary of plural comparisons (a combination of such comparisons) are reported to the network as a monitoring metric in a monitoring report such that the network may adapt its BM based on the monitoring report.

[0029] Summary

[0030] It is an object to improve the prior art.

[0031] According to a first aspect, there is provided an apparatus, comprising

[0032] first means for triggering a terminal by transmitting a first trigger message to the terminal at a first trigger time, wherein the first trigger message triggers the terminal to measure a set of one or more monitoring reference signals at each of one or more time instances to obtain a monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances, a monitoring metric based on the monitoring measurement instances for the one or more time instances without triggering the terminal to report the monitoring metric;

[0033] second means for triggering the terminal by transmitting a second trigger message to the terminal at a second trigger time being later than the first trigger time, wherein the second trigger message triggers the terminal to report the monitoring metric for the one or more time instances in a monitoring report.Thus, the following technical advantages may be achieved:

[0034] • increased flexibility in DCI scheduling;

[0035] • computing requirements to the UE may be relaxed;

[0036] • different time domain behaviors for the measurement and preparing the monitoring report may be supported.

[0037] The apparatus may further comprise

[0038] means for configuring the terminal with at least one of a first trigger state or a second trigger state, wherein

[0039] if the means for configuring configures the terminal with the first trigger state, the first trigger message indicates the first trigger state;

[0040] if the means for configuring configures the terminal with the second trigger state, the second trigger message indicates the second trigger state;

[0041] the first trigger state triggers the terminal to measure the set of the one or more monitoring reference signals at each of the one or more time instances to obtain the monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances, the monitoring metric based on the monitoring measurement instances for the one or more time instances without triggering the terminal to report the monitoring metric;

[0042] the second trigger state triggers the terminal to report the monitoring metric for the one or more time instances in the monitoring report.

[0043] Thus, signaling effort may be reduced. For example, the existing trigger state framework may be exploited.

[0044] If the means for configuring is configured to configure the terminal with the first trigger state, the second trigger message may comprise a reference to the first trigger state.

[0045] Thus, the signaling effort may be reduced.

[0046] The second trigger state may comprise a reference to the first trigger state.

[0047] Thus, the signaling effort for the second trigger message may be reduced.The second trigger state may comprise a reference to the set of the one or more monitoring reference signals.

[0048] Thus, the terminal can derive which monitoring metric to report.

[0049] The second trigger state may comprise a reference to a time series of inference reports, and the first trigger state may trigger the terminal to calculate, for each of the one or more time instances, a monitoring sub-metric for the respective time instance based on the monitoring measurement instance for the respective time instance and a prediction of the one or more monitoring reference signals reported in a respective one of the inference reports.

[0050] Thus, the terminal is informed about the inference reports to be used for the (sub-)metric calculation.

[0051] The first trigger message may indicate a monitoring window, wherein the one or more time instances are within the monitoring window;

[0052] the second trigger time may be at a time instance after the monitoring window has passed after the first trigger time.

[0053] Thus, network may determine the monitoring window for the metric.

[0054] The second trigger time may be at a time instance after the monitoring window and a metric calculation time have passed after the first trigger time;

[0055] the metric calculation time may be predefined; or the apparatus may comprise means for determining the metric calculation time based on the monitoring metric to be calculated by the terminal.

[0056] Thus, network may take the metric calculation time into account such that the terminal may respond with the monitoring report shortly after the second trigger time.

[0057] The apparatus may further comprise

[0058] means for receiving a readiness indication from the terminal after the monitoring window has passed after the first trigger time; wherein

[0059] the second trigger time may be when the readiness indication has been received.Thus, network gets informed that the monitoring report is ready such that the terminal may safely respond with the monitoring report shortly after the second trigger time.

[0060] The apparatus may further comprise

[0061] means for indicating, to the terminal, a criterion to be fulfilled by the monitoring metric; means for checking whether a fulfilment indication is received from the terminal; wherein

[0062] the fulfillment indication may indicate that the criterion is fulfilled by the monitoring metric;

[0063] the second means for triggering may be configured to transmit the second trigger message in response to checking that the fulfillment indication is received.

[0064] Thus, monitoring report may be more useful for the network and, nevertheless, the terminal may safely respond with the monitoring report shortly after the second trigger time.

[0065] The apparatus may further comprise

[0066] means for inhibiting the second means for triggering from transmitting the second trigger message if the fulfillment indication is not received.

[0067] Thus, it may be avoided that the terminal needs a long time after receipt of the second trigger message to provide the monitoring report, causing obstacles in PLISCH transmission.

[0068] The criterion may comprise at least one of a total number of calculated monitoring submetrics or a minimum number of calculated monitoring sub-metrics within a given time span. Thus, the monitoring report comprises useful information for the network.

[0069] According to a second aspect, there is provided an apparatus, comprising

[0070] first means for receiving a first trigger message from a network at a first trigger time, wherein the first trigger message triggers a terminal to measure a set of one or more monitoring reference signals at each of one or more time instances to obtain a monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances, a monitoring metric based on the monitoring measurement instances for the time instances without triggering the terminal to report the monitoring metric;

[0071] first means for instructing the terminal, if the first trigger message is received, to measure the set of the one or more monitoring reference signals at each of the one or moretime instances to obtain the monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances, the monitoring metric based on the monitoring measurement instances for the one or more time instances;

[0072] second means for receiving a second trigger message from the network at a second trigger time later than the first trigger time, wherein the second trigger message triggers the terminal to report the monitoring metric for the one or more time instances in a monitoring report;

[0073] second means for instructing the terminal, if the second trigger message is received, to report the monitoring metric for the one or more time instances to the network in the monitoring report.

[0074] Thus, the following technical advantages may be achieved:

[0075] • increased flexibility in DCI scheduling;

[0076] • computing requirements to the UE may be relaxed;

[0077] • different time domain behaviors for the measurement and preparing the monitoring report may be supported.

[0078] The terminal may be configured with at least one of a first trigger state or a second trigger state;

[0079] if the terminal is configured with the first trigger state, the first trigger message may indicate the first trigger state and the first means for instructing may be configured to instruct the terminal according to the first trigger state;

[0080] if the terminal is configured with the second trigger state, the second trigger message may indicate the second trigger state and the second means for instructing may be configured to instruct the terminal according to the second trigger state;

[0081] the first trigger state may trigger the terminal to measure the set of the one or more monitoring reference signals at each of the one or more time instances to obtain the monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances, the monitoring metric based on the monitoring measurement instances for the one or more time instances without triggering the terminal to report the monitoring metric in the monitoring report;

[0082] the second trigger state may trigger the terminal to report the monitoring metric for the one or more time instances in the monitoring report.Thus, signaling effort may be reduced. For example, the existing trigger state framework may be exploited.

[0083] If the first trigger message indicates the first trigger state, the second trigger message may comprise a reference to the first trigger state; and the apparatus may further comprise means for identifying the monitoring metric to be reported in the monitoring report based on the reference to the first trigger state.

[0084] Thus, the signaling effort may be reduced.

[0085] The second trigger state may comprise a reference to the first trigger state, and the apparatus may further comprise

[0086] means for identifying the monitoring metric to be reported in the monitoring report based on the reference to the first trigger state.

[0087] Thus, the signaling effort for the second trigger message may be reduced.

[0088] The second trigger state may comprise a reference to the set of the one or more monitoring reference signals, and the apparatus may further comprise

[0089] means for identifying the monitoring metric to be reported in the monitoring report based on the reference to the set of the one or more monitoring reference signals.

[0090] Thus, the terminal can derive which monitoring metric to report.

[0091] The first trigger state may comprise a reference to a time series of inference reports;

[0092] the second trigger state may comprise the reference to the inference report;

[0093] the first means for instructing may be configured to instruct the terminal to calculate, for each of the one or more time instances, a monitoring sub-metric for the respective time instance based on the monitoring measurement instance for the respective time instance and a prediction of the one or more monitoring reference signals reported in one of the inference reports;

[0094] the means for identifying may be configured to identify the monitoring metric to be reported in the monitoring report based on the reference to the set of the one or more monitoring reference signals and the reference to the time series of inference reports.Thus, the terminal is informed about the inference reports to be used for the (sub-)metric calculation.

[0095] The first trigger message may indicate a monitoring window;

[0096] the first means for instructing may be configured to instruct the terminal, if the first trigger message is received, to measure the set of the one or more monitoring reference signals at each of the one or more time instances in the monitoring window to obtain the monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances in the monitoring window, the monitoring metric for the one or more time instances based on the monitoring measurement instances for the one or more time instances in the monitoring window;

[0097] the second trigger time may be a time instance being after the monitoring window and a metric calculation time have passed after the first trigger time.

[0098] Thus, network may determine the monitoring window for the metric.

[0099] The first trigger message may indicate a monitoring window;

[0100] the first means for instructing may be configured to instruct the terminal, if the first trigger message is received, to measure the set of the one or more monitoring reference signals at each of the one or more time instances in the monitoring window to obtain the monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances in the monitoring window, the monitoring metric for the one or more time instances based on the monitoring measurement instances for the one or more time instances in the monitoring window; and the apparatus may further comprise means for monitoring whether the terminal is ready to report the monitoring metric for the one or more time instances in the monitoring window in the monitoring report;

[0101] means for indicating to the network, by providing a readiness indication, that the terminal is ready to report the monitoring metric for the one or more time instances in the monitoring window in the monitoring report, wherein

[0102] the second trigger time may be later than the providing the readiness indication.

[0103] Thus, network gets informed that the monitoring report is ready such that the terminal may safely respond with the monitoring report shortly after the second trigger time.

[0104] The apparatus may further comprisemeans for receiving, from the network, a criterion to be fulfilled by the monitoring metric for the one or more time instances;

[0105] means for checking whether the criterion is fulfilled by the monitoring metric for the one or more time instances;

[0106] means for indicating to the network, by transmitting a fulfillment indication, that the criterion is fulfilled by the monitoring metric for the one or more time instances; wherein the second means for receiving may be configured to receive the second trigger message in response to the transmitting the fulfillment indication.

[0107] Thus, monitoring report may be more useful for the network and, nevertheless, the terminal may safely respond with the monitoring report shortly after the second trigger time.

[0108] The apparatus may further comprise

[0109] means for prohibiting the second means for receiving to receive the second trigger message if the fulfillment indication is not transmitted.

[0110] Thus, it may be avoided that the terminal needs a long time after receipt of the second trigger message to provide the monitoring report, causing obstacles in PLISCH transmission.

[0111] The criterion may comprise at least one of a total number of calculated monitoring submetrics or a minimum number of calculated monitoring sub-metrics within a given time span.

[0112] Thus, the monitoring report comprises useful information for the network.

[0113] The apparatus may further comprise

[0114] means for inhibiting the second means for instructing from instructing the terminal to report the monitoring metric for the one or more time instances to the network if the second trigger message is not received.

[0115] Thus, PLISCH scheduling of the reporting a long time in advance may be avoided, causing difficulties in other PLISCH scheduling.

[0116] According to a third aspect, there is provided an method, comprising

[0117] triggering a terminal by transmitting a first trigger message to the terminal at a first trigger time, wherein the first trigger message triggers the terminal to measure a set of one ormore monitoring reference signals at each of one or more time instances to obtain a monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances, a monitoring metric based on the monitoring measurement instances for the one or more time instances without triggering the terminal to report the monitoring metric;

[0118] triggering the terminal by transmitting a second trigger message to the terminal at a second trigger time being later than the first trigger time, wherein the second trigger message triggers the terminal to report the monitoring metric for the one or more time instances in a monitoring report.

[0119] The method may further comprise

[0120] configuring the terminal with at least one of a first trigger state or a second trigger state, wherein

[0121] if the terminal is configured with the first trigger state, the first trigger message indicates the first trigger state;

[0122] if the terminal is configured with the second trigger state, the second trigger message indicates the second trigger state;

[0123] the first trigger state triggers the terminal to measure the set of the one or more monitoring reference signals at each of the one or more time instances to obtain the monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances, the monitoring metric based on the monitoring measurement instances for the one or more time instances without triggering the terminal to report the monitoring metric;

[0124] the second trigger state triggers the terminal to report the monitoring metric for the one or more time instances in the monitoring report.

[0125] If the terminal is configured with the first trigger state, the second trigger message may comprise a reference to the first trigger state.

[0126] The second trigger state may comprise a reference to the first trigger state.

[0127] The second trigger state may comprise a reference to the set of the one or more monitoring reference signals.

[0128] The second trigger state may comprise a reference to a time series of inference reports, andthe first trigger state may trigger the terminal to calculate, for each of the one or more time instances, a monitoring sub-metric for the respective time instance based on the monitoring measurement instance for the respective time instance and a prediction of the one or more monitoring reference signals reported in a respective one of the inference reports.

[0129] The first trigger message may indicate a monitoring window, wherein the one or more time instances are within the monitoring window;

[0130] the second trigger time may be at a time instance after the monitoring window has passed after the first trigger time.

[0131] The second trigger time may be at a time instance after the monitoring window and a metric calculation time have passed after the first trigger time;

[0132] the metric calculation time may be predefined; or the method may comprise determining the metric calculation time based on the monitoring metric to be calculated by the terminal.

[0133] The method may further comprise

[0134] receiving a readiness indication from the terminal after the monitoring window has passed after the first trigger time; wherein

[0135] the second trigger time may be when the readiness indication has been received.

[0136] The method may further comprise

[0137] indicating, to the terminal, a criterion to be fulfilled by the monitoring metric; checking whether a fulfilment indication is received from the terminal; wherein the fulfillment indication may indicate that the criterion is fulfilled by the monitoring metric;

[0138] the second trigger message may be transmitted in response to checking that the fulfillment indication is received.

[0139] The method may further comprise

[0140] inhibiting the triggering the transmitting the second trigger message if the fulfillment indication is not received.The criterion may comprise at least one of a total number of calculated monitoring submetrics or a minimum number of calculated monitoring sub-metrics within a given time span.

[0141] According to a fourth aspect, there is provided an method, comprising

[0142] receiving a first trigger message from a network at a first trigger time, wherein the first trigger message triggers a terminal to measure a set of one or more monitoring reference signals at each of one or more time instances to obtain a monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances, a monitoring metric based on the monitoring measurement instances for the time instances without triggering the terminal to report the monitoring metric;

[0143] instructing the terminal, if the first trigger message is received, to measure the set of the one or more monitoring reference signals at each of the one or more time instances to obtain the monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances, the monitoring metric based on the monitoring measurement instances for the one or more time instances;

[0144] receiving a second trigger message from the network at a second trigger time later than the first trigger time, wherein the second trigger message triggers the terminal to report the monitoring metric for the one or more time instances in a monitoring report;

[0145] instructing the terminal, if the second trigger message is received, to report the monitoring metric for the one or more time instances to the network in the monitoring report.

[0146] The terminal may be configured with at least one of a first trigger state or a second trigger state;

[0147] if the terminal is configured with the first trigger state, the first trigger message may indicate the first trigger state and the terminal may be instructed according to the first trigger state;

[0148] if the terminal is configured with the second trigger state, the second trigger message may indicate the second trigger state and the terminal may be instructed according to the second trigger state;

[0149] the first trigger state may trigger the terminal to measure the set of the one or more monitoring reference signals at each of the one or more time instances to obtain the monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances, the monitoring metric based on the monitoring measurement instances for the one or more time instances without triggering the terminal to report the monitoring metric in the monitoring report;the second trigger state may trigger the terminal to report the monitoring metric for the one or more time instances in the monitoring report.

[0150] If the first trigger message indicates the first trigger state, the second trigger message may comprise a reference to the first trigger state; and the method may further comprise identifying the monitoring metric to be reported in the monitoring report based on the reference to the first trigger state.

[0151] The second trigger state may comprise a reference to the first trigger state, and the method may further comprise

[0152] identifying the monitoring metric to be reported in the monitoring report based on the reference to the first trigger state.

[0153] The second trigger state may comprise a reference to the set of the one or more monitoring reference signals, and the method may further comprise

[0154] identifying the monitoring metric to be reported in the monitoring report based on the reference to the set of the one or more monitoring reference signals.

[0155] The first trigger state may comprise a reference to a time series of inference reports;

[0156] the second trigger state may comprise the reference to the inference report;

[0157] the terminal may be instructed to calculate, for each of the one or more time instances, a monitoring sub-metric for the respective time instance based on the monitoring measurement instance for the respective time instance and a prediction of the one or more monitoring reference signals reported in one of the inference reports;

[0158] the monitoring metric to be reported in the monitoring report may be identified based on the reference to the set of the one or more monitoring reference signals and the reference to the time series of inference reports.

[0159] The first trigger message may indicate a monitoring window;

[0160] the terminal may be instructed, if the first trigger message is received, to measure the set of the one or more monitoring reference signals at each of the one or more time instances in the monitoring window to obtain the monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances in the monitoring window, the monitoring metric for the one or more time instances based on themonitoring measurement instances for the one or more time instances in the monitoring window;

[0161] the second trigger time may be a time instance being after the monitoring window and a metric calculation time have passed after the first trigger time.

[0162] The first trigger message may indicate a monitoring window;

[0163] the terminal may be instructed, if the first trigger message is received, to measure the set of the one or more monitoring reference signals at each of the one or more time instances in the monitoring window to obtain the monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances in the monitoring window, the monitoring metric for the one or more time instances based on the monitoring measurement instances for the one or more time instances in the monitoring window; and the method may further comprise

[0164] monitoring whether the terminal is ready to report the monitoring metric for the one or more time instances in the monitoring window in the monitoring report;

[0165] indicating to the network, by providing a readiness indication, that the terminal is ready to report the monitoring metric for the one or more time instances in the monitoring window in the monitoring report, wherein

[0166] the second trigger time may be later than the providing the readiness indication.

[0167] The method may further comprise

[0168] receiving, from the network, a criterion to be fulfilled by the monitoring metric for the one or more time instances;

[0169] checking whether the criterion is fulfilled by the monitoring metric for the one or more time instances;

[0170] indicating to the network, by transmitting a fulfillment indication, that the criterion is fulfilled by the monitoring metric for the one or more time instances; wherein

[0171] the second trigger message may be received in response to the transmitting the fulfillment indication.

[0172] The method may further comprise

[0173] prohibiting to receive the second trigger message if the fulfillment indication is not transmitted.The criterion may comprise at least one of a total number of calculated monitoring submetrics or a minimum number of calculated monitoring sub-metrics within a given time span.

[0174] The method may further comprise

[0175] inhibiting instructing the terminal to report the monitoring metric for the one or more time instances to the network if the second trigger message is not received.

[0176] Each of the methods of the third or fourth aspects may be a method of reporting for aperiodic AI / ML model performance monitoring.

[0177] According to a fifth aspect, there is provided a computer program product comprising a set of instructions which, when executed on an apparatus, is configured to cause the apparatus to carry out the method according to any of the third or fourth aspects. The computer program product may be embodied as a computer-readable medium or directly loadable into a computer.

[0178] According to some example embodiments, at least one of the following advantages may be achieved:

[0179] • increased flexibility in DCI scheduling;

[0180] • computing requirements to the UE may be relaxed;

[0181] • different time domain behaviors for the measurement and preparing the monitoring report may be supported.

[0182] Brief description of the drawings

[0183] Further details, features, objects, and advantages are apparent from the following detailed description of the preferred example embodiments which is to be taken in conjunction with the appended drawings, wherein:

[0184] Fig. 1 illustrates an example of UE-sided assisted AI / ML model performance monitoring; Fig. 2 illustrates UE-sided assisted AI / ML model performance monitoring according to an example embodiment;

[0185] Fig. 3 illustrates UE-sided assisted AI / ML model performance monitoring according to an example embodiment;

[0186] Fig. 4 illustrates a signaling diagram according to an example embodiment;Fig. 5 illustrates association of trigger message 1 and trigger message 2 according to an example embodiment;

[0187] Fig. 6 illustrates association of trigger message 1 and trigger message 2 according to an example embodiment;

[0188] Fig. 7 shows an apparatus according to an example embodiment;

[0189] Fig. 8 shows a method according to an example embodiment;

[0190] Fig. 9 shows an apparatus according to an example embodiment;

[0191] Fig. 10 shows a method according to an example embodiment; and

[0192] Fig. 11 shows an apparatus according to an example embodiment.

[0193] Detailed description of certain example embodiments

[0194] Herein below, certain example embodiments are described in detail with reference to the accompanying drawings, wherein the features of the example embodiments can be freely combined with each other unless otherwise described. However, it is to be expressly understood that the description of certain example embodiments is given by way of example only, and that it is by no way intended to be understood as limiting the disclosure to the disclosed details.

[0195] Moreover, it is to be understood that the apparatus is configured to perform the corresponding method, although in some cases only the apparatus or only the method are described.

[0196] AI / ML based beam management

[0197] In Rel-18 / Rel-19, a framework for AI-ML for beam prediction is under development. It leverages AI / ML models to predict the best beam(s) based on a limited set of measurements of reference signals. There are two Sub-Use cases:

[0198] • Case 1: Spatial-Domain Prediction

[0199] • Beam prediction based on a limited set of measurements that does not contain any historical information

[0200] • Case 2: Time-Domain Prediction

[0201] • Beam prediction into the future based on a limited set of measurements that contains historical informationThe measurements and prediction may be based on two Beam Sets (and the corresponding reference signals):

[0202] • Set A: the complete set of beams over which the prediction will operate

[0203] • Set B: the set of beams whose measurements of reference signals are inputted to the AI / ML model (e.g., L1-RSRP, etc.)

[0204] Set B can be:

[0205] • Different from Set A (space-domain and time-domain prediction);

[0206] • A subset of Set A (space-domain and time-domain prediction); or

[0207] • Same as Set A (time-domain prediction)

[0208] UE-sided assisted AI / ML model monitoring

[0209] Some aspects concerning UE-sided assisted AI / ML model monitoring include the configuration of monitoring resource sets, the definition and calculation of AI / ML model performance monitoring metrics (e.g. beam prediction accuracy) and the linkage between monitoring and inference RS resources for calculating the AI / ML model performance monitoring metrics.

[0210] • Configuration of monitoring RS resources: Due to the significant burden imposed on both the NW (for RS transmission) and to the UE (for RS measurements), the measurement of monitoring RS resources is a quite demanding operation. A basic approach requires the NW to transmit the full Set A of beams, which may not be practical for large codebooks. The UE, in turn, may measure each beam of Set A before determining the ground truth values, leading to significant RS overhead. Selection of the subset of Set A beams to configure as monitoring RS resources is not straightforward. One possibility is to use the Top-K beams predicted by the UE and reported to the NW with a beam prediction report. However, configuring Top-K beams as monitoring RS resources introduces complications because the Top-K beams may change after each inference report. For aperiodic CSI reports, the UE is provided with a higher-layer configuration of a Trigger State List, where each trigger is linked to specific CSI resource settings.

[0211] Linkage between monitoring and inference report configurations: Dedicated resource set(s) for monitoring and report configuration for monitoring may be configured in adedicated CSI report configuration used for monitoring. Separate resource sets may be configured for monitoring and inference. Typically, the resources (RSs) in the two resource sets are different from each other. To identify the connection between the monitoring and inference resource sets, the UE may receive linking information between the two-resource set(s). In one example, the UE may receive a first linking identifier corresponding to the CSI reporting configuration for monitoring and a second linking identifier corresponding to the CSI reporting configuration for inference. If the first and second linking identifiers match, the UE can establish a connection between the RSs for monitoring and predicted Set A beams. This connection allows the UE to compare measurements from monitoring RS set and prediction obtained over Set A.

[0212] • Definition and reporting of AI / ML performance monitoring metric: For the definition and calculation of performance monitoring metrics (e.g. beam prediction accuracy) one defines at first the metric. In one example of beam prediction accuracy metric, this involves for each monitoring instance, determining whether the beams predicted (e.g. Top-1 beams predicted) to be reported with an inference report matches with the ground truth sample obtained from measurements of monitoring RS . The accuracy metric can be defined by considering the ratio of correct predictions to the total number of monitoring instances. When to report the monitoring results could involve periodic or semi-persistent reporting, or aperiodic reporting or event-triggered reporting and it may follow the reporting behavior defined in the CSI report configuration specific to the monitoring.

[0213] For the UE-sided assisted AI / ML performance monitoring of UE-side AI / ML model, when a dedicated report configuration for monitoring is configured to the UE, the dedicated report configuration for monitoring is linked / associated to an inference report configuration (e.g. for BM)-Case 1, BM-Case 2, CSI prediction and any other use case).

[0214] In Figs. 1 to 3, arrows downwards indicate signals from the network to the UE, and arrows upwards indicate signals from the UE to the network. The time advances from left to right.

[0215] Fig. 1 illustrates UE-sided assisted AI / ML model performance monitoring for an AP CSI report. For aperiodic (AP) CSI report, the UE may receive a trigger, such as a DCI trigger (e.g. DCI 0_1)) on PDCCH indicating that the UE should report a monitoring metric overPLISCH in uplink. As shown in Fig. 1, after the monitoring report is triggered by DCI, the UE may determine, for one or more time instances, a respective monitoring measurement instance by measuring the RSs of the monitoring resource set at these time instances. For example, as shown in Fig. 1, the time instances may be defined by a monitoring window.

[0216] Independently from the transmitting of the monitoring RSs by the network and the measuring of the monitoring RSs by the UE, the network transmits RSs (on set B of beams, for example) which are used for inference by the AI / ML model (indicated by dashed arrows in Figs. 1 to 3). The UE measures these RSs and sends respective inference reports to the network (a time series of inference reports). Typically, the inference reports are sent more frequently than the monitoring reports.

[0217] After each measurement of the monitoring RSs, the UE may compare the monitoring measurement instances with the prediction of the RSs in one of the inference reports from the AI / ML model employed on the UE to check if the prediction substantially fits to the monitoring measurement instance. Thus, a monitoring sub-metric is obtained for each of the time instances. These monitoring sub-metrics may be transmitted as a monitoring metric in a monitoring report to the network for further use. In order to save signalling effort, the UE may combine the monitoring sub-metrics of several time instances into a single monitoring metric which is reported in the monitoring report instead of or in addition to the monitoring submetrics. An example of such a monitoring metric obtained from plural monitoring sub-metrics is that x% of the measurements were substantially (i.e. within a certain range) equal to the predictions, and y% of the measurements were substantially different from the predictions.

[0218] Typically, only when the monitoring measurement instance is valid, the UE calculates monitoring metric(s) by comparing the prediction results and the measurements from monitoring RS resources. In one example, for the validation, the UE determines if a predefined monitoring measurement (e.g. the first or the last or some other monitoring measurement) occurs within a pre-defined time window from the latest inference report. In another example, predicted instances may be earlier than the inference report, in such case, a CSI reference resource may be used to determine the earliest time instance of the prediction. For the validation, the UE may determine if a predefined monitoring measurement occurs within a pre-defined time window from the CSI reference resource of the latest inference report.A AI / ML model performance monitoring metric (herein also shortly denoted “monitoring metric”) may be calculated either per time instance or per set of time instances (spanning a time window). In the latter case, the UE may be configured for calculating the AI / ML performance monitoring metric with a monitoring window, which entails calculating multiple monitoring RS measurements before the report is created. Therefore, the UE-sided assisted model monitoring may require a significant time gap (>500 ms) for measuring over plural time instances (the monitoring window), calculating the monitoring metric for the time instances, and preparing the monitoring report. However, the DCI scheduling of a CSI report (such as the monitoring report) has a timing dependency indicated by the K2 parameter, which is configured by RRC signalling. The K2 parameter indicates the time when the transmission of the report from the UE to the network is scheduled after the measurements are triggered by the DCI message. Typically, the duration indicated by the K2 parameter is much shorter (<10 ms) than the time gap required for the above described activities for obtaining the monitoring report. Increasing the K2 value to a very large value is not practical in this case as is it may cause scheduling problem on PUSCH transmission.

[0219] Some example embodiments tackle this problem and provide a method that supports Aperiodic (AP) reporting for US-sided assisted AI / ML model performance monitoring. It may support many (if not all) combinations on time domain behavior of the monitoring report.

[0220] According to some example embodiments, the network sends two triggers to obtain the monitoring report:

[0221] 1. a first trigger for measuring the monitoring RSs and the calculation of the monitoring metric without providing a monitoring report; and

[0222] 2. a second trigger for providing the monitoring report. The second trigger is sent when the monitoring report is ready for transmission or when the network may assume that the monitoring report is ready for transmission.

[0223] The first and second triggers may be explicit triggers, for example in a DCI message. However, in some example embodiments, at least on of the first trigger and the second trigger may refer to a trigger state contained in a trigger state list configured in the UE. Many of the following examples are described for the case that the triggers refer to respective trigger states, but the same behavior may be achieved by explicit triggers not referring to the trigger states.Trigger State Configuration:

[0224] The UE is configured with a trigger State List that contains at least one trigger state (first Trigger State, for example) that is associated with monitoring RS measurement, and the at least one trigger state can trigger one or more measurements of a monitoring RS resource set without triggering the actual CSI reporting of monitoring metric(s),

[0225] o The first Trigger State may be denoted as “monitoring RS measurement Trigger State”.

[0226] o In one variant a pre-defined Trigger State Id (e.g., ID = 1) is defined as the monitoring RS measurement Trigger State. In another variant, the network may configure, by a respective information element, which of the trigger states is the monitoring RS measurement Trigger State. That is, the respective trigger state may comprise an indication that it does not trigger the providing of a monitoring report.

[0227] o The monitoring RS measurement Trigger State configures the UE to measure multiple monitoring RS measurement instances within a monitoring window. o The monitoring RS measurement Trigger State also configures the UE perform metric calculation (based on the linked / associated inference report configuration) for each valid monitoring RS measurement instances.

[0228] The UE is also configured with another trigger state (second Trigger State) included in the trigger State List containing the CSI reporting configuration for monitoring metric reporting, and the CSI reporting configuration is associated (linked) with the monitoring RS resource set that may be triggered by the first Trigger State. Figs. 5 and 6 illustrate examples how the second trigger (second trigger state) may be associated with the first trigger (first trigger state).

[0229] As shown in the example of Fig. 5, the second trigger message or the second trigger state may comprise an explicit link to the first trigger message or first trigger state. Thus, the UE receiving the second trigger message (pointing to the second trigger state) knows from the explicit link, which monitoring report to transmit to the network.

[0230] As shown in the example of Fig. 6, both the first trigger state and the second trigger state may comprise a link to a respective CSI report configuration (CSI-Report #1 and CSI-Report#2). Each of these CSI report configurations refers to the inference CSI Report indicating a time series of inference reports. Each of the CSI report configurations contains a link to a Inference CSI report, also contained in the list of CSI report configurations. TheInference CSI report configuration configures inference reports of a time series. Each of these inference reports contains a prediction for the measurement of the monitoring RS. The monitoring metric to be reported to the network contains a comparison of the measurement of the monitoring RS and the prediction thereof or is based on such a comparison. Thus, the UE receiving the trigger message 2 pointing to the trigger state 2 may check its CSI report configuration and determine which other CSI report configuration refers to the same Inference CSI report and, thus, identify the monitoring report to be transmitted to the network. I.e., the first and second Trigger States may be implicitly associated by the monitoring metric such that CSI monitoring report triggered by the second Trigger State can be linked to the monitoring metric determined in monitoring RS measurements triggered by the first Trigger State.

[0231] In addition, the UE may check if the CSI report configurations are related to the same monitoring RS set, as indicated by the dashed arrow in Fig. 6.

[0232] For example, the CSI report triggered by the second Trigger State may be for Top-1 or Top-K beam prediction accuracy monitoring metric.

[0233] In Fig. 6, the trigger states contain a link to the respective CSI reports #1 and #2, respectively. Instead, in some examples, the trigger states contain the respective CSI reports #1 and #2 themselves, respectively. A link is considered to be equivalent to containing the information itself.

[0234] An advantage of the example of Fig. 6 over that of Fig. 5 is that the trigger states (or the respective trigger messages) need not to comprise an additional information element such as the linked CSI AP trigger state ID in Fig. 5. On the other hand, the computation effort for identifying the associated monitoring report may be higher in the example of Fig. 6.

[0235] Two example embodiments are described at greater detail with reference to Figs. 2 and 3. Figs. 2 and 3 are similar to Fig. 1 except that the triggering is different. Therefore, redundant description is omitted, and it is referred to the description of Fig. 1.

[0236]

[0237] The UE receives a first DCI message including monitoring RS measurement Trigger State that instructs the UE to perform one or more measurements of a monitoring RS resource set, depending on the indicated length of the monitoring window.

[0238] Based on the linked / associated inference report configuration, the UE may determine whether each monitoring RS measurement instances is valid or not. For each valid monitoring RS measurement instance, the UE performs metric calculation, otherwise the UE counts the monitoring RS measurement as a monitoring error instance in metric calculation.

[0239] The first DCI message does not instruct the UE to provide the monitoring report to the network.

[0240] After expiry of the monitoring window and of some metric calculation time for calculating the monitoring metric (based on comparison of RS measurement instances with predictions), the network sends and the UE receives a second DCI message including a reference to a second Trigger State that instructs the UE to report AI / ML model performance monitoring metric. First and second Trigger States are linked / associated as described with respect to Figs. 5 or 6, for example.

[0241] The metric calculation time may be predefined (e.g. by a standard defining a maximum time allowed for the UE to prepare the monitoring report). As another option, the network may determine the metric calculation time, based on the metric to be calculated.

[0242] Since the monitoring report should be finalized when the network transmits the second trigger message, K2 parameter indicating the time until the monitoring report is scheduled for transmission to the network on PUSCH may indicate a short time. Thus, scheduling of PUSCH is not unnecessarily impaired.

[0243] Example Embodiment 2 (AP Trigger for monitoring metric report after UE indication,

[0244] Example embodiment 2 is applicable when NW transmits periodically monitoring RSs (Periodic Monitoring RS resources).The UE receives a first DCI message including monitoring RS measurement Trigger State that instructs the UE to perform one or more measurements of a monitoring RS resource set without instructing to provide a monitoring report to the NW. However, the monitoring window is not fixed as in Fig. 2 but the network sets a condition when to prepare and provide the monitoring report. Typically, the condition is related to measurement of the monitoring RSs and / or the inference reports. For example, the condition may be that the total number of valid monitoring samples is equal to or larger than a certain threshold value. As another example, the condition may be that the frequency of valid monitoring samples (i.e. the number of valid monitoring samples over a certain time period) is larger than another threshold value. In this case, the monitoring time window duration varies depending on when the condition is fulfilled.

[0245] For the conditions mentioned above, the UE may determine for each of the measurement samples whether or not the measurement sample is valid as described above. However, other conditions, for example related to a maximum value, a minimum value, or an average value of the metric, are feasible, too.

[0246] When the UE determines that the condition is fulfilled, the UE indicates to the NW, by a fulfillment indication, that the condition is fulfilled and that it is ready to report monitoring RS measurement (to provide the monitoring report). For example, the fulfillment indication can be made with an UL MAC-CE. As another option, an event may be configured for this purpose.

[0247] When the network receives the fulfillment indication, the network transmits and the UE receives a second DCI message including a link to the second Trigger State that instructs the UE to report AI / ML model performance monitoring metric (to provide the monitoring report).

[0248] First and second Trigger States are linked / associated as their monitoring report configurations are both linked / associated to the same inference report configuration. In addition, in some example embodiments, for an association the ReportQuantity of the CSI Report for Monitoring configured by second Trigger States has to match the ReportQuantity of the CSI Report for Monitoring configured by first Trigger States. Upon receipt of the second trigger message pointing to the second trigger state, the UE reports the monitoring metric determined in the monitoring window configured, dependent on the condition, by the first Trigger State.In example embodiments, such as the example embodiments 1 and 2, the UE does not report the monitoring report before the second trigger message is received. Since the network transmits the second trigger message only when the network may assume (based on the fixed monitoring window and some metric calculation time, for example) or knows (based on the fulfillment indication, for example) that the monitoring report is ready for transmission, the value of K2 may be small such that PLISCH scheduling is not unnecessarily impaired.

[0249] Fig. 4 illustrates a signaling diagram according to some example embodiments. For simplicity, the inference report configuration and inference reporting are not represented in Fig. 4 because these activities run independently from the monitoring illustrated in Fig. 4. The actions are as follows:

[0250] Action 1 The UE receives an RRC configuration containing the higher layer parameter CSI- AperiodicTriggerStateList, which includes a list of aperiodic trigger states (CSI- AperiodicT riggerState). At least one of the trigger states in the CSI-AperiodicTriggerStateList contains a monitoring CSI-ReportConfig indicating the Resource Set ID for monitoring RS resources. The monitoring CSI-ReportConfig is linked / associated to an inference CSI-ReportConfig, and the UE may determine a monitoring RS measurement instance by associating an inference report to the received monitoring RS resources.

[0251] The UE is configured with a Trigger State List that associates each Trigger State to a specific RS resource set. In one variant, a pre-defined Trigger State Id (e.g., ID = 1) is defined as the monitoring RS measurement Trigger State. In another variant, the a Trigger State comprises an information element marking it as a monitoring RS measurement Trigger State. The information element is set by the network. The monitoring RS measurement Trigger State does not trigger reporting of the measurement results or of the monitoring metric calculated based on the measurement results.

[0252] A second Trigger State includes the CSI reporting configuration for monitoring metric reporting and triggers the reporting of monitoring metric related to what has been measured by the monitoring RS measurement Trigger State. A link between these two trigger states may be derived as described with respect to Figs. 5 and 6, for example.Action 2 First trigger message (first DCI Message)

[0253] The network sends the first DCI message to the UE with the CSI request field set to a Trigger State Id (e.g., ID = 1) to activate the monitoring RS measurement Trigger State.

[0254] Example: DCI Message: CSI request field = 1 (monitoring RS measurement Trigger State).

[0255] Action 3 The NW transmits monitoring RS resources as configured These RS resources are measured by the UE.

[0256] With a single Aperiodic Trigger, the NW may configure the UE to measure multiple time instances of the monitoring RS for a monitoring window duration. The monitoring window duration may be fixed or variable depending on a condition.

[0257] Action 4 Based on the first DCI message, the UE performs monitoring RS measurements and monitoring metric calculation without reporting any monitoring metric results.

[0258] Action 5 The UE shall repeat measurements for monitoring RS over multiple time instances and determines a monitoring metric when the monitoring window duration (may be fixed or variable depending on a condition) ends.

[0259]

[0260] In some example embodiments, the UE indicates to the NW that it is ready to report monitoring RS measurement (a monitoring report). In some of these example embodiments, the indication (fulfillment indication) is based on UE determination that a condition (such as a sufficient number of valid monitoring RS measurements for performing monitoring metric calculation) is fulfilled. The condition typically depends on the measurements of the monitoring reference signals and / or the inference reports.

[0261] In some of these example embodiments, the UE may send a similar indication (readiness indication) if the monitoring report is ready for transmission to the network, even in case the monitoring window is fixed (does not depend on a condition). Thus, the network need not to make an assumption on the metric calculation time needed after the end of the monitoring window to prepare the monitoring report. In that sense, the fulfillment indication includes a readiness indication.An advantage of the fulfillment indication and the readiness indication is that the network actually knows that the monitoring report is ready for transmission and need not make assumptions thereabout. On the other hand, additional signaling is required.

[0262] Action 7 Second DCI Message

[0263] Due to the receipt of the fulfillment indication or the readiness indication, or because the monitoring window and the metric calculation time have been passed after the first DCI message, the network sends a second DCI message including the second Trigger State that instructs the UE to report AI / ML model performance monitoring metric.

[0264] Action 8 The UE links the first and second DCI Messages (as described with reference to Figs. 5 and 6, for example) and determines the AI / ML model performance monitoring metric to report based on first and second DCI Messages.

[0265] For example, the UE may link the first and second DCI messages based on the Trigger State Id from the second DCI message and the Trigger State Id from the first DCI message. In one example, at least two conditions may hold:

[0266] • First and second Trigger States are linked / associated as their CSI-ReportConfig are both linked / associated to the same inference CSI-ReportConfig.

[0267] • The ReportQuantity (i.e., the resource set of monitoring RSs) of the CSI report for monitoring configured by the second Trigger State should be the same as the ReportQuantity of the CSI report for monitoring configured by the first Trigger State. In other words, the second Trigger State will trigger the UE to report the monitoring metric calculated in the monitoring window configured by the first Trigger State.

[0268] Report of AI / ML model performance monitoring metric

[0269] The UE reports the AI / ML model performance monitoring metric on PUSCH as instructed by the second DCI message.

[0270] An example of a reported AI / ML performance monitoring metric is: Top-1 or Top-K beam prediction accuracy determined over configured monitoring window.

[0271] Fig. 7 shows an apparatus according to an example embodiment. The apparatus may be a network (represented by a base station such as a gNB or eNB, for example) or an elementthereof. Fig. 8 shows a method according to an example embodiment. The apparatus according to Fig. 7 may perform the method of Fig. 8 but is not limited to this method. The method of Fig. 8 may be performed by the apparatus of Fig. 7 but is not limited to being performed by this apparatus.

[0272] The apparatus comprises first means for triggering 110 and second means for triggering 120. The first means for triggering 110 and second means for triggering 120 may be a first triggering means and second triggering means, respectively. The means for first triggering 110 and second means for triggering 120 may be a first triggerer and a second triggerer, respectively. The first means for triggering 110 and second means for triggering 120 may be a first triggering processor and second triggering processor, respectively.

[0273] The first means for triggering 110 triggers a terminal by transmitting a first trigger message to the terminal at a first trigger time (S110). The first trigger message triggers the terminal to measure a set of one or more monitoring reference signals at each of one or more time instances to obtain a monitoring measurement instance for the respective time instance. Furthermore, the first trigger message triggers the terminal to calculate, for the one or more time instances, a monitoring metric based on the monitoring measurement instances for the one or more time instances. The first trigger message does not trigger the terminal to report the monitoring metric. In other words, upon receipt of the first trigger message only, the terminal is inhibited to report the monitoring metric.

[0274] The second means for triggering 120 triggers the terminal by transmitting a second trigger message to the terminal at a second trigger time being later than the first trigger time (S120). The second trigger message triggers the terminal to report the monitoring metric for the one or more time instances in a monitoring report.

[0275] Fig. 9 shows an apparatus according to an example embodiment. The apparatus may be terminal (such as a UE) or an element thereof. Fig. 10 shows a method according to an example embodiment. The apparatus according to Fig. 9 may perform the method of Fig. 10 but is not limited to this method. The method of Fig. 10 may be performed by the apparatus of Fig. 9 but is not limited to being performed by this apparatus.

[0276] The apparatus comprises first means for receiving 210, first means for instructing 220, second means for receiving 230, and second means for instructing 240. The first means forreceiving 210, first means for instructing 220, second means for receiving 230, and second means for instructing 240 may be a first receiving means, first instructing means, second receiving means, and second instructing means, respectively. The first means for receiving 210, first means for instructing 220, second means for receiving 230 and second means for instructing 240 may be a first receiver, first instructor, second receiver, and second instructor, respectively. The first means for receiving 210, first means for instructing 220, second means for receiving 230 and second means for instructing 240 may be a first receiving processor, first instructing processor, second receiving processor, and second instructing processor, respectively.

[0277] The first means for receiving 210 receives a first trigger message from a network at a first trigger time (S210). The first trigger message triggers a terminal to measure a set of one or more monitoring reference signals at each of one or more time instances to obtain a monitoring measurement instance for the respective time instance. Furthermore, the first trigger message triggers the terminal to calculate, for the one or more time instances, a monitoring metric based on the monitoring measurement instances for the time instances. The first trigger message does not trigger the terminal to report the monitoring metric. In other words, the first trigger message inhibits the terminal for reporting the monitoring metric.

[0278] If the first trigger message is received in S210, the first means for instructing 220 instructs the terminal to measure the set of the one or more monitoring reference signals at each of the one or more time instances to obtain the monitoring measurement instance for the respective time instance (S220). Furthermore, if the first trigger message is received in S210, the first means for instructing 220 instructs the terminal to calculate, for the one or more time instances, the monitoring metric based on the monitoring measurement instances for the one or more time instances (S220). The first means for instructing does not instruct the terminal to report the monitoring metric.

[0279] The second means for receiving 230 receives a second trigger message from the network at a second trigger time later than the first trigger time (S230). The second trigger message triggers the terminal to report the monitoring metric for the one or more time instances in a monitoring report.If the second trigger message is received in S230, the second means for instructing 240 instructs the terminal to report the monitoring metric for the one or more time instances to the network in the monitoring report (S240).

[0280] Fig. 11 shows an apparatus according to an example embodiment. The apparatus comprises at least one processor 810, at least one memory 820 storing instructions that, when executed by the at least one processor 810, cause the apparatus at least to perform at least one of the methods according to at least one of Figs. 8 or 10 and related description.

[0281] Example embodiments may be applied to mobile communication networks, such as 3GPP networks of any generation (4G, 5G, 6G, 7G, for example). The reference signals measured for inference or monitoring may be arbitrary among the reference signals available in the respective network, such as CSI-RS, DMRS, SRS, PTRS, orTRS.

[0282] It is preferred that each of the communications between the different entities is performed in a secure way, e.g. by tunneling and / or encrypting etc..

[0283] One piece of information may be transmitted in one or plural messages from one entity to another entity. Each of these messages may comprise further (different) pieces of information.

[0284] Names of network elements, network functions, protocols, and methods are based on current standards, or are current proposals. These names are not limiting. For example, in other versions or other technologies, the names of corresponding entities may be different or the same as in the present description.

[0285] If not otherwise stated or otherwise made clear from the context, the statement that two entities are different means that they perform different functions. It does not necessarily mean that they are based on different hardware. That is, each of the entities described in the present description may be based on a different hardware, or some or all of the entities may be based on the same hardware. It does not necessarily mean that they are based on different software. That is, each of the entities described in the present description may be based on different software, or some or all of the entities may be based on the samesoftware. Each of the entities described in the present description may be deployed in the cloud.

[0286] According to the above description, it should thus be apparent that example embodiments provide, for example, a network (e.g. represented by a base station such as a gNB, eNB etc.) or an element thereof, an apparatus embodying the same, a method for controlling and / or operating the same, and computer program(s) controlling and / or operating the same as well as mediums carrying such computer program(s) and forming computer program product(s). According to the above description, it should thus be apparent that example embodiments provide, for example, a terminal (such as a UE, a MTC device, etc.) or an element thereof, an apparatus embodying the same, a method for controlling and / or operating the same, and computer program(s) controlling and / or operating the same as well as mediums carrying such computer program(s) and forming computer program product(s).

[0287] Implementations of any of the above described blocks, apparatuses, systems, techniques or methods include, as non-limiting examples, implementations as hardware, software, firmware, special purpose circuits or logic, general purpose hardware or controller or other computing devices, or some combination thereof. Each of the entities described in the present description may be embodied in the cloud.

[0288] It is to be understood that what is described above is what is presently considered the preferred example embodiments. However, it should be noted that the description of the preferred example embodiments is given by way of example only and that various modifications may be made without departing from the scope of the disclosure as defined by the appended claims.

[0289] The terms “first X” and “second X” include the options that “first X” is the same as “second X” and that “first X” is different from “second X”, unless otherwise specified. As used herein, “at least one of the following: ” and “at least one of ” and similar wording, where the list of two or more elements are joined by “and” or “or”, mean at least any one of the elements, or at least any two or more of the elements, or at least all the elements. The term "or" refers to a non-exclusive “or” unless otherwise indicated (e.g., use of “or else” or “or in the alternative”).

Claims

32Claims:

1. Apparatus, comprisingfirst means for triggering a terminal by transmitting a first trigger message to the terminal at a first trigger time, wherein the first trigger message triggers the terminal to measure a set of one or more monitoring reference signals at each of one or more time instances to obtain a monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances, a monitoring metric based on the monitoring measurement instances for the one or more time instances without triggering the terminal to report the monitoring metric;second means for triggering the terminal by transmitting a second trigger message to the terminal at a second trigger time being later than the first trigger time, wherein the second trigger message triggers the terminal to report the monitoring metric for the one or more time instances in a monitoring report.

2. The apparatus according to claim 1, further comprisingmeans for configuring the terminal with at least one of a first trigger state or a second trigger state, whereinif the means for configuring configures the terminal with the first trigger state, the first trigger message indicates the first trigger state;if the means for configuring configures the terminal with the second trigger state, the second trigger message indicates the second trigger state;the first trigger state triggers the terminal to measure the set of the one or more monitoring reference signals at each of the one or more time instances to obtain the monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances, the monitoring metric based on the monitoring measurement instances for the one or more time instances without triggering the terminal to report the monitoring metric;the second trigger state triggers the terminal to report the monitoring metric for the one or more time instances in the monitoring report.

3. The apparatus according to claim 2, wherein, if the means for configuring is configured to configure the terminal with the first trigger state, the second trigger message comprises a reference to the first trigger state.

334. The apparatus according to claim 2, wherein the second trigger state comprises a reference to the first trigger state.

5. The apparatus according to claim 2, wherein the second trigger state comprises a reference to the set of the one or more monitoring reference signals.

6. The apparatus according to claim 5, whereinthe second trigger state comprises a reference to a time series of inference reports, andthe first trigger state triggers the terminal to calculate, for each of the one or more time instances, a monitoring sub-metric for the respective time instance based on the monitoring measurement instance for the respective time instance and a prediction of the one or more monitoring reference signals reported in a respective one of the inference reports.

7. The apparatus according to any of claims 1 to 6, whereinthe first trigger message indicates a monitoring window, wherein the one or more time instances are within the monitoring window;the second trigger time is at a time instance after the monitoring window has passed after the first trigger time.

8. The apparatus according to claim 7, whereinthe second trigger time is at a time instance after the monitoring window and a metric calculation time have passed after the first trigger time;the metric calculation time is predefined; or the apparatus comprisesmeans for determining the metric calculation time based on the monitoring metric to be calculated by the terminal.

9. The apparatus according to claim 7, further comprisingmeans for receiving a readiness indication from the terminal after the monitoring window has passed after the first trigger time; whereinthe second trigger time is when the readiness indication has been received.

10. The apparatus according to any of claims 1 to 6, further comprisingmeans for indicating, to the terminal, a criterion to be fulfilled by the monitoring metric;means for checking whether a fulfilment indication is received from the terminal; whereinthe fulfillment indication indicates that the criterion is fulfilled by the monitoring metric; the second means for triggering is configured to transmit the second trigger message in response to checking that the fulfillment indication is received.

11. Apparatus, comprisingfirst means for receiving a first trigger message from a network at a first trigger time, wherein the first trigger message triggers a terminal to measure a set of one or more monitoring reference signals at each of one or more time instances to obtain a monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances, a monitoring metric based on the monitoring measurement instances for the time instances without triggering the terminal to report the monitoring metric;first means for instructing the terminal, if the first trigger message is received, to measure the set of the one or more monitoring reference signals at each of the one or more time instances to obtain the monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances, the monitoring metric based on the monitoring measurement instances for the one or more time instances;second means for receiving a second trigger message from the network at a second trigger time later than the first trigger time, wherein the second trigger message triggers the terminal to report the monitoring metric for the one or more time instances in a monitoring report;second means for instructing the terminal, if the second trigger message is received, to report the monitoring metric for the one or more time instances to the network in the monitoring report.

12. The apparatus according to claim 11, whereinthe terminal is configured with at least one of a first trigger state or a second trigger state;if the terminal is configured with the first trigger state, the first trigger message indicates the first trigger state and the first means for instructing is configured to instruct the terminal according to the first trigger state;if the terminal is configured with the second trigger state, the second trigger message indicates the second trigger state and the second means for instructing is configured to instruct the terminal according to the second trigger state;the first trigger state triggers the terminal to measure the set of the one or more monitoring reference signals at each of the one or more time instances to obtain the monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances, the monitoring metric based on the monitoring measurement instances for the one or more time instances without triggering the terminal to report the monitoring metric in the monitoring report;the second trigger state triggers the terminal to report the monitoring metric for the one or more time instances in the monitoring report.

13. The apparatus according to claim 12, wherein,if the first trigger message indicates the first trigger state, the second trigger message comprises a reference to the first trigger state; and the apparatus further comprises means for identifying the monitoring metric to be reported in the monitoring report based on the reference to the first trigger state.

14. The apparatus according to claim 12, wherein the second trigger state comprises a reference to the first trigger state, and the apparatus further comprisesmeans for identifying the monitoring metric to be reported in the monitoring report based on the reference to the first trigger state.

15. The apparatus according to claim 12, wherein the second trigger state comprises a reference to the set of the one or more monitoring reference signals, and the apparatus further comprisesmeans for identifying the monitoring metric to be reported in the monitoring report based on the reference to the set of the one or more monitoring reference signals.

16. The apparatus according to claim 15, whereinthe first trigger state comprises a reference to a time series of inference reports; the second trigger state comprises the reference to the inference report;the first means for instructing is configured to instruct the terminal to calculate, for each of the one or more time instances, a monitoring sub-metric for the respective time instance based on the monitoring measurement instance for the respective time instance and a prediction of the one or more monitoring reference signals reported in one of the inference reports;36the means for identifying is configured to identify the monitoring metric to be reported in the monitoring report based on the reference to the set of the one or more monitoring reference signals and the reference to the time series of inference reports.

17. The apparatus according to any of claims 11 to 16, whereinthe first trigger message indicates a monitoring window;the first means for instructing is configured to instruct the terminal, if the first trigger message is received, to measure the set of the one or more monitoring reference signals at each of the one or more time instances in the monitoring window to obtain the monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances in the monitoring window, the monitoring metric for the one or more time instances based on the monitoring measurement instances for the one or more time instances in the monitoring window;the second trigger time is a time instance being after the monitoring window and a metric calculation time have passed after the first trigger time.

18. The apparatus according to any of claims 11 to 16, whereinthe first trigger message indicates a monitoring window;the first means for instructing is configured to instruct the terminal, if the first trigger message is received, to measure the set of the one or more monitoring reference signals at each of the one or more time instances in the monitoring window to obtain the monitoring measurement instance for the respective time instance and to calculate, for the one or more time instances in the monitoring window, the monitoring metric for the one or more time instances based on the monitoring measurement instances for the one or more time instances in the monitoring window; and the apparatus further comprisesmeans for monitoring whether the terminal is ready to report the monitoring metric for the one or more time instances in the monitoring window in the monitoring report;means for indicating to the network, by providing a readiness indication, that the terminal is ready to report the monitoring metric for the one or more time instances in the monitoring window in the monitoring report, whereinthe second trigger time is later than the providing the readiness indication.

19. The apparatus according to any of claims 11 to 16, further comprisingmeans for receiving, from the network, a criterion to be fulfilled by the monitoring metric for the one or more time instances;means for checking whether the criterion is fulfilled by the monitoring metric for the one or more time instances;means for indicating to the network, by transmitting a fulfillment indication, that the criterion is fulfilled by the monitoring metric for the one or more time instances; wherein the second means for receiving is configured to receive the second trigger message in response to the transmitting the fulfillment indication.

20. The apparatus according to any of claims 11 to 19, further comprisingmeans for inhibiting the second means for instructing from instructing the terminal to report the monitoring metric for the one or more time instances to the network if the second trigger message is not received.

Images