WO2024013851A1 - Terminal, wireless communication method, and base station - Google Patents

Abstract

A terminal according to one aspect of the present disclosure comprises: a reception unit that receives information pertaining to a scenario setting which requires collection of a dataset; and a control unit that trains an artificial intelligence (AI) model on the basis of a dataset collected based on a specific scenario setting. According to one aspect of the present disclosure, suitable overhead reduction, channel estimation, and resource utilization can be achieved.

Description

Terminal, wireless communication method and base station

The present disclosure relates to a terminal, a wireless communication method, and a base station in a next-generation mobile communication system.

In the Universal Mobile Telecommunications System (UMTS) network, Long Term Evolution (LTE) has been specified for the purpose of higher data rates, lower delays, etc. (Non-Patent Document 1). In addition, LTE-Advanced (3GPP Rel. 10-14) is a specification for the purpose of further increasing capacity and sophistication of LTE (Third Generation Partnership Project (3GPP (registered trademark)) Release (Rel. 8, 9). was made into

Successor systems to LTE (for example, also referred to as 5th generation mobile communication system (5G), 5G+ (plus), 6th generation mobile communication system (6G), New Radio (NR), 3GPP Rel. 15 or later) are also being considered. .

Regarding future wireless communication technology, the use of artificial intelligence (AI) technology such as machine learning (ML) is being considered for network/device control and management.

In the field of AI/ML, Generalization Capability (GC) refers to the ability to adapt not only to the training data given during training but also to unknown data (test data) (able to produce desired outputs, be able to predict well) ) Refers to the ability of an AI model. GC performance is also called GC performance (or generalization performance).

The NR standard also discusses how to evaluate GC performance, but it is required to efficiently secure GC for the AI model without spending too much effort on monitoring/updating the AI model. However, no progress has been made in considering specific methods. If the method is not properly defined, it will not be possible to achieve appropriate overhead reduction, highly accurate channel estimation, and highly efficient resource utilization based on the AI model, and there is a risk that improvements in communication throughput and communication quality will be suppressed. .

Therefore, one of the purposes of the present disclosure is to provide a terminal, a wireless communication method, and a base station that can realize suitable overhead reduction/channel estimation/resource utilization.

A terminal according to an aspect of the present disclosure includes a receiving unit that receives information regarding a scenario setting that requires collection of a data set, and an artificial intelligence (Artificial Intelligence) based on a data set collected under a specific scenario setting. AI)) A controller for training the model.

According to one aspect of the present disclosure, suitable overhead reduction/channel estimation/resource utilization can be achieved.

FIG. 1 is a diagram illustrating an example of an AI model management framework. FIG. 2 is a diagram illustrating an example of data set collection/reporting for securing GC of an AI model according to an embodiment. FIG. 3 is a diagram illustrating an example of the correspondence between use cases and scenario setting formats in the first embodiment. FIG. 4 is a diagram illustrating an example of the correspondence between the AI model and the scenario setting format in the second embodiment. FIG. 5 is a diagram illustrating an example of receiving information regarding scenario settings according to Embodiment 3.1. FIG. 6 is a diagram illustrating an example of data set collection according to Embodiment 3.2. FIG. 7 is a diagram illustrating an example of data set transmission according to Embodiment 3.3. FIG. 8 is a diagram illustrating an example of training on the UE side according to the fourth embodiment. FIG. 9 is a diagram illustrating an example of model selection according to Embodiment 5.1. 10A and 10B are diagrams illustrating an example of GC performance evaluation in Embodiment 5.2. FIG. 11 is a diagram illustrating an example of GC performance evaluation in Embodiment 5.2. FIG. 12 is a diagram illustrating an example of a schematic configuration of a wireless communication system according to an embodiment. FIG. 13 is a diagram illustrating an example of the configuration of a base station according to an embodiment. FIG. 14 is a diagram illustrating an example of the configuration of a user terminal according to an embodiment. FIG. 15 is a diagram illustrating an example of the hardware configuration of a base station and a user terminal according to an embodiment. FIG. 16 is a diagram illustrating an example of a vehicle according to an embodiment.

(Application of Artificial Intelligence (AI) technology to wireless communications)
Regarding future wireless communication technology, the use of AI technology such as machine learning (ML) is being considered for network/device control and management.

For example, improved Channel State Information (CSI) feedback (e.g., reduced overhead, improved accuracy, prediction), improved beam management (e.g., improved accuracy, prediction in the time/spatial domain), position A terminal, user terminal, user equipment (UE)/Base Station (BS) utilizes AI technology to improve measurements (e.g., improve position estimation/prediction), etc. It is being considered to do so.

Based on the input information, the AI model may output at least one information such as an estimated value, a predicted value, a selected action, a classification, etc. The UE/BS inputs channel state information, reference signal measurements, etc. to the AI model, and provides highly accurate channel state information/measurements/beam selection/position, future channel state information/radio link quality, etc. may be output.

Note that in this disclosure, AI may be read as an object (also referred to as a target, object, data, function, program, etc.) that has (implements) at least one of the following characteristics:
・Estimation based on observed or collected information;
- Selection based on observed or collected information;
- Predictions based on observed or collected information.

In the present disclosure, estimation, prediction, and inference may be used interchangeably. Furthermore, in the present disclosure, the terms "estimate," "predict," and "infer" may be used interchangeably.

In the present disclosure, an object may be, for example, an apparatus, a device, etc., such as a UE or a BS. Furthermore, in the present disclosure, an object may correspond to a program/model/entity that operates on the device.

Furthermore, in this disclosure, the AI model may be replaced by an object that has (implements) at least one of the following characteristics:
・Produce estimates by feeding information,
・Predict the estimated value by giving information,
・Discover characteristics by providing information,
・Select an action by providing information.

Additionally, in this disclosure, an AI model may refer to a data-driven algorithm that applies AI technology and generates a set of outputs based on a set of inputs.

In addition, in this disclosure, AI models, models, ML models, predictive analytics, predictive analysis models, tools, autoencoders (autoencoders), encoders, decoders, neural network models, AI algorithms, Schemes etc. may be read interchangeably. Further, the AI model may be derived using at least one of regression analysis (eg, linear regression analysis, multiple regression analysis, logistic regression analysis), support vector machine, random forest, neural network, deep learning, etc.

In the present disclosure, the autoencoder may be interchanged with any autoencoder such as a stacked autoencoder or a convolutional autoencoder. The encoder/decoder of the present disclosure may adopt models such as Residual Network (ResNet), DenseNet, RefineNet, etc.

In addition, in this disclosure, an encoder, encoding, encode/encoded, modification/change/control by an encoder, compressing, compress/compressed, generation ( "generate", "generate/generated", etc. may be used interchangeably.

In addition, in this disclosure, a decoder, decoding, decode/decoded, modification/change/control by a decoder, decompressing, decompress/decompressed, re- Reconstructing, reconstruct/reconstructed, etc. may be used interchangeably.

In this disclosure, layers (with respect to the AI model) may be interchanged with layers (input layer, intermediate layer, etc.) used in the AI model. The layers of the present disclosure include an input layer, an intermediate layer, an output layer, a batch normalization layer, a convolution layer, an activation layer, a dense layer, a normalization layer, a pooling layer, an attention layer, a dropout layer, It may correspond to at least one of the fully connected layers.

In the present disclosure, AI model training methods may include supervised learning, unsupervised learning, reinforcement learning, federated learning, and the like. Supervised learning may refer to the process of training a model from input and corresponding labels. Unsupervised learning may refer to the process of training a model without labeled data. Reinforcement learning is the process of training a model from inputs (in other words, states) and feedback signals (in other words, rewards) resulting from the model's outputs (in other words, actions) in the environment in which the models are interacting. It can also mean

In this disclosure, generation, calculation, derivation, etc. may be read interchangeably. In this disclosure, implementation, operation, operation, execution, etc. may be read interchangeably. In this disclosure, training, learning, updating, retraining, etc. may be used interchangeably. In this disclosure, inference, after-training, production use, actual use, etc. may be read interchangeably. In this disclosure, a signal may be interchanged with a signal/channel.

FIG. 1 is a diagram illustrating an example of an AI model management framework. In this example, each stage related to the AI model is shown as a block. This example is also expressed as AI model life cycle management.

The data collection stage corresponds to the stage of collecting data for generating/updating an AI model. The data collection stage includes data reduction (e.g., deciding which data to transfer for model training/model inference), data transfer (e.g., to entities performing model training/model inference (e.g., UE, gNB)), and transfer data).

Note that data collection may refer to a process in which data is collected by a network node, management entity, or UE for the purpose of AI model training/data analysis/inference. In the present disclosure, the terms "process" and "procedure" may be interchanged with each other. Also, in this disclosure, collection refers to datasets (e.g., available as input/output) for training/inference of an AI model based on measurements (channel measurements, beam measurements, radio link quality measurements, location estimation, etc.) It may also mean obtaining.

In the present disclosure, offline field data may be data collected from the field (real world) and used for offline training of an AI model. Also, in the present disclosure, online field data may be data collected from the field (real world) and used for online training of an AI model.

In the model training stage, model training is performed based on the data (training data) transferred from the collection stage. This stage includes data preparation (e.g., performing data preprocessing, cleaning, formatting, transformation, etc.), model training/validation, and model testing (e.g., whether the trained model meets performance thresholds). verification), model exchange (e.g., transferring a model for distributed learning), model deployment/updating (deploying/updating a model to an entity that performs model inference), etc.

Note that AI model training may refer to processing for training an AI model in a data-driven manner and obtaining a trained AI model for inference.

Furthermore, AI model validation may refer to a training sub-process for evaluating the quality of an AI model using a data set different from the data set used for model training. This sub-processing helps select model parameters that generalize beyond the dataset used to train the model.

Also, AI model testing refers to a sub-process of training to evaluate the performance of the final AI model using a dataset different from the dataset used for model training/validation. You may. Note that unlike validation, testing does not have to be based on subsequent model tuning.

At the model inference stage, model inference is performed based on the data (inference data) transferred from the collection stage. This stage includes data preparation (e.g., performing data preprocessing, cleaning, formatting, transformation, etc.), model inference, model monitoring (e.g., monitoring the performance of model inference), and model performance feedback (the entity performing model training). (feedback of model performance to actors), output (provide model output to actors), etc.

Note that AI model inference may refer to processing for producing a set of outputs from a set of inputs using a trained AI model.

Additionally, a UE side model may refer to an AI model whose inference is completely performed in the UE. A network side model may refer to an AI model whose inference is performed entirely in the network (eg, gNB).

Also, the one-sided model may mean a UE-side model or a network-side model. A two-sided model may refer to a pair of AI models in which joint inference is performed. Here, joint inference may include AI inference where the inference is performed jointly across the UE and the network, e.g., the first part of the inference is performed by the UE first and the remaining part is performed by the gNB. (or vice versa).

Furthermore, AI model monitoring may mean processing for monitoring the inference performance of an AI model, and may be interchanged with model performance monitoring, performance monitoring, etc.

Note that model registration means giving a version identifier to a model and making it executable by compiling it on the specific hardware used in the inference stage. You may. Model deployment also delivers a fully developed and tested model runtime image (or execution environment image) to the target (e.g., UE/gNB) where inference is performed. (or enable on the target).

The actor stage includes action triggers (e.g., deciding whether to trigger an action on other entities), feedback (e.g., feeding back information necessary for training data/inference data/performance feedback), etc. May include.

Note that, for example, training of a model for mobility optimization may be performed, for example, in Operation, Administration and Maintenance (Management) (OAM) in a network (Network (NW)) / gNodeB (gNB). good. The former has advantages in interoperability, large storage capacity, operator manageability, and model flexibility (e.g., feature engineering). In the latter case, the advantage is that there is no need for model update latency or data exchange for model development. Inference of the above model may be performed in the gNB, for example.

Depending on the use case (in other words, the functionality of the AI model), the entity that performs the training/inference may be different. Functions of the AI model may include beam management, beam prediction, autoencoder (or information compression), CSI feedback, position positioning, etc.

For example, for AI-assisted beam management based on measurement reports, the OAM/gNB may perform model training and the gNB may perform model inference.

For AI-assisted UE assisted positioning, a Location Management Function (LMF) may perform model training, and the LMF may perform model inference.

For CSI feedback/channel estimation using an autoencoder, the OAM/gNB/UE may perform model training and the gNB/UE (jointly) may perform model inference.

For AI-assisted beam management based on beam measurements or AI-assisted UE-based positioning, the OAM/gNB/UE may perform model training and the UE may perform model inference.

Note that model activation may mean activating an AI model for a specific function. Model deactivation may mean disabling an AI model for a particular function. Model switching may mean deactivating the currently active AI model for a particular function and activating a different AI model.

Additionally, model transfer may mean distributing the AI model over the air interface. This distribution may include distributing one or both of the parameters of the model structure known at the receiving end, or a new model with the parameters. This distribution may also include complete models or partial models. Model download may refer to model transfer from the network to the UE. Model upload may refer to model transfer from the UE to the network.

(Generalization Capability (GC))
In the field of AI/ML, GC refers to the ability of an AI model to adapt not only to the training data given during training but also to unknown data (test data) (achieve desired outputs and predict well). . GC performance is also called GC performance (or generalization performance).

The NR standard also discusses how to evaluate GC performance. For example, the following is discussed from the perspective of training datasets and test/inference sets (test/inference datasets) (A, B, C, etc. are datasets/configurations/scenarios). ), and if multiple are listed, it means the dataset corresponding to the multiple. The same applies to the discussion below):
- Training dataset is constructed by mixing training inputs from multiple settings/scenarios, and testing/inference is performed on a single setting/scenario (e.g. training dataset = AB and inference dataset is B, or the training dataset = CDE and the inference dataset is B),
A training dataset is constructed by mixing training inputs from a single setting/scenario, and testing/inference is performed against another single setting/scenario (e.g. training dataset = A and The inference dataset is B),
A training dataset is constructed by mixing training inputs from a single configuration/scenario, and testing/inference is performed for the same configuration/scenario (e.g., training dataset = A and inference dataset is A).

Note that the scenario may be one of a scenario, a frequency range, a new merology, an inter-BS distance, an outdoor/indoor UE distribution, and a UE speed. The configuration may be one or more of antenna port number, antenna configuration, bandwidth, CSI feedback payload.

Additionally, the following is discussed in terms of how to consider the combination/prioritization of various scenarios and settings for configuring training datasets and test/inference sets:
・Scenario-based: Fix the scenario and verify the GC performance of the AI model. This AI model is trained by a dataset with one specific scenario and one or more settings and applied to a test/inference set with the same scenario and same/different settings (e.g. the training dataset is = A and setting = ab, and the inference dataset is scenario = A and setting = c),
・Scenario generalization: Verify the GC performance of the AI model. This AI model is trained by a dataset with one or more scenarios and one or more settings and is applied to a test/inference set with the same/different scenarios and the same/different settings (e.g. training data The set is scenario = A and setting = ab and the inference dataset is scenario = B and setting = c, or the training dataset is scenario = A and setting = ab and the inference dataset is scenario = B and setting = ab).

Additionally, the following is being discussed from the perspective of AI model update/selection:
- A single AI model is trained based on scenario/setting #A and applied to inference/testing of the same scenario/setting #A or a different scenario/setting #B,
An AI model is trained based on scenario/setting #A, further fine-tuned using additional datasets from scenario/setting #B, and applied to inference/testing of scenario/setting #B;
- Multiple AI models are learned based on scenario/setting #A, scenario/setting #B, etc., and applied to inference/testing of scenario/setting #A, scenario/setting #B, etc., respectively.

From the above, the AI model has the following problems regarding GC:
・The model was trained offline using a simulation-oriented dataset, which may not be consistent with real-world field data;
- A model is trained offline/online in one scenario/setting and used for inference in another scenario/setting. Frequent model changes/updates can cause excessive signal overhead.
- Models trained offline/online in mixed scenarios/settings will become outdated and incorrect if the wireless communication environment changes.

Therefore, there is a need to efficiently secure the GC of the AI model without spending much effort on monitoring/updating the AI model, but no progress has been made in considering a specific method. If the method is not properly defined, it will not be possible to achieve appropriate overhead reduction, highly accurate channel estimation, and highly efficient resource utilization based on the AI model, and there is a risk that improvements in communication throughput and communication quality will be suppressed. .

Therefore, the present inventors came up with a method of constructing a dataset, a method of GC evaluation/verification, a model management method that takes into account GC performance (also called GC level), etc. in order to ensure GC of AI models. did. Note that each embodiment of the present disclosure may be used regardless of GC evaluation or the like (simply for acquiring a data set or the like).

Hereinafter, embodiments according to the present disclosure will be described in detail with reference to the drawings. The wireless communication methods according to each embodiment may be applied singly or in combination.

In the present disclosure, "A/B" and "at least one of A and B" may be read interchangeably. Furthermore, in the present disclosure, "A/B/C" may mean "at least one of A, B, and C."

In the present disclosure, "activate", "deactivate", "indicate", "select", "configure", "update", "determine", etc. may be read interchangeably. In this disclosure, supporting, controlling, being able to control, operating, capable of operating, etc. may be read interchangeably.

In the present disclosure, Radio Resource Control (RRC), RRC parameters, RRC messages, upper layer parameters, fields, Information Elements (IEs), settings, etc. may be read interchangeably. In the present disclosure, the terms Medium Access Control Element (CE), update command, activation/deactivation command, etc. may be read interchangeably.

In the present disclosure, the upper layer signaling may be, for example, Radio Resource Control (RRC) signaling, Medium Access Control (MAC) signaling, broadcast information, etc., or a combination thereof.

In the present disclosure, MAC signaling may use, for example, a MAC Control Element (MAC CE), a MAC Protocol Data Unit (PDU), or the like. Broadcast information includes, for example, a master information block (MIB), a system information block (SIB), a minimum system information (RMSI), and other system information ( Other System Information (OSI)) may also be used.

In the present disclosure, the physical layer signaling may be, for example, downlink control information (DCI), uplink control information (UCI), etc.

In this disclosure, an index, an identifier (ID), an indicator, a resource ID, etc. may be read interchangeably. In this disclosure, sequences, lists, sets, groups, groups, clusters, subsets, etc. may be used interchangeably.

In the present disclosure, channel measurement/estimation includes, for example, a channel state information reference signal (CSI-RS), a synchronization signal (SS), a synchronization signal/broadcast channel (Synchronization Signal/Physical It may be performed using at least one of a Broadcast Channel (SS/PBCH) block, a demodulation reference signal (DMRS), a measurement reference signal (Sounding Reference Signal (SRS)), and the like.

In this disclosure, CSI includes a channel quality indicator (CQI), a precoding matrix indicator (PMI), and a CSI-RS resource indicator (CRI). , SS/PBCH Block Resource Indicator (SSBRI), Layer Indicator (LI), Rank Indicator (RI), L1-RSRP (Reference in Layer 1) Signal received power (Layer 1 Reference Signal Received Power), L1-RSRQ (Reference Signal Received Quality), L1-SINR (Signal to Interference plus Noise Ratio), L1-SNR (Signal to Noise Ratio), channel matrix (or channel information regarding the precoding matrix (or precoding coefficients), and the like.

In the present disclosure, CSI-RS, Non Zero Power (NZP) CSI-RS, Zero Power (ZP) CSI-RS, and CSI Interference Measurement (CSI-IM) are: They may be read interchangeably. Additionally, the CSI-RS may include other reference signals.

In the following embodiments, the relevant entities are the UE and the BS in order to explain an AI model regarding communication between the UE and the BS, but the application of each embodiment of the present disclosure is not limited to this. For example, for communication between different entities (for example, communication between UE and UE), the UE and BS in the embodiment below may be replaced with a first UE and a second UE. In other words, the UE, BS, etc. of the present disclosure may be replaced with any UE/BS.

Note that in the present disclosure, data set and data may be read interchangeably. Further, in the present disclosure, channel state, channel status, channel, channel environment, etc. may be read interchangeably.

(Wireless communication method)
FIG. 2 is a diagram illustrating an example of data set collection/reporting for securing GC of an AI model according to an embodiment. Note that although similar environments are illustrated in subsequent drawings, it may be assumed that contents that are not repeatedly described are the same as those in FIG. 2.

According to the embodiment below, as shown in FIG. ) can be collected/reported.

If the data of scenario setting ID #1 can be collected, the UE/BS can perform model training based on the data of scenario setting ID #1.

If the data of scenario setting ID #1 and #2 can be collected, the UE/BS can perform model training based on the data of scenario setting ID #1, the data of scenario setting ID #2, or both of these data.

If the data of scenario setting ID #1, #2, and #3 can be collected, the UE/BS collects the data of scenario setting ID #1, the data of scenario setting ID #2, the data of scenario setting ID #3, or a combination thereof ( #1 and #2, #2 and #3, #1 and #3, or #1, #2 and #3).

The scenario setting will be mainly explained in the first/second embodiment. Training on the NW side/UE side will be mainly explained in the third/fourth embodiment.

<First embodiment>
The first embodiment relates to a scenario configuration format.

In a first embodiment, a use case using an AI model is associated with a scenario configuration format consisting of long-term features.

Note that long-term characteristics may be interchanged with short-term/medium-term/long-term characteristics, or simply characteristics. Further, the scenario configuration format may be interchanged with scenario and configuration format, scenario configuration format, scenario format, configuration format, use case format, environment format, metaformat, etc. Further, the format may be interchanged with type, mode, data, setting, etc.

The above features may include a combination of one or more of the following elements:
・Scenarios/models (Urban Macro (UMa), Urban Micro (Umi), indoor, outdoor, indoor hotspot, etc.),
・Frequency/frequency range,
- Numerology (or subcarrier spacing),
- Distribution/set of common channel parameters (e.g. inter-site distances (ISD), gNB height, delay spread, angular spread, Doppler spread, etc.) in a single scenario/model;
・UE distribution,
・UE speed,
・UE orbit,
・Number of transmit beams/receive beams,
・UE rotation pattern,
gNB/UE antenna configuration (e.g. transmit and receive antenna vectors),
・Number of cells/number of sectors,
・Bandwidth,
・UE payload,
-Channel quality (e.g. RSRP, SINR),
・Beam configuration ID,
・Physical Cell ID (PCI),
・Global Cell ID (GCI),
・Absolute Radio Frequency Channel Number (ARFCN),
・Probability of Line Of Site (LOS)/Non-Line Of Site (NLOS).

In the first embodiment, the UE may be expected to be configured/registered with a model whose associated scenario configuration format matches the UE's configuration/status.

Also, in the first embodiment, the UE may be expected to activate a model whose associated scenario configuration format matches the UE's configuration/status.

The correspondence between use cases and scenario setting formats may be specified in the standard, or information regarding the correspondence may be notified to the UE. Further, the features included in the scenario setting format corresponding to the use case may be specified in a standard, or information regarding the features may be notified to the UE.

FIG. 3 is a diagram showing an example of the correspondence between use cases and scenario setting formats in the first embodiment. In this example, "AI4CSI" (an AI model for CSI feedback) and "AI4BM" (an AI model for beam management (BM)) are shown as use cases. Each use case may be assigned a corresponding index, and the UE/BS may decide which AI model to use based on the notified index. Note that the use case names are just examples and are not limited to these.

In this example, AI4CSI is associated with a scenario configuration format that includes features such as location, AoA/ZoA, etc. AI4BM is also associated with a scenario configuration format that includes characteristics such as frequency, beam configuration ID, etc.

According to the first embodiment described above, the UE can suitably determine the scenario setting corresponding to the data set for training/inference based on the scenario setting format.

<Second embodiment>
The second embodiment relates to a list of scenario setting formats.

The second embodiment is the same as the first embodiment in that a use case using an AI model is associated with a scenario setting format, but for each AI model for the use case, one or more scenario settings The difference is that a list of formats is associated.

Note that one or more AI models may be associated with a use case. One or more AI models associated with one use case may correspond to the same model structure or different model structures.

The scenario configuration formats included in the list may be identified, for example, by a scenario configuration ID, and the included features are used to generate a dataset for training an AI model (or correspond to a dataset). ), a range of values of (or corresponding to) a data set that can be used (or collected) for inference, and/or the like.

The correspondence between the AI model and the list of scenario setting formats may be specified in the standard, or information regarding the correspondence may be notified to the UE.

FIG. 4 is a diagram showing an example of the correspondence between the AI model and the scenario setting format in the second embodiment. In this example, use case ID #1 is associated with AI models A and B, and use case ID #2 is associated with AI models AC.

AI models A and B associated with use case ID #1 correspond to the same model configuration. AI model A and AI models B and C associated with use case ID #2 correspond to different model configurations.

It may be possible to allow duplicate scenario setting IDs for each use case (even if the scenario setting ID is the same, it may indicate different scenario settings for each use case). Further, the configuration may be such that scenario setting IDs do not overlap for each use case. In this example, the scenario settings referenced (associated with the scenario setting ID) differ for each use case.

In the present disclosure, the scenario settings may define the value or value range of the feature. For example, in FIG. 4, for scenario setting ID #1 for use case ID #1, the location corresponds to range 1 (for example, the latitude and longitude are within range 1), and the AoA/ZoA corresponds to range X (for example, AoA /ZoA is within range X).

Note that hereinafter the scenario setting ID, one or more scenario setting IDs, and the list of scenario setting IDs may be read interchangeably.

Furthermore, in the present disclosure, the use case ID/model ID/scenario setting ID may be mutually read as another information (for example, another ID) associated with the use case ID/model ID/scenario setting ID. The correspondence between the other information and the use case ID/model ID/scenario setting ID may be defined in the standard, or information regarding the correspondence may be notified to the UE.

According to the second embodiment described above, the UE can suitably determine the scenario setting corresponding to the data set for training/inference based on the list of scenario setting formats.

<Third embodiment>
The third embodiment relates to training on the NW side. The third embodiment may relate to training an AI model where the training process is completely performed only on the NW side.

[Embodiment 3.1: Receiving information regarding scenario settings]
The UE may receive information from the NW regarding scenario settings for which data sets need to be collected for a particular AI model for a particular use case.

The information regarding the scenario setting may include at least one of a corresponding scenario setting ID, a corresponding model ID, a corresponding use case ID, etc.

Information regarding the scenario settings may include information regarding the format of the data set that is (should be) collected. Further, the information regarding the scenario setting may be information regarding the scenario setting independent of the format of the data set to be collected (to be collected).

The format of the data set may correspond to the input/output data of the AI model. The UE infers/obtains the format of the data set from AI model information (described later in the Supplementary section of this disclosure. The same applies to subsequent model information) or dedicated signaling (e.g., information regarding the format described above). Good too.

FIG. 5 is a diagram illustrating an example of receiving information regarding scenario settings according to Embodiment 3.1. Duplicate explanations regarding points similar to previous drawings will not be repeated (the same applies hereafter).

In this example, the UE receives from the NW information regarding scenario settings for a list of scenario settings IDs (indicating scenario settings IDs #1 and #2) corresponding to AI model A with use case ID #1. Further, the information regarding the scenario setting may include information indicating that the format of the dataset is a channel/precoding matrix, as information regarding the format of the dataset.

[Embodiment 3.2: Data set collection]
The UE may collect training datasets for a particular AI model under particular scenario settings.

The UE may receive a data collection request (data collection request signal) for the specific scenario setting. For example, the data collection request may include a scenario setting ID indicating a specific scenario setting. The UE may collect a data set under the scenario configuration indicated by the data collection request if the scenario configuration indicated by the information regarding the scenario configuration received in Embodiment 3.1 is included. .

The UE may determine the particular scenario configuration based on measurements/sensing of the channel environment. The UE compares the measured channel environment and the scenario setting indicated by the information regarding the scenario setting received in Embodiment 3.1, and determines that they match (the environment of the scenario setting includes the measured channel environment). If so, the UE may collect the matching data set of the scenario configuration.

Note that only the model ID/use case ID corresponding to the data set to be collected is notified to the UE, and the UE selects the model ID/use case ID based on channel environment measurement/sensing from among the scenario settings associated with the model ID/use case ID. The specific scenario settings may be determined based on the above.

The UE may assign a corresponding scenario setting ID to the collected data set. The UE may transmit information indicating that a data set for a specific scenario setting is available (collected) to the NW.

Control based on data collection requests and control based on measurement/sensing of the channel environment may be used in combination.

Note that in the present disclosure, the data set may be collected from the data server based on the scenario setting ID, or may be collected based on channel measurements corresponding to the format of the data set.

Note that in the present disclosure, the data server may be interchanged with a repository, an uploader, a library, a cloud server, or simply a server. Further, the data server in the present disclosure may be provided by any platform such as GitHub (registered trademark), and may be operated by any company/organization.

The UE may collect datasets for specific AI model training based solely on NW instructions (e.g. request/trigger/terminate). In this case, the scenario configuration may be transparent to the UE. Since the UE only needs to collect data sets based on NW instructions without knowing the current environment, a reduction in the UE load can be expected.

FIG. 6 is a diagram illustrating an example of data set collection according to Embodiment 3.2.

The UE may receive the trigger information for the scenario setting ID #1 from the NW, and collect the data set for the scenario setting ID #1 based on the information. Note that in the present disclosure, collecting a data set for a certain scenario setting ID may mean associating (or labeling) the collected data set as being for a certain scenario setting ID.

Furthermore, even if the UE does not receive the above trigger information, the measured channel environment/sensing result corresponds to scenario setting ID #1 (for example, the position corresponds to range 1 and the AoA/ZoA corresponds to range X). ), the data set for scenario setting ID #1 may be collected.

Furthermore, when the UE collects the data set for scenario setting ID #1, it may transmit information indicating that the data set for scenario setting ID #1 is available.

[Embodiment 3.3: Transmission of data set]
The UE may send information regarding the collected training dataset for a particular AI model (which may be referred to as a dataset report) to the NW.

The information regarding the data set may include at least one of the data set itself, the corresponding scenario setting ID/model ID/use case ID, etc.

After receiving the data set request (data set request signal) (or in response to the reception of the request), the UE may transmit information regarding the data set. The data set request may include information indicating a scenario setting ID/model ID/use case ID for which the data set is desired to be requested. When the UE receives a dataset request, the UE may transmit information regarding all datasets collected since sending (information regarding datasets) in response to the previous dataset request, or may transmit information regarding all datasets collected since the transmission of information regarding the dataset in response to the previous dataset request, or may send information regarding the specific scenario configuration ID/model ID. / Information regarding the dataset associated with the use case ID may be sent (eg, if the dataset request includes information indicating the particular scenario configuration ID/model ID/use case ID).

When the UE transmits information regarding a dataset in response to receiving a dataset request indicating a scenario configuration ID/model ID/use case ID, the UE shall include the scenario configuration ID/model ID/use case ID in the information regarding the dataset. It is not necessary (because the NW can know the ID corresponding to the data set).

Note that communications for (reporting) information regarding datasets (e.g., UE-BS communications, BS-core network communications, communications within the core network, etc.) are subject to the new quality of service for dataset transmission ( It may also be performed based on the Quality of Service (QoS) level. For example, new 5G QoS Identifier (5QI)/QoS Class Identifier (QCI) values for transmitting AI model data sets may be defined.

The device included in the UE/BS/core network may control communication for (reporting) information regarding the data set request based on the new 5QI/QCI value. For example, a device included in the UE/BS/core network may receive the new 5QI/QCI value for communication of a packet indicating the new 5QI/QCI value or a bearer (or network slice) corresponding to the new 5QI/QCI value. QoS control corresponding to the /QCI value may be performed.

The UE may transmit (upload) information regarding the data set to the data server. The transmission may include transmission from the UE to the NW and transfer from the NW to the data server.

FIG. 7 is a diagram illustrating an example of data set transmission according to Embodiment 3.3.

Upon receiving a data set request from the NW, the UE may transmit information regarding the data set to the NW. This information may include information indicating scenario setting ID #1. This information may be transferred from the NW to a specific data server. Information regarding the data set may be provided through QoS communication corresponding to 5QI=XX.

According to the third embodiment described above, it is possible to appropriately train the AI model on the NW side.

<Fourth embodiment>
The fourth embodiment relates to training on the UE side. The fourth embodiment may relate to training an AI model, where the training process is performed entirely on the UE side.

In the fourth embodiment, the UE may receive information regarding scenario settings as in Embodiment 3.1, or may collect data sets as in Embodiment 3.2.

The UE may train the AI model based on the group of collected datasets associated with the scenario configuration ID. In this disclosure, "training an AI model based on a group of datasets collected in association with a scenario setting ID" may be interchanged with "training an AI model for a scenario setting ID."

The UE may receive model training information (which may be referred to as training instructions, training requests, etc.) from the NW, and may train the AI model using the dataset corresponding to the information. The model training information may include information on (a list of) use case ID/model ID/scenario setting ID associated with the dataset used for training. For example, if the NW determines that a model under a certain scenario setting is useful/has poor performance, the NW transmits model training information including the ID of the scenario setting to the UE, and the model is associated with the scenario setting. The AI model may be trained on the UE.

The UE may train the AI model according to the measurement results of the channel conditions. For example, if a channel condition measurement matches a particular scenario configuration (e.g., the environment of the scenario configuration includes the measured channel condition), the UE may train an AI model associated with the particular scenario configuration. Good too.

The UE may train the AI model for all combinations of available scenario settings of the dataset. For example, as described in the third embodiment, when the UE receives a model ID and a scenario setting ID, the UE trains an AI model for all possible combinations corresponding to the model ID/scenario setting ID. You may.

The UE may monitor the GC performance (described later in the fifth embodiment) and, if the GC performance is poor under a particular scenario setting, train the AI model for the particular scenario setting.

After training the AI model, the UE may transmit information regarding the availability of the AI model (which may be referred to as model availability information) to the NW, including the use case ID/model ID/scenario setting ID. . The use case ID and model ID included in the information may indicate the applied use case and AI model, respectively. The scenario setting ID included in the information may indicate a trained scenario setting. Model availability information may correspond to, for example, information indicating that a model is available.

FIG. 8 is a diagram showing an example of training on the UE side according to the fourth embodiment. In this example, it is assumed that the UE has already collected data sets for scenario setting ID #1 and scenario setting ID #2.

When the UE receives model training information including the use case ID/model ID/scenario setting ID from the NW, the UE may train the model corresponding to the model ID for the scenario setting corresponding to the scenario setting ID.

For example, upon receiving model training information indicating model ID #1 and scenario setting ID #1, the UE may train the model with model ID #1 using the data set with scenario setting ID #1. Further, upon receiving model training information indicating model ID #1 and scenario setting ID #2, the UE may train the model with model ID #1 using the data set with scenario setting ID #2. Further, upon receiving the model training information indicating the model ID #1 and the scenario setting IDs #1 and #2, the UE transmits the model with the model ID #1 to the data set of the scenario setting ID #1 and the data set of the scenario setting ID #2. You can also train using sets.

According to the fourth embodiment described above, it is possible to appropriately train an AI model on the UE side.

<Fifth embodiment>
The fifth embodiment relates to a UE side model. The fifth embodiment may relate to an AI model where the inference is performed entirely on the UE side.

[Embodiment 5.1: Receiving information on list of scenario setting IDs]
The UE may receive information on the list of scenario configuration IDs in the model deployment procedure of the AI model. The information on the list may be included in the model information of the AI model, or may be information independent of the model information of the AI model. In the latter case, the UE may refer to the scenario configuration ID associated with the AI model ID. The correspondence between the scenario setting ID and the model ID may be defined in the standard, or information regarding the correspondence may be notified to the UE.

For example, the UE uses/activates/deploys a model with a corresponding AI model ID in an environment of a scenario setting with a certain scenario setting ID based on the information in the received AI model information/scenario setting ID list. You may decide (select).

When the UE receives information on multiple AI models that include (or are associated with) the same scenario configuration ID, the UE uses/activates/deploys from among these AI models based on at least one of the following: You may choose a model:
- Scenario settings and measured channel conditions associated with the AI model;
- UE state (e.g. UE speed);
・Settings/activation from NW.

Note that the UE may perform the above model selection even when receiving information on multiple AI models that include (or are associated with) different scenario setting IDs.

When selecting (determining) a model to be applied to a certain scenario setting, the UE may report model information (and corresponding scenario setting ID) regarding the determined model to the NW.

FIG. 9 is a diagram showing an example of model selection according to Embodiment 5.1. In this example, the UE received AI model information indicating model ID=00001 and 00002. The AI models corresponding to these IDs are defined as having the function of estimating the highest CSI-RS beam, as shown in the figure. Note that this is just an example, and the details of the model are not limited to this.

In this example, model ID = 00001 is associated with scenario setting ID #1 and #2, and model ID = 00002 is associated with scenario setting ID #2 and #3, both of which include the same scenario setting ID #2. .

The UE may decide which model to use, for example, based on the measured channel conditions. The UE may decide to use the model with model ID = 00001 corresponding to the scenario configuration ID (#1-#2) that is closer to the current state at the first position of the UE's movement path illustrated; In the central position, the UE may decide to use the model with model ID=00002, which corresponds to the scenario configuration ID (#2-#3) that is closer to the current state.

According to Embodiment 5.1 described above, the UE can appropriately select a model.

[Embodiment 5.2: Monitoring GC performance]
The UE may monitor the state of GC performance during model inference. A Key Performance Indicator (KPI) of GC performance may be at least one of the following (in other words, GC performance may be represented/evaluated by at least one KPI below): :
Final/experimental performance KPIs for specific use cases (e.g. throughput) that can be obtained after model inference;
- KPIs of AI model performance that can be obtained after model training/testing (e.g., generalization error/bias/variance (dispersion error), layer rotation, AI model performance for each scenario).

The generalization error/bias/variance may be, for example, a value obtained by comparing the output of an AI model in different scenario settings. Layer rotation is the variation (or change) in the cosine of the angle between the weight vector and a particular vector (e.g. trained weight vector, untrained weight vector, etc.) for each layer of the neural network. There may be.

The UE may check (evaluate) whether at least one of the following conditions is met for one or more GC performance:
・Option 1 (absolute method): The GC performance of the active/registered/configured model under a certain scenario setting is lower/higher than the threshold,
- Option 2 (relative method): GC performance of active/registered/configured model under one scenario setting is lower/higher than GC performance under another scenario setting,
・Option 3: The GC performance of a certain model falls below the threshold a certain number of times over a certain period of time.

Note that in the present disclosure, GC performance may be read as performance obtained by adding offset X (X is, for example, a real number) to GC performance. Offset X may be determined based on factors other than pure performance (reproducibility performance) (eg, unmonitored/monitor-free performance). By introducing an offset, it is possible to evaluate a model that comprehensively considers other factors.

Here, the unmonitored/unnecessary performance may correspond to at least one of overhead, reliability (of the model/calculated value), model complexity, power consumption for calculation, etc.

Values such as X and threshold values (or information regarding the values) may be specified in advance in the standard, may be determined based on the UE capabilities, may be notified from the NW to the UE, or may be determined based on scenario settings. (may be determined based on the scenario settings) or may be included in the AI model information (may be determined based on the model). Information regarding values such as X and threshold values may be defined/notified for each model/model group/scenario setting/scenario setting group.

Which (or which combination) of options 1-3 the UE should check may be prescribed/notified for each model/model group/scenario setting/scenario setting group.

FIGS. 10A and 10B are diagrams showing an example of GC performance evaluation in Embodiment 5.2.

FIG. 10A shows an example of option 1 above. In this example, the UE evaluates that the GC performance of model ID #1 and scenario setting ID #1 is lower than the threshold value.

FIG. 10B shows an example of option 2 above. In this example, the UE assumes that offset X (X>0) is applied to the GC performance of model ID #1 and scenario setting ID #1. Further, the UE evaluates that the GC performance of model ID #1 and scenario setting ID #2 is lower than the GC performance of model ID #1 and scenario setting ID #1 to which offset X is applied.

Option 3 above will be explained in more detail. Option 3 may include, for example, the following steps for a model/scenario configuration:
- start a timer when the first counter counts that the GC performance is less than the first value for a first number of times or more;
- Stop the timer when the second counter counts a second number of times or more that the GC performance is greater than the second value while the timer is running;
- Resetting the second counter if the GC performance is less than the first value while the timer is running;
- if the GC performance is greater than the first value, resetting the first counter;
- When the timer expires, the GC performance is evaluated as "poor".

Note that the first value may be a first threshold value (threshold _out ), or a value that is lower by a first offset (offset _out ) from a reference value (baseline value) for a specific model/scenario setting. It may be.

Additionally, the second value may be a second threshold (threshold _in ) or may be a second offset (offset _in ) greater than the reference value (baseline value) for the particular model/scenario setting. There may be.

Note that resetting the counter may mean setting the counter to a specific value (for example, 0).

Here, values (or information regarding the values) such as the first/second threshold, baseline value, first/second offset, first/second counter, counter granularity, timer time length, etc. may be defined in advance in the standard, may be determined based on the UE capabilities, may be notified from the NW to the UE, or may be included in the AI model information (determined based on the model). ). Information regarding these values may be defined/notified for each model/scenario setting.

FIG. 11 is a diagram showing an example of GC performance evaluation in Embodiment 5.2. In this example, the GC performance is initially good, but when the first counter counts that it is less than a first value a first number of times, a timer is started. After that, while the timer was running, the second counter counted that the GC performance was greater than the second value several times, but the second counter did not exceed the second number and the timer expired. However, the GC performance of this model/scenario setting is evaluated to be low.

In embodiment 5.2, the UE may evaluate one or more GC performances and select (determine) the top K (K is an integer) performances.

The value of K (or information regarding the value) may be specified in advance in the standard, may be determined based on the UE capability, may be notified to the UE from the NW, or may be determined based on the model. It may be information (may be determined based on a model).

Note that in the present disclosure, the UE may derive GC performance based on one or more GC performances and one or more unmonitored/monitor-unnecessary performances. Also, in this disclosure, GC performance may be averaged/weighted over a period of time when evaluated/compared. Information regarding the period, averaging/weighting method, etc. may be specified in advance in the standard, may be determined based on the UE capability, may be notified from the NW to the UE, or may be linked to the model. (It may also be determined based on a model.)

Note that if the GC performance of a certain model is not sufficient, the UE may switch to another model. For example, the UE determines the model to be applied (or the model to be used) to a certain scenario setting based on the GC performance evaluated under the conditions (at least one of options 1-3) described in Embodiment 5.2. You can.

According to Embodiment 5.2 described above, the UE can appropriately perform GC performance evaluation.

According to the fifth embodiment described above, inference/evaluation based on the AI model can be appropriately performed on the UE side.

<Sixth embodiment>
The sixth embodiment relates to a UE side model. The sixth embodiment may relate to an AI model where the inference is performed entirely on the UE side.

First, we will explain the performance report. The performance report in the sixth embodiment may be a report regarding GC performance.

[Timing of report]
The UE may transmit a performance report based on information notified from the NW. For example, the UE may transmit performance reports in periodically/semi-persistent/aperiodically scheduled uplink resources based on RRC/MAC CE/DCI. In this case, the performance report may be included in the UCI. At this time, the UE may determine the reporting period/offset based on the RRC/MAC CE/DCI.

The UE may itself determine the trigger related to the performance report and transmit the performance report when the trigger is triggered. For example, the UE may send a performance report if the conditions mentioned in embodiment 5.2 (eg, at least one of options 1-3) are met. In this case, the performance report may be included in the MAC CE (because it can be transmitted if the PUSCH is scheduled).

The UE may send performance reports when a new model is activated/registered/configured. The UE may also send a performance report when a timer based on configured/specified parameters (eg, the timer shown in option 3) expires.

[Contents of report]
The performance report may include at least one of the information shown in Embodiment 5.2, such as information indicating one or more GC performances, information indicating the performance of a model corresponding to (a list of) scenario setting IDs, etc. good.

The performance report may include information indicating the model/scenario setting corresponding to the reported GC performance (for example, model ID, scenario setting ID). At least one of the number of information indicating GC performance included in the performance report, the number (list of) scenario setting IDs included in the performance report, etc. is determined based on K shown in Embodiment 5.2. Good too.

The UE may determine the reported performance based on at least one of the following:
- Performance evaluated under the conditions described in Embodiment 5.2 (for example, at least one of options 1-3),
- Performance selected based on K shown in Embodiment 5.2,
-Performance determined to be reported based on notification from NW.

The above notification may be, for example, a model activation command, or may be a notification including information indicating a report/monitor target. The UE may report the GC performance of activated/monitored models.

For example, if the UE has models #1 and #2, and only model #2 is activated, the UE may report the performance of model #2. The UE may not report the performance of model #1.

According to the sixth embodiment described above, the UE can appropriately transmit a performance report.

<Seventh embodiment>
The seventh embodiment relates to a NW side model. The seventh embodiment may relate to an AI model in which inference is completely performed only on the NW side.

The UE may transmit information on the list of scenario setting IDs in the model transfer procedure of the AI model. The information on the list may be included in the model information of the AI model, or may be information independent of the model information of the AI model.

In other words, the UE may transmit the information on the list of scenario setting IDs as part of the model information, or as an RRC information element independent of the model information.

For example, the UE may transmit model information indicating model ID #1 and scenario setting ID #1, model information indicating model ID #1 and scenario setting ID #2, or model information indicating model ID #1 and scenario setting ID #2. Model information indicating ID #1 and scenario setting IDs #1 and #2 may be transmitted.

According to the seventh embodiment described above, the UE can appropriately transmit the scenario setting ID to the NW.

<Supplement>
[AI model information]
In this disclosure, AI model information may mean information including at least one of the following:
・AI model input/output information,
・Pre-processing/post-processing information for AI model input/output,
・Information on AI model parameters,
・Training information for AI models (training information),
・Inference information for AI models,
・Performance information regarding AI models.

Here, the input/output information of the AI model may include information regarding at least one of the following:
- Contents of input/output data (e.g. RSRP, SINR, amplitude/phase information in channel matrix (or precoding matrix), information on angle of arrival (AoA), angle of departure (AoD)) ), location information),
・Data auxiliary information (may be called meta information),
- type of input/output data (e.g. immutable value, floating point number),
- bit width of input/output data (e.g. 64 bits for each input value),
- Quantization interval (quantization step size) of input/output data (for example, 1 dBm for L1-RSRP),
- The range that input/output data can take (for example, [0, 1]).

Note that in the present disclosure, the information regarding AoA may include information regarding at least one of the azimuth angle of arrival and the zenith angle of arrival (ZoA). Further, the information regarding the AoD may include, for example, information regarding at least one of a radial azimuth angle of departure and a radial zenith angle of depth (ZoD).

In the present disclosure, the location information may be location information regarding the UE/NW. Location information includes information (e.g., latitude, longitude, altitude) obtained using a positioning system (e.g., Global Navigation Satellite System (GNSS), Global Positioning System (GPS), etc.), and information (e.g., latitude, longitude, altitude) adjacent to the UE. Information on the serving (or serving) BS (e.g., BS/cell identifier (ID), BS-UE distance, direction/angle of the BS (UE) as seen from the UE (BS), The information may include at least one of the coordinates of the BS (UE) as seen from the BS (e.g., X/Y/Z axis coordinates, etc.), the specific address of the UE (e.g., Internet Protocol (IP) address), etc. . The location information of the UE is not limited to information based on the location of the BS, but may be information based on a specific point.

The location information may include information regarding its own implementation (for example, location/position/orientation of antennas, location/orientation of antenna panels, number of antennas, number of antenna panels, etc.).

The location information may include mobility information. The mobility information may include information indicating at least one of the mobility type, the moving speed of the UE, the acceleration of the UE, the moving direction of the UE, and the like.

Here, the mobility types are fixed location UE, movable/moving UE, no mobility UE, low mobility UE, and medium mobility UE. (middle mobility UE), high mobility UE (high mobility UE), cell-edge UE (cell-edge UE), and non-cell-edge UE (not-cell-edge UE).

In this disclosure, environmental information (for data) may be information regarding the environment in which the data is acquired/used, such as frequency information (band ID, etc.), environment type information (indoor, etc.). , outdoor, Urban Macro (UMa), Urban Micro (Umi), etc.), Line Of Site (LOS)/Non-Line Of Site (NLOS), etc. Good too.

Here, LOS may mean that the UE and BS are in an environment where they can see each other (or there is no shielding), and NLOS may mean that the UE and BS are not in an environment where they can see each other (or there is a shield). It can also mean The information indicating LOS/NLOS may indicate a soft value (for example, probability of LOS/NLOS) or may indicate a hard value (for example, either LOS/NLOS).

In the present disclosure, meta information may mean, for example, information regarding input/output information suitable for an AI model, information regarding acquired/obtainable data, etc. Specifically, the meta information includes information regarding beams of RS (for example, CSI-RS/SRS/SSB, etc.) (for example, the pointing angle of each beam, 3 dB beam width, the shape of the pointing beam, (number of beams), gNB/UE antenna layout information, frequency information, environment information, meta information ID, etc. Note that the meta information may be used as input/output of the AI model.

The pre-processing/post-processing information for the input/output of the AI model may include information regarding at least one of the following:
- whether to apply normalization (e.g., Z-score normalization, min-max normalization);
- Parameters for normalization (e.g. mean/variance for Z-score normalization, minimum/maximum for min-max normalization),
- Whether to apply a specific numerical conversion method (e.g. one hot encoding, label encoding, etc.);
- Selection rules for whether or not to be used as training data.

For example, normalized input information x that has been subjected to Z-score normalization (x _new = (x-μ)/σ, where μ is the average of x and σ is the standard deviation) as a preprocessing for input information x. _new may be input into the AI model, and the output y _out from the AI model may be post-processed to obtain the final output y.

The information on the parameters of the AI model may include information regarding at least one of the following:
・Weight information in the AI model (e.g. neuron coefficients (coupling coefficients)),
・Structure of the AI model,
・Type of AI model as a model component (e.g. ResNet, DenseNet, RefineNet, Transformer model, CRBlock, Recurrent Neural Network (RNN), Long Short Memory -Term Memory (LSTM)), Gated Recurrent Unit (GRU)),
- Functions of the AI model as a model component (e.g. decoder, encoder).

Note that the weight information in the AI model may include information regarding at least one of the following:
・Bit width (size) of weight information,
・Quantization interval of weight information,
- Granularity of weight information,
・The range that weight information can take,
・Weight parameters in the AI model,
・Difference information from the AI model before update (if updating),
・Weight initialization methods (e.g. zero initialization, random initialization (based on normal distribution/uniform distribution/truncated normal distribution), Xavier initialization (for sigmoid functions), He initialization (rectified) For Rectified Linear Units (ReLU)).

Further, the structure of the AI model may include information regarding at least one of the following:
・Number of layers,
- Type of layer (e.g. convolution layer, activation layer, dense layer, normalization layer, pooling layer, attention layer),
・Layer information,
- Time series specific parameters (e.g. bidirectionality, time step),
- Parameters for training (e.g. type of function (L2 regularization, dropout function, etc.), where to put this function (e.g. after which layer)).

The layer information may include information regarding at least one of the following:
・Number of neurons in each layer,
・Kernel size,
・Stride for pooling layer/convolution layer,
・Pooling method (MaxPooling, AveragePooling, etc.),
・Residual block information,
・Number of heads,
・Normalization methods (batch normalization, instance normalization, layer normalization, etc.),
- Activation function (sigmoid, tanh function, ReLU, leaky ReLU information, Maxout, Softmax).

One AI model may be included as a component of another AI model. For example, an AI model may be an AI model in which processing proceeds in the following order: ResNet as model component #1, a transformer model as model component #2, a dense layer, and a normalization layer.

The training information for the AI model may include information regarding at least one of the following:
・Information for the optimization algorithm (e.g., optimization type (Stochastic Gradient Descent (SGD), AdaGrad, Adam, etc.), optimization parameters (learning rate, momentum, etc.) information, etc.),
・Information on the loss function (for example, information on the metrics of the loss function (Mean Absolute Error (MAE), Mean Square Error (MSE)), cross entropy loss, NLLLoss, Kullback- Leibler (KL) divergence, etc.),
parameters to be frozen for training (e.g. layers, weights),
- parameters to be updated (e.g. layers, weights),
・Parameters (for example, layers, weights) that should be initial parameters for training (should be used as initial parameters),
- How to train/update the AI model (e.g. (recommended) number of epochs, batch size, number of data used for training).

The inference information for the AI model may include information regarding branch pruning of a decision tree, parameter quantization, functions of the AI model, and the like. Here, the function of the AI model may correspond to at least one of, for example, time domain beam prediction, spatial domain beam prediction, an autoencoder for CSI feedback, an autoencoder for beam management, etc.

An autoencoder for CSI feedback may be used as follows:
- The UE inputs the CSI/channel matrix/precoding matrix into the encoder's AI model and transmits the output encoded bits as CSI feedback (CSI report).
- The BS inputs the received encoded bits into the decoder's AI model and reconstructs the output CSI/channel matrix/precoding matrix.

In spatial domain beam prediction, the UE/BS inputs sparse (or thick) beam-based measurements (beam quality, e.g. RSRP) into an AI model and outputs dense (or thin) beam quality. You can.

In time-domain beam prediction, the UE/BS may input time-series (past, current, etc.) measurement results (beam quality, e.g. RSRP) to the AI model and output future beam quality.

The performance information regarding the AI model may include information regarding the expected value of a loss function defined for the AI model.

The AI model information in the present disclosure may include information regarding the applicable range (applicable range) of the AI model. The applicable range may be indicated by a physical cell ID, a serving cell index, etc. Information regarding the scope of application may be included in the above-mentioned environmental information.

AI model information regarding a specific AI model may be predefined in a standard, or may be notified to the UE from a network (NW). The AI model defined in the standard may be called a reference AI model. AI model information regarding the reference AI model may be referred to as reference AI model information.

Note that the AI model information in the present disclosure may include an index (for example, may be referred to as an AI model index, AI model ID, model ID, etc.) for identifying an AI model. The AI model information in the present disclosure may include an AI model index in addition to/instead of the input/output information of the AI model described above. The association between the AI model index and the AI model information (for example, input/output information of the AI model) may be predetermined in the standard, or may be notified from the NW to the UE.

The AI model information in the present disclosure may be associated with the AI model, and may also be referred to as AI model relevant information, simply related information, or the like. The AI model related information does not need to explicitly include information for identifying the AI model. The AI model related information may be information containing only meta information, for example.

In the present disclosure, a model ID may be mutually read as an ID corresponding to a set of AI models (model set ID). Further, in the present disclosure, the model ID may be interchanged with the meta information ID. The meta information (or meta information ID) may be associated with beam-related information (beam settings) as described above. For example, the meta information (or meta information ID) may be used by the UE to select an AI model considering which beams the BS is using, or the UE may apply the deployed AI model. It may be used to inform the BS which beam to use for this purpose. Note that in the present disclosure, the meta information ID may be interchanged with an ID corresponding to a set of meta information (meta information set ID).

[Notification of information to UE]
In the embodiments described above, the notification of any information to the UE (from the NW) (in other words, the reception of any information from the BS in the UE) is performed using physical layer signaling (e.g., DCI), upper layer signaling (e.g., RRC). MAC CE), specific signals/channels (eg, PDCCH, PDSCH, reference signals), or a combination thereof.

When the above notification is performed by a MAC CE, the MAC CE may be identified by including a new logical channel ID (LCID), which is not specified in the existing standard, in the MAC subheader.

When the above notification is performed by a DCI, the above notification includes a specific field of the DCI, a radio network temporary identifier (Radio Network Temporary Identifier (RNTI)), the format of the DCI, etc.

Additionally, notification of any information to the UE in the above embodiments may be performed periodically, semi-persistently, or aperiodically.

[Notification of information from UE]
The notification of any information from the UE (to the NW) in the above embodiments (in other words, the transmission/reporting of any information to the BS in the UE) is performed using physical layer signaling (e.g. UCI), upper layer signaling (e.g. , RRC signaling, MAC CE), specific signals/channels (eg, PUCCH, PUSCH, reference signals), or a combination thereof.

When the above notification is performed by a MAC CE, the MAC CE may be identified by including a new LCID that is not defined in the existing standard in the MAC subheader.

When the above notification is performed by UCI, the above notification may be transmitted using PUCCH or PUSCH.

Further, notification of arbitrary information from the UE in the above embodiments may be performed periodically, semi-persistently, or aperiodically.

[About application of each embodiment]
At least one of the embodiments described above may be applied if certain conditions are met. The specific conditions may be specified in the standard, or may be notified to the UE/BS using upper layer signaling/physical layer signaling.

At least one of the embodiments described above may be applied only to UEs that have reported or support a particular UE capability.

The particular UE capability may indicate at least one of the following:
- supporting specific processing/operation/control/information for at least one of the above embodiments;
・Supported use cases,
・Number of AI models supported per use case,
- Number of supported scenario configurations per AI model per use case.

Further, the specific UE capability may be a capability that is applied across all frequencies (commonly regardless of frequency) or a capability that is applied across all frequencies (e.g., cell, band, band combination, BWP, component carrier, etc.). or a combination thereof), or it may be a capability for each frequency range (for example, Frequency Range 1 (FR1), FR2, FR3, FR4, FR5, FR2-1, FR2-2). Alternatively, it may be a capability for each subcarrier spacing (SCS), or a capability for each Feature Set (FS) or Feature Set Per Component-carrier (FSPC).

Furthermore, the above-mentioned specific UE capability may be a capability that is applied across all duplex schemes (commonly regardless of the duplex scheme), or may be a capability that is applied across all duplex schemes (for example, Time Division Duplex). The capability may be for each frequency division duplex (TDD)) or frequency division duplex (FDD)).

In addition, at least one of the embodiments described above may be configured such that the UE configures/activates specific information related to the embodiment described above (or performs the operation of the embodiment described above) by upper layer signaling/physical layer signaling. / May be applied when triggered. For example, the specific information may be information indicating that the use of the AI model is enabled, arbitrary RRC parameters for a specific release (for example, Rel. 18/19), or the like.

If the UE does not support at least one of the specific UE capabilities or is not configured with the specific information, for example, Rel. 15/16 operations may be applied.

(Additional note)
Regarding one embodiment of the present disclosure, the following invention will be added.
[Additional note 1]
a receiver for receiving a request for a dataset for a particular Artificial Intelligence (AI) model;
a control unit that collects the data set under specific scenario settings;
A terminal comprising: a transmitting unit that transmits information regarding the data set indicating an identifier (ID) of the specific scenario setting.
[Additional note 2]
The terminal according to supplementary note 1, wherein the control unit collects the data set under the specific scenario setting in response to receiving a data collection request signal indicating an ID of the specific scenario setting.
[Additional note 3]
The terminal according to Supplementary Note 1 or 2, wherein the receiving unit receives information regarding scenario settings that require collection of data sets.

(Additional note)
Regarding one embodiment of the present disclosure, the following invention will be added.
[Additional note 1]
a receiving unit that receives information regarding scenario settings that require data set collection;
A terminal having a control unit that trains an artificial intelligence (AI) model based on a data set collected under a specific scenario setting.
[Additional note 2]
The terminal according to supplementary note 1, wherein the control unit trains the AI model based on the data set in response to receiving model training information indicating an identifier (ID) of the specific scenario setting.
[Additional note 3]
The terminal according to appendix 1 or 2, wherein the control unit determines the specific scenario setting based on a measurement result of a channel state.
[Additional note 4]
The terminal according to any one of Supplementary Notes 1 to 3, wherein the control unit determines the specific scenario setting based on the generalization ability of the monitored AI model.

(wireless communication system)
The configuration of a wireless communication system according to an embodiment of the present disclosure will be described below. In this wireless communication system, communication is performed using any one of the wireless communication methods according to the above-described embodiments of the present disclosure or a combination thereof.

FIG. 12 is a diagram illustrating an example of a schematic configuration of a wireless communication system according to an embodiment. The wireless communication system 1 (also simply referred to as system 1) uses Long Term Evolution (LTE), 5th generation mobile communication system New Radio (5G NR), etc. specified by the Third Generation Partnership Project (3GPP). It may also be a system that realizes communication using

Additionally, the wireless communication system 1 may support dual connectivity between multiple Radio Access Technologies (RATs) (Multi-RAT Dual Connectivity (MR-DC)). MR-DC has dual connectivity between LTE (Evolved Universal Terrestrial Radio Access (E-UTRA)) and NR (E-UTRA-NR Dual Connectivity (EN-DC)), and dual connectivity between NR and LTE (NR-E -UTRA Dual Connectivity (NE-DC)).

In EN-DC, the LTE (E-UTRA) base station (eNB) is the master node (Master Node (MN)), and the NR base station (gNB) is the secondary node (Secondary Node (SN)). In NE-DC, the NR base station (gNB) is the MN, and the LTE (E-UTRA) base station (eNB) is the SN.

The wireless communication system 1 has dual connectivity between multiple base stations within the same RAT (for example, dual connectivity (NR-NR Dual Connectivity (NN-DC) where both the MN and SN are NR base stations (gNB)). )) may be supported.

The wireless communication system 1 includes a base station 11 that forms a macro cell C1 with relatively wide coverage, and base stations 12 (12a-12c) that are located within the macro cell C1 and form a small cell C2 that is narrower than the macro cell C1. You may prepare. User terminal 20 may be located within at least one cell. The arrangement, number, etc. of each cell and user terminal 20 are not limited to the embodiment shown in the figure. Hereinafter, when base stations 11 and 12 are not distinguished, they will be collectively referred to as base station 10.

The user terminal 20 may be connected to at least one of the plurality of base stations 10. The user terminal 20 may use at least one of carrier aggregation (CA) using a plurality of component carriers (CC) and dual connectivity (DC).

Each CC may be included in at least one of a first frequency band (Frequency Range 1 (FR1)) and a second frequency band (Frequency Range 2 (FR2)). Macro cell C1 may be included in FR1, and small cell C2 may be included in FR2. For example, FR1 may be a frequency band below 6 GHz (sub-6 GHz), and FR2 may be a frequency band above 24 GHz (above-24 GHz). Note that the frequency bands and definitions of FR1 and FR2 are not limited to these, and FR1 may correspond to a higher frequency band than FR2, for example.

Further, the user terminal 20 may communicate using at least one of time division duplex (TDD) and frequency division duplex (FDD) in each CC.

The plurality of base stations 10 may be connected by wire (for example, optical fiber, X2 interface, etc. compliant with Common Public Radio Interface (CPRI)) or wirelessly (for example, NR communication). For example, when NR communication is used as a backhaul between base stations 11 and 12, base station 11, which is an upper station, is an Integrated Access Backhaul (IAB) donor, and base station 12, which is a relay station, is an IAB donor. May also be called a node.

The base station 10 may be connected to the core network 30 via another base station 10 or directly. The core network 30 may include, for example, at least one of Evolved Packet Core (EPC), 5G Core Network (5GCN), Next Generation Core (NGC), and the like.

Core Network 30 is, for example, User Plane Function (UPF), Access and Mobility Management Function (AMF), Session Management (SMF), Unified Data Management. T (UDM), ApplicationFunction (AF), Data Network (DN), Location Management Network Functions (NF) such as Function (LMF) and Operation, Administration and Maintenance (Management) (OAM) may also be included. Note that multiple functions may be provided by one network node. Further, communication with an external network (eg, the Internet) may be performed via the DN.

The user terminal 20 may be a terminal compatible with at least one of communication systems such as LTE, LTE-A, and 5G.

In the wireless communication system 1, an orthogonal frequency division multiplexing (OFDM)-based wireless access method may be used. For example, in at least one of the downlink (DL) and uplink (UL), Cyclic Prefix OFDM (CP-OFDM), Discrete Fourier Transform Spread OFDM (DFT-s-OFDM), Orthogonal Frequency Division Multiple Access (OFDMA), Single Carrier Frequency Division Multiple Access (SC-FDMA), etc. may be used.

A wireless access method may also be called a waveform. Note that in the wireless communication system 1, other wireless access methods (for example, other single carrier transmission methods, other multicarrier transmission methods) may be used as the UL and DL radio access methods.

In the wireless communication system 1, the downlink channels include a physical downlink shared channel (PDSCH) shared by each user terminal 20, a broadcast channel (physical broadcast channel (PBCH)), and a downlink control channel (physical downlink control). Channel (PDCCH)) or the like may be used.

In the wireless communication system 1, uplink channels include a physical uplink shared channel (PUSCH) shared by each user terminal 20, an uplink control channel (PUCCH), and a random access channel. (Physical Random Access Channel (PRACH)) or the like may be used.

User data, upper layer control information, System Information Block (SIB), etc. are transmitted by the PDSCH. User data, upper layer control information, etc. may be transmitted by PUSCH. Furthermore, a Master Information Block (MIB) may be transmitted via the PBCH.

Lower layer control information may be transmitted by PDCCH. The lower layer control information may include, for example, downlink control information (DCI) that includes scheduling information for at least one of PDSCH and PUSCH.

Note that the DCI that schedules PDSCH may be called DL assignment, DL DCI, etc., and the DCI that schedules PUSCH may be called UL grant, UL DCI, etc. Note that PDSCH may be replaced with DL data, and PUSCH may be replaced with UL data.

A control resource set (CONtrol REsource SET (CORESET)) and a search space may be used to detect the PDCCH. CORESET corresponds to a resource for searching DCI. The search space corresponds to a search area and a search method for PDCCH candidates (PDCCH candidates). One CORESET may be associated with one or more search spaces. The UE may monitor the CORESET associated with a certain search space based on the search space configuration.

One search space may correspond to PDCCH candidates corresponding to one or more aggregation levels. One or more search spaces may be referred to as a search space set. Note that "search space", "search space set", "search space setting", "search space set setting", "CORESET", "CORESET setting", etc. in the present disclosure may be read interchangeably.

The PUCCH allows channel state information (CSI), delivery confirmation information (for example, may be called Hybrid Automatic Repeat Request ACKnowledgement (HARQ-ACK), ACK/NACK, etc.), and scheduling request ( Uplink Control Information (UCI) including at least one of SR)) may be transmitted. A random access preamble for establishing a connection with a cell may be transmitted by PRACH.

Note that in this disclosure, downlinks, uplinks, etc. may be expressed without adding "link". Furthermore, various channels may be expressed without adding "Physical" at the beginning.

In the wireless communication system 1, a synchronization signal (SS), a downlink reference signal (DL-RS), and the like may be transmitted. In the wireless communication system 1, the DL-RS includes a cell-specific reference signal (CRS), a channel state information reference signal (CSI-RS), and a demodulation reference signal (DeModulation). Reference Signal (DMRS)), Positioning Reference Signal (PRS), Phase Tracking Reference Signal (PTRS), etc. may be transmitted.

The synchronization signal may be, for example, at least one of a primary synchronization signal (PSS) and a secondary synchronization signal (SSS). A signal block including SS (PSS, SSS) and PBCH (and DMRS for PBCH) may be called an SS/PBCH block, SS Block (SSB), etc. Note that SS, SSB, etc. may also be called reference signals.

In addition, in the wireless communication system 1, measurement reference signals (Sounding Reference Signal (SRS)), demodulation reference signals (DMRS), etc. are transmitted as uplink reference signals (UL-RS). good. Note that DMRS may be called a user terminal-specific reference signal (UE-specific reference signal).

(base station)
FIG. 13 is a diagram illustrating an example of the configuration of a base station according to an embodiment. The base station 10 includes a control section 110, a transmitting/receiving section 120, a transmitting/receiving antenna 130, and a transmission line interface 140. Note that one or more of each of the control unit 110, the transmitting/receiving unit 120, the transmitting/receiving antenna 130, and the transmission path interface 140 may be provided.

Note that this example mainly shows functional blocks that are characteristic of the present embodiment, and it may be assumed that the base station 10 also has other functional blocks necessary for wireless communication. A part of the processing of each unit described below may be omitted.

The control unit 110 controls the entire base station 10. The control unit 110 can be configured from a controller, a control circuit, etc., which will be explained based on common recognition in the technical field related to the present disclosure.

The control unit 110 may control signal generation, scheduling (e.g., resource allocation, mapping), and the like. The control unit 110 may control transmission and reception, measurement, etc. using the transmitting/receiving unit 120, the transmitting/receiving antenna 130, and the transmission path interface 140. The control unit 110 may generate data, control information, a sequence, etc. to be transmitted as a signal, and may transfer the generated data to the transmitting/receiving unit 120. The control unit 110 may perform communication channel call processing (setting, release, etc.), status management of the base station 10, radio resource management, and the like.

The transmitting/receiving section 120 may include a baseband section 121, a radio frequency (RF) section 122, and a measuring section 123. The baseband section 121 may include a transmission processing section 1211 and a reception processing section 1212. The transmitter/receiver unit 120 includes a transmitter/receiver, an RF circuit, a baseband circuit, a filter, a phase shifter, a measurement circuit, a transmitter/receiver circuit, etc., which are explained based on common understanding in the technical field related to the present disclosure. be able to.

The transmitting/receiving section 120 may be configured as an integrated transmitting/receiving section, or may be configured from a transmitting section and a receiving section. The transmitting section may include a transmitting processing section 1211 and an RF section 122. The reception section may include a reception processing section 1212, an RF section 122, and a measurement section 123.

The transmitting/receiving antenna 130 can be configured from an antenna described based on common recognition in the technical field related to the present disclosure, such as an array antenna.

The transmitter/receiver 120 may transmit the above-mentioned downlink channel, synchronization signal, downlink reference signal, etc. The transmitter/receiver 120 may receive the above-mentioned uplink channel, uplink reference signal, and the like.

The transmitter/receiver 120 may form at least one of a transmit beam and a receive beam using digital beam forming (e.g., precoding), analog beam forming (e.g., phase rotation), or the like.

The transmitting/receiving unit 120 (transmission processing unit 1211) performs Packet Data Convergence Protocol (PDCP) layer processing, Radio Link Control (RLC) layer processing (for example, RLC retransmission control), Medium Access Control (MAC) layer processing (for example, HARQ retransmission control), etc. may be performed to generate a bit string to be transmitted.

The transmitting/receiving unit 120 (transmission processing unit 1211) performs channel encoding (which may include error correction encoding), modulation, mapping, filter processing, and discrete Fourier transform (DFT) on the bit string to be transmitted. A baseband signal may be output by performing transmission processing such as processing (if necessary), Inverse Fast Fourier Transform (IFFT) processing, precoding, and digital-to-analog conversion.

The transmitting/receiving unit 120 (RF unit 122) may perform modulation, filter processing, amplification, etc. on the baseband signal in a radio frequency band, and may transmit the signal in the radio frequency band via the transmitting/receiving antenna 130. .

On the other hand, the transmitting/receiving section 120 (RF section 122) may perform amplification, filter processing, demodulation into a baseband signal, etc. on the radio frequency band signal received by the transmitting/receiving antenna 130.

The transmitting/receiving unit 120 (reception processing unit 1212) performs analog-to-digital conversion, fast Fourier transform (FFT) processing, and inverse discrete Fourier transform (IDFT) on the acquired baseband signal. )) processing (if necessary), applying reception processing such as filter processing, demapping, demodulation, decoding (which may include error correction decoding), MAC layer processing, RLC layer processing and PDCP layer processing, User data etc. may also be acquired.

The transmitting/receiving unit 120 (measuring unit 123) may perform measurements regarding the received signal. For example, the measurement unit 123 may perform Radio Resource Management (RRM) measurement, Channel State Information (CSI) measurement, etc. based on the received signal. The measurement unit 123 is the receiving power (for example, Reference Signal Received Power (RSRP)), Receive Quality (eg, Reference Signal Received Quality (RSRQ), Signal To InterfERENCE PLUS NOI. SE RATIO (SINR), Signal to Noise Ratio (SNR) , signal strength (for example, Received Signal Strength Indicator (RSSI)), propagation path information (for example, CSI), etc. may be measured. The measurement results may be output to the control unit 110.

The transmission path interface 140 transmits and receives signals (backhaul signaling) between devices included in the core network 30 (for example, network nodes providing NF), other base stations 10, etc., and provides information for the user terminal 20. User data (user plane data), control plane data, etc. may be acquired and transmitted.

Note that the transmitting unit and receiving unit of the base station 10 in the present disclosure may be configured by at least one of the transmitting/receiving unit 120, the transmitting/receiving antenna 130, and the transmission path interface 140.

Note that the transmitting/receiving unit 120 may transmit a request for a data set for a specific artificial intelligence (AI) model to the user terminal 20. The transmitting/receiving unit 120 may receive information regarding the data set indicating an identifier (ID) of a particular scenario setting.

Note that the transmitting/receiving unit 120 may transmit information regarding scenario settings for which data sets need to be collected to the user terminal 20. The control unit 110 stores an identifier ( The transmission of model training information indicating the Identifier (ID) may also be controlled.

(user terminal)
FIG. 14 is a diagram illustrating an example of the configuration of a user terminal according to an embodiment. The user terminal 20 includes a control section 210, a transmitting/receiving section 220, and a transmitting/receiving antenna 230. Note that one or more of each of the control unit 210, the transmitting/receiving unit 220, and the transmitting/receiving antenna 230 may be provided.

Note that this example mainly shows functional blocks that are characteristic of the present embodiment, and it may be assumed that the user terminal 20 also has other functional blocks necessary for wireless communication. A part of the processing of each unit described below may be omitted.

The control unit 210 controls the entire user terminal 20. The control unit 210 can be configured from a controller, a control circuit, etc., which will be explained based on common recognition in the technical field related to the present disclosure.

The control unit 210 may control signal generation, mapping, etc. The control unit 210 may control transmission and reception using the transmitting/receiving unit 220 and the transmitting/receiving antenna 230, measurement, and the like. The control unit 210 may generate data, control information, sequences, etc. to be transmitted as a signal, and may transfer the generated data to the transmitting/receiving unit 220.

The transmitting/receiving section 220 may include a baseband section 221, an RF section 222, and a measuring section 223. The baseband section 221 may include a transmission processing section 2211 and a reception processing section 2212. The transmitting/receiving unit 220 can be configured from a transmitter/receiver, an RF circuit, a baseband circuit, a filter, a phase shifter, a measuring circuit, a transmitting/receiving circuit, etc., which are explained based on common recognition in the technical field related to the present disclosure.

The transmitting/receiving section 220 may be configured as an integrated transmitting/receiving section, or may be configured from a transmitting section and a receiving section. The transmitting section may include a transmitting processing section 2211 and an RF section 222. The reception section may include a reception processing section 2212, an RF section 222, and a measurement section 223.

The transmitting/receiving antenna 230 can be configured from an antenna, such as an array antenna, as described based on common recognition in the technical field related to the present disclosure.

The transmitter/receiver 220 may receive the above-mentioned downlink channel, synchronization signal, downlink reference signal, etc. The transmitter/receiver 220 may transmit the above-mentioned uplink channel, uplink reference signal, and the like.

The transmitting/receiving unit 220 may form at least one of a transmitting beam and a receiving beam using digital beamforming (e.g., precoding), analog beamforming (e.g., phase rotation), or the like.

The transmission/reception unit 220 (transmission processing unit 2211) performs PDCP layer processing, RLC layer processing (e.g. RLC retransmission control), MAC layer processing (e.g. , HARQ retransmission control), etc., to generate a bit string to be transmitted.

The transmitting/receiving unit 220 (transmission processing unit 2211) performs channel encoding (which may include error correction encoding), modulation, mapping, filter processing, DFT processing (as necessary), and IFFT processing on the bit string to be transmitted. , precoding, digital-to-analog conversion, etc., and output a baseband signal.

Note that whether or not to apply DFT processing may be based on the settings of transform precoding. When transform precoding is enabled for a certain channel (for example, PUSCH), the transmitting/receiving unit 220 (transmission processing unit 2211) performs the above processing in order to transmit the channel using the DFT-s-OFDM waveform. DFT processing may be performed as the transmission processing, or if not, DFT processing may not be performed as the transmission processing.

The transmitting/receiving unit 220 (RF unit 222) may perform modulation, filter processing, amplification, etc. on the baseband signal in a radio frequency band, and may transmit the signal in the radio frequency band via the transmitting/receiving antenna 230. .

On the other hand, the transmitting/receiving section 220 (RF section 222) may perform amplification, filter processing, demodulation into a baseband signal, etc. on the radio frequency band signal received by the transmitting/receiving antenna 230.

The transmission/reception unit 220 (reception processing unit 2212) performs analog-to-digital conversion, FFT processing, IDFT processing (if necessary), filter processing, demapping, demodulation, and decoding (error correction) on the acquired baseband signal. (which may include decoding), MAC layer processing, RLC layer processing, and PDCP layer processing may be applied to obtain user data and the like.

The transmitting/receiving unit 220 (measuring unit 223) may perform measurements regarding the received signal. For example, the measurement unit 223 may perform RRM measurement, CSI measurement, etc. based on the received signal. The measurement unit 223 may measure received power (for example, RSRP), reception quality (for example, RSRQ, SINR, SNR), signal strength (for example, RSSI), propagation path information (for example, CSI), and the like. The measurement results may be output to the control unit 210.

Note that the transmitting unit and receiving unit of the user terminal 20 in the present disclosure may be configured by at least one of the transmitting/receiving unit 220 and the transmitting/receiving antenna 230.

Note that the transmitting/receiving unit 220 may receive a request for a data set for a specific artificial intelligence (AI) model. The control unit 210 may collect the data set under specific scenario settings. The transmitting/receiving unit 220 may transmit information regarding the data set indicating an identifier (ID) of the specific scenario setting.

The control unit 210 may collect the data set under the specific scenario setting in response to receiving a data collection request signal indicating the ID of the specific scenario setting.

The transmitting/receiving unit 220 may receive information regarding scenario settings that require collection of data sets.

Note that the transmitting/receiving unit 220 may receive information regarding scenario settings that require collection of data sets. The control unit 210 may train an artificial intelligence (AI) model based on a data set collected under a specific scenario setting.

The control unit 210 may train the AI model based on the data set in response to receiving model training information indicating an identifier (ID) of the specific scenario setting.

The control unit 210 may determine the specific scenario setting based on the measurement results of the channel state.

The control unit 210 may determine the specific scenario setting based on the generalization ability of the monitored AI model.

(Hardware configuration)
It should be noted that the block diagram used to explain the above embodiment shows blocks in functional units. These functional blocks (components) are realized by any combination of at least one of hardware and software. Furthermore, the method for realizing each functional block is not particularly limited. That is, each functional block may be realized using one physically or logically coupled device, or may be realized using two or more physically or logically separated devices directly or indirectly (e.g. , wired, wireless, etc.) and may be realized using a plurality of these devices. The functional block may be realized by combining software with the one device or the plurality of devices.

Here, functions include judgment, decision, judgement, calculation, calculation, processing, derivation, investigation, exploration, confirmation, reception, transmission, output, access, solution, selection, selection, establishment, comparison, assumption, expectation, and consideration. , broadcasting, notifying, communicating, forwarding, configuring, reconfiguring, allocating, mapping, assigning, etc. Not limited. For example, a functional block (configuration unit) that performs transmission may be called a transmitting unit, a transmitter, or the like. In either case, as described above, the implementation method is not particularly limited.

For example, a base station, a user terminal, etc. in an embodiment of the present disclosure may function as a computer that performs processing of the wireless communication method of the present disclosure. FIG. 15 is a diagram illustrating an example of the hardware configuration of a base station and a user terminal according to an embodiment. The base station 10 and user terminal 20 described above may be physically configured as a computer device including a processor 1001, a memory 1002, a storage 1003, a communication device 1004, an input device 1005, an output device 1006, a bus 1007, etc. .

Note that in this disclosure, words such as apparatus, circuit, device, section, unit, etc. can be read interchangeably. The hardware configuration of the base station 10 and the user terminal 20 may be configured to include one or more of each device shown in the figure, or may be configured not to include some of the devices.

For example, although only one processor 1001 is illustrated, there may be multiple processors. Also, the processing may be performed by one processor, or the processing may be performed by two or more processors simultaneously, sequentially, or using other techniques. Note that the processor 1001 may be implemented using one or more chips.

Each function in the base station 10 and the user terminal 20 is performed by, for example, loading predetermined software (program) onto hardware such as a processor 1001 and a memory 1002, so that the processor 1001 performs calculations and communicates via the communication device 1004. This is achieved by controlling at least one of reading and writing data in the memory 1002 and storage 1003.

The processor 1001, for example, operates an operating system to control the entire computer. The processor 1001 may be configured by a central processing unit (CPU) that includes interfaces with peripheral devices, a control device, an arithmetic unit, registers, and the like. For example, at least a portion of the above-mentioned control unit 110 (210), transmitting/receiving unit 120 (220), etc. may be realized by the processor 1001.

Furthermore, the processor 1001 reads programs (program codes), software modules, data, etc. from at least one of the storage 1003 and the communication device 1004 to the memory 1002, and executes various processes in accordance with these. As the program, a program that causes a computer to execute at least part of the operations described in the above embodiments is used. For example, the control unit 110 (210) may be realized by a control program stored in the memory 1002 and operated in the processor 1001, and other functional blocks may also be realized in the same way.

The memory 1002 is a computer-readable recording medium, and includes at least one of Read Only Memory (ROM), Erasable Programmable ROM (EPROM), Electrically EPROM (EEPROM), Random Access Memory (RAM), and other suitable storage media. It may be composed of one. Memory 1002 may be called a register, cache, main memory, or the like. The memory 1002 can store executable programs (program codes), software modules, and the like to implement a wireless communication method according to an embodiment of the present disclosure.

The storage 1003 is a computer-readable recording medium, such as a flexible disk, a floppy (registered trademark) disk, a magneto-optical disk (for example, a compact disk (CD-ROM), etc.), a digital versatile disk, removable disk, hard disk drive, smart card, flash memory device (e.g., card, stick, key drive), magnetic stripe, database, server, or other suitable storage medium. It may be configured by Storage 1003 may also be called an auxiliary storage device.

The communication device 1004 is hardware (transmission/reception device) for communicating between computers via at least one of a wired network and a wireless network, and is also referred to as, for example, a network device, network controller, network card, communication module, etc. The communication device 1004 includes, for example, a high frequency switch, a duplexer, a filter, a frequency synthesizer, etc. in order to realize at least one of frequency division duplex (FDD) and time division duplex (TDD). It may be configured to include. For example, the above-described transmitting/receiving section 120 (220), transmitting/receiving antenna 130 (230), etc. may be realized by the communication device 1004. The transmitter/receiver 120 (220) may be physically or logically separated into a transmitter 120a (220a) and a receiver 120b (220b).

The input device 1005 is an input device (eg, keyboard, mouse, microphone, switch, button, sensor, etc.) that accepts input from the outside. The output device 1006 is an output device (for example, a display, a speaker, a light emitting diode (LED) lamp, etc.) that performs output to the outside. Note that the input device 1005 and the output device 1006 may have an integrated configuration (for example, a touch panel).

Further, each device such as the processor 1001 and the memory 1002 is connected by a bus 1007 for communicating information. The bus 1007 may be configured using a single bus, or may be configured using different buses for each device.

The base station 10 and user terminal 20 also include a microprocessor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a programmable logic device (PLD), a field programmable gate array (FPGA), etc. It may be configured to include hardware, and a part or all of each functional block may be realized using the hardware. For example, processor 1001 may be implemented using at least one of these hardwares.

(Modified example)
Note that terms explained in this disclosure and terms necessary for understanding this disclosure may be replaced with terms having the same or similar meanings. For example, channel, symbol and signal may be interchanged. Also, the signal may be a message. The reference signal may also be abbreviated as RS, and may be called a pilot, pilot signal, etc. depending on the applicable standard. Further, a component carrier (CC) may be called a cell, a frequency carrier, a carrier frequency, or the like.

A radio frame may be composed of one or more periods (frames) in the time domain. Each of the one or more periods (frames) constituting a radio frame may be called a subframe. Furthermore, a subframe may be composed of one or more slots in the time domain. A subframe may have a fixed time length (eg, 1 ms) that does not depend on numerology.

Here, the numerology may be a communication parameter applied to at least one of transmission and reception of a certain signal or channel. Numerology includes, for example, subcarrier spacing (SCS), bandwidth, symbol length, cyclic prefix length, transmission time interval (TTI), number of symbols per TTI, and radio frame configuration. , a specific filtering process performed by the transceiver in the frequency domain, a specific windowing process performed by the transceiver in the time domain, etc.

A slot may be composed of one or more symbols (Orthogonal Frequency Division Multiplexing (OFDM) symbols, Single Carrier Frequency Division Multiple Access (SC-FDMA) symbols, etc.) in the time domain. Furthermore, a slot may be a time unit based on numerology.

A slot may include multiple mini-slots. Each minislot may be made up of one or more symbols in the time domain. Furthermore, a mini-slot may also be called a sub-slot. A minislot may be made up of fewer symbols than a slot. PDSCH (or PUSCH) transmitted in time units larger than minislots may be referred to as PDSCH (PUSCH) mapping type A. PDSCH (or PUSCH) transmitted using minislots may be referred to as PDSCH (PUSCH) mapping type B.

Radio frames, subframes, slots, minislots, and symbols all represent time units when transmitting signals. Other names may be used for the radio frame, subframe, slot, minislot, and symbol. Note that time units such as frames, subframes, slots, minislots, and symbols in the present disclosure may be read interchangeably.

For example, one subframe may be called a TTI, a plurality of consecutive subframes may be called a TTI, and one slot or one minislot may be called a TTI. In other words, at least one of the subframe and TTI may be a subframe (1ms) in existing LTE, a period shorter than 1ms (for example, 1-13 symbols), or a period longer than 1ms. It may be. Note that the unit representing the TTI may be called a slot, minislot, etc. instead of a subframe.

Here, TTI refers to, for example, the minimum time unit for scheduling in wireless communication. For example, in the LTE system, a base station performs scheduling to allocate radio resources (frequency bandwidth, transmission power, etc. that can be used by each user terminal) to each user terminal on a TTI basis. Note that the definition of TTI is not limited to this.

The TTI may be a transmission time unit of a channel-coded data packet (transport block), a code block, a codeword, etc., or may be a processing unit of scheduling, link adaptation, etc. Note that when a TTI is given, the time interval (for example, the number of symbols) to which transport blocks, code blocks, code words, etc. are actually mapped may be shorter than the TTI.

Note that when one slot or one minislot is called a TTI, one or more TTIs (that is, one or more slots or one or more minislots) may be the minimum time unit for scheduling. Further, the number of slots (minislot number) that constitutes the minimum time unit of the scheduling may be controlled.

A TTI having a time length of 1 ms may be called a normal TTI (TTI in 3GPP Rel. 8-12), normal TTI, long TTI, normal subframe, normal subframe, long subframe, slot, etc. A TTI that is shorter than the normal TTI may be referred to as an abbreviated TTI, short TTI, partial or fractional TTI, shortened subframe, short subframe, minislot, subslot, slot, etc.

Note that long TTI (for example, normal TTI, subframe, etc.) may be read as TTI with a time length exceeding 1 ms, and short TTI (for example, short TTI, etc.) It may also be read as a TTI having the above TTI length.

A resource block (RB) is a resource allocation unit in the time domain and frequency domain, and may include one or more continuous subcarriers (subcarriers) in the frequency domain. The number of subcarriers included in an RB may be the same regardless of the numerology, and may be 12, for example. The number of subcarriers included in an RB may be determined based on numerology.

Additionally, an RB may include one or more symbols in the time domain, and may have a length of one slot, one minislot, one subframe, or one TTI. One TTI, one subframe, etc. may each be composed of one or more resource blocks.

Note that one or more RBs include a physical resource block (Physical RB (PRB)), a sub-carrier group (SCG), a resource element group (REG), a PRB pair, and an RB. They may also be called pairs.

Additionally, a resource block may be configured by one or more resource elements (REs). For example, 1 RE may be a radio resource region of 1 subcarrier and 1 symbol.

Bandwidth Part (BWP) (also called partial bandwidth, etc.) refers to a subset of consecutive common resource blocks (RB) for a certain numerology in a certain carrier. Good too. Here, the common RB may be specified by an RB index based on a common reference point of the carrier. PRBs may be defined in a BWP and numbered within that BWP.

BWP may include UL BWP (BWP for UL) and DL BWP (BWP for DL). One or more BWPs may be configured within one carrier for a UE.

At least one of the configured BWPs may be active and the UE may not expect to transmit or receive a given signal/channel outside of the active BWP. Note that "cell", "carrier", etc. in the present disclosure may be replaced with "BWP".

Note that the structures of the radio frame, subframe, slot, minislot, symbol, etc. described above are merely examples. For example, the number of subframes included in a radio frame, the number of slots per subframe or radio frame, the number of minislots included in a slot, the number of symbols and RBs included in a slot or minislot, the number of symbols included in an RB, The number of subcarriers, the number of symbols within a TTI, the symbol length, the cyclic prefix (CP) length, and other configurations can be changed in various ways.

In addition, the information, parameters, etc. described in this disclosure may be expressed using absolute values, relative values from a predetermined value, or using other corresponding information. may be expressed. For example, radio resources may be indicated by a predetermined index.

The names used for parameters and the like in this disclosure are not limiting in any respect. Furthermore, the mathematical formulas etc. using these parameters may differ from those explicitly disclosed in this disclosure. Since the various channels (PUCCH, PDCCH, etc.) and information elements can be identified by any suitable designation, the various names assigned to these various channels and information elements are not in any way exclusive designations. .

The information, signals, etc. described in this disclosure may be represented using any of a variety of different technologies. For example, data, instructions, commands, information, signals, bits, symbols, chips, etc., which may be referred to throughout the above description, may refer to voltages, currents, electromagnetic waves, magnetic fields or magnetic particles, light fields or photons, or any of these. It may also be represented by a combination of

Additionally, information, signals, etc. may be output from the upper layer to the lower layer and from the lower layer to at least one of the upper layer. Information, signals, etc. may be input and output via multiple network nodes.

Input/output information, signals, etc. may be stored in a specific location (for example, memory) or may be managed using a management table. Information, signals, etc. that are input and output can be overwritten, updated, or added. The output information, signals, etc. may be deleted. The input information, signals, etc. may be transmitted to other devices.

Notification of information is not limited to the aspects/embodiments described in this disclosure, and may be performed using other methods. For example, the notification of information in this disclosure may be physical layer signaling (e.g., Downlink Control Information (DCI), Uplink Control Information (UCI)), upper layer signaling (e.g., Radio Resource Control (RRC) signaling, broadcast information (Master Information Block (MIB), System Information Block (SIB), etc.), Medium Access Control (MAC) signaling), other signals, or a combination thereof It may be carried out by

Note that the physical layer signaling may also be called Layer 1/Layer 2 (L1/L2) control information (L1/L2 control signal), L1 control information (L1 control signal), etc. Further, RRC signaling may be called an RRC message, and may be, for example, an RRC Connection Setup message, an RRC Connection Reconfiguration message, or the like. Further, MAC signaling may be notified using, for example, a MAC Control Element (CE).

Further, notification of prescribed information (for example, notification of "X") is not limited to explicit notification, but may be made implicitly (for example, by not notifying the prescribed information or by providing other information) (by notification).

The determination may be made by a value expressed by 1 bit (0 or 1), or by a boolean value expressed by true or false. , may be performed by numerical comparison (for example, comparison with a predetermined value).

Software includes instructions, instruction sets, code, code segments, program code, programs, subprograms, software modules, whether referred to as software, firmware, middleware, microcode, hardware description language, or by any other name. , should be broadly construed to mean an application, software application, software package, routine, subroutine, object, executable, thread of execution, procedure, function, etc.

Additionally, software, instructions, information, etc. may be sent and received via a transmission medium. For example, if the software uses wired technology (such as coaxial cable, fiber optic cable, twisted pair, Digital Subscriber Line (DSL), etc.) and/or wireless technology (such as infrared, microwave, etc.) to , a server, or other remote source, these wired and/or wireless technologies are included within the definition of a transmission medium.

The terms "system" and "network" used in this disclosure may be used interchangeably. "Network" may refer to devices (eg, base stations) included in the network.

In this disclosure, "precoding", "precoder", "weight (precoding weight)", "quasi-co-location (QCL)", "Transmission Configuration Indication state (TCI state)", "space "spatial relation", "spatial domain filter", "transmission power", "phase rotation", "antenna port", "antenna port group", "layer", "number of layers", Terms such as "rank", "resource", "resource set", "resource group", "beam", "beam width", "beam angle", "antenna", "antenna element", and "panel" are interchangeable. can be used.

In the present disclosure, "Base Station (BS)", "Wireless base station", "Fixed station", "NodeB", "eNB (eNodeB)", "gNB (gNodeB)", "Access point", "Transmission Point (TP)", "Reception Point (RP)", "Transmission/Reception Point (TRP)", "Panel" , "cell," "sector," "cell group," "carrier," "component carrier," and the like may be used interchangeably. A base station is sometimes referred to by terms such as macrocell, small cell, femtocell, and picocell.

A base station can accommodate one or more (eg, three) cells. If a base station accommodates multiple cells, the overall coverage area of the base station can be partitioned into multiple smaller areas, and each smaller area is connected to a base station subsystem (e.g., an indoor small base station (Remote Radio Communication services can also be provided by the Head (RRH)). The term "cell" or "sector" refers to part or all of the coverage area of a base station and/or base station subsystem that provides communication services in this coverage.

In the present disclosure, a base station transmitting information to a terminal may be interchanged with the base station instructing the terminal to control/operate based on the information.

In this disclosure, terms such as "Mobile Station (MS)," "user terminal," "User Equipment (UE)," and "terminal" are used interchangeably. can be done.

A mobile station is a subscriber station, mobile unit, subscriber unit, wireless unit, remote unit, mobile device, wireless device, wireless communication device, remote device, mobile subscriber station, access terminal, mobile terminal, wireless terminal, remote terminal. , handset, user agent, mobile client, client, or some other suitable terminology.

At least one of a base station and a mobile station may be called a transmitting device, a receiving device, a wireless communication device, etc. Note that at least one of the base station and the mobile station may be a device mounted on a moving object, the moving object itself, or the like.

The moving body refers to a movable object, and the moving speed is arbitrary, and naturally includes cases where the moving body is stopped. The mobile objects include, for example, vehicles, transport vehicles, automobiles, motorcycles, bicycles, connected cars, excavators, bulldozers, wheel loaders, dump trucks, forklifts, trains, buses, carts, rickshaws, and ships (ships and other watercraft). , including, but not limited to, airplanes, rockets, artificial satellites, drones, multicopters, quadcopters, balloons, and items mounted thereon. Furthermore, the mobile object may be a mobile object that autonomously travels based on a travel command.

The moving object may be a vehicle (for example, a car, an airplane, etc.), an unmanned moving object (for example, a drone, a self-driving car, etc.), or a robot (manned or unmanned). ). Note that at least one of the base station and the mobile station includes devices that do not necessarily move during communication operations. For example, at least one of the base station and the mobile station may be an Internet of Things (IoT) device such as a sensor.

FIG. 16 is a diagram illustrating an example of a vehicle according to an embodiment. The vehicle 40 includes a drive unit 41, a steering unit 42, an accelerator pedal 43, a brake pedal 44, a shift lever 45, left and right front wheels 46, left and right rear wheels 47, an axle 48, an electronic control unit 49, various sensors (current sensor 50, (including a rotation speed sensor 51, an air pressure sensor 52, a vehicle speed sensor 53, an acceleration sensor 54, an accelerator pedal sensor 55, a brake pedal sensor 56, a shift lever sensor 57, and an object detection sensor 58), an information service section 59, and a communication module 60. Be prepared.

The drive unit 41 is composed of, for example, at least one of an engine, a motor, and a hybrid of an engine and a motor. The steering unit 42 includes at least a steering wheel (also referred to as a steering wheel), and is configured to steer at least one of the front wheels 46 and the rear wheels 47 based on the operation of the steering wheel operated by the user.

The electronic control unit 49 includes a microprocessor 61, a memory (ROM, RAM) 62, and a communication port (for example, an input/output (IO) port) 63. Signals from various sensors 50-58 provided in the vehicle are input to the electronic control unit 49. The electronic control section 49 may be called an electronic control unit (ECU).

The signals from the various sensors 50 to 58 include a current signal from the current sensor 50 that senses the current of the motor, a rotation speed signal of the front wheel 46/rear wheel 47 obtained by the rotation speed sensor 51, and a signal obtained by the air pressure sensor 52. air pressure signals of the front wheels 46/rear wheels 47, a vehicle speed signal acquired by the vehicle speed sensor 53, an acceleration signal acquired by the acceleration sensor 54, a depression amount signal of the accelerator pedal 43 acquired by the accelerator pedal sensor 55, and a brake pedal sensor. 56, a shift lever 45 operation signal obtained by the shift lever sensor 57, and an object detection sensor 58 for detecting obstacles, vehicles, pedestrians, etc. There are signals etc.

The information service department 59 includes various devices such as car navigation systems, audio systems, speakers, displays, televisions, and radios that provide (output) various information such as driving information, traffic information, and entertainment information, and these devices. It consists of one or more ECUs that control the The information service unit 59 provides various information/services (for example, multimedia information/multimedia services) to the occupants of the vehicle 40 using information acquired from an external device via the communication module 60 or the like.

The information service unit 59 may include an input device (for example, a keyboard, a mouse, a microphone, a switch, a button, a sensor, a touch panel, etc.) that accepts input from the outside, and an output device that performs output to the outside (for example, display, speaker, LED lamp, touch panel, etc.).

The driving support system unit 64 includes millimeter wave radar, Light Detection and Ranging (LiDAR), a camera, a positioning locator (for example, Global Navigation Satellite System (GNSS), etc.), and map information (for example, High Definition (HD)). maps, autonomous vehicle (AV) maps, etc.), gyro systems (e.g., inertial measurement units (IMUs), inertial navigation systems (INS), etc.), artificial intelligence ( Artificial Intelligence (AI) chips, AI processors, and other devices that provide functions to prevent accidents and reduce the driver's driving burden, as well as one or more devices that control these devices. It consists of an ECU. Further, the driving support system section 64 transmits and receives various information via the communication module 60, and realizes a driving support function or an automatic driving function.

The communication module 60 can communicate with the microprocessor 61 and components of the vehicle 40 via the communication port 63. For example, the communication module 60 communicates via the communication port 63 with a drive unit 41, a steering unit 42, an accelerator pedal 43, a brake pedal 44, a shift lever 45, left and right front wheels 46, left and right rear wheels 47, which are included in the vehicle 40. Data (information) is transmitted and received between the axle 48, the microprocessor 61 and memory (ROM, RAM) 62 in the electronic control unit 49, and various sensors 50-58.

The communication module 60 is a communication device that can be controlled by the microprocessor 61 of the electronic control unit 49 and can communicate with external devices. For example, various information is transmitted and received with an external device via wireless communication. The communication module 60 may be located either inside or outside the electronic control unit 49. The external device may be, for example, the base station 10, user terminal 20, etc. described above. Further, the communication module 60 may be, for example, at least one of the base station 10 and the user terminal 20 described above (it may function as at least one of the base station 10 and the user terminal 20).

The communication module 60 receives signals from the various sensors 50 to 58 described above that are input to the electronic control unit 49, information obtained based on the signals, and input from the outside (user) obtained via the information service unit 59. At least one of the information based on the information may be transmitted to an external device via wireless communication. The electronic control unit 49, various sensors 50-58, information service unit 59, etc. may be called an input unit that receives input. For example, the PUSCH transmitted by the communication module 60 may include information based on the above input.

The communication module 60 receives various information (traffic information, signal information, inter-vehicle information, etc.) transmitted from an external device, and displays it on the information service section 59 provided in the vehicle. The information service unit 59 is an output unit that outputs information (for example, outputs information to devices such as a display and a speaker based on the PDSCH (or data/information decoded from the PDSCH) received by the communication module 60). may be called.

The communication module 60 also stores various information received from external devices into a memory 62 that can be used by the microprocessor 61. Based on the information stored in the memory 62, the microprocessor 61 controls the drive unit 41, steering unit 42, accelerator pedal 43, brake pedal 44, shift lever 45, left and right front wheels 46, and left and right rear wheels provided in the vehicle 40. 47, axle 48, various sensors 50-58, etc. may be controlled.

Additionally, the base station in the present disclosure may be replaced by a user terminal. For example, communication between a base station and a user terminal is replaced with communication between multiple user terminals (for example, it may be called Device-to-Device (D2D), Vehicle-to-Everything (V2X), etc.). Regarding the configuration, each aspect/embodiment of the present disclosure may be applied. In this case, the user terminal 20 may have the functions that the base station 10 described above has. Further, words such as "uplink" and "downlink" may be replaced with words corresponding to inter-terminal communication (for example, "sidelink"). For example, uplink channels, downlink channels, etc. may be replaced with sidelink channels.

Similarly, the user terminal in the present disclosure may be replaced with a base station. In this case, the base station 10 may have the functions that the user terminal 20 described above has.

In this disclosure, the operations performed by the base station may be performed by its upper node in some cases. In a network that includes one or more network nodes having a base station, various operations performed for communication with a terminal may be performed by the base station, one or more network nodes other than the base station (e.g. It is clear that this can be performed by a Mobility Management Entity (MME), a Serving-Gateway (S-GW), etc. (though not limited thereto), or a combination thereof.

Each aspect/embodiment described in this disclosure may be used alone, in combination, or may be switched and used in accordance with execution. Further, the order of the processing procedures, sequences, flowcharts, etc. of each aspect/embodiment described in this disclosure may be changed as long as there is no contradiction. For example, the methods described in this disclosure use an example order to present elements of the various steps and are not limited to the particular order presented.

Each aspect/embodiment described in this disclosure includes Long Term Evolution (LTE), LTE-Advanced (LTE-A), LTE-Beyond (LTE-B), SUPER 3G, IMT-Advanced, 4th generation mobile communication system ( 4G), 5th generation mobile communication system (5G), 6th generation mobile communication system (6G), xth generation mobile communication system (xG (x is an integer or decimal number, for example)), Future Radio Access (FRA), New-Radio Access Technology (RAT), New Radio (NR), New radio access (NX), Future generation radio access (FX), Global System for Mobile communications (GSM (registered trademark)), CDMA2000, Ultra Mobile Broadband (UMB), IEEE 802 .11 (Wi-Fi (registered trademark)), IEEE 802.16 (WiMAX (registered trademark)), IEEE 802.20, Ultra-WideBand (UWB), Bluetooth (registered trademark), and other appropriate wireless communication methods. The present invention may be applied to systems to be used, next-generation systems expanded, modified, created, or defined based on these systems. Furthermore, a combination of multiple systems (for example, a combination of LTE or LTE-A and 5G) may be applied.

As used in this disclosure, the phrase "based on" does not mean "based solely on" unless explicitly stated otherwise. In other words, the phrase "based on" means both "based only on" and "based at least on."

As used in this disclosure, any reference to elements using the designations "first," "second," etc. does not generally limit the amount or order of those elements. These designations may be used in this disclosure as a convenient way to distinguish between two or more elements. Thus, reference to a first and second element does not imply that only two elements may be employed or that the first element must precede the second element in any way.

The term "determining" as used in this disclosure may encompass a wide variety of actions. For example, "judgment" can mean judging, calculating, computing, processing, deriving, investigating, looking up, search, inquiry ( For example, searching in a table, database, or other data structure), ascertaining, etc. may be considered to be "determining."

In addition, "judgment (decision)" includes receiving (e.g., receiving information), transmitting (e.g., sending information), input (input), output (output), access ( may be considered to be "determining", such as accessing data in memory (eg, accessing data in memory).

In addition, "judgment" is considered to mean "judging" resolving, selecting, choosing, establishing, comparing, etc. Good too. In other words, "judgment (decision)" may be considered to be "judgment (decision)" of some action.

Furthermore, "judgment (decision)" may be read as "assuming", "expecting", "considering", etc.

The "maximum transmit power" described in this disclosure may mean the maximum value of transmit power, the nominal maximum transmit power (the nominal UE maximum transmit power), or the rated maximum transmit power (the It may also mean rated UE maximum transmit power).

As used in this disclosure, the terms "connected", "coupled", or any variations thereof refer to any connection or coupling, direct or indirect, between two or more elements. can include the presence of one or more intermediate elements between two elements that are "connected" or "coupled" to each other. The coupling or connection between elements may be physical, logical, or a combination thereof. For example, "connection" may be replaced with "access."

In this disclosure, when two elements are connected, they may be connected using one or more electrical wires, cables, printed electrical connections, etc., as well as in the radio frequency domain, microwave can be considered to be "connected" or "coupled" to each other using electromagnetic energy having wavelengths in the light (both visible and invisible) range.

In the present disclosure, the term "A and B are different" may mean "A and B are different from each other." Note that the term may also mean that "A and B are each different from C". Terms such as "separate" and "coupled" may also be interpreted similarly to "different."

Where "include", "including" and variations thereof are used in this disclosure, these terms are inclusive, as is the term "comprising". It is intended that Furthermore, the term "or" as used in this disclosure is not intended to be exclusive or.

In this disclosure, when articles are added by translation, such as a, an, and the in English, the present disclosure may include that the nouns following these articles are plural.

In the present disclosure, "less than or equal to", "less than", "more than", "more than", "equal to", etc. may be read interchangeably. In addition, in the present disclosure, words meaning "good", "bad", "large", "small", "high", "low", "early", "slow", etc. may be read interchangeably. (Not limited to original, comparative, and superlative). In addition, in this disclosure, words meaning "good", "bad", "large", "small", "high", "low", "early", "slow", etc. are replaced with "i-th". They may be interchanged as expressions (not limited to the original, comparative, and superlative) (for example, "the highest" may be interchanged with "the i-th highest").

In this disclosure, "of", "for", "regarding", "related to", "associated with", etc. are used to refer to each other. It may be read differently.

Although the invention according to the present disclosure has been described in detail above, it is clear for those skilled in the art that the invention according to the present disclosure is not limited to the embodiments described in the present disclosure. The invention according to the present disclosure can be implemented as modifications and variations without departing from the spirit and scope of the invention as determined based on the claims. Therefore, the description of the present disclosure is for the purpose of illustrative explanation and does not have any limiting meaning on the invention according to the present disclosure.

Claims (6)

1. a receiving unit that receives information regarding scenario settings that require data set collection;
   A terminal having a control unit that trains an artificial intelligence (AI) model based on a data set collected under a specific scenario setting.
2. The terminal according to claim 1, wherein the control unit trains the AI model based on the data set in response to receiving model training information indicating an identifier (ID) of the specific scenario setting.
3. The terminal according to claim 1, wherein the control unit determines the specific scenario setting based on a measurement result of a channel state.
4. The terminal according to claim 1, wherein the control unit determines the specific scenario setting based on the generalization ability of the monitored AI model.
5. receiving information about scenario settings for which datasets need to be collected;
   1. A method for wireless communication of a terminal, comprising: training an artificial intelligence (AI) model based on a data set collected under a specific scenario setting.
6. a transmitting unit that transmits information regarding scenario settings that require data set collection to the terminal;
   Indicating an identifier (ID) of the specific scenario setting to train an Artificial Intelligence (AI) model based on a dataset collected at the terminal under the specific scenario setting. A base station comprising: a control unit that controls transmission of model training information.

Images