Communication terminal, base station, communication method, and program
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2025-12-26
- Publication Date
- 2026-08-13
Smart Images
Figure JP2025045858_13082026_PF_FP_ABST
Abstract
Description
Communication Terminal, Base Station, Communication Method, and Program
[0001] The present disclosure relates to a communication terminal, a base station, a communication method, and a program.
[0002] In 3GPP (3rd Generation Partnership Project) (registered trademark), which formulates communication standards related to mobile networks, functions to support AI (Artificial Intelligence) and ML (Machine Learning) in 5G are being studied. Non-Patent Document 1 shows CSI (Channel State Information) feedback enhancement, beam management, and positioning accuracy enhancement as use cases related to AI / ML.
[0003] The use case of beam management includes downlink (DL) spatial-domain beam prediction and temporal beam prediction. Downlink spatial-domain beam prediction is called BM-Case1. Downlink temporal beam prediction is called BM-Case2.
[0004] Downlink (DL) beam prediction in the spatial domain predicts Set A of beams based on the measurement results of Set B of beams. Set B of beams is a set of beams in which the measurements are used as inputs to the AI / ML model. Set A of beams is different from Set B of beams or Set B of beams is a subset of Set A of beams.
[0005] Temporal downlink (DL) beam prediction is the prediction of beamset A based on the historical measurement results of beamset B. Beamset B is the set of beams whose measurements are used as inputs for an AI / ML model. In temporal DL beam prediction, beamset A may be different from beamset B, beamset B may be a subset of beamset A, or beamset A and beamset B may be identical.
[0006] Network-side AI / ML refers to the network performing AI / ML inference. UE (User Equipment)-side AI / ML refers to the UE performing AI / ML inference.
[0007] Monitoring AI / ML inference is called performance monitoring. Non-patent document 1 discusses Type 1 and Type 2 performance monitoring.
[0008] Type 1 performance monitoring includes Option 1 NW-side performance monitoring and Option 2 UE-assistant performance monitoring. In NW-side performance monitoring, the UE sends reports to the network, and the network calculates performance metrics. The UE reports to the network for the calculation of performance metrics. In UE-assistant performance monitoring, the UE calculates performance metrics and reports them to the network, or reports events to the network based on those performance metrics. The UE performs lifecycle management (LCM) operations based on instructions from the network.
[0009] In Type 2 performance monitoring, the gNB (g Node B) configures / signals to the UE, and the UE sends instructions, requests, or reports to the gNB. In the case of UE-side model monitoring, the UE makes decisions regarding model selection, activation, deactivation, switching, or fallback operations.
[0010] Performance monitoring for BM-Case 1 and BM-Case 2 includes four options for performance metrics (metric(s)): • Key Performance Indicators (KPIs) related to beam prediction accuracy, e.g., Top-1 or Top-K beam prediction accuracy; • Link quality-related KPIs, e.g., throughput, L1-RSRP (Layer 1 (L1) Reference Signal Received Power), L1 SINR (Signal-to-Interference-plus-Noise Ratio), virtual BLER (Block Error Rate); • Performance metrics based on AI / ML input / output data distribution; • L1-RSRP difference evaluated by comparing measured RSRP with predicted RSRP.
[0011] Non-patent document 2 discloses that Option 2 UE-assistant performance monitoring includes the following two options for monitoring configuration: Option 1: A resource set and a report configuration for monitoring are configured within the CSI report configuration used for inference. Option 2: A resource set and a report configuration specifically for monitoring are configured in a dedicated CSI report configuration used for monitoring.
[0012] Non-patent document 3 agrees that, for option 2, the CSI framework will be reused for the settings to monitor result reports with L1 signaling. It is also agreed that the inference report settings and the monitoring report settings will be linked by including the ID of the inference report settings in the monitoring settings.
[0013] The CSI framework was introduced in 3GPP Release 15. For example, the CSI framework is defined in Non-Patent Document 4 and Patent Document 5 (e.g., Sections 6.3.2, 5.2.1.1, and 5.2.1.2).
[0014] 3rd Generation Partnership Project; Technical Specification Group Radio Access Network; Study on Artificial Intelligence (AI) / Machine Learning (ML) for NR air interface (Release 18), 3GPP TR 38.843 V18.0.0 (2023-12)RAN1 Chair’s Notes, "Chair notes RAN1#118bis eom0", 3GPP TSG RAN WG1 #118b, Hefei, China, October 14th - 18th, 2024RAN1 Chair’s Notes, "Chair notes RAN1#119 eom0", 3GPP TSG RAN WG1 #119, Orlando, US, November 18th - 22nd, 20243GPP TS 38.331 V18.4.0 (2024-12) "3rd Generation Partnership Project; Technical Specification Group Radio Access Network; NR; Radio Resource Control (RRC) protocol specification (Release 18)", December 20243GPP TS 38.214 V18.5.0 (2024-12) "3rd Generation Partnership Project; Technical Specification Group Radio Access Network; NR; Physical layer procedures for data (Release 18)", December 2024Samsung, "Discussion for supporting AI / ML based beam management",R1-2409581, 3GPP TSG RAN WG1 #119, Orlando, US, November 18-22, 2024
[0015] In UE-assistant performance monitoring, there is a problem in that it is unclear how the UE correlates and reports multiple performance metrics to the network when it calculates several performance metrics. Specifically, Non-Patent Literature 6 states that metrics should be determined based on a comparison of each inference result in the CSI report of a particular CSI-ReportConfig with the associated measurement. However, Non-Patent Literature 6 does not clearly explain how to correlate multiple inference results obtained at different time points (e.g., time instances) with the monitoring report.
[0016] One of the purposes of this disclosure is to provide a communication terminal, base station, communication method, and program that enables a UE to correlate and report multiple performance metrics to the network.
[0017] The communication terminal according to this disclosure includes a receiving unit that receives from a base station first configuration information and second configuration information used in a learning model for performing beam inference in beam management, and third configuration information used for monitoring a beam to be monitored, and a transmitting unit that transmits to the base station a first performance metric based on the inference result using the first configuration information and a second performance metric based on the inference result using the second configuration information at a communication timing specified in the third configuration information.
[0018] The base station according to this disclosure includes a transmitting unit that transmits to a communication terminal first configuration information and second configuration information used in a learning model for performing beam inference in beam management, and third configuration information used for monitoring a beam to be monitored, and a receiving unit that receives from the communication terminal a first performance metric based on the inference result using the first configuration information and a second performance metric based on the inference result using the second configuration information at a communication timing specified in the third configuration information.
[0019] The communication method relating to this disclosure receives from a base station first configuration information and second configuration information used in a learning model that performs beam inference in beam management, and third configuration information used for monitoring the beam to be monitored, and transmits to the base station a first performance metric based on the inference result using the first configuration information and a second performance metric based on the inference result using the second configuration information at a communication timing specified in the third configuration information.
[0020] The program relating to this disclosure receives from a base station first configuration information and second configuration information used in a learning model for performing beam inference in beam management, and third configuration information used for monitoring the beam to be monitored, and causes a computer to transmit to the base station a first performance metric based on the inference result using the first configuration information and a second performance metric based on the inference result using the second configuration information at the communication timing specified in the third configuration information.
[0021] This disclosure provides a communication terminal, base station, communication method, and program that enable a UE to correlate and report multiple performance metrics to the network.
[0022] Figure 1 shows an example of a communication terminal configuration. Figure 2 shows the flow of communication processing performed at the communication terminal. Figure 3 shows an example of a base station configuration. Figure 4 shows the flow of communication processing performed at the base station. Figure 5 shows the flow of inference report transmission processing between the gNB and the UE. Figure 6 shows the flow of monitoring result notification processing between the gNB and the UE. Figure 7 shows the inference report time instance related to CSI-ReportConfig. Figure 8 shows the inference report time instance related to CSI-ReportConfig. Figure 9 shows the resources used to activate the beam. Figure 10 shows the flow of monitoring result notification processing between the gNB and the UE. Figure 11 shows the resources used to activate the beam. Figure 12 shows the CSI-ReportConfig used when the gNB sets information regarding the monitoring report time instance to the UE. Figure 13 is a block diagram showing an example of a base station and gNB configuration. Figure 14 is a block diagram showing an example of a communication terminal and UE configuration.
[0023] Embodiment 1 Figure 1 shows an example of the configuration of a communication terminal 10. The communication terminal 10 may be a computer device that operates by having a processor execute a program stored in memory.
[0024] The communication terminal 10 may be a smartphone, an IoT (Internet of Things) terminal, etc. The communication terminal 10 may also be a UE (User Equipment), which is a general term for communication terminals in 3GPP.
[0025] The communication terminal 10 has a receiving unit 11 and a transmitting unit 12. The receiving unit 11 and the transmitting unit 12 may be software or modules whose processing is performed by a processor executing a program stored in memory. Alternatively, the receiving unit 11 and the transmitting unit 12 may be hardware such as a circuit or chip.
[0026] The receiving unit 11 receives from the base station 20 first configuration information and second configuration information used in a learning model that performs beam inference in beam management, and third configuration information used for monitoring the beam to be monitored.
[0027] The learning model may be, for example, a learning model generated by performing supervised learning using training data during the learning phase. Alternatively, the learning model may be a learning model generated by performing unsupervised learning without using training data. Alternatively, the learning model may be a learning model generated by performing reinforcement learning to determine the optimal control that the communication terminal 10 should perform. The learning model may also be a model that performs AI / ML inference.
[0028] The training data used when performing supervised learning may include, for example, data containing information about multiple beams used between the base station 20 and the communication terminal 10, and information about the optimal beam. The beam information may include, for example, identification information to identify the beam, and quality information about the beam. The quality information about the beam may include information about the received power of the signal transmitted through the beam, the throughput when communication was performed using the beam, etc. Quality information may also be referred to as measurement information. The optimal beam may be the beam with the best quality among the multiple beams. The training data may be data generated by simulating communication between the base station 20 and the communication terminal 10, or it may be data collected when the base station 20 and the communication terminal 10 actually communicate.
[0029] In the inference phase, the learning model takes information about multiple beams used, for example, between the base station 20 and the communication terminal 10, as input and outputs information about the optimal beam. The information about the optimal beam may be, for example, information about the ranking of beams arranged in optimal order, or information that expresses the optimality of each beam using a numerical value. Alternatively, the information about the optimal beam may be quality information estimated when each beam is used. Alternatively, the information about the optimal beam may be information indicating the beam with the best quality.
[0030] The first and second configuration information may, for example, be information about at least one beam used as input to the learning model. For example, the information about at least one beam used as input to the learning model may indicate beamset B. The beam information may include, for example, a beam ID that identifies the beam. By inputting beamset B to the learning model, beamset A is output as the inference result. Beamset A may include information indicating the probability that each beam included in beamset A is Top-1. The Top-1 beam is the most appropriate beam and may also be the beam with the best communication quality. Beamset A may also include information about the top K beams sorted in order of quality, i.e., information about Top-K beams. The Top-K beams may be the top K beams when each beam is sorted in order of decreasing probability of being Top-1. Alternatively, they may be the top K beams sorted in order of quality. Beamset A may include at least one beam.
[0031] Furthermore, the first and second configuration information may include information indicating the timing for transmitting the inference results from the communication terminal 10 to the base station 20. The inference results transmitted from the communication terminal 10 to the base station 20 may be referred to as an inference report.
[0032] The beam to be monitored may be the beam actually monitored by the communication terminal 10. The beam to be monitored may, for example, represent beamset A, or it may be a subset of beamset A, which is a part of beamset A. The third configuration information may be information about at least one beam used as the beam to be monitored. Furthermore, the third configuration information may include information indicating the timing for transmitting the beam monitoring results from the communication terminal 10 to the base station 20.
[0033] The transmitting unit 12 transmits the first performance metric and the second performance metric to the base station 20 at the communication timing specified in the third configuration information. The first performance metric is generated based on the inference result using the first configuration information. The second performance metric is generated based on the inference result using the second configuration information.
[0034] The performance metric may be information indicating the accuracy of the inference result, obtained by comparing the inference result at the communication terminal 10 with the monitoring result regarding the monitored beam. The transmission unit 12 transmits to the base station 20 multiple performance metrics generated based on different inference results obtained based on different configuration information at a single communication timing specified in the third configuration information. The transmission unit 12 may also transmit to the base station 20 the monitoring result regarding the beam based on the third configuration information. In other words, the transmission unit 12 may transmit the monitoring result to the base station 20 in order for the base station 20 to generate the performance metric.
[0035] Figure 2 shows the flow of communication processing performed in the communication terminal 10. First, the receiving unit 11 receives from the base station 20 first configuration information and second configuration information used in a learning model that performs beam inference in beam management, and third configuration information used for monitoring the beam to be monitored (S11). Next, the transmitting unit 12 transmits the first performance metric and the second performance metric to the base station 20 at the communication timing specified in the third configuration information (S12). The first performance metric is generated based on the inference result using the first configuration information. The second performance metric is generated based on the inference result using the second configuration information.
[0036] Figure 3 shows an example configuration of base station 20. Base station 20 may be a computer device that operates by a processor executing a program stored in memory. Base station 20 may be a base station system that includes a device for wireless communication and a device for baseband processing. In other words, the components constituting base station 20 may be distributed across multiple devices. The multiple devices may be connected via a network. Base station 20 may be, for example, a gNB (g Node B) that supports a wireless communication standard known as 5G (5th Generation) in 3GPP, or an eNB (evolved Node B) that supports a wireless communication standard known as 4G (4th Generation). Alternatively, base station 20 may be, for example, a base station that supports a wireless communication standard known as 6G (6th Generation).
[0037] The base station 20 has a transmitting unit 21 and a receiving unit 22. The transmitting unit 21 and the receiving unit 22 may be software or modules whose processing is performed by a processor executing a program stored in memory. Alternatively, the transmitting unit 21 and the receiving unit 22 may be hardware such as a circuit or chip.
[0038] The transmitting unit 21 transmits to the communication terminal first configuration information and second configuration information used in a learning model that performs beam inference in beam management, and third configuration information used for monitoring the beam to be monitored.
[0039] The receiving unit 22 receives the first performance metric and the second performance metric from the communication terminal 10 at the communication timing specified in the third configuration information. The first performance metric is generated based on the inference result using the first configuration information. The second performance metric is generated based on the inference result using the second configuration information.
[0040] Figure 4 shows the flow of communication processing performed at the base station 20. First, the transmitting unit 21 transmits to the communication terminal a first configuration information and a second configuration information used in a learning model that performs beam inference in beam management, and a third configuration information used for monitoring the beam to be monitored (S21). Next, the receiving unit 22 receives the first performance metric and the second performance metric from the communication terminal 10 at the communication timing specified in the third configuration information (S22).
[0041] As described above, the communication terminal 10 transmits the first performance metric and the second performance metric to the base station 20 at the communication timing specified in the third configuration information. This allows the communication terminal 10 to report multiple performance metrics in association to the network.
[0042] Embodiment 2 Figure 5 shows the flow of the inference report transmission process between gNB 30 and UE 40. gNB 30 corresponds to the base station 20 in Figure 3. UE 40 corresponds to the communication terminal 10 in Figure 1. Here, UE 40 is assumed to have a learning model used for selecting a beam corresponding to CSI-RS (Channel State Information-Reference Signal) by having a neural network or the like perform machine learning. The beam corresponding to CSI-RS may also be called a CSI-RS beam.
[0043] Here, we will describe the learning model used by UE40 to select an appropriate CSI-RS beam. In the learning phase, the learning model used to select an appropriate CSI-RS beam may learn information about the CSI-RS beam of beamset A and information about the CSI-RS beam of beamset B as training data or training data. The information about the CSI-RS beam of beamset A may include quality information about the CSI-RS beam of beamset A and quality information about the optimal CSI-RS beam or the optimal CSI-RS beam. Furthermore, the information about the CSI-RS beam of beamset B may include quality information about the CSI-RS beam of beamset B and quality information about the optimal CSI-RS beam or the optimal CSI-RS beam. Furthermore, the training data may include identification information of the CSI-RS beam. Quality information for the CSI-RS beam may be, for example, RSRP or SINR (Signal-to-Interference-plus-Noise Ratio). The training data used in the learning phase may be generated by performing simulations, or it may be measurement results in a communication environment that actually has gNB and UE.
[0044] Furthermore, the learning model used for selecting an appropriate CSI-RS beam may be a learning model that takes, as input, quality information regarding the CSI-RS beams in beam set B and outputs the optimal CSI-RS beam within beam set A during the inference phase. Alternatively, the learning model used for selecting an appropriate CSI-RS beam may be a learning model that outputs quality information of the optimal CSI-RS beam. The optimal CSI-RS beam may be, for example, the CSI-RS beam with the highest quality among a plurality of CSI-RS beams, for example, the CSI-RS beam with the highest RSRP. Or, the learning model used for selecting an appropriate CSI-RS beam may output a ranking in which a plurality of CSI-RS beams are arranged in order from the CSI-RS beam with high quality. Also, the learning model may output information regarding the Top-K beams in which the CSI-RS beams included in beam set A are arranged in the Top-K order and are arranged in order of the probability that each beam becomes Top-1 being high. The Top-K beams may be the top K CSI-RS beams arranged in order of good quality. Or, the learning model used for selecting an appropriate CSI-RS beam may include information indicating the probability that each beam included in beam set A becomes Top-1.
[0045] Beam set A indicates a set of CSI-RS beams that gNB 30 can generate. Beam set B may be a subset indicating some of the beams in beam set A, may be a beam different from beam set A, or may be the same as the beam in beam set A.
[0046] Here, the communication processing flow shown in Figure 5 will be explained. First, gNB30 sets a dedicated resource for beamset A, which is the target of monitoring, in UE40 (S31). Setting a dedicated resource in UE40 may also mean notifying UE40 of the dedicated resource. The resource for beamset A may be the resource allocated to the CSI-RS in each beam included in beamset A. The resource may be represented using, for example, the frequency domain and the time domain. The frequency domain may be represented using, for example, subcarriers. The time domain may be represented using slots, symbols, etc.
[0047] Next, gNB30 sends configuration information regarding inference to UE40 (S32). The configuration information may be shown, for example, in CSI-Report Config. gNB30 may also send CSI-Report Config#1 and CSI-Report Config#2 to UE40. CSI-Report Config#1 and CSI Report Config#2 may each contain resources and inference report time instances for different beamset B. The different beamset B may be different beams included in beamset B for CSI-Report Config#1 and beams included in beamset B for CSI-Report Config#2. The inference report time instance may be information indicating the timing when UE40 sends an inference report showing the inference results to gNB30.
[0048] For example, the inference using CS-Report Config#1 may be executed before the inference report time instance. That is, after the UE 40 executes the inference using CS-Report Config#1, it transmits an inference report to the gNB 30 at the inference report time instance. The execution of the inference and the transmission of the inference report may be repeatedly executed. That is, the inference using CS-Report Config#1 may be repeatedly executed at different timings.
[0049] Next, the UE 40 executes an inference using a learning model using the resources related to the beam set B shown in CS-Report Config#1 and CS-Report Config#2, and transmits an inference report to the gNB 30 at the inference report time instance (S33 to S36). Here, as the timings of the inference report time instances of CSI-Report Config#1, times t1 to tm are shown. Further, as the timings of the inference report time instances of CSI-Report Config#2, times tm+1 to tn are shown. m and n are integers of 1 or more, and n is a number larger than m. In step S33, the UE 40 transmits an inference report indicating the inference result using CSI-Report Config#1 at time t1 to the gNB 30. In step S34, the UE 40 transmits an inference report indicating the inference result using CSI-Report Config#1 at time tm to the gNB 30. In step S35, the UE 40 transmits an inference report indicating the inference result using CSI-Report Config#2 at time tm+1 to the gNB 30. In step S36, the UE 40 transmits an inference report indicating the inference result using CSI-Report Config#2 at time tn to the gNB 30.
[0050] UE40 may include information indicating the most appropriate beam in each inference report time instance, or it may include information indicating the top K (where K is an integer greater than or equal to 2) beams when the beams are arranged in appropriate order. The most appropriate beam may be called, for example, the Top-1 beam, and the top K beams may be called the Top-K beams. The Top-K beams may be the top K beams when each beam is arranged in order of the probability of being Top-1. Alternatively, the Top-K beams may be the top K beams when arranged in order of quality.
[0051] Figure 6 shows the flow of the monitoring result notification process between gNB30 and UE40. The monitoring result notification process shown in Figure 6 may be executed after the inference report transmission process shown in Figure 5 has been executed. The monitoring result may also be information indicating the accuracy of the inference result in UE40.
[0052] First, gNB30 sends configuration information related to monitoring to UE40 (S41). The configuration information related to monitoring may be shown, for example, in CSI-Report Config. The configuration information related to monitoring may include, for example, a monitoring report time instance. The monitoring report time instance may be, for example, information indicating when UE40 sends performance metrics to gNB30. Furthermore, the configuration information may include information indicating the inference report time instance used to calculate the performance metrics. The inference report time instance used to calculate the performance metrics may be an inference report time instance at consecutive timings or an inference report time instance at non-consecutive timings.
[0053] Here, Figure 7 shows an overview of the inference report time instances for CSI-Report Config. For example, suppose timings 1 to n (where n is an integer greater than or equal to 2) are defined as inference report time instances for CS-Report Config#1. Timings 1 to n may be represented using symbols or other units of time. In this case, timings 1 to m (where m is an integer greater than or equal to 2 and less than n) may be defined as inference report time instances used to calculate performance metrics. Alternatively, non-consecutive timings such as timings 1, 3, and m may be defined as inference report time instances used to calculate performance metrics, as shown in Figure 8.
[0054] The gNB 30 may determine the inference report time instance used to calculate performance metrics based, for example, on the quality of communication at the time the inference report received in steps S33 to S36 of Figure 5 was received. For example, the gNB 30 may select an inference report time instance in communication where the quality, such as throughput and received power, is higher than a predetermined standard as the inference report time instance used to calculate performance metrics. Furthermore, the gNB 30 may select inference report time instances at consecutive timings by selecting an inference report time instance at a timing close to the selected inference report time instance.
[0055] Returning to Figure 6, gNB30 generates a Top-K merged set (S42). The Top-K merged set may be a beamset that combines the Top-K beams or Top-1 beams included in the inference report received from UE40 in steps S33 to S36 of Figure 5.
[0056] Combining Top-K beams or Top-1 beams can also be rephrased as merging Top-K beams or Top-1 beams, or including all Top-K beams or Top-1 beams. Furthermore, a Top-K merged set may also be a set containing the top K beams of the best quality from a beamset that includes all Top-K beams or Top-1 beams included in the inference report. Alternatively, a Top-K merged set may be a set containing the top K beams that are most likely to become Top-1 beams from a beamset that includes all Top-K beams or Top-1 beams included in the inference report.
[0057] Next, the gNB30 transmits a MAC (Medium Access Control) CE (Control Element) to activate the beams included in the Top-K merged set (S43). In other words, the gNB30 activates the beams included in the Top-K merged set by transmitting a MAC CE. In step S31 of Figure 5, the beams related to resource set A have already been configured. Therefore, the gNB30 transmits a MAC CE to activate the beams included in the Top-K merged set from among the pre-configured beams.
[0058] Figure 9 shows the resources used to activate the beams. Figure 9 shows the identification information of the RS transmitted using each beam. Assume that each RS corresponds to a beam. If a 1 is set in the field of each RS shown in Figure 9, the corresponding beam is activated; if a 0 is set, it is deactivated.
[0059] Returning to Figure 6, the next step is to send a Reference Signal (RS) to the UE40 via the beam included in the Top-K merged set (S44). The UE40 then generates a performance metric (S45).
[0060] The process for generating performance metrics is described below. UE40 measures the quality of the beams included in the Top-K merged set. Specifically, UE40 may measure the RSRP of the RS transmitted in each beam. By measuring the RSRP of the RS transmitted in each beam, UE may identify the Top-1 beam or the Top-K beam. The Top-1 beam may be the beam containing the highest quality RS. The Top-K beams may be the top K beams when the beams containing RS are arranged in descending order of quality.
[0061] Next, UE40 compares the inference results in the inference report time instance used to calculate the performance metrics shown in step S41 with the measurement results for the beams included in the Top-K merged set.
[0062] UE40 may determine whether the Top-1 beam included in the inference report time instance matches the Top-1 beam in the measurement result. For example, if UE40 uses five inference report time instances to calculate the performance metric, and all of the Top-1 beams match, the accuracy of the inference result may be considered to be 100%. The accuracy of the inference result may be referred to as, for example, prediction accuracy. Alternatively, if the Top-1 beams in three of the five inference report time instances match the Top-1 beam in the measurement result, the accuracy of the inference result may be considered to be 60%.
[0063] Alternatively, UE40 may determine whether the Top-1 beams included in the inference report time instance are included in the Top-K beams in the measurement results. For example, UE40 uses five inference report time instances to calculate the performance metric. In this case, if all of the Top-1 beams are included in the Top-K beams in the measurement results, the accuracy of the inference results may be considered to be 100%. Or, if the Top-1 beams in three of the five inference report time instances used to calculate the performance metric are included in the Top-K beams in the measurement results, the accuracy of the inference results may be considered to be 60%.
[0064] Alternatively, UE40 may determine whether any of the Top-K beams included in the inference report time instance match the Top-1 beam in the measurement result. For example, UE40 uses five inference report time instances to calculate the performance metric. In this case, if any of the Top-K beams in each instance match the Top-1 beam in the measurement result, the accuracy of the inference result may be considered to be 100%. Or, suppose that any of the Top-K beams in three of the five inference report time instances used to calculate the performance metric match the Top-1 beam in the measurement result. In this case, the accuracy of the inference result may be considered to be 60%.
[0065] Alternatively, UE40 may determine whether any of the Top-K beams included in the inference report time instance are included in the Top-K beams in the measurement results. For example, UE40 uses five inference report time instances to calculate the performance metric. In this case, if any of the Top-K beams in each instance are included in the Top-K beams in the measurement results, the accuracy of the inference results may be considered to be 100%. Or, suppose that any of the Top-K beams in three of the five inference report time instances used to calculate the performance metric are included in the Top-K beams in the measurement results. In this case, the accuracy of the inference results may be considered to be 60%.
[0066] UE40 sends performance metrics to gNB30 at the monitoring report time instance (S46). For example, UE40 sends a value indicating the accuracy of the inference result to gNB30 as a performance metric.
[0067] As explained above, UE40 sends performance metrics generated based on the inference results for each of the multiple configuration pieces of information to the monitoring report time instance. As a result, the performance metrics generated based on the inference results for each configuration piece of information performed by UE40 are associated with a single monitoring report time instance and sent to gNB30.
[0068] Furthermore, due to limitations in the radio resources required for transmitting the reference signal and the increased computational load, there is a problem in that it is difficult to transmit the Reference Signal (RS) to UE40 through the entire beam set A. As a result, there is a possibility that a predicted Top-K may occur in the monitoring resource set, and the gNB needs to adjust the selection of monitoring beams and the method of transmitting the reference signal. In contrast, gNB30 transmits the Reference Signal to UE40 through the beams included in the Top-K merged set. As a result, the increase in radio resources related to the Reference Signal can be suppressed, as can the increase in computational load on UE40.
[0069] Embodiment 3 Figure 10 shows the flow of the monitoring result notification process between gNB30 and UE40. The monitoring result notification process shown in Figure 10 may be executed after the inference report transmission process shown in Figure 5 has been executed.
[0070] First, gNB30 sends configuration information related to monitoring to UE40 (S51). The configuration information related to monitoring may be shown, for example, in CSI-Report Config. The configuration information related to monitoring may include, for example, a monitoring report time instance. The monitoring report time instance may be, for example, information indicating when UE40 sends performance metrics to gNB30.
[0071] Next, gNB30 transmits MAC CE to UE40 to activate the beams included in the subset of beamset A (S52). The subset of beamset A may be arbitrarily selected by gNB30, for example. Alternatively, the subset of beamset A may be a Top-K merged set.
[0072] Figure 11 shows the resources used to activate beams. Figure 11 illustrates that each octet is used to configure the activation or deactivation of beams within a subset. In this figure, the Subset ID represents the corresponding subset ID within Set A, and each subset ID is associated with the placement of each beam. For example, Set A contains 128 beams, which are divided into 8 subsets, each containing 16 beams. In this structure, the Subset ID is represented by a binary code to activate a specific subset. For example, to activate the first subset, the Subset ID is set to 0000 0001, and to activate the next subset, the Subset ID is set to 0000 0010. In this way, the network can efficiently manage and configure beams.
[0073] Returning to Figure 10, the gNB30 then transmits a monitoring RS to the UE40 via a beam included in a subset of beamset A (S53). The UE40 then selects an inference report time instance to use for calculating performance metrics (S54).
[0074] For example, UE40 compares the Top-1 beam or Top-K beam included in each inference report time instance in steps S33 to S36 of Figure 5 with the beams included in the subset of beamset A to determine whether there are any overlapping beams. UE40 may select the inference report time instances in which it has determined that there are overlapping beams as the inference report time instances to be used to calculate the performance metric.
[0075] Steps S55 and S56 are the same as steps S45 and S46 in Figure 6, so a detailed explanation is omitted.
[0076] Here, UE40 may send inference report time instances that were not used to calculate performance metrics (inference report time instances not used in the calculation) to gNB30 in step S56. Inference report time instances not used in the calculation may include, for example, an inference report time instance that includes a Top-1 beam or a Top-K beam that does not overlap with beams included in a subset of beamset A. Furthermore, inference report time instances not used in the calculation may include inference results that include errors, or inference results that include invalid results.
[0077] As explained above, UE40 can select the inference report time instance to use for calculating performance metrics. This eliminates the need to send information indicating the inference report time instance to be used for calculating performance metrics from, for example, gNB30 to UE40, thus reducing the amount of data transmitted.
[0078] Embodiment 4 This section describes an example of setting weights for inference report time instances used to calculate performance metrics. For example, the closer an inference report time instance is to the timing of a monitoring report time instance, the greater its impact on the performance metrics may be. In other words, the closer an inference report time instance is to the timing of a monitoring report time instance, the larger the weight value may be.
[0079] For example, in the example in Figure 7, the weight value assigned to the comparison between the inference result and the monitoring result may increase as the timing approaches the monitoring report instance, from timing 1 to timing m.
[0080] For example, let tmon be the monitoring report time instance, and let t1 to tm be the inference report time instances used to calculate performance metrics. Assume that tm is closest to tmon and t1 is furthest from tmon. In such a case, the weight wi at ti (where i is an integer between 1 and m) may be expressed by the following formula.
[0081] wi=1 / {1+α|tmon-(ti+N)|}
[0082] α is a value between 0 and 1, and may be determined, for example, by gNB30 or by UE40. N is a constant. In this equation, as the value of ti approaches tmon, the value of wi increases.
[0083] Here, Pi is a value that indicates the accuracy of the inference result included in the inference report time instance. For example, if the inference result is correct, Pi is set to 1, and if the inference result is incorrect, Pi is set to 0. The case where the inference result, the Top-1 beam, matches the measured Top-1 beam, or the inference result, the Top-1 beam, is included in the measured Top-K beam. Alternatively, the case where the inference result, the Top-K beam, and the measured Top-1 beam or Top-K beam have overlapping beams. The case where the inference result is incorrect may be any other case.
[0084] Here, the performance metric may also be expressed as follows:
[0085]
[0086] M represents the number of inference report time instances used to calculate performance metrics.
[0087] As explained above, the accuracy of performance metrics can be improved by adding weights to the inference report time instances used in performance metrics.
[0088] Embodiment 5 Figure 12 shows the CSI-ReportConfig used when gNB30 sets information regarding the monitoring report time instance to UE40. gNB30 may send the CSI Report Config shown in Figure 12 before step S41 in Figure 6, or together with step S41. Alternatively, gNB30 may send the CSI Report Config shown in Figure 12 before step S51 in Figure 10, or together with step S51.
[0089] The AssociatedType may be set to a value that indicates, for example, whether gNB30 or UE40 determines the inference report time instance used for performance metrics.
[0090] The linkedInferenceinstancesList may contain information that identifies, for example, the inference report time instances used for performance metrics.
[0091] The associated InstanceMode may, for example, specify whether consecutive timing instances or non-consecutive timing instances are used as inference reporting time instances for performance metrics.
[0092] maxInferenceInstances may be set to the number of instances, for example, when consecutively timed instances are used.
[0093] The offset may contain information that identifies the first instance, for example, when consecutively timed instances are used.
[0094] The calculationmode setting may specify whether to calculate performance metrics for each inference report time instance or for each of multiple inference report time instances.
[0095] The `timeweightmode` parameter may also be configured to determine whether or not to set weights for the inference report time instances used in performance metrics.
[0096] alphaforweight may contain the value of α used when assigning weights to the inference report time instance used for performance metrics.
[0097] The irregularInstanceReport may contain information instructing it to report inference report time instances that were not used to calculate performance metrics.
[0098] As explained above, the gNB30 can use CSI-ReportConfig when configuring information regarding monitoring report time instances.
[0099] Figure 13 is a block diagram showing an example configuration of a base station 20 and a gNB 30 (hereinafter referred to as "base station 20, etc."). Referring to Figure 13, the base station 20, etc. includes an RF transceiver 1001, a network interface 1003, a processor 1004, and a memory 1005. The RF transceiver 1001 performs analog RF signal processing to communicate with UEs. The RF transceiver 1001 may include multiple transceivers. The RF transceiver 1001 is coupled with an antenna 1002 and a processor 1004. The RF transceiver 1001 receives modulation symbol data (or OFDM symbol data) from the processor 1004, generates a transmit RF signal, and supplies the transmit RF signal to the antenna 1002. The RF transceiver 1001 also generates a baseband receive signal based on the received RF signal received by the antenna 1002 and supplies this to the processor 1004.
[0100] The network interface 1003 is used to communicate with network nodes (e.g., other core network nodes). The network interface 1003 may include, for example, a network interface card (NIC) compliant with the IEEE 802.3 series.
[0101] The processor 1004 performs data plane processing and control plane processing, including digital baseband signal processing for wireless communication.
[0102] The processor 1004 may include multiple processors. For example, the processor 1004 may include a modem processor (e.g., DSP) that performs digital baseband signal processing and a protocol stack processor (e.g., CPU or MPU) that performs control plane processing.
[0103] Memory 1005 is composed of a combination of volatile memory and non-volatile memory. Memory 1005 may include a plurality of physically independent memory devices. Volatile memory is, for example, Static Random Access Memory (SRAM) or Dynamic RAM (DRAM), or a combination thereof. Non-volatile memory is Mask Read Only Memory (MROM), Electrically Erasable Programmable ROM (EEPROM), flash memory, or hard disk drive, or any combination thereof. Memory 1005 may include storage located away from the processor 1004. In this case, the processor 1004 may access memory 1005 via the network interface 1003 or an I / O interface not shown.
[0104] The memory 1005 may store a software module (computer program) containing a set of instructions and data for processing by the base station 20, etc., as described in the above-described embodiments. In some implementations, the processor 1004 may be configured to read the software module from the memory 1005 and execute it to perform the processing of the base station 20, etc., as described in the above-described embodiments.
[0105] Figure 14 is a block diagram showing an example configuration of the communication terminal 10 and UE 40 (hereinafter referred to as the communication terminal 10, etc.). The Radio Frequency (RF) transceiver 1101 performs analog RF signal processing to communicate with the gNB 40. The analog RF signal processing performed by the RF transceiver 1101 includes frequency upconversion, frequency downconversion, and amplification. The RF transceiver 1101 is coupled with the antenna 1102 and the baseband processor 1103. That is, the RF transceiver 1101 receives modulation symbol data (or OFDM symbol data) from the baseband processor 1103, generates a transmit RF signal, and supplies the transmit RF signal to the antenna 1102. The RF transceiver 1101 also generates a baseband receive signal based on the received RF signal received by the antenna 1102 and supplies this to the baseband processor 1103.
[0106] The baseband processor 1103 performs digital baseband signal processing (data plane processing) and control plane processing for wireless communication. Digital baseband signal processing includes (a) data compression / decompression, (b) data segmentation / concatenation, (c) generation / decomposition of transmission format (transmission frame), (d) transmission path coding / decoding, (e) modulation (symbol mapping) / demodulation, and (f) generation of OFDM symbol data (baseband OFDM signal) by Inverse Fast Fourier Transform (IFFT). Meanwhile, control plane processing includes communication management at Layer 1, Layer 2, and Layer 3.
[0107] The baseband processor 1103 may include a modem processor (e.g., Digital Signal Processor (DSP)) for performing digital baseband signal processing and a protocol stack processor (e.g., Central Processing Unit (CPU) or Micro Processing Unit (MPU)) for performing control plane processing. In this case, the protocol stack processor for performing control plane processing may be shared with the application processor 1104 described later.
[0108] The application processor 1104 is also called a CPU, MPU, microprocessor, or processor core. The application processor 1104 may include multiple processors (multiple processor cores). The application processor 1104 implements various functions of the communication terminal 10, etc., by executing system software programs (Operating System (OS)) and various application programs (for example, a calling application, a web browser, a mailer, a camera operation application, a music playback application) read from memory 1106 or memory not shown.
[0109] In some implementations, the baseband processor 1103 and the application processor 1104 may be integrated on a single chip, as shown by the dashed line (1105) in Figure 14. In other words, the baseband processor 1103 and the application processor 1104 may be implemented as a single System on Chip (SoC) device 1105. An SoC device is sometimes called a System Large Scale Integration (LSI) or chipset.
[0110] Memory 1106 is volatile memory, non-volatile memory, or a combination thereof. Memory 1106 may include multiple physically independent memory devices. Volatile memory is, for example, Static Random Access Memory (SRAM) or Dynamic RAM (DRAM), or a combination thereof. Non-volatile memory is Mask Read Only Memory (MROM), Electrically Erasable Programmable ROM (EEPROM), flash memory, or hard disk drive, or any combination thereof. For example, memory 1106 may include an external memory device accessible from the baseband processor 1103, the application processor 1104, and the SoC 1105. Memory 1106 may also include an internal memory device integrated within the baseband processor 1103, the application processor 1104, or the SoC 1105. Furthermore, memory 1106 may include memory within a Universal Integrated Circuit Card (UICC).
[0111] The memory 1106 may store a software module (computer program) containing a set of instructions and data for processing by the communication terminal 10, etc., as described in the above-described embodiments. In some implementations, the baseband processor 1103 or application processor 1104 may be configured to read the software module from the memory 1106 and execute it to perform the processing of the communication terminal 10, etc., as described in the above-described embodiments.
[0112] In the examples described above, the program includes a set of instructions (or software code) that, when loaded into a computer, cause the computer to perform one or more of the functions described in the embodiments. The program may be stored on a non-temporary computer-readable medium or a physical storage medium. Examples, but not limited to, include random-access memory (RAM), read-only memory (ROM), flash memory, solid-state drive (SSD) or other memory technologies, CD-ROM, digital versatile disc (DVD), Blu-ray® disc or other optical disc storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage devices. The program may be transmitted over a temporary computer-readable medium or a communication medium. Examples, but not limited to, include temporary computer-readable medium or a communication medium that includes electrical, optical, acoustic or other forms of propagating signals.
[0113] Although the present disclosure has been described above with reference to embodiments, the present disclosure is not limited to the embodiments described above. Various modifications to the structure and details of the present disclosure can be made as can be understood by those skilled in the art within the scope of the present disclosure. Furthermore, each embodiment can be combined with other embodiments as appropriate.
[0114] Each drawing is merely illustrative to illustrate one or more embodiments. Each drawing may be associated with one or more other embodiments, rather than being associated with only one specific embodiment. As those skilled in the art will understand, various features or steps described with reference to any one drawing can be combined with features or steps shown in one or more other drawings, for example, to create embodiments not explicitly shown or described. Not all features or steps shown in any one drawing to illustrate an exemplary embodiment are necessarily required, and some features or steps may be omitted. The order of steps described in any of the drawings may be changed as appropriate.
[0115] Some or all of the above embodiments may also be described as follows, but are not limited to the following: (Note 1) A communication terminal comprising: a receiving unit that receives from a base station first configuration information and second configuration information used in a learning model for performing beam inference in beam management, and third configuration information used for monitoring a beam to be monitored; and a transmitting unit that transmits to the base station, at a communication timing specified in the third configuration information, a first performance metric based on the inference result using the first configuration information and a second performance metric based on the inference result using the second configuration information. (Note 2) A base station comprising: a transmitting unit that transmits to a communication terminal first configuration information and second configuration information used in a learning model that performs beam inference in beam management, and third configuration information used for monitoring a beam to be monitored; and a receiving unit that receives from the communication terminal a first performance metric based on the inference result using the first configuration information and a second performance metric based on the inference result using the second configuration information at a communication timing specified in the third configuration information. (Note 3) A communication method that receives from a base station first configuration information and second configuration information used in a learning model that performs beam inference in beam management, and third configuration information used for monitoring a beam to be monitored, and transmits to the base station a first performance metric based on the inference result using the first configuration information and a second performance metric based on the inference result using the second configuration information at a communication timing specified in the third configuration information.(Note 4) A communication method comprising: transmitting first configuration information and second configuration information used in a learning model for performing beam inference in beam management, and third configuration information used for monitoring a target beam, to a communication terminal; and receiving from the communication terminal, at a communication timing specified in the third configuration information, a first performance metric based on the inference result using the first configuration information and a second performance metric based on the inference result using the second configuration information. (Note 5) A program that causes a computer to receive first configuration information and second configuration information used in a learning model for performing beam inference in beam management, and third configuration information used for monitoring a target beam, from a base station; and transmitting to the base station, at a communication timing specified in the third configuration information, a first performance metric based on the inference result using the first configuration information and a second performance metric based on the inference result using the second configuration information. (Note 6) A program that causes a computer to transmit first configuration information and second configuration information used in a learning model for performing beam inference in beam management, and third configuration information used for monitoring the beam to be monitored, to a communication terminal, and to receive from the communication terminal a first performance metric based on the inference result using the first configuration information and a second performance metric based on the inference result using the second configuration information at the communication timing specified in the third configuration information.
[0116] Some or all of the elements described in any appendix may apply to various hardware, software, recording means, systems, and methods for recording software.
[0117] This application claims priority based on Japanese Patent Application No. 2025-017724, filed on 5 February 2025, and incorporates all of its disclosures herein.
[0118] 10 Communication terminal 11 Receiving unit 12 Transmitting unit 20 Base station 21 Transmitting unit 22 Receiving unit 30 gNB 40 UE
Claims
1. A communication terminal comprising: a receiving means for receiving from a base station first configuration information and second configuration information used in a learning model for performing beam inference in beam management, and third configuration information used for monitoring a beam to be monitored; and a transmitting means for transmitting to the base station, at a communication timing specified in the third configuration information, a first performance metric based on the inference result using the first configuration information and a second performance metric based on the inference result using the second configuration information.
2. A base station comprising: a transmission means for transmitting to a communication terminal first configuration information and second configuration information used in a learning model for performing beam inference in beam management, and third configuration information used for monitoring a beam to be monitored; and a receiving means for receiving from the communication terminal a first performance metric based on the inference result using the first configuration information and a second performance metric based on the inference result using the second configuration information at a communication timing specified in the third configuration information.
3. A communication method comprising: receiving from a base station first configuration information and second configuration information used in a learning model for performing beam inference in beam management, and third configuration information used for monitoring a beam to be monitored; and transmitting to the base station, at a communication timing specified in the third configuration information, a first performance metric based on the inference result using the first configuration information and a second performance metric based on the inference result using the second configuration information.
4. A communication method comprising: transmitting first configuration information and second configuration information used in a learning model for performing beam inference in beam management, and third configuration information used for monitoring a target beam, to a communication terminal; and receiving from the communication terminal, at a communication timing specified in the third configuration information, a first performance metric based on the inference result using the first configuration information and a second performance metric based on the inference result using the second configuration information.
5. A program that causes a computer to receive from a base station first configuration information and second configuration information used in a learning model for performing beam inference in beam management, and third configuration information used for monitoring the beam to be monitored, and to transmit to the base station a first performance metric based on the inference results using the first configuration information and a second performance metric based on the inference results using the second configuration information at the communication timing specified in the third configuration information.
6. A program that causes a computer to transmit first and second configuration information used in a learning model for performing beam inference in beam management, and third configuration information used for monitoring the target beam, to a communication terminal, and to receive from the communication terminal a first performance metric based on the inference results using the first configuration information and a second performance metric based on the inference results using the second configuration information at the communication timing specified in the third configuration information.