Method for monitoring CSI prediction and device therefor

Abstract

A method according to an embodiment of the present specification comprises the steps of: receiving configuration information related to CSI; reporting first CSI including information related to prediction; and reporting second CSI related to prediction accuracy. The configuration information includes a first report configuration related to prediction and a second report configuration related to prediction accuracy. The second report configuration includes information indicating a time instance for monitoring among one or more time instances related to prediction.

Description

Method and device for monitoring CSI prediction

[0001] The present specification relates to a method and device for monitoring CSI prediction.

[0002] Mobile communication systems were developed to provide voice services while ensuring user activity. However, they have expanded beyond voice to include data services. Currently, explosive growth in traffic is leading to resource shortages and users are demanding faster services, necessitating a more advanced mobile communication system.

[0003] Next-generation mobile communication systems must support explosive data traffic growth, dramatically increasing data rates per user, a vastly increased number of connected devices, ultra-low end-to-end latency, and high energy efficiency. To achieve these goals, various technologies are being studied, including dual connectivity, massive multiple input multiple output (MIMO), in-band full duplex, non-orthogonal multiple access (NOMA), super wideband support, and device networking.

[0004] Performance monitoring for CSI prediction may be performed. Specifically, the terminal may transmit a report related to monitoring the accuracy of predicted information (e.g., predicted CRI, predicted SSBRI). The report may include performance metric(s). For example, to monitor the performance of CSI prediction for a single time instance, a performance metric (e.g., a prediction accuracy indicator) for the CSI prediction for that single time instance may be reported.

[0005] Meanwhile, CSI prediction can be performed for more than one time instance. In such cases, it's unclear how performance metrics should be determined and reported. For example, calculating and determining all performance metrics for each time instance and then reporting them could be considered. However, it's unlikely that the efficiency of model management through performance monitoring will increase proportionally with the number of reported performance metrics. Therefore, the efficiency improvement due to monitoring may not be significant compared to the increased signaling overhead.

[0006] The purpose of this specification is to propose a method to solve the above-mentioned problems.

[0007] The technical problems to be achieved in this specification are not limited to the technical problems mentioned above, and other technical problems not mentioned can be clearly understood by a person having ordinary skill in the technical field to which this specification pertains from the description below.

[0008] According to one embodiment of the present disclosure for solving the above-described problem, a method includes the steps of receiving configuration information related to channel state information (CSI), reporting a first CSI including information related to prediction, and reporting a second CSI related to prediction accuracy. The configuration information includes i) a first reporting configuration related to prediction, and ii) a second reporting configuration related to prediction accuracy. The first CSI includes predicted information for each of one or more time instances. The second CSI includes information related to a time instance for monitoring among the one or more time instances. The second reporting configuration is characterized in that it includes information indicating the time instance for monitoring among the one or more time instances.

[0009] Therefore, since monitoring is performed based on an explicitly designated time instance among one or more time instances, the ambiguity regarding which time instance a performance metric should be calculated / determined for monitoring related to CSI prediction can be resolved.

[0010] According to embodiments of the present disclosure, when prediction is performed for one or more time instances, information based on a time instance designated for monitoring among the one or more time instances is reported. Accordingly, performance monitoring can be performed more effectively. In other words, when prediction is performed for one or more time instances, a performance metric (e.g., a prediction accuracy indicator) can be calculated / determined / reported based on the most suitable time instance in terms of model management efficiency / signaling overhead according to performance monitoring.

[0011] The effects that can be obtained from this specification are not limited to the effects mentioned above, and other effects that are not mentioned can be clearly understood by a person having ordinary skill in the technical field to which this specification belongs from the description below.

[0012] Figure 1 is a diagram to explain overall functions from an AI / ML model perspective.

[0013] Figure 2 illustrates a general form of AI / ML related procedures performed between a network and a terminal.

[0014] Figure 3 illustrates an example of AI / ML-based beam management operations.

[0015] Figure 4 illustrates an example of AI / ML-based CSI measurement / reporting operations.

[0016] Figure 5 illustrates an example of AI / ML-based positioning operation.

[0017] Figure 6 is a flowchart showing an example of a CSI-related procedure.

[0018] Figure 7 illustrates an AI / ML-based CSI reporting operation.

[0019] Figure 8 illustrates AI / ML-based CSI prediction.

[0020] Figure 9 illustrates a signaling procedure based on an AI / ML model.

[0021] Figure 10 is a diagram for explaining CSI prediction for multiple time instances.

[0022] FIG. 11 is a flowchart illustrating a method according to one embodiment of the present specification.

[0023] FIG. 12 is a flowchart illustrating a method according to another embodiment of the present specification.

[0024] FIG. 13 is a drawing showing the configuration of a first device and a second device according to an embodiment of the present specification.

[0025] As used herein, "A or B" can mean "only A," "only B," or "both A and B." In other words, as used herein, "A or B" can be interpreted as "A and / or B." For example, as used herein, "A, B or C" can mean "only A," "only B," "only C," or "any combination of A, B and C."

[0026] As used herein, a slash ( / ) or a comma can mean "and / or." For example, "A / B" can mean "A and / or B." Accordingly, "A / B" can mean "only A," "only B," or "both A and B." For example, "A, B, C" can mean "A, B, or C."

[0027] In this specification, "at least one of A and B" may mean "only A", "only B" or "both A and B". Additionally, in this specification, the expressions "at least one of A or B" or "at least one of A and / or B" may be interpreted identically to "at least one of A and B".

[0028] Additionally, in this specification, “at least one of A, B and C” can mean “only A,” “only B,” “only C,” or “any combination of A, B and C.” Additionally, “at least one of A, B or C” or “at least one of A, B and / or C” can mean “at least one of A, B and C.”

[0029] Additionally, parentheses used herein may mean "for example." Specifically, when indicated as "control information (PDCCH)", "PDCCH" may be proposed as an example of "control information." In other words, "control information" in this specification is not limited to "PDCCH," and "PDCCH" may be proposed as an example of "control information." Furthermore, even when indicated as "control information (i.e., PDCCH)", "PDCCH" may be proposed as an example of "control information."

[0030] In the following explanation, ‘when, if, in case of’ can be replaced with ‘based on’.

[0031] Technical features individually described in a single drawing in this specification may be implemented individually or simultaneously.

[0032] Hereinafter, preferred embodiments according to the present disclosure will be described in detail with reference to the accompanying drawings. The detailed description set forth below, together with the accompanying drawings, is intended to illustrate exemplary embodiments of the present disclosure and is not intended to represent the only embodiments in which the present disclosure may be implemented. The following detailed description includes specific details to provide a thorough understanding of the present disclosure.

[0033] In this specification, a terminal is a user equipment (UE) or a consumer-side device, and may also be referred to as a base station / second node / IAB node / first node that receives / transmits signals from / to a Transmission-Reception Point (TRP). A terminal may correspond to a physical node or a logical node. A terminal may correspond to an endpoint on the user side, or may correspond to an intermediate point between other endpoints. In communication between two points that are not limited to endpoints (including one-to-one / many-to-one / one-to-many / many-to-many communication), a terminal may correspond to a served node. A terminal may be a node with a fixed location, or a node with an unfixed location (or mobile).

[0034] In this specification, a base station (BS) is a device on the network side, and may also be called a second node / IAB node / x-NodeB (x-NodeB, x may be an abbreviation related to radio access technology (RAT)) / Transmission-Reception Point (TRP). A BS may correspond to a physical node or a logical node. A BS may correspond to an endpoint on the network side, or may correspond to an intermediate point between other endpoints. In communication between two points that are not limited to endpoints (including one-to-one / many-to-one / one-to-many / many-to-many communication), a BS may correspond to a serving node. A BS may be a node with a fixed location, or a node with an unfixed location.

[0035] In this specification, higher layer parameters may be set for the terminal, preset, or predefined. For example, the base station may transmit higher layer parameters to the terminal. For example, the terminal may transmit parameters such as capabilities to the base station as higher layer parameters. For example, the higher layer parameters may be transmitted via radio resource control (RRC) signaling or medium access control (MAC) signaling.

[0036] In this specification, the information / state / parameter being “configured or pre-configured” can be interpreted as the information / state / parameter being provided / pre-provided to the terminal through pre-defined signaling (e.g., SIB, MAC, RRC) from the base station. In this specification, the information / state / parameter being “defined or pre-defined” can be interpreted as the information / state / parameter being known in advance or pre-stored at the base station and the terminal without signaling between the base station and the terminal.

[0037] < AI / ML for Wireless Communication >

[0038] With the advancement of computing technology, artificial intelligence (AI) and machine learning (ML) are being adopted in various industries and technical fields. In the field of wireless communications, various discussions are underway to apply AI models trained based on ML, and the 3GPP standardization process refers to this as AI / ML. This specification describes "AI / ML" according to the terminology used in the 3GPP standardization process. However, "AI / ML" may be referred to by various other terms depending on the progress and implementation of future standards. For example, it may be referred to as "transmission / reception mode" or "signal / channel / operation / transmission / reception configuration" configured for AI / ML, but is not limited thereto. The meanings of the terms currently used in the 3GPP standardization process are briefly summarized as follows.

[0039] - AI / ML model: A data-driven algorithm that applies AI / ML technology to generate a set of outputs containing prediction information and / or decision parameters based on a set of inputs.

[0040] - Data collection: The process of collecting data required for AI / ML model training, data analysis, and inference from network nodes, management entities, or terminals.

[0041] - AI / ML Training: An online or offline process of training an AI model by learning features and patterns that best represent the data and obtain a trained AI / ML model for inference.

[0042] - Offline training: The process of training a model based on a previously collected data set, and the trained model is used or provided for future inference.

[0043] - Online training: This is a method in which the model is trained in real time when new training sample data is acquired and used for inference.

[0044] - AI / ML Inference: This is the process of making predictions or inducing decisions based on collected data and the AI ​​model using a trained AI model. Meanwhile, depending on whether the AI / ML model is set up on both the transmitting and receiving devices or only on one of them, it can be divided into (i) a two-sided model and (ii) a one-sided model. (i) In the case of the two-sided model, collaborative inference is performed through paired AI / ML models. Collaborative inference refers to cooperation between the network and the UE, in which one party performs part of the inference and the other party performs the rest of the inference. (ii) The one-sided model is divided into a UE-side model and a network-side model. In the one-sided model, inference is performed entirely by the UE / network-side model.

[0045] 1. Life Cycle Management (LCM) for AI / ML models

[0046] LCM for AI / ML models is a concept that encompasses all the overall procedures for AI / ML models, including data collection, model training, model deployment, model inference, model monitoring, and model updates.

[0047] LCM for AI / ML models can be broadly categorized into functionality-based LCM and model ID-based LCM. In functionality-based LCM, the network can instruct the activation / deactivation / fallback / switching of specific functions, even if the target AI / ML model may not be identified by the network. In model ID-based LCM, the network can instruct the activation / deactivation / selection / switching of AI / ML models identified by their AI / ML model ID.

[0048] Figure 1 is a diagram to explain overall functions from an AI / ML model perspective.

[0049] Referring to Fig. 1, a general AI / ML functional framework can be configured to include a data collection function (10), a model training function (20), a management function (30), an inference function (40), and a model storage function (50).

[0050] The Data Collection function (10) is a function that provides input data to the Model Training function (20), Management function (30), and Inference function (40). The Data Collection function (10) can perform data preparation and provide input data processed through data preparation.

[0051] Here, training data (11) refers to data required as input for the AI / ML Model Training function (20). Monitoring data (12) refers to data required as input for the Management (30) of the AI / ML model or AI / ML function. Inference data (13) refers to data required as input for the AI / ML Inference function (30).

[0052] The Model Training function (20) is a function that performs AI / ML model training, validation, and testing, and can generate model performance metrics that can be used as part of the AI / ML model testing procedure. If necessary, the Model Training function (20) can perform data preparation (e.g., data pre-processing and cleaning, forming, and transformation) based on the Training Data (11) transmitted from the Data Collection function (10).

[0053] Trained / Updated Model (21): If there is a Model Storage function (50), it is used to pass a trained, validated and tested AI / ML model to the Model Storage function (50) or to pass an updated version of the model to the Model Storage function (50).

[0054] The Management function (30) is a function that monitors the operation of the AI / ML model or AI / ML function. In addition, the Management function (30) may perform a decision to ensure appropriate inference operation based on data received from the Data Collection function (10) (i.e., Monitoring Data (12)) and / or data received from the Inference function (40) (i.e., Inference Output (41)).

[0055] Management Instruction (32) is information required as input to manage the Inference function (40). The relevant information may include selection / (de)activation / switching of an AI / ML model or AI / ML-based function, and may also include fallback to non-AI / ML operations (i.e., not relying on the inference process).

[0056] A Model Transfer / Delivery Request (33) can be used to request model(s) from Model Storage (50).

[0057] A Performance Feedback / Retraining Request (31) refers to information required as input to the Model Training function (20) (e.g., for the purpose of (re)training or updating the model).

[0058] The Inference function (40) is a function that provides output from the process of applying an AI / ML model or AI / ML function using data (i.e., Inference Data (13)) provided by Data Collection (10) as input. Data preparation (e.g., data preprocessing and cleaning, formatting, and transformation) may also be performed based on the Inference Data (13) delivered by Data Collection (10). If necessary, the Inference function (40) may also perform data preparation (e.g., data preprocessing and cleaning, forming, and transformation) based on the Inference Data (13) provided by Data Collection function (10).

[0059] Inference Output (41) is data used in the Management function (30) to monitor the performance of an AI / ML model or AI / ML function. Inference Output (41) may include the inference output of the AI / ML model generated by the Inference function (30), and the details of the inference output may vary depending on the use case.

[0060] The Model Storage function (50) stores a learned / updated model that can be used to perform the Inference function (40). The Model Storage function (50) illustrated in Fig. 1 can be used as a reference point (if any) when applicable to protocol termination, model transmission / delivery, and related processes. Furthermore, the Model Storage function (50) is merely an example and is not intended to limit the storage location of actual AI / ML models, and may be omitted.

[0061] Model Transfer / Delivery (51) is used to transfer AI / ML models to inference functions.

[0062] 2. General AI / ML-related procedures between networks and terminals

[0063] Figure 2 illustrates the general form of AI / ML-related procedures performed between a network and a terminal. While Figure 1 examined LCM from an AI / ML model perspective, Figure 2 describes the general form of procedures performed from a signaling / protocol perspective between a terminal and the network.

[0064] (1) Setting procedures related to AI / ML

[0065] Referring to FIG. 2, an AI / ML-related configuration procedure may be performed between the network and the terminal (B05). The AI / ML-related configuration procedure may include information exchange between the terminal and the network via at least one upper-layer signaling, and / or preparatory / follow-up operations at the terminal / network, respectively, before / after the upper-layer signaling.

[0066] Specifically, the AI / ML-related configuration procedure may include, but is not limited to, at least one of (i) AI / ML-related terminal capability reporting, (ii) data collection, (iii) model training, (iv) model transfer / transfer, (v) AI / ML function / model selection, and (vi) configuration for various operations performed based on AI / ML models (e.g., AI / ML-based CSI / Positioning / Beam Management).

[0067] (i) A terminal can report to the network its capabilities, such as models / functionalities supported by the terminal in relation to AI / ML, through UE Capability Reporting. The network can provide AI / ML-related settings to the terminal based on the AI / ML-related capabilities reported by the terminal.

[0068] (ii) AI / ML-related configuration procedures may include data collection and / or provision of configuration information related to AI / ML model training / inference, etc. Configuration information related to data collection may relate to how to configure the method / action of data collection, etc.

[0069] (iii) AI / ML-related configuration procedures may include online or offline AI / ML model training and / or providing configuration information for AI / ML model training. Configuration information for AI / ML model training may relate to how to configure the method / operation of training the AI / ML model, etc.

[0070] (iv) The AI / ML-related configuration procedure may include transmitting / delivering configuration information for the model. The configuration information for the model may include parameters configuring the AI / ML model and / or an identifier (ID) for the AI / ML model.

[0071] The AI / ML model provided can be either a network-trained model or a model that requires self-training on the terminal. Even if a network-trained model is provided, the terminal can perform fine-tuning / retraining processes as needed. Meanwhile, if a network-trained model is provided, the terminal can provide training data to the network.

[0072] Meanwhile, AI / ML models can be categorized into Type A models, which can be identified without over-the-air (OTA) signaling, and Type B models, which are identified through OTA signaling. A model ID can be assigned during the model identification process, which can be further subdivided into terminal-initiated and network-initiated methods.

[0073] (v) The AI / ML-related configuration procedure may include a configuration of how to select an AI / ML Functionality / model and / or a selection process for the AI / ML Functionality / model. Selection of the UE part in a UE-side AI / ML model or a two-sided AI / ML model may be performed through instructions / signaling from the network or may be performed by the UE itself. Selection of the AI / ML Functionality / model may be performed when multiple AI / ML Functionality / models are configured / provided.

[0074] (vi) The AI / ML-related setup procedure may include setup information for various inference operations performed based on AI / ML models, for example, AI / ML-based CSI measurement / reporting, AI / ML-based Positioning, and / or AI / ML-based Beam Management.

[0075] (2) Actions based on inference from AI / ML models

[0076] Referring back to FIG. 2, the network and / or the terminal may perform inference of the AI / ML model through the trained AI / ML model, and perform various operations based on the inference of the AI / ML model (B10). If the AI / ML model is a one-sided model, the inference of the AI / ML model may be performed on either the network or the terminal where the AI / ML model is configured. If the AI / ML model is a two-sided model, each part of the inference of the AI / ML model may be performed on the network and the terminal, and such inference may be performed cooperatively between the network and the terminal depending on the implementation.

[0077] (i) Actions performed based on the inference of the AI / ML model may include AI / ML-based CSI measurement / reporting. The AI / ML-based CSI measurement / reporting may be for improving CSI feedback, and may be related to overhead reduction / CSI compression, accuracy improvement, and / or CSI prediction.

[0078] (ii) The actions performed based on the inference of the AI / ML model may include AI / ML-based beam management. The AI / ML-based beam management may be related to beam prediction in the time domain, reducing overhead / delay in the spatial domain, and / or improving beam selection accuracy.

[0079] (iii) Actions performed based on the inference of the AI / ML model may include AI / ML-based positioning. AI / ML-based positioning may be relevant to improving positioning accuracy in various scenarios, such as non-line-of-sight environments.

[0080] (3) Procedures for AI / ML management

[0081] The network and / or terminal can perform procedures for managing AI / ML Functionality / model or its settings (B15).

[0082] The network and / or terminal may perform monitoring of AI / ML Functionality / model during the process of AI / ML model inference or operation based thereon (B10) for management procedures (B15).

[0083] Management procedures may include, for example, at least one of activation / deactivation, switching, model update, and / or fallback operations for AI / ML Functionality / models. Signaling for management procedures may use various 3GPP signaling methods, such as RRC, MAC-CE, and DCI.

[0084] As an example of model switching, multiple model groups are formed, and switching between them, groups can be performed based on models having a common model structure or partially common sub-structures, and models within the same group can be related to different input / output formats or processing.

[0085] Model updating is the process of changing the parameters used by the model to adapt them to changing channel conditions over time, and fine-tuning is an example of model updating.

[0086] fallback: In a wireless communication system using an AI / ML model, when the reliability of the AI / ML model is reduced due to internal / external environmental factors, it can mean not using the AI / ML model or operating in a default operation mode that is set / defined in advance.

[0087] For example, the decision to perform a management procedure may be made by the network. For example, the network may decide to perform a management procedure upon network initiation, or upon terminal initiation and request.

[0088] As another example, the decision to perform a management procedure can be made by the terminal. For example, the terminal's decision to perform a management procedure can be triggered by the satisfaction of an event condition set by the network, by reporting the terminal's decision to the network, or by the terminal performing the decision autonomously.

[0089] 3. Specific examples of actions based on AI / ML model inference

[0090] (1) Beam management

[0091] Figure 3 illustrates an example of AI / ML-based beam management operations.

[0092] Referring to FIG. 3, the network / terminal can perform a configuration procedure related to AI / ML-based beam management (C05). The network / terminal can exchange configuration information for upper layer signaling for AI / ML-based beam management and perform a configuration procedure for an AI / ML model to be used for AI / ML-based beam management. For example, at least one of information related to model inference, configuration for the first set / second set of beams, monitoring performance, data collection, and assistance information for beam measurement can be signaled.

[0093] The network / terminal can perform measurements on the first set of beams (C10). The beam measurements can be related to RSRP measurements.

[0094] The network / terminal can obtain information about the second set of beams based on the measurement results for the first set of beams (C15). For example, the network / terminal can perform AI / ML inference using the measurement results for the first set of beams as AI / ML input data. Beam ID information may be additionally provided as AI / ML input data. The information about the second set of beams may correspond to AI / ML output data. The AI / ML output data may be related to prediction of future beam quality, such as, but not limited to, the probability that each beam will become a top-N beam, and the predicted RSRP.

[0095] In some embodiments, the network / terminal may transmit and receive information about the acquired second set of beams.

[0096] Specifically, the AI / ML-based beam management operation may include at least one of the following BM-Case 1 and BM-Case 2.

[0097] - BM-Case 1: Prediction of the second set of DL beams in the spatial domain using the first set of beam measurements.

[0098] - BM-Case 2: Prediction of the second set of DL beams in the time domain using the first set of beam measurements.

[0099] In BM-Case 1 and / or 2, both AI / ML model training and inference can be performed in the network or at the terminal. The first set of beams and the second set of beams can be different beams. Alternatively, the first set of beams can be a subset of the second set of beams. Alternatively, particularly in BM-Case 2, the first set of beams and the second set of beams can be the same beams.

[0100] The report corresponding to the inference of the UE-side model for BM-Case 1 may relate to the predicted RSRP for the upper N beams. The report may include, for example, the predicted RSRP value, and for example, the predicted RSRP value may be reported together with the actually measured RSRP.

[0101] UE-side AI / ML model inference for BM-Case 2 can report inference results for N future time points in a single report. Each time point report can correspond to the report for BM-Case 1.

[0102] For performance monitoring of the UE-side model for BM-Case 1 / 2, (i) network side performance monitoring and / or (ii) UE-assisted performance monitoring may be supported. (i) For network side performance monitoring, the UE may report information necessary for the network to calculate performance metrics, for example, measurement results (e.g., RSRP) and / or RS index for a resource set for monitoring. (ii) For UE-assisted performance monitoring, the UE may also calculate performance metrics.

[0103] Regarding the NW-side model for BM-Case 1 / 2, quantization of reported RSRPs may be supported, e.g., differential RSRP reporting may be supported along with existing quantization steps and ranges. The reported content may include information about the RSRP and the corresponding upper N beams, where N may be configured by the network.

[0104] Regarding the configuration of the first set of beams and the second set of beams in the UE-side model of BM Case-1, two resource sets may be separately configured for each of the first set and the second set, and the corresponding resource sets may be provided through the CSI reporting configuration. The terminal may perform inference / measurement for the resource set of the first set of beams. The terminal may not be expected to perform measurement / inference for the resource set of the second set of beams. The beam information of the inference report may include resource set information for the first set.

[0105] In relation to the UE-side model, relevant IDs can be provided via the CSI framework. UEs can assume identical / similar characteristics for DL ​​transmission beams / sets (lists) with the same relevant ID.

[0106] Regarding UE-assisted performance monitoring for the UE-side models of BM-Case 1 and 2, the following methods can be considered.

[0107] i) Compare prediction results based on resources for monitoring and use the top 1 or top K beam prediction accuracy.

[0108] ii) Use RSRP difference information based on RSRP measurements of resources for monitoring and actual RSRP measurements for at least one of the top N predicted beams.

[0109] iii) Use the difference information between the measured RSRP and the predicted RSRP for the corresponding beam of the resources for monitoring.

[0110] iv) Probability information about the predicted beam being one of the top 1 or N beams.

[0111] Quantization of RSRP can be supported for reporting inference results for UE-side models, and differential RSRP with existing quantization steps can be supported. The scope of RSRP reporting is that differential RSRP can be supported among multiple beams in the case of BM-case 1, and differential RSRP can be supported among multiple beams at multiple viewpoints in the case of BM-case 2.

[0112] For BM-Case 2 of the UE-side model, the network can be configured to report inferences for N future time points to the terminal.

[0113] (2) CSI prediction and / or compression (CSI prediction / compression)

[0114] Figure 4 illustrates an example of AI / ML-based CSI measurement / reporting operations.

[0115] Referring to FIG. 4, the network / terminal may perform a configuration procedure related to AI / ML-based CSI (D05). The network / terminal may exchange configuration information for upper-layer signaling for AI / ML-based CSI measurement / reporting, and perform a configuration procedure for an AI / ML model to be used for AI / ML-based CSI. For example, at least one of information related to model inference, configuration of RS / resources to be used for CSI measurement, monitoring performance, data collection, and conditions / resources for CSI reporting may be signaled.

[0116] The terminal can perform CSI measurements based on AI / ML model inference (D10). The AI / ML model used by the terminal for CSI measurements may be a UE-side AI / ML model corresponding to a one-side AI / ML model, or an AI / ML model corresponding to the terminal portion of a two-side AI / ML model.

[0117] A terminal may report CSI to the network based on the CSI measurement results (D15). CSI reporting may be performed periodically or aperiodically depending on the configuration, and in the case of aperiodic CSI reporting, a network instruction (not shown) such as DCI that triggers it may be additionally signaled. CSI reporting may include AI / ML-based CSI content, and additionally (depending on the configuration / scheduling) may further include legacy CSI content (e.g., non-AI / ML-based RI, PMI, CQI, etc.). AI / ML-based CSI content may be related to at least one of 1) CSI compression to reduce the overhead of CSI reporting, and 2) CSI prediction for future time points in the time domain.

[0118] The network can obtain CSI based on the CSI report of the terminal.

[0119] If a two-sided AI / ML model is configured, the network can reconstruct CSI using the terminal's CSI report as input data to the network-configured AI / ML model (D20). The inference (output) of the network-configured AI / ML model can be the reconstructed CSI. In such a two-sided AI / ML model, the terminal-side AI / ML model can be understood as a CSI encoder, while the network-side AI / ML model can be understood as a concept similar to a CSI decoder.

[0120] CSI compression is a spatial-frequency domain CSI compression, which can be primarily based on two-sided AI / ML models. CSI prediction can primarily be based on one-sided, specifically, UE-side AI / ML models.

[0121] In CSI compression based on two-side AI / ML models, AI / ML model training may include at least one of (i) Type 1, in which either the terminal or the network jointly trains two-side AI / ML models, (ii) Type 2, in which the terminal and the network each jointly train their respective two-side AI / ML model parts, and (iii) Type 3, in which the terminal and the network each separately train their respective two-side AI / ML model parts, with the terminal training being primarily related to CSI generation and the network training being primarily related to CSI reconstruction. Joint training means that the CSI generation / reconstruction model is trained in the same loop for forward / backward delays, and separate training may mean a sequential method in which either the terminal or the network starts training first and then the other performs training.

[0122] (3) Positioning

[0123] Figure 5 illustrates an example of AI / ML-based positioning operation.

[0124] Referring to FIG. 5, the network / terminal may perform a configuration procedure related to AI / ML-based positioning (E05). The network / terminal may exchange configuration information for upper-layer signaling for AI / ML-based positioning and perform a configuration procedure for an AI / ML model to be used for AI / ML-based positioning. For example, at least one of information related to model inference, positioning configuration for RS, monitoring performance, data collection, and assistance information for positioning measurement may be signaled.

[0125] The network / terminal can perform measurements for positioning (E10). The measurements for positioning may be related to PRS and / or SRS measurements.

[0126] Based on the measurement results, the network / terminal can obtain information about terminal positioning (E15). For example, the network / terminal can use the PRS / SRS measurement results as AI / ML input data to perform AI / ML inference. Information about terminal positioning may correspond to AI / ML output data. The AI / ML output data may be, for example, terminal location or assistance information that serves as the basis for determining terminal location, but is not limited thereto.

[0127] Depending on the embodiment, the network / terminal may transmit and receive information about the acquired terminal positioning.

[0128] Specifically, the AI / ML-based positioning may include at least one of (i) direct AI / ML positioning and / or (ii) AI / ML-assisted positioning. (i) In direct AI / ML positioning, the output of the AI / ML model includes the UE location. (ii) In AI / ML-assisted positioning, the output of the AI / ML model may be a new measurement result and / or an improvement to an existing measurement (e.g., LoS / NLoS identification, timing and / or angle of the measurement, likelihood of the measurement, etc.).

[0129] The data sample collected for training data collection may include at least one of the following Parts:

[0130] - Part A: Channel measurements, quality indicators of channel measurements, timestamps of channel measurements

[0131] - Part B: ground truth label (or its approximation), label quality indicator, label timestamp

[0132] The training data samples for Part A and Part B may be for the same terminal (e.g., PRU or Non-PRU UE) and for the same location associated with Part B.

[0133] Measurements for positioning can be divided into sample-based measurements and path-based measurements.

[0134] - In sample-based measurements, the measurement consists of Nt' samples of the estimated channel response in the time domain. Timing information for the Nt' samples is reported with a timing granularity T, where T = 2k x Tc. k denotes a timing reporting granularity factor, and Tc denotes the fundamental time unit of the wireless communication system. Nt' and k can be signaled parameters. Timing information can be defined as a value relative to a reference time.

[0135] - Path-based measurement refers to the measurement defined in existing wireless communication systems.

[0136] AI / ML-based positioning may relate to at least one of the following specific cases:

[0137] - (i) Case 1: UE-based positioning using a UE-side model, in the case of direct AI / ML or AI / ML assisted positioning.

[0138] - (ii) Case 2a: UE-assisted / LMF-based positioning using the UE-side model, in the case of AI / ML assisted positioning

[0139] - (iii) Case 2b: UE-assisted / LMF-based positioning using the LMF-side model, in the case of direct AI / ML positioning

[0140] - (iv) Case 3a: NG-RAN node assisted positioning using the gNB-side model, in the case of AI / ML assisted positioning

[0141] - (v) Case 3b: NG-RAN node assisted positioning using LMF-side model, in case of direct AI / ML positioning

[0142] (i) In relation to model performance monitoring in Case 1, the following options may be considered for calculating model performance metrics in label-based model monitoring.

[0143] 1) Option A. Calculate monitoring metrics on the target terminal side.

[0144] - Option A-1: ​​At least some of the information about the target terminal's ground truth label is generated by the LMF and provided to the target terminal. For example, the target terminal and / or base station may send measurement results to the LMF, which may then derive information about the ground truth label.

[0145] - Option A-2: At least part of the information for position calculation assistance data is provided from the LMF to the target UE.

[0146] - Option A-3: This is a method of reusing assistance data previously provided from the LMF to the target terminal, in which PRU measurement results and corresponding PRU location information are provided from the LMF to the target terminal.

[0147] - Option A-4: PRU measurements and PRU locations are provided from the PRU to the target UE.

[0148] 2) Option B. LMF calculates monitoring metrics

[0149] - Option B-1: At least the inference results of the target terminal (i.e., the model output corresponding to the channel measurements of the target terminal) can be transmitted to the LMF by the target terminal.

[0150] - Option B-2: The channel measurement of the PRU is provided to the target terminal through the LMF, and the inference result (i.e., the model output corresponding to the channel measurement of the PRU) can be transmitted from the target terminal to the LMF.

[0151] (iv) With regard to generating training data in Case 3a, at least LMF can generate labels and data related to them (e.g., timestamps).

[0152] (iv) For calculating model performance monitoring metrics in label-based model monitoring in Case 3a, the following options A and B may be considered.

[0153] - Option A: NG-RAN nodes perform monitoring metric calculations for their own models.

[0154] - Option B: LMF performs monitoring metric calculations for models located on NG-RAN nodes.

[0155] (iv) For generating training data for Case 3a and (v) Case 3b, measurements and related data (e.g., timestamp) can be generated at the TRP / base station.

[0156] (v) For base station channel measurements reported to the LMF (location management server) in Case 3b, timing information can be expressed as a relative value to the UL RTOA reference time T0+tSRS.

[0157] (iii) Time domain channel measurements supported for reporting in Case 2b and (v) Case 3b may include timing information and / or power information corresponding to timing information.

[0158] For the definition of a sample-based measurement, Nt' samples can be selected from a list of consecutive Nt samples, and the Nt samples can have a time granularity of T.

[0159] In relation to sample-based measurements, the Nt' samples selected for measurement of the network may be those having the highest power.

[0160] For positioning based on Case 3b, for sample-based measurements, LMF can signal parameter values ​​such as Nt, Nt', k, etc. to the base station.

[0161] < CSI-related actions >

[0162] The channel state information (CSI) may include at least one of a channel quality indicator (CQI), a precoding matrix indicator (PMI), a CSI-RS resource indicator (CRI), an SS / PBCH block resource indicator (SSBRI), a layer indicator (LI), a rank indicator (RI), a Layer 1-Reference Signal Received Power (L1-RSRP), and / or a Layer 1-Signal-to-Interference-plus-Noise Ratio (L1-SINR).

[0163] In the case of the CSI prediction described below, the CSI related to the prediction may include at least one of predicted CQI (predicted CQI, P-CQI), predicted PMI (predicted PMI, P-PMI), predicted CRI (predicted CRI, P-CRI), predicted SSBRI (predicted SSBRI, P-SSBRI), predicted LI (predicted LI, P-LI), predicted RI (predicted RI, P-RI), predicted L1-RSRP (predicted L1-RSRP, P-L1-RSRP), and / or predicted L1-SINR (predicted L1-SINR, P-L1-SINR).

[0164] When monitoring the performance / accuracy of the CSI prediction described below, the CSI related to the prediction accuracy may include a Prediction Accuracy Indicator (PAI). For example, the PAI may indicate the accuracy of predicted downlink reference signal(s) (e.g., predicted CRI(s) and / or predicted SSBRI(s)), and thus the PAI may be interpreted / replaced with a Reference Signal-Prediction Accuracy Indicator (RS-PAI). For example, the PAI may indicate the accuracy of predicted CSI (e.g., predicted PMI), and thus the PAI may be interpreted / replaced with a Channel State Information-Prediction Accuracy Indicator (CSI-PAI).

[0165] Figure 6 is a flowchart showing an example of a CSI-related procedure.

[0166] Referring to FIG. 6, in order to perform one of the purposes of CSI-RS, a terminal (e.g., user equipment, UE) receives configuration information related to CSI from a base station (e.g., general Node B, gNB) through RRC (radio resource control) signaling (S610).

[0167] The configuration information related to the CSI may include at least one of CSI-IM (interference management) resource related information, CSI measurement configuration related information, CSI resource configuration related information (e.g., M≥1 CSI-ResourceConfig resource setting), CSI-RS resource related information, or CSI report configuration related information (e.g., N≥1 CSI-ReportConfig reporting setting). For example, the configuration information may include at least one of one or more CSI resource settings and / or one or more CSI reporting settings.

[0168] For example, the configuration information may include a first CSI resource configuration for measurement and a second CSI resource configuration for prediction. As a specific example, a measurement related to prediction of CSI, which will be described later, may be performed based on the first CSI resource configuration. The terminal may perform L1-RSRP measurements on CSI-RS resources or SS / PBCH block resources associated with the first CSI resource configuration. As a specific example, the prediction of CSI, which will be described later, may be performed based on the second CSI resource configuration. The terminal may perform prediction on CSI-RS resources or SS / PBCH block resources associated with the second CSI resource configuration based on the L1-RSRP measurements. In other words, best CRI / best SSBRI (e.g., P-CRI(s), P-SSBRI(s)) may be predicted using L1-RSRPs as measurement metrics.

[0169] For example, the configuration information may include a first CSI reporting setting related to prediction and a second CSI reporting setting related to prediction accuracy.

[0170] CSI resource configuration related information can be expressed as CSI-ResourceConfig IE. The CSI resource configuration related information defines a group including at least one of a non-zero power (NZP) CSI-RS resource set, a CSI-IM resource set, or a CSI-SSB resource set. That is, the CSI resource configuration related information includes a CSI-RS resource set list, and the CSI-RS resource set list can include at least one of an NZP CSI-RS resource set list, a CSI-IM resource set list, or a CSI-SSB resource set list. A CSI-RS resource set is identified by a CSI-RS resource set ID, and one resource set includes at least one CSI-RS resource. Each CSI-RS resource is identified by a CSI-RS resource ID.

[0171] Information related to CSI report configuration (e.g., CSI-ReportConfig IE) includes a report configuration type (reportConfigType) parameter indicating time domain behavior and a report quantity (reportQuantity) parameter indicating the CSI-related quantity to be reported. The time domain behavior may be periodic, aperiodic, or semi-persistent.

[0172] The above reportQuantity parameter is a channel quality indicator (CQI), a precoding matrix indicator (PMI), a CSI-RS resource indicator (CRI), an SS / PBCH block resource indicator (SSBRI), a layer indicator (LI), a rank indicator (RI), a Layer 1-Reference Signal Received Power (L1-RSRP), a Layer 1-Signal-to-Interference-plus-Noise Ratio (L1-SINR), a predicted CQI (predicted CQI, P-CQI), a predicted PMI (predicted PMI, P-PMI), a predicted CRI (predicted CRI, P-CRI), a predicted SSBRI (predicted SSBRI, P-SSBRI), a predicted LI (predicted LI, P-LI), a predicted RI (predicted RI, P-RI), a predicted L1-RSRP (predicted L1-RSRP, P-L1-RSRP) and / or a predicted It may be set to a value indicating at least one of L1-SINR (predicted L1-SINR, P-L1-SINR) and / or Prediction Accuracy Indicator (PAI) (e.g., CSI-PAI or RS-PAI).

[0173] For example, the reportQuantity parameter can be set to cri, ssb-Index, cri-RSRP, or ssb-Index-RSRP. cri represents the CSI-RS resource indicator (CRI). RSRP represents the Layer 1-Reference Signal Received Power (L1-RSRP). ssb-Index represents the SS / PBCH block resource indicator (SSBRI).

[0174] For example, the reportQuantity parameter can be set to p-cri, p-ssb-index, p-cri-RSRP, or p-ssb-index-RSRP. p-cri represents predicted CRI (P-CRI). p-ssb-index represents predicted SSBRI (P-SSBRI). p-cri-RSRP represents predicted CRI (P-CRI) and predicted L1-RSRP (P-L1-RSRP). p-ssb-index-RSRP represents predicted SSBRI (P-SSBRI) and predicted L1-RSRP (P-L1-RSRP).

[0175] For example, the reportQuantity parameter can be set to pai (or rs-pai), where pai (e.g., csi-pai or rs-pai) represents PAI (e.g., CSI-PAI or RS-PAI).

[0176] Measurement resources may include configurations for downlink signals and / or downlink resources on which a terminal will perform measurements to determine feedback information. Measurement resources may be configured as ZP and / or NZP CSI-RS resource sets associated with CSI reporting configurations. The NZP CSI-RS resource set may include a CSI-RS set or an SSB set. For example, L1-RSRP may be measured for a CSI-RS set or an SSB set.

[0177] The terminal measures CSI based on configuration information related to the CSI (S620). The CSI measurement may include (1) a process of receiving a CSI-RS by the terminal (S621) and (2) a process of computing CSI using the received CSI-RS (S622). The terminal reports the CSI to the base station (S630).

[0178] Resource setting

[0179] Each CSI resource setting 'CSI-ResourceConfig' contains a configuration for S≥1 CSI resource sets (given by the higher layer parameter csi-RS-ResourceSetList). A CSI resource setting corresponds to a CSI-RS-resourcesetlist, where S represents the number of configured CSI-RS resource sets. Wherein, the list of S≥1 CSI resource sets contains either or both of NZP CSI-RS resource set(s) and SS / PBCH block (SSB) set(s) used for L1-RSRP computation, or contains CSI-IM resource set(s).

[0180] One or more CSI resource settings for channel measurement (CM) and interference measurement (IM) are configured via higher layer signaling.

[0181] - CSI-IM resource for interference measurement.

[0182] - NZP CSI-RS resources for interference measurement.

[0183] - NZP CSI-RS resources for channel measurement.

[0184] That is, the CMR (channel measurement resource) can be NZP CSI-RS for CSI acquisition, and the IMR (Interference measurement resource) can be NZP CSI-RS for CSI-IM and IM.

[0185] Here, CSI-IM (or ZP CSI-RS for IM) is mainly used for inter-cell interference measurement.

[0186] And, NZP CSI-RS for IM is mainly used for intra-cell interference measurement from multi-user.

[0187] A UE may assume that the CSI-RS resource(s) configured for channel measurement for one CSI reporting and the CSI-IM / NZP CSI-RS resource(s) for interference measurement (when NZP CSI-RS resource(s) are used for interference measurement) are in a QCL relationship with respect to 'QCL-TypeD' per resource.

[0188] As we have seen, resource setting can mean a resource set list.

[0189] For aperiodic CSI, each trigger state set using the higher layer parameter CSI-AperiodicTriggerState is associated with one or more CSI-ReportConfigs, and each CSI-ReportConfig is linked to a periodic or semi-persistent or aperiodic resource setting.

[0190] A single reporting setting (e.g., CSI-ReportConfig) can be associated with up to three resource settings (e.g., CSI-ResourceConfig). For example, a single CSI reporting setting can include the ID of at least one CSI resource setting (e.g., CSI-ResourceConfigId). The at least one CSI resource setting can include a CSI resource setting related to a measurement.

[0191] < Beam Management (BM) >

[0192] BM procedures are L1 (layer 1) / L2 (layer 2) procedures for acquiring and maintaining a set of base station (e.g., gNB, TRP, etc.) and / or terminal (e.g., UE) beams that can be used for downlink (DL) and uplink (UL) transmission / reception, and may include the following procedures and terminology.

[0193] - Beam measurement: An operation in which a base station or UE measures the characteristics of a received beam-forming signal.

[0194] - Beam determination: An operation in which a base station or UE selects its own transmit beam (Tx beam) / receive beam (Rx beam).

[0195] - Beam sweeping: The operation of covering a spatial area using a transmit and / or receive beam over a predetermined time interval in a predetermined manner.

[0196] - Beam report: An operation in which a UE reports information about a beam-formed signal based on beam measurement.

[0197] The BM procedure can be divided into (1) a DL BM procedure using SS (synchronization signal) / PBCH (physical broadcast channel) Block or CSI-RS, and (2) a UL BM procedure using SRS (sounding reference signal).

[0198] Additionally, each BM procedure may include Tx beam sweeping to determine the Tx beam and Rx beam sweeping to determine the Rx beam.

[0199] DL BM

[0200] The DL BM procedure may include (1) transmission of beamformed DL RSs (reference signals) (e.g., CSI-RS or SS Block (SSB)) of the base station and (2) beam reporting of the terminal.

[0201] Here, beam reporting may include preferred DL RS ID(identifier)(s) and corresponding L1-RSRP (Reference Signal Received Power).

[0202] The above DL RS ID may be an SSBRI (SSB Resource Indicator) or a CRI (CSI-RS Resource Indicator).

[0203] An example of beamforming using SSB and CSI-RS is described in detail below.

[0204] Both SSB and CSI-RS beams can be used for beam measurement. The measurement metric is L1-RSRP per resource / block. SSB is used for coarse beam measurement, and CSI-RS can be used for fine beam measurement. SSB can be used for both Tx beam sweeping and Rx beam sweeping.

[0205] Rx beam sweeping using SSB can be performed by the UE changing the Rx beam for the same SSBRI across multiple SSB bursts, where one SS burst contains one or more SSBs, and one SS burst set contains one or more SSB bursts.

[0206] Below we will look at the DL BM procedure.

[0207] The configuration for beam report using SSB is performed during CSI / beam configuration in RRC connected state (or RRC connected mode).

[0208] - The terminal receives configuration information from the base station. As a specific example, the terminal receives a CSI-ResourceConfig IE containing a CSI-SSB-ResourceSetList containing SSB resources used for BM from the base station.

[0209] Table 1 shows an example of the CSI-ResourceConfig IE. As shown in Table 1, BM configuration using SSB is not defined separately, and SSB is configured as a CSI-RS resource.

[0210]

[0211] In Table 1, the csi-SSB-ResourceSetList parameter indicates a list of SSB resources used for beam management and reporting in a single CSI-RS resource set. Here, the SSB resource set can be set to {SSBx1, SSBx2, SSBx3, SSBx4, …}. For example, the SSB index can be defined from 0 to 63.

[0212] - The terminal receives a downlink reference signal (DL RS) from the base station. As a specific example, the terminal receives an SSB from the base station based on the CSI-SSB-ResourceSetList.

[0213] - The terminal transmits a beam report to the base station. For example, if CSI-ReportConfig related to reporting on SSBRI (SSB Resource Indicator) and L1-RSRP is set, the terminal reports the best SSBRI and its corresponding L1-RSRP to the base station.

[0214] That is, when the reportQuantity of the above CSI-ReportConfig IE is set to 'ssb-Index-RSRP', the terminal reports the best SSBRI and the corresponding L1-RSRP to the base station.

[0215] And, if the terminal sets the CSI-RS resource in the same OFDM symbol(s) as the SSB (SS / PBCH Block) and 'QCL-TypeD' is applicable, the terminal can assume that the CSI-RS and SSB are quasi co-located from the 'QCL-TypeD' perspective.

[0216] Here, the QCL TypeD may mean that the antenna ports are QCL-connected from a spatial Rx parameter perspective. When a terminal receives multiple DL antenna ports in a QCL Type D relationship, the same reception beam may be applied. In addition, the terminal does not expect the CSI-RS to be configured in an RE that overlaps with the SSB RE.

[0217] The configuration for beam reporting using CSI-RS is performed in the same manner as the configuration for beam reporting using SSB described above, and thus, redundant description is omitted. The operation of the beam reporting procedure using CSI is described below.

[0218] - The terminal receives configuration information from the base station. As a specific example, the terminal receives a CSI-ResourceConfig IE containing a CSI-SSB-ResourceSetList containing CSI resources used for BM (e.g., NZP CSI-RS resource set IE) from the base station.

[0219] - The terminal receives CSI-RS resources within the NZP CSI-RS resource set through different Tx beams (DL spatial domain transmission filters) of the base station.

[0220] - The terminal selects (or decides) the best beam.

[0221] - The terminal reports the ID and related quality information (e.g., L1-RSRP) for the selected beam to the base station. In this case, the reportQuantity of the CSI report config can be set to 'cri-RSRP'.

[0222] < Description of Rel-17 / 18 beam management >

[0223] In Rel-17, both the DL TCI state and the UL TCI state can be indicated through DL DCI (e.g., DCI format 1-1 or 1-2), or only the UL TCI state can be indicated without indicating the DL TCI state. Therefore, the methods used for UL beam and power control (PC) configuration in the existing R15 / R16 are replaced in Rel-17 with the above UL TCI state indication method. More specifically, in R17, one UL TCI state can be indicated through the TCI field of DL DCI, and the UL TCI state is applied to all PUSCHs and all PUCCHs after a certain time called the beam application time, and can be applied to some or all of the indicated SRS resource sets. In addition, the base station can perform a terminal common beam update by using DCI and / or MAC-CE to perform indication / update with one beam in common (using a joint or separate TCI state) for specific DL / UL channel / RS combinations of multiple terminals. The target channels / RS of common beam update include UE-dedicated CORESET, UE-dedicated reception on PDSCH for DL, DG / CG-PUSCH, all or a subset of dedicated PUCCH for UL, and additionally, AP CSI-RS for tracking / BM, SRS can be set as target channels / RS.In Rel-18, considering the M-TRP environment, the method of indicating multiple UL TCI states (and / or DL ​​TCI states) through the TCI field of DL DCI has been standardized, and the uplink / downlink resources to which multiple indicated TCIs are applied can be defined / configured depending on the S-DCI based M-TRP environment and the M-DCI based M-TRP environment.

[0224] In this specification, 'beam' may mean a source RS for a 'spatial filter' or a 'spatial relation', and may be interpreted as a QCL (type-D) RS or a TCI state or (in the case of uplink) a spatial relation RS.

[0225] For example, in this specification, 'beam' may mean a spatial filter determined based on the reference RS or the source RS. The spatial filter may include a spatial domain filter, a spatial domain transmission filter, and a spatial domain receive filter. For example, in this specification, 'beam' may be interpreted / replaced with a reference signal index (RS index), a reference signal resource index (RS resource index), and / or a resource indicator (e.g., RS index, SSB index, CSI-RS resource index, SRS resource index, SSB Resource Indicator (SSBRI), CSI-RS Resource Indicator (CRI), etc.).

[0226] For example, a beam associated with UL may be referred to as i) a spatial filter (for uplink transmission or uplink reception), ii) a spatial domain filter (for uplink transmission or uplink reception), iii) an uplink spatial domain transmission filter, iv) an uplink spatial domain receive filter, v) an uplink transmit spatial filter (UL Tx spatial filter), or vi) an uplink receive spatial filter (UL Rx spatial filter).

[0227] For example, a beam associated with DL may be referred to as i) a spatial filter (for downlink transmission or downlink reception), ii) a spatial domain filter (for downlink transmission or downlink reception), iii) a downlink spatial domain transmission filter, iv) a downlink spatial domain receive filter, v) a downlink transmit spatial filter (DL Tx spatial filter) or vi) a downlink receive spatial filter (DL Rx spatial filter).

[0228] In this specification, ' / ' means 'and', 'or', or 'and / or' depending on the context.

[0229] This specification considers AI / ML-based CSI reporting, which is described below with reference to Figure 7.

[0230] Figure 7 illustrates an AI / ML-based CSI reporting operation.

[0231] Referring to Fig. 7, the terminal is equipped with an AI encoder (e.g., a CSI encoder), and the base station is equipped with an AI decoder (e.g., a CSI decoder). The terminal and the base station each perform AI / ML model inference. A model in which AI / ML model inference is performed at each of the two nodes (terminal and base station) in this way is called a two-sided model. Fig. 7 exemplifies CSI compression for the purpose of overhead reduction.

[0232] In other words, two-sided AI / ML means that AI / ML models are deployed / configured in the terminal and the base station (or network), respectively, and each performs inference. For example, an AI / ML model in the form of an auto-encoder can be configured. The terminal side (CSI encoder) uses channel information (e.g., channel matrix / channel covariance matrix / channel eigenvector) or information that has gone through pre-processing for the corresponding channel information as input, and calculates the AI / ML model inference output. The terminal feeds back information about the output (e.g., CSI) to the base station. For example, the information about the output may be generated based on specific post-processing. For example, the information about the output may be generated without post-processing.

[0233] The base station (CSI decoder) uses the feedback information (e.g., CSI) as input to the AI / ML model. For example, the feedback information with preprocessing applied may be used as the input. For example, the feedback information may be used as the input without preprocessing. The base station calculates the inference output and decodes the final CSI without applying or postprocessing.

[0234] Another use case, the Rel-18 AI / ML study, explored CSI prediction based on a UE (User Equipment)-sided model. The CSI prediction use case is described with reference to Figure 8.

[0235] Figure 8 illustrates AI / ML-based CSI prediction.

[0236] Specifically, Fig. 8 shows an inference operation based on a UE-sided model. An AI / ML (Artificial Intelligence / Machine Learning) model (8B) is provided only on the terminal side, and model inference is performed by the model (8B). Historical measurements are applied as AI / ML input (8A) of the model (8B). For example, the historical measurements may be performed based on a CSI resource configuration (e.g., CSI-ResourceConfig) linked to a CSI reporting configuration (e.g., CSI-ReportConfig) related to prediction. t -x , t -x+1,.. t0 represents historical time instances related to multiple measurements. The AI / ML model output (8C) is one or more CSIs (e.g., one or more RSRPs and / or one or more beam indices). t N , t N+1 ,.. t N+K represents future time instances related to the CSI. The CSI type of the input may be a raw channel matrix or a precoder type (e.g., eigenvector).

[0237] Below, we examine the non-AI-based CSI reporting behavior. For example, in Rel-15 Type I / II CSI reporting, the following actions can be performed for inter-cell interference management purposes. The base station can transmit a configuration / instruction to the terminal, based on RRC signaling, that restricts the PMIs used in CSI calculation. The terminal excludes the corresponding PMI(s) from CSI calculation, calculates preferred CSI (e.g., CQI / RI / PMI), and reports it to the base station.

[0238] Even in AI / ML-based CSI reporting, operations such as codebook subset restriction (CBSR) can be implemented to control inter-cell interference. Furthermore, AI / ML can be used to predict rank restrictions, CBSR, and codebook types, effectively reducing feedback overhead. This specification presents an effective method for configuring and operating restrictions in AI / ML-based CSI reporting.

[0239] First, let's look at legacy CBSR.

[0240] 1) Type 1 CSI

[0241] A. PMI restriction - bitmap based ( ) indication

[0242] i. In the above bitmap , In this case, it indicates the number of antennas in the 1st domain and 2nd domain of the base station antenna port (x-pol antenna port). * is the length of the DFT-vector (1D / 2D) corresponding to the ports corresponding to one slant of the x-pol antenna. The length of the final DL precoder is The final DL precoder consists of two * The DFT vectors are connected in a cophase form. The Rank1 codebook is It has the form of . is length * is a DFT vector, has the value {1,j,-1,-j} as the co-phase value.

[0243] ii. O1 and O2 are oversampling factors applied to the 1st domain and 2nd domain, respectively.

[0244] iii. represents the total number of DFT vectors, and a specific bit corresponds to a specific DFT vector.

[0245] B. RI restriction - 8 bit bitmap is used.

[0246] i. The size of the RI restriction can be determined based on the max rank value set by the base station. For example, if the max rank set by the base station is 8, the rank value that the terminal should not use is indicated based on 8 bits. For example, even if the max rank is 8, if "11110000" is indicated as the RI restriction, the terminal can only calculate and report CSI corresponding to ranks 1, 2, 3, and 4.

[0247] 2) Type 2 CSI

[0248] A. PMI restriction

[0249] i. It does not restrict the entire beam. Specifically, it selects a specific beam group and restricts the DFT vector within that beam group. For that beam, the amplitude that can be used for Type 2 CSI configuration for each beam being combined is also limited to 2 bits based on Table 2 below.

[0250] 1. Above The dog's beam It is divided into groups of dogs. Each group is -by- It includes adjacent beams.

[0251] 2. Select P beam groups based on the indicator.

[0252] 3. , here, is the length Bitmap(length- bitmap), -by- The beam is Limited by the beam ( -by- beam is restricted by beam), each beam is 2-bit soft power restricted.

[0253]

[0254] B. RI restriction - 4-bit bitmap is used.

[0255] From Release 18 CSI to Doppler domain By compressing the basis vector of the dog A codebook (e.g., codebook for predicted PMI) is introduced that can predict future time instances(s). In this case, the reportQuantity parameter is set to cri-RI-PMI-CQI (or a value representing predicted PMI). Similarly, AI / ML-based CSI prediction can be performed for time instance(s) of a dog.

[0256] The information included in the CSI related to the prediction described below and the information included in the CSI related to monitoring are explained based on the background of the CSI-related operations described above as follows.

[0257] For example, CSI reported based on existing reporting configuration may include at least one of a CSI-RS Resource Indicator (CRI), an SSB Resource block Indicator (SSBRI), a Layer Indicator (LI), a Layer 1-Reference Signal Received Strength (L1-RSRP), a rank indicator (RI), a Precoding Matrix Indicator (PMI), and / or a channel quality indicator (CQI). In other words, the reportQuantity parameter in CSI-Reportconfig may be set to a value indicating at least one of the CRI, the LI, the SSBRI, the L1-RSRP, the RI, the PMI, and / or the CQI.

[0258] For example, CSI reported based on prediction-related reporting settings may include predicted information / predicted parameter(s) / indicator(s). As a specific example, the prediction-related CSI may include at least one of predicted CRI, predicted SSBRI, predicted L1-RSRP, and / or predicted PMI. In other words, the reportQuantity parameter in CSI-Reportconfig may be set to a value indicating at least one of predicted CRI, predicted SSBRI, predicted L1-RSRP, and / or predicted PMI. For convenience of explanation, a predicted parameter may be expressed as a P-parameter in the present specification. For example, predicted CRI, predicted SSBRI, and predicted L1-RSRP may be expressed as P-CRI, P-SSBRI, and P-L1-RSRP. For example, in the case of PMI, it can be expressed as predicted PMI.

[0259] For example, CSI reported based on monitoring-related reporting settings may include parameter(s) / indicators(s) related to monitoring / prediction accuracy. As a specific example, the CSI related to monitoring / prediction accuracy may include a prediction accuracy indicator (e.g., a metric based on the embodiments described below). In other words, the reportQuantity parameter within CSI-Reportconfig may be set to a value indicating the prediction accuracy indicator. For convenience of explanation, the prediction accuracy indicator may be referred to as PAI in this specification.

[0260] This specification proposes a performance metric calculation and reporting method to effectively support performance monitoring when predictions are performed for multiple time instances in AI / ML-based CSI prediction. The operations related to CSI prediction are described below with reference to FIG. 10.

[0261] Figure 10 is a diagram for explaining CSI prediction for multiple time instances.

[0262] Referring to Figure 10, there are multiple time instances (e.g., N) that need to be predicted on the prediction window. -4=4) can be set. When the number of time instances to be considered is 1, performance monitoring can be performed based on a comparison between the predicted CSI (e.g., RS indicator(s), P-CRI(s), P-SSBRI(s), predicted PMI determined based on prediction) for the time instance and the ground-truth CSI (e.g., RS indicator(s), CRI(s), SSBRI(s), PMI determined based on measurement). As described above, when the number of time instances is 1, performance monitoring and reporting can be performed simply. However, when the number of time instances is multiple, various methods can be considered for performance monitoring and reporting. For example, a performance metric (e.g., PAI) can be calculated based on the predicted CSI (e.g., RS indicator(s), P-CRI(s), P-SSBRI(s), predicted PMI determined based on prediction) for each of the multiple time instances. For example, a performance metric (e.g., PAI) can be calculated based on predicted CSI (e.g., RS indicator(s) determined based on prediction, P-CRI(s), P-SSBRI(s), predicted PMI) for a specific time instance among multiple time instances.

[0263] Below, we examine methods for monitoring the performance of predictions (i.e., prediction accuracy) and calculating / reporting performance metrics (e.g., PAI) when predicted information (e.g., P-CRI(s), P-SSBRI(s), predicted PMI) is reported for one or more time instances (e.g., Multiple time instances) based on AI / Ml (e.g., UE sided model).

[0264] Proposal 1

[0265] When a terminal performs CSI prediction (e.g., reports CSI including predicted information) for one or more time instances, a method of determining / calculating and reporting a single metric for performance monitoring may be considered. For example, the terminal may report CSI including predicted information (e.g., P-CRI(s), P-SSBRI(s), P-L1-RSRP(s) and / or predicted PMI) for one or more time instances to the base station. In this case, the terminal may report CSI including a single metric (e.g., PAI) determined based on the time instance for monitoring to the base station.

[0266] In one embodiment, the base station may configure / instruct the terminal to configure / instruct information related to calculation / determination of metric / output (e.g., report configuration including information related to determination of PAI).

[0267] For example, the reporting settings may include at least one of a number of groups, a weight, and / or a threshold.

[0268] For example, the single metric may be a weighted metric. Specifically, a terminal may group one or more instances (e.g., all instances). The terminal may determine the weighted metric by multiplying metrics based on instances within the same group by the same weight, and then adding up all metrics multiplied by the weights for each group.

[0269] For example, for convenience of explanation =metric value of the nth instance, = If we define the nth weight value, the weighted metric can be expressed as follows.

[0270]

[0271] At this time, the number of groups for the grouping can be set / indicated / defined as one of the values ​​up to the maximum number of instances. For example, if the maximum number of instances is 8, the base station can transmit information about the number of groups to the terminal (e.g., information indicating one of the values ​​2 to 8). The terminal can classify all instances related to the prediction into up to 8 groups.

[0272] For example, as the accuracy of the predicted values ​​is likely to decrease over time, the weights It can be set / instructed as follows.

[0273] For example, the weighted metric, and fallback can be performed based on a threshold. As a specific example, the terminal may use a weighted metric, can determine / calculate and report it to the base station. The base station can compare the weighted metric with a specific threshold to make a fallback decision. As a specific example, the terminal can determine the weighted metric, The base station can determine / calculate the weighted metric and compare the result with a specific threshold and report it to the base station. The base station can make a fallback decision based on the result of the comparison.

[0274] For example, a specific threshold may be a predefined value. For example, a specific threshold may be a value set / instructed by a base station.

[0275] For example, the value of each weight may be determined by the terminal and reported to the base station. For example, the value of each weight may be set / instructed to the terminal by the base station.

[0276] The above fallback actions may include i) AI / ML model de-activation, ii) fallback to legacy CSI prediction, and / or iii) no filter update in non-AI based prediction.

[0277] For example, performance monitoring can be performed by setting a specific weight to 1 (e.g., w1=1).

[0278] For example, even if CSI prediction and reporting are performed for multiple monitoring occasions (or multiple time instances), performance monitoring may be performed for a single monitoring occasion (or single time instance). As a specific example, only the predicted CSI (e.g., P-CRI(s), P-SSBRI(s), predicted PMI) of a single time instance may be monitored. The single metric (e.g., PAI) described above may be determined / calculated based on the single time instance.

[0279] In one embodiment, it can be defined in advance which predicted CSI among predicted CSI (e.g., P-CRI(s), P-SSBRI(s), predicted PMI) for a plurality of monitoring occasions (or a plurality of time instances) will be monitored (e.g., 1st predicted CSI, average predicted CSI, last predicted CSI). For example, the predicted CSI for the first time instance or the last time instance among the plurality of monitoring occasions (or a plurality of time instances) can be monitored. In other words, a metric for monitoring (e.g., PAI) can be determined based on the predicted CSI (e.g., P-CRI(s), P-SSBRI(s), predicted PMI) for the first time instance or the last time instance.

[0280] In one embodiment, information indicating which predicted CSI among predicted CSI (e.g., P-CRI(s), P-SSBRI(s), predicted PMI) for which monitoring is to be performed may be set by the base station to the terminal. For example, the base station may set information indicating a time instance for monitoring among one or more time instances related to prediction to the terminal. More specifically, the base station may transmit a reporting configuration including information indicating a time instance for monitoring (e.g., a reporting configuration related to monitoring / prediction accuracy) to the terminal.

[0281] In one embodiment, information indicating which predicted CSI (e.g., P-CRI(s), P-SSBRI(s), predicted PMI) among predicted CSI for a plurality of monitoring occasions (or a plurality of time instances) was monitored may be reported. For example, when a terminal performs a monitoring report (when the terminal performs a CSI report including a PAI), the terminal may transmit to the base station a monitoring report (CSI) including information about a time instance among the plurality of time instances for which monitoring was performed.

[0282] For example, the terminal may determine the number of groups and report this to the base station. The terminal may report information about the number of groups along with a monitoring metric. In other words, the terminal may report CSI including the number of groups and a monitoring metric (e.g., PAI) to the base station.

[0283] Reporting of the above monitoring metrics (e.g., PAI) can be performed based on Uplink Control Information (UCI). In this case, both one-part encoding and two-part encoding may be considered for encoding the UCI associated with the report.

[0284] For single-part encoding, the UCI size can be determined based on the (configured) maximum number of monitoring occasions (or time instances) or the number of groups described above. If the number of monitoring occasions (or time instances) associated with an actual report is less than the maximum, zero-padding can be applied to the remaining bits used for reporting. This can eliminate ambiguity in the UCI payload.

[0285] For 2-part encoding, part 1 CSI has a fixed payload, and part 2 CSI has a variable payload. Part 1 CSI may include at least one of the number of metrics (# of metrics), the number of groups (# of groups), and / or the first metric. Part 2 CSI may include the remaining metric(s) (e.g., metrics other than the first metric among the metrics to be reported). For example, if part 1 CSI does not include the first metric, part 2 CSI may include all metrics. For example, if only a single metric is reported, reporting of part 2 CSI may be omitted.

[0286] For example, the terminal may report monitoring results (e.g., PAI) and inference results (e.g., P-CRI(s), P-SSBRI(s), P-L1-RSRP(s) and / or predicted PMI).

[0287] As a specific example, the terminal may report inference results and monitoring results based on a single payload (e.g., a single CSI report). In other words, the terminal may report CSI that includes predicted information (e.g., at least one of P-CRI(s), P-SSBRI(s), P-L1-RSRP(s), and / or predicted PMI) and information related to prediction accuracy (e.g., PAI).

[0288] As a specific example, the terminal may report an inference result and report a monitoring result related to the inference result. In other words, the terminal may report a first CSI including predicted information (e.g., at least one of P-CRI(s), P-SSBRI(s), P-L1-RSRP(s), and / or predicted PMI) and a second CSI including information related to prediction accuracy (e.g., PAI).

[0289] For example, part 1 CSI may include an indicator indicating whether monitoring-related reporting is performed.

[0290] Proposal 2

[0291] When a terminal performs CSI prediction (e.g., reports CSI including predicted information) for one or more time instances, a method of determining / calculating and reporting multiple metrics for performance monitoring may be considered. For example, the terminal may report CSI including predicted information (e.g., P-CRI(s), P-SSBRI(s), P-L1-RSRP(s) and / or predicted PMI) for one or more time instances to the base station. In this case, the terminal may report CSI including multiple metrics (e.g., PAIs) determined based on the one or more time instances for monitoring to the base station.

[0292] For performance monitoring, the terminal calculates and reports metrics (e.g., PAI) based on each time instance. When considering monitoring predictions for multiple time instances, the terminal must report multiple metrics. In this case, the number of prediction instances to be predicted ( ) may cause an issue where the number of ground-truth CSIs is not equal to the number of ground-truth CSIs set.

[0293] Below, we specifically describe methods for solving the above problem. Methods for determining the number of metrics to be reported among the multiple metrics are sequentially described in Methods 2-1 through 2-3. For convenience of explanation, the number of ground-truth CSIs set is referred to as S.

[0294] Method 2-1

[0295] The terminal can determine / report S metrics.

[0296] For example, the terminal may be configured to predict the number of prediction instances to be predicted (or the number of time instances for the prediction to be monitored) (e.g. ) is different from the number of ground-truth CSIs (S), the terminal can determine / report S metrics.

[0297] For example, the terminal may be configured to predict the number of prediction instances to be predicted (or the number of time instances for the prediction to be monitored) (e.g. ) may not be expected to be different from the number of ground-truth CSIs (S). In other words, the terminal may not expect the number of prediction instances to be predicted (or the number of time instances for the prediction to be monitored) (e.g., ) can be expected to be equal to the number of ground-truth CSIs (S). In other words, the terminal can be expected to be configured / instructed to report S metrics (e.g., S PAIs).

[0298] Method 2-2

[0299] The terminal Ability to determine / report dog metrics.

[0300] For example, the terminal may be configured to predict the number of prediction instances to be predicted (or the number of time instances for the prediction to be monitored) (e.g. ) is different from the number of ground-truth CSIs (S), the terminal Ability to determine / report dog metrics.

[0301] For example, the terminal may be configured to predict the number of prediction instances to be predicted (or the number of time instances for the prediction to be monitored) (e.g. ) may not be expected to be different from the number of ground-truth CSIs (S). In other words, the terminal may not expect the number of prediction instances to be predicted (or the number of time instances for the prediction to be monitored) (e.g., ) can be expected to be equal to the number of ground-truth CSIs (S). In other words, the terminal Dog metrics (e.g.: You can expect to be set / instructed to report your PAIs (Personal AIs).

[0302] Method 2-3

[0303] The terminal can determine / report specific M metrics.

[0304] For example, the terminal may be configured to predict the number of prediction instances to be predicted (or the number of time instances for the prediction to be monitored) (e.g. ) is different from the number of ground-truth CSIs (S), the terminal Among the metrics, M metrics can be determined / reported.

[0305] For example, the terminal may be configured to predict the number of prediction instances to be predicted (or the number of time instances for the prediction to be monitored) (e.g. ) can be expected to be set / instructed to report M metrics based on the M values. As a specific example, the base station can set / instruct the terminal to the M value. M is It can be set to a value less than or equal to .

[0306] In method 2-2 / method 2-3, S If the value is smaller, there may be a problem where prediction instances do not have ground-truth CSI. The following methods can be considered to address this issue.

[0307] To calculate metrics related to predicted CSI for a given instance, the terminal can predict / determine ground-truth CSI for the instance. The terminal can calculate / determine metrics based on the predicted / determined ground-truth CSI.

[0308] Method 2-4

[0309] The terminal is S and Metrics can be determined / reported based on small numbers.

[0310] For example, the terminal may be configured to predict the number of prediction instances to be predicted (or the number of time instances for the prediction to be monitored) (e.g. ) is different from the number of ground-truth CSIs (S), the terminal Metrics can be determined / reported based on the smaller number of S. As a specific example, < In case of S, the terminal is It can determine / report the metrics of dogs. For example, S < In this case, the terminal can determine / report S metrics.

[0311] For example, the terminal It can be expected that the base station will be configured / instructed to determine / report metrics based on the smaller number of S and S. As a specific example, the base station can be configured with settings related to how the number of metrics is determined (e.g., (decided as the smaller number among S and S) can be transmitted to the terminal.

[0312] Proposal 2-1

[0313] It may be assumed that the terminal reports a smaller number of metrics (hereinafter referred to as m metrics) than the number of metrics calculated / determined by the terminal. For example, it may be assumed that the methods of Proposal 2 above require the terminal to report a smaller number of metrics than the number of metrics calculated by the terminal.

[0314] In cases like the above, criteria are required to select / determine the metrics to report.

[0315] The criteria for selecting / determining the metrics to be reported may be based on at least one of the following (1) to (4), which are explained in turn below.

[0316] (1) Selection based on the ID order of the metric

[0317] A base station can explicitly set / instruct the terminal to report metrics in order. The terminal can report m metrics based on the order.

[0318] For example, a base station can set / instruct a terminal to assign a priority for a monitoring occasion (or time instance). Based on the priority, the terminal calculates and reports m metrics.

[0319] For example, a new ID for a monitoring occasion (or time instance) can be introduced / defined. The base station can configure / instruct / manage the monitoring occasion (or time instance) based on the ID. The terminal can calculate and report m metrics based on the order of the IDs.

[0320] (2) Selection based on metric calculation time

[0321] Based on the point in time at which metrics are calculated, the terminal can determine / report m metrics. For example, the terminal can report the m most recently calculated metrics (the most recent m metrics). For example, the terminal can determine / report m metrics based on the center of the prediction window.

[0322] For example, the terminal can determine / report m metrics based on the start and end points of the prediction window. Specifically, the terminal can first determine metrics with high priority and then determine the remaining metrics. The terminal can determine metrics (e.g., two metrics) based on the start and end points. The terminal can determine the remaining metrics (e.g., m-2 metrics) at equal intervals from after the start point to before the end point.

[0323] (3) Selection based on metric accuracy / probability / confidence

[0324] The terminal can determine / report the best m metrics based on accuracy / probability / confidence. After calculating the metrics, the terminal can determine / report the best m metrics in descending order of accuracy / probability / confidence for each metric.

[0325] (4) Selection based on metric priority

[0326] The terminal can determine / report the best m metrics based on the priority. Specifically, the priority of the metric for each instance can be set / determined / defined. The terminal can determine / report the best m metrics in descending order of priority. For example, the priority for an instance can be set to the terminal by the base station. For example, the priority for an instance can be configured / calculated / determined as a function of specific parameters. For example, the base station can set / instruct the terminal to specific parameters for calculating / determining the priority. For example, the specific parameters can include at least one of the metric calculation time, the metric ID, and / or the metric accuracy / probability / confidence.

[0327] Proposal 3

[0328] When a terminal performs CSI prediction (e.g., reports CSI including predicted information) for one or more time instances, the terminal may report other information for performance monitoring. For example, the other information may be related to a preferred window size.

[0329] For UE-sided monitoring, the terminal can determine the preferred window size by comparing the above metric with a threshold. The terminal can report the preferred window size to the base station.

[0330] This embodiment takes into consideration the following technical matters.

[0331] The number of time instances for which predictions can be made may vary depending on the terminal's environment or conditions. A terminal with a favorable environment or conditions (e.g., terminal 1) can predict many instances. A terminal with a poor environment or conditions (e.g., terminal 2) may not be able to predict the same number of time instances as terminal 1.

[0332] In summary, this embodiment supports efficient inference by considering that the number of time instances supported for prediction by a terminal varies depending on the terminal's environment / conditions. Specifically, the terminal can report a preferred window size (e.g., observation window / prediction window) for efficient inference to the base station. The number of time instances for prediction of the terminal can be set based on the preferred window size reported by the terminal, thereby obtaining improved results. Specifically, compared to cases where time instances for prediction are set without considering the terminal's environment / conditions, prediction can be performed by maximizing the capability of each terminal. In addition, it can prevent the number of time instances from being set exceeding the capability of each terminal. In other words, it can prevent ambiguity in terminal operation from occurring.

[0333] As an example of determining the preferred window size, when the terminal calculates the metric for each instance, if the number of instances with a metric exceeding a specific threshold is T, the terminal can report the preferred window size as T to the base station.

[0334] For example, if the prediction is good, T= In this case, the terminal sends a preferred window size to the base station. can be reported. Additionally, the terminal can report to the base station. A separate signal can be sent requesting an increase in the number of instances. This allows predictions to be made for more instances within the limits allowed by the terminal.

[0335] Information about the preferred window size can be reported to the base station together with monitoring or inference reports. Based on this information, the base station can dynamically set the window size for the terminal.

[0336] Below, a signaling procedure based on at least one of proposals 1 to 3 is specifically described in terms of terminal operation / base station operation.

[0337] For the UE-sided model, the following terminal operations can be performed.

[0338] Step 1: The terminal reports capability information to the base station, including the maximum number of CSI-RS resources that can be supported, the number of CSI-RS ports, the total number of CSI-RS ports that can be supported simultaneously, and the number of Rx antenna groups.

[0339] Step 2: The terminal receives configuration information related to CSI-RS and configuration information related to CSI reporting from the base station.

[0340] Step 3: The terminal receives a CSI-RS from the base station. Based on the CSI-RS, the terminal measures / predicts / calculates CSI using an AI / ML model.

[0341] Step 3-1: The terminal performs CSI omission based on configuration information and CSI priority. In other words, the information included in the CSI (CSI report) and the information omitted from the CSI (CSI report) are determined based on configuration information and CSI priority.

[0342] Step 4: The terminal reports the measured / predicted / calculated CSI to the base station.

[0343] Step 5: The terminal receives information (e.g., DCI) for scheduling downlink channels (e.g., PDCCH, PDSCH) from the base station.

[0344] Step 6: The terminal receives a downlink channel / signal from the base station.

[0345] For the UE-sided model, the following base station operations can be performed:

[0346] Step 1: The base station receives capability information from the terminal, including the maximum number of CSI-RS resources that can be supported, the number of CSI-RS ports, the total number of CSI-RS ports that can be supported simultaneously, and the number of Rx antenna groups.

[0347] Step 2: The base station transmits configuration information related to CSI-RS and configuration information related to CSI reporting to the terminal.

[0348] Step 3: The base station transmits CSI-RS to the terminal.

[0349] Step 4: The base station receives CSI (measured / predicted / calculated by the terminal) from the terminal.

[0350] Step 5: The base station transmits information for scheduling downlink channels (e.g., PDCCH, PDSCH) to the terminal. At this time, the information for scheduling the downlink channel may be based on the CSI.

[0351] Step 6: The base station transmits a downlink channel / signal to the terminal.

[0352] For the NW-sided model, the following terminal operations can be performed:

[0353] Terminal operation:

[0354] Step 1: The terminal reports capability information to the base station, including the maximum number of CSI-RS resources that can be supported, the number of CSI-RS ports, the total number of CSI-RS ports that can be supported simultaneously, and the number of Rx antenna groups.

[0355] Step 2: The terminal receives configuration information related to CSI-RS and configuration information related to CSI reporting from the base station.

[0356] Step 3: The terminal receives the CSI-RS from the base station. Based on the CSI-RS, the terminal measures / predicts / calculates CSI.

[0357] Step 3-1: The terminal performs CSI omission based on configuration information and CSI priority.

[0358] Step 4: The terminal reports the measured / predicted / calculated CSI to the base station.

[0359] Step 5: The terminal receives information (e.g., DCI) for scheduling downlink channels (e.g., PDCCH, PDSCH) from the base station.

[0360] Step 6: The terminal receives a downlink channel / signal from the base station.

[0361] For the NW-sided model, the following base station operations can be performed:

[0362] Step 1: The base station receives capability information from the terminal, including the maximum number of CSI-RS resources that can be supported, the number of CSI-RS ports, the total number of CSI-RS ports that can be supported simultaneously, and the number of Rx antenna groups.

[0363] Step 2: The base station transmits configuration information related to CSI-RS and configuration information related to CSI reporting to the terminal.

[0364] Step 3: The base station transmits CSI-RS to the terminal.

[0365] Step 4: The base station receives CSI (measured / predicted / calculated by the terminal) from the terminal.

[0366] Step 5: The base station transmits information to the terminal for scheduling downlink channels (e.g., PDCCH, PDSCH). This information for scheduling the downlink channel may be based on CSI predicted by the base station. Specifically, the base station can predict CSI using an AI / ML model based on the reported CSI. The base station can then perform downlink scheduling based on the predicted CSI.

[0367] In the above terminal / base station operation, (some) specific steps may be omitted.

[0368] The above suggestions 1 / 2 / 3 can be used alone or in combination.

[0369] The AI / ML model-based signaling operation is specifically described with reference to FIG. 9 below.

[0370] Figure 9 illustrates a signaling procedure based on an AI / ML model.

[0371] Step 1: In the description of this specification, signaling (e.g., information / data / channel / signal, etc.) or a set of signaling between a specific node (e.g., a terminal, a network, etc.) and another node can be interpreted as the signaling or a set of signaling of Step 1 used to perform an operation based on an AI / ML model, even if there is no separate mention. For example, the signaling may correspond to training data for training (i.e., generation and / or reconstruction) the AI / ML model of FIG. 1, or correspond to inference data used for inference of the AI / ML model, or correspond to feedback for the AI / ML model, etc. If signaling between nodes is not required prior to an operation based on an AI / ML model in this specification, Step 1 may be omitted.

[0372] If the one-side model is used in this specification, the one-way / two-way signaling (set) in this specification may correspond to one stage of signaling. In addition, if the two-side model is used in this specification, the one-way / two-way signaling in this specification may correspond to one stage of signaling, and also, a repetitive signaling operation may correspond to one stage of signaling.

[0373] For example, in AI / ML model-based beam management (BM), if a base station predicts (i.e., infers) beam(s) with good quality based on an AI / ML model, the base station can receive quality / intensity information for multiple beams from a terminal. Furthermore, if a terminal predicts (i.e., infers) beam(s) with good quality based on an AI / ML model, the terminal can receive multiple beams from the base station.

[0374] Step 2: In the description of this specification, an operation (e.g., calculation, selection, prediction, etc.) in a specific node (e.g., terminal, network, etc.) or a joint operation (e.g., calculation, selection, prediction, etc.) in multiple nodes (e.g., terminal, network, etc.) may correspond to a step 2 operation based on one or more functions in the functional framework of the AI / ML model, even if not mentioned separately. For example, the operation may correspond to training (i.e., generation and / or reconstruction) of the AI / ML model of FIG. 2, or may correspond to inference of the AI / ML model, etc. When a one-side model is used, an operation performed by a single node in this specification may correspond to a step 2 operation. In addition, when a two-side model is used, a joint operation performed by multiple nodes in this specification may correspond to a step 2 operation.

[0375] For example, in an AI / ML model-based BM, the base station can use quality / intensity information for multiple beams received from the terminal as inference data to predict (i.e., infer) beam(s) with good quality based on the AI / ML model. Furthermore, the terminal can measure multiple beams received from the base station and use the measurement results as inference data to predict (i.e., infer) beam(s) with good quality based on the AI / ML model.

[0376] Step 3: In the description of this specification, signaling (e.g., information / data / channel / signal, etc.) or a set of signaling between a specific node (e.g., terminal, network, etc.) and another node can be interpreted as a three-step signaling or a set of signaling generated due to (as a result of) an operation based on an AI / ML model, even if not otherwise stated. For example, the signaling may correspond to an output resulting from inference of the AI / ML model of FIG. 2. If signaling between nodes is not required as a result of an operation based on an AI / ML model in this specification, Step 3 may be omitted. If a One-side model is used in this specification, a one-way / two-way signaling (set) in this specification may correspond to the three-step signaling. In addition, if a Two-side model is used in this specification, a one-way / two-way signaling in this specification may correspond to the three-step signaling, and a repetitive signaling operation may also correspond to the three-step signaling.

[0377] For example, in an AI / ML model-based BM, the base station can transmit to the terminal the beam(s) predicted based on the AI / ML model as candidates so that the terminal can determine the optimal beam. Furthermore, the terminal can report to the base station the beam(s) predicted based on the AI / ML model to request the base station to transmit the candidate beams as candidates for determining the optimal beam.

[0378] In terms of implementation, the operations of the base station / terminal according to the embodiments described above (e.g., operations based on at least one of Proposals 1 to 3) can be processed by the device of FIG. 13 (e.g., the processor (110, 210) of FIG. 13).

[0379] In addition, the operations of the base station / terminal according to the above-described embodiment (e.g., operations based on at least one of proposals 1 to 3) may be stored in a memory (e.g., 140, 240 of FIG. 13) in the form of commands / programs (e.g., instructions, executable codes) for driving at least one processor (e.g., 110, 210 of FIG. 13).

[0380] The embodiments described below are specifically described with reference to FIGS. 11 and 12 in terms of the operation of the terminal and base station. The methods described below are distinguished for convenience of explanation, and it is understood that some components of one method may be substituted for or combined with some components of another method.

[0381] FIG. 11 is a flowchart illustrating a method according to one embodiment of the present specification.

[0382] Referring to FIG. 11, a method according to one embodiment of the present specification includes a step of receiving setting information related to CSI (S1110), a first CSI reporting step related to prediction (S1120), and a second CSI reporting step related to prediction accuracy (S1130).

[0383] In S1110, the terminal receives configuration information related to channel state information (CSI) from the base station.

[0384] For example, the configuration information may include information based on the above-described CSI-related operation and at least one of Proposals 1 to 3. As a specific example, the configuration information may include at least one of one or more resource settings (e.g., M≥1 CSI-ResourceConfig resource settings) and / or one or more reporting settings (e.g., N≥1 CSI-ReportConfig reporting settings). Each of the one or more reporting settings may be associated with at most three resource settings. In other words, each reporting setting may include IDs (e.g., CSI-ResourceConfigId) of at most three resource settings.

[0385] For example, the one or more reporting settings may include reporting settings related to prediction and / or reporting settings related to prediction accuracy. As a specific example, the configuration information may include i) a first reporting setting related to prediction and ii) a second reporting setting related to prediction accuracy.

[0386] In one embodiment, the number of one or more time instances may be set based on the first report setting. For example, the first report setting may include information indicating the number of one or more time instances (e.g., information indicating the above-described N4 or the upper layer parameter nroftimeinstance).

[0387] In S1120, the terminal reports a first CSI including information related to the prediction to the base station.

[0388] In one embodiment, the first CSI may include predicted information for each of one or more time instances. Specifically, the first CSI may include predicted CSI parameter(s) (e.g., predicted PMI, P-CRI(s), P-SSBRI(s), and / or P-L1-RSRP(s)) for each of one or more time instances.

[0389] In one embodiment, the predicted information may include at least one of i) at least one predicted CSI-RS Resource Indicator (P-CRI), ii) at least one predicted SSB Resource Indicator (P-SSBRI), iii) at least one predicted Layer 1-Reference Signal Received Power (L1-RSRP), and / or iv) a predicted Precoding Matrix Indicator (PMI).

[0390] In one embodiment, the first CSI may include predicted CSI parameter(s) (e.g., P-CRI(s), P-SSBRI(s), P-L1-RSRP(s) and / or predicted PMI) based on a report quantity of the first report configuration.

[0391] For example, the report quantity of the first report configuration may be set to p-cri, p-cri-RSRP, p-ssb-index, p-ssb-index-RSRP, or cri-RI-PMI-CQI (or a value indicating predicted PMI). p-cri indicates a predicted channel state information-reference signal resource indicator (P-CRI). p-ssb-index indicates a predicted SSB resource indicator (P-SSBRI). In p-cri-RSRP or p-ssb-index-RSRP, RSRP indicates a predicted Layer 1-Reference Signal Received Power (L1-RSRP, P-L1-RSRP). In cri-RI-PMI-CQI, PMI may indicate the predicted PMI.

[0392] In one embodiment, the prediction may be performed based on a measurement. The measurement may be performed based on a resource configuration (e.g., CSI-ResourceConfig) associated with the first reporting configuration. For example, the first reporting configuration may be associated with a first resource configuration associated with measurement and a second resource configuration associated with prediction. The first reporting configuration may include an ID of the first resource configuration and an ID of the second resource configuration. Each ID may be based on CSI-ResourceConfigId.

[0393] For example, the measurement may include Layer 1-Reference Signal Received Power (L1-RSRP) measurements. Based on the L1-RSRP measurements, i) at least one predicted channel state information-reference signal resource indicator (CSI-RS Resource Indicator, P-CRI), ii) at least one predicted SSB Resource Indicator (P-SSBRI), and / or iii) at least one predicted Layer 1-Reference Signal Received Power (P-L1-RSRP) may be determined.

[0394] More specifically, predictions can be made for CSI-RS resources or SSB resources associated with the second resource configuration based on the L1-RSRP measurements. Specifically, predicted L1-RSRPs of the CSI-RS resources or SSB resources associated with the second resource configuration can be determined. For example, best CRI(s) or best SSBRI(s) can be determined based on an order or ranking of the predicted L1-RSRPs. For example, predicted P-CRI(s) or predicted SSBRI(s) reported through the first CSI can be based on the best CRI(s) or the best SSBRI(s).

[0395] At S1130, the terminal reports the second CSI related to the prediction accuracy to the base station.

[0396] For example, the second CSI may include information (e.g., PAI) related to a time instance for monitoring among the one time instances.

[0397] In one embodiment, the second reporting configuration may include information indicating the time instance for monitoring among the one or more time instances (e.g., a higher layer parameter timeinstanceformmonitoring). As an example, the information indicating the time instance for monitoring may be based on a value indicating which time instance for monitoring among the one or more time instances set based on the first reporting configuration is the time instance (e.g., first, second, etc.).

[0398] For example, the second CSI may include a CSI parameter (e.g., RS-PAI) based on the report quantity of the second report configuration. For example, the report quantity of the second report configuration may be set to rs-pai.

[0399] In one embodiment, the second CSI may include a Prediction Accuracy Indicator (PAI) determined based on the time instance for the monitoring. The PAI may include a Reference Signal-Prediction Accuracy Indicator (RS-PAI) or a Channel State Information-Prediction Accuracy Indicator (CSI-PAI).

[0400] For example, a metric (PAI) for monitoring may be determined based on a comparison of predicted CSI and ground-truth CSI. Specifically, the PAI may be determined based on CSI parameters determined for the time instance for monitoring and at least one predicted CSI parameter within the first CSI. More specifically, the PAI may be based on a comparison result (e.g., cosine similarity, number of identical CSI parameters) between the determined CSI parameters and the at least one predicted CSI parameter.

[0401] As a specific example of the CSI-PAI, the CSI-PAI may be determined based on CSI parameters determined for the time instance for the monitoring (e.g., non-predicted PMI) and at least one predicted CSI parameter (e.g., predicted PMI) within the first CSI. More specifically, a cosine similarity (e.g., Square Generalized Cosine Similarity (SGCS)) may be determined / calculated based on i) the determined CSI parameters (e.g., non-predicted PMI) and ii) at least one predicted CSI parameter (e.g., predicted PMI) within the first CSI. The CSI-PAI may be based on the cosine similarity (e.g., SGCS).

[0402] As a specific example of the RS-PAI, the RS-PAI may be determined based on whether the CSI parameters (e.g., CRIs, SSBRIs) determined for the time instance for the monitoring are mapped to at least one predicted CSI parameter (e.g., P-CRI(s) and / or P-SSBRI(s)) within the first CSI. As a specific example, the RS-PAI may be related to the number of CSI parameters mapped to the at least one predicted CSI parameter among the determined CSI parameters.

[0403] In one embodiment, an inference report (first CSI report) and a monitoring report (second CSI report) may be configured to be performed together. Specifically, the second report configuration may be linked to the first report configuration. The second report configuration may include information (e.g., a defined ID or inferenceReportConfigId) that links the report of the first CSI with the report of the second CSI. As a specific example, the configuration information may include first report configurations related to prediction and second report configurations related to prediction accuracy. Each second report configuration may be linked to one of the first report configurations based on the information (ID).

[0404] For example, the 1st / 2nd CSI can be interpreted / replaced with the 1st / 2nd CSI report.

[0405] In one embodiment, the method may further include a step of transmitting information related to a preferred window size. For example, the terminal may report / transmit information related to the preferred window size to the base station. This embodiment may be based on Proposal 3. For example, the number of the one or more time instances may be less than or equal to the preferred window size.

[0406] The operations based on S1110 to S1130 described above can be implemented by the device of FIG. 13. For example, referring to FIG. 13, the terminal (200) can control one or more transceivers (230) and / or one or more memories (240) to perform the operations based on S1110 to S1130.

[0407] The embodiments described below are specifically described in terms of base station operation.

[0408] S1210 to S1230 described below correspond to S1110 to S1130 described in FIG. 11. Considering the above correspondence, redundant descriptions are omitted. That is, the specific description of the base station operation described below may be replaced with the description / example of FIG. 11 corresponding to the corresponding operation.

[0409] FIG. 12 is a flowchart illustrating a method according to another embodiment of the present specification.

[0410] Referring to FIG. 12, a method according to another embodiment of the present specification includes a step of transmitting configuration information related to CSI (S1210), a step of receiving a first CSI related to prediction (S1220), and a step of receiving a second CSI related to prediction accuracy (S1230).

[0411] In S1210, the base station transmits configuration information related to channel state information (CSI) to the terminal. For example, the configuration information may include i) a first report configuration related to prediction and ii) a second report configuration related to prediction accuracy.

[0412] In S1220, the base station receives a first CSI including information related to the prediction from the terminal.

[0413] At S1230, the base station receives second CSI related to the prediction accuracy from the terminal.

[0414] For example, the first CSI may include predicted information for each of one or more time instances. The second CSI may include information related to a time instance for monitoring among the one or more time instances. The second reporting configuration may include information indicating the time instance for monitoring among the one or more time instances.

[0415] The operations based on S1210 to S1230 described above can be implemented by the device of FIG. 13. For example, referring to FIG. 13, the base station (100) can control one or more transceivers (130) and / or one or more memories (140) to perform operations based on S1210 to S1230.

[0416] The operations / terms based on the embodiments described above have been described assuming an existing system (e.g., a 5G system). However, this is for convenience of explanation and is not intended to limit the scope of application of the technical problems and problem-solving means to be solved by this specification to a specific system. That is, the technical problems / technical issues / problems mentioned in this specification may equally exist in other systems (e.g., a 6G system). It is clear that the embodiments of this specification can be expanded and applied to solve problems that equally exist in the other systems. Therefore, for the expanded application of the embodiments of this specification to other systems, the terms defined / described based on the 5G system may be replaced / changed with terms defined in the other systems (or generalized terms that are not specific to one system). For example, PRACH, PUSCH, PUCCH, or SRS may be replaced / changed with uplink signals (or uplink channels). For example, SSB, CSI-RS, PDSCH, and PDCCH may be replaced / changed with downlink signals (or downlink channels).

[0417] Hereinafter, a device to which an embodiment of the present specification can be applied (a device that implements a method / operation according to an embodiment of the present specification) is described with reference to FIG. 13.

[0418] FIG. 13 is a drawing showing the configuration of a first device and a second device according to an embodiment of the present specification.

[0419] The first device (100) may include a processor (110), an antenna unit (120), a transceiver (130), and a memory (140).

[0420] The processor (110) performs baseband-related signal processing and may include a higher layer processing unit (111) and a physical layer processing unit (115). The higher layer processing unit (111) may process operations of a MAC layer, an RRC layer, or higher layers. The physical layer processing unit (115) may process operations of a PHY layer. For example, when the first device (100) is a base station device in base station-terminal communication, the physical layer processing unit (115) may perform uplink reception signal processing, downlink transmission signal processing, etc. For example, when the first device (100) is a first terminal device in terminal-to-terminal communication, the physical layer processing unit (115) may perform downlink reception signal processing, uplink transmission signal processing, sidelink transmission signal processing, etc. In addition to performing baseband-related signal processing, the processor (110) may also control the overall operation of the first device (100).

[0421] The antenna unit (120) may include one or more physical antennas, and when it includes multiple antennas, it may support MIMO transmission and reception. The transceiver (130) may include an RF (Radio Frequency) transmitter and an RF receiver. The memory (140) may store information processed by the processor (110), and software, an operating system, applications, etc. related to the operation of the first device (100), and may also include components such as a buffer.

[0422] The processor (110) of the first device (100) may be configured to implement the operation of the base station in the base station-to-terminal communication (or the operation of the first terminal device in the terminal-to-terminal communication) in the embodiments described in the present disclosure.

[0423] The second device (200) may include a processor (210), an antenna unit (220), a transceiver (230), and a memory (240).

[0424] The processor (210) performs baseband-related signal processing and may include a higher layer processing unit (211) and a physical layer processing unit (215). The higher layer processing unit (211) may process operations of a MAC layer, an RRC layer, or higher layers. The physical layer processing unit (215) may process operations of a PHY layer. For example, when the second device (200) is a terminal device in base station-terminal communication, the physical layer processing unit (215) may perform downlink reception signal processing, uplink transmission signal processing, etc. For example, when the second device (200) is a second terminal device in terminal-to-terminal communication, the physical layer processing unit (215) may perform downlink reception signal processing, uplink transmission signal processing, sidelink reception signal processing, etc. In addition to performing baseband-related signal processing, the processor (210) may also control the overall operation of the second device (210).

[0425] The antenna unit (220) may include one or more physical antennas, and when it includes multiple antennas, it may support MIMO transmission and reception. The transceiver (230) may include an RF transmitter and an RF receiver. The memory (240) may store information processed by the processor (210), software, an operating system, applications, etc. related to the operation of the second device (200), and may also include components such as a buffer.

[0426] The processor (210) of the second device (200) may be configured to implement operations of the terminal in base station-to-terminal communication (or operations of the second terminal device in terminal-to-terminal communication) in the embodiments described in the present disclosure.

[0427] In the operation of the first device (100) and the second device (200), the same explanations given for the base station and the terminal (or the first terminal and the second terminal in the terminal-to-terminal communication) in the examples of the present disclosure may be applied, and redundant explanations are omitted.

[0428] Here, the wireless communication technology implemented in the device of the present disclosure may include LTE, NR, and 6G, as well as Narrowband Internet of Things (NB-IoT) for low-power communication. For example, NB-IoT technology may be an example of LPWAN (Low Power Wide Area Network) technology and may be implemented in standards such as LTE Cat NB1 and / or LTE Cat NB2, and is not limited to the above-described names.

[0429] Additionally or alternatively, the wireless communication technology implemented in the device of the present disclosure may perform communication based on LTE-M technology. For example, LTE-M technology may be an example of LPWAN technology and may be called by various names such as eMTC (enhanced Machine Type Communication). For example, LTE-M technology may be implemented by at least one of various standards such as 1) LTE CAT 0, 2) LTE Cat M1, 3) LTE Cat M2, 4) LTE non-BL (non-Bandwidth Limited), 5) LTE-MTC, 6) LTE Machine Type Communication, and / or 7) LTE M, and is not limited to the above-described names.

[0430] Additionally or alternatively, the wireless communication technology implemented in the device of the present disclosure may include at least one of ZigBee, Bluetooth, and Low Power Wide Area Network (LPWAN), which take low-power communication into account, and is not limited to the above-described names. For example, ZigBee technology can create personal area networks (PANs) related to small / low-power digital communication based on various standards such as IEEE 802.15.4, and may be called by various names.

Claims

1. In the method, A step of receiving configuration information related to channel state information (CSI), the configuration information including i) a first report configuration related to prediction and ii) a second report configuration related to prediction accuracy; A step of reporting a first CSI including information related to the above prediction; and A step of reporting a second CSI related to the above prediction accuracy; including, The first CSI includes predicted information for each of one or more time instances, The above second CSI includes information related to a time instance for monitoring among the one time instances, A method characterized in that the second report setting includes information indicating the time instance for monitoring among the one or more time instances.

2. In paragraph 1, A method characterized in that the predicted information comprises at least one of i) at least one predicted channel state information-reference signal resource indicator (CSI-RS Resource Indicator, P-CRI), ii) at least one predicted SSB Resource Indicator (P-SSBRI), iii) at least one predicted Layer 1-Reference Signal Received Power (L1-RSRP, P-L1-RSRP) and / or iv) a predicted Precoding Matrix Indicator (PMI).

3. In paragraph 1, The above prediction is performed based on measurements, A method characterized in that the above measurement is performed based on resource settings related to the first report settings.

4. In paragraph 3, The above measurements include L1-RSRP measurements (Layer1-Reference Signal Received Power, L1-RSRP, measurements), A method characterized in that, based on the above L1-RSRP measurements, i) at least one predicted channel state information-reference signal resource indicator (CSI-RS Resource Indicator, P-CRI), ii) at least one predicted SSB Resource Indicator (P-SSBRI) and / or iii) at least one predicted L1-RSRP (predicted Layer1-Reference Signal Received Power, P-L1-RSRP) is determined.

5. In paragraph 1, A method characterized in that the number of said one or more time instances is set based on said first report setting.

6. In paragraph 1, The second CSI includes a Prediction Accuracy Indicator (PAI) determined based on the time instance for the monitoring, A method characterized in that the above PAI includes a Reference Signal-Prediction Accuracy Indicator (RS-PAI) or a Channel State Information-Prediction Accuracy Indicator (CSI-PAI).

7. In paragraph 6, A method characterized in that the PAI is determined based on the CSI parameters determined for the time instance for the monitoring and at least one predicted CSI parameter within the first CSI.

8. In paragraph 7, A method characterized in that the PAI is based on a comparison result between the determined CSI parameters and the at least one predicted CSI parameter.

9. In paragraph 1, A method characterized in that the second report setting is linked to the first report setting.

10. In paragraph 9, A method characterized in that the second report setting includes information linking the report of the first CSI with the report of the second CSI.

11. In paragraph 10, The report quantity of the above first report setting is set to p-cri, p-cri-RSRP, p-ssb-index, p-ssb-index-RSRP or cri-RI-PMI-CQI, A method characterized in that the report quantity of the above second report setting is set to rs-pai or csi-pai.

12. At the terminal, One or more transmitters and receivers; one or more processors; and One or more memories connected to said one or more processors and storing instructions, A terminal characterized in that the instructions, based on being executed by the one or more processors, cause the terminal to perform all steps of the method according to any one of claims 1 to 11.

13. In a device comprising one or more memories and one or more processors connected to the one or more memories, A device characterized in that said one or more memories store instructions that cause said device to perform all steps of a method according to any one of claims 1 to 11, based on being executed by said one or more processors.

14. In a non-transitory computer-readable storage medium storing instructions, A non-transitory computer-readable storage medium characterized in that the instructions executable by one or more processors cause a terminal to perform all steps of a method according to any one of claims 1 to 11.

15. In the method, A step of transmitting configuration information related to channel state information (CSI), wherein the configuration information includes i) a first report configuration related to prediction and ii) a second report configuration related to prediction accuracy; A step of receiving a first CSI including information related to the above prediction; and A step of receiving a second CSI related to the above prediction accuracy; including: The first CSI includes predicted information for each of one or more time instances, The above second CSI includes information related to a time instance for monitoring among the one time instances, A method characterized in that the second report setting includes information indicating the time instance for monitoring among the one or more time instances.

16. At the base station, One or more transmitters and receivers; one or more processors; and One or more memories connected to said one or more processors and storing instructions, A base station characterized in that the instructions, based on being executed by the one or more processors, cause the base station to perform all steps of the method according to claim 15.

Images