Method and device for ai / ml inference performance monitoring in wireless communication system

Abstract

A method of a UE, according to one embodiment of the present disclosure, may comprise the steps of: receiving, from a base station, a first message including transmission information of a monitoring RS and monitoring report information; measuring monitoring RSs received from the base station on the basis of the transmission information of the monitoring RS; on the basis that a time difference between a first RS used for inference of the AI model and the monitoring RS satisfies a preset time condition, generating a monitoring report message by using an inference result of an AI model and a measurement result of the monitoring RSs; and transmitting, to the base station, the generated monitoring report message through an uplink (UL) channel.

Description

Method and device for monitoring AI / ML inference performance in a wireless communication system

[0001] The present disclosure relates to artificial intelligence (AI) / machine learning (ML) technology in wireless communication systems, and more specifically, to technology for monitoring the inference performance of AI / LM in wireless communication systems.

[0002] Along with the advancement of information and communication technology, various wireless communication technologies are being developed. Representative wireless communication technologies include LTE (long term evolution) and NR (new radio), which are defined in the 3GPP (3rd generation partnership project) standards. LTE can be one of the wireless communication technologies among 4G (4th Generation) wireless communication technologies, and NR can be one of the wireless communication technologies among 5G (5th Generation) wireless communication technologies.

[0003] In order to process the rapidly increasing amount of wireless data following the commercialization of 4G communication systems (e.g., communication systems supporting LTE), 5G communication systems (e.g., communication systems supporting NR) that use frequency bands higher than those of 4G communication systems (e.g., frequency bands below 6 GHz) as well as frequency bands above 6 GHz have been developed and are reaching the commercialization stage. 5G communication systems can support eMBB (enhanced Mobile BroadBand), URLLC (Ultra-Reliable and Low Latency Communication), and mMTC (massive Machine Type Communication).

[0004] In 4G and 5G communication systems, MIMO (multiple input multiple output) was introduced to improve network capacity, performance, and frequency efficiency between the network (e.g., base station (gNB)) and user equipment (UE). MIMO technology is a beamforming technology that uses multiple antennas, allowing the network and each of the UEs to communicate with one another through beamforming.

[0005] Meanwhile, 3GPP, which establishes technical standards for 5G communication systems, is proposing various methods for communication using artificial intelligence (AI) and machine learning (ML) technologies. However, specific procedures for verifying whether the inference of AI / ML models is appropriate have not been proposed.

[0006] The objective of the present disclosure to address the above-mentioned requirements is to provide a method and apparatus for monitoring the performance of inference of an AI / ML model in a wireless communication system.

[0007] A method of user equipment (UE) according to one embodiment of the disclosure for achieving the above-mentioned purpose may include: receiving a first message from a base station comprising one or more monitoring reference signal (monitoring RS) transmission information and monitoring report setting information; measuring one or more monitoring RSs received from the base station based on the RS transmission information; selecting a monitoring RS for evaluating the performance of the AI ​​model from among one or more monitoring RSs having a temporal correlation with inference-related information by the UE-side artificial intelligence (AI) model; generating a monitoring report message using the measurement value of the selected monitoring RS and the inference result of the AI ​​model; and transmitting the monitoring report message to the base station through an uplink channel.

[0008] The selection of the above monitoring RS allows selecting the monitoring RS closest to the inference RS time point within preset time conditions and a specific RS transmission time point set in association with the inference RS or inference report used for inference.

[0009] The selected monitoring RS may be the monitoring RS closest to the inference target instance among the monitoring RSs existing within a preset time condition and a future inference target time instance determined based on the inference of the AI ​​model.

[0010] The above temporal association may include one or more of temporal proximity, whether a temporal offset is satisfied, or time conditions set at a base station.

[0011] The above monitoring RS can be transmitted through beams based on the inference results of the above AI model.

[0012] The transmission type of the above monitoring RS is associated with the transmission type of the RS for inference of the above AI model, and the transmission type of the above monitoring RS can be set to any one of periodic, semi-permanent, or non-periodic.

[0013] The first message above may further include first information configured to link the monitoring RS set to a periodic transmission type and the RS for inference of the AI ​​model set to a periodic type.

[0014] The above monitoring report message may include performance evaluation indicators of the AI ​​model configured based on monitoring RS values ​​measured by the same criteria as the quality indicators of the inference results of the AI ​​model.

[0015] The above monitoring report message includes measurement results of monitoring RS corresponding to the top K beams based on the quality of the beams among the multiple beams included in the inference result of the AI ​​model, and K may be a natural number greater than or equal to 1.

[0016] The above inference-related information includes one or more of the predicted value of beam quality in a future time instance, the beam quality metric in a future time instance, or the identifier for the top X beams available in a future time instance, wherein X may be a natural number greater than or equal to 1.

[0017] User equipment (UE) according to one embodiment of the present disclosure comprises at least one processor, wherein the at least one processor may cause the UE to: receive a first message from a base station comprising one or more monitoring reference signal (monitoring RS) transmission information and monitoring report setting information; measure one or more monitoring RS received from the base station based on the RS transmission information; select a monitoring RS for evaluating the performance of the AI ​​model from among one or more monitoring RSs having a temporal correlation with inference-related information by the UE-side artificial intelligence (AI) model; generate a monitoring report message using the measurement value of the selected monitoring RS and the inference result of the AI ​​model; and cause the monitoring report message to be transmitted to a base station through an uplink channel.

[0018] The selection of the above monitoring RS allows selecting the monitoring RS closest to the inference RS time point within preset time conditions and a specific RS transmission time point set in association with the inference RS or inference report used for inference.

[0019] The selected monitoring RS may be the monitoring RS closest to the inference target instance among the monitoring RSs existing within a preset time condition and a future inference target time instance determined based on the inference of the AI ​​model.

[0020] The above temporal association may include one or more of temporal proximity, whether a temporal offset is satisfied, or time conditions set at a base station.

[0021] The above monitoring RS can be transmitted through beams based on the inference results of the above AI model.

[0022] The transmission type of the above monitoring RS is associated with the transmission type of the RS for inference of the above AI model, and the transmission type of the above monitoring RS can be set to any one of periodic, semi-permanent, or non-periodic.

[0023] The first message above may further include first information configured to link the monitoring RS set to a periodic transmission type and the RS for inference of the AI ​​model set to a periodic type.

[0024] The above monitoring report message may include performance evaluation indicators of the AI ​​model configured based on monitoring RS values ​​measured by the same criteria as the quality indicators of the inference results of the AI ​​model.

[0025] The above monitoring report message includes measurement results of monitoring RS corresponding to the top K beams based on the quality of the beams among the multiple beams included in the inference result of the AI ​​model, and K may be a natural number greater than or equal to 1.

[0026] The above inference-related information includes one or more of the predicted value of beam quality in a future time instance, the beam quality metric in a future time instance, or the identifier for the top X beams available in a future time instance, wherein X may be a natural number greater than or equal to 1.

[0027] According to one embodiment of the present disclosure, a method and apparatus for monitoring the performance of an AI model in a network using an AI model may be provided. In particular, the present disclosure may provide transmission methods of RS for monitoring the performance of an AI model and reporting methods based on transmission methods of RS. Through this, the network may verify the performance of an AI model in operation and determine the deactivation and / or update of the AI ​​model based on the performance of the AI ​​model.

[0028] In addition, the present disclosure has the advantage of reducing the load caused by the transmission of monitoring report messages by ensuring that the monitoring report messages reported by the UE to the network do not include measurement information for all beams, but only report minimal information.

[0029] In addition, when monitoring reporting is set up in a trigger manner according to the present disclosure, the UE has the advantage of being able to reduce the transmission of unnecessary monitoring report messages when a trigger for monitoring does not occur.

[0030] FIG. 1 is a conceptual diagram illustrating an embodiment of a communication system.

[0031] FIG. 2 is a block diagram illustrating an example of a communication node constituting a communication system.

[0032] FIG. 3 is a conceptual diagram illustrating the configuration in which AI / ML is mounted on a base station and a UE to which the present disclosure applies, and the case in which beamforming is performed.

[0033] FIG. 4 is a flowchart illustrating a first embodiment in which a UE transmits a monitoring report to a base station when periodic RS monitoring is set by upper layer setting information.

[0034] FIG. 5 is a flowchart illustrating a second embodiment in which a UE transmits a monitoring report to a base station when periodic RS monitoring is set by upper layer setting information.

[0035] FIG. 6 is a flowchart illustrating a third embodiment in which a UE transmits a monitoring report to a base station when periodic RS monitoring is set by upper layer setting information.

[0036] Figure 7 is a conceptual diagram illustrating a case where monitoring RS is transmitted and measured from an instance that is the prediction target through a UE side model.

[0037] FIG. 8a is a conceptual diagram of a first embodiment of a transmission method of a monitoring instance in which an inference instance and a monitoring RS are transmitted.

[0038] FIG. 8b is a conceptual diagram of a second embodiment of a transmission method of a monitoring instance in which an inference instance and a monitoring RS are transmitted.

[0039] FIG. 8c is a conceptual diagram of a third embodiment of a transmission method of a monitoring instance in which an inference instance and a monitoring RS are transmitted.

[0040] Figure 9 is a flowchart illustrating the case where a UE transmits a monitoring report to a base station when non-periodic RS monitoring is set by upper layer setting information.

[0041] The present disclosure is capable of various modifications and may have various embodiments, and specific embodiments are illustrated in the drawings and described in detail. However, this is not intended to limit the present disclosure to specific embodiments, and it should be understood that it includes all modifications, equivalents, and substitutions that fall within the spirit and scope of the present disclosure.

[0042] Terms such as "first," "second," etc., may be used to describe various components, but said components should not be limited by said terms. Such terms are used solely for the purpose of distinguishing one component from another. For example, without departing from the scope of the present disclosure, the first component may be named the second component, and similarly, the second component may be named the first component. The term "and / or" includes a combination of a plurality of related described items or any of a plurality of related described items.

[0043] When it is stated that one component is "connected" or "connected" to another component, it should be understood that while it may be directly connected or connected to that other component, there may also be other components in between. On the other hand, when it is stated that one component is "directly connected" or "directly connected" to another component, it should be understood that there are no other components in between.

[0044] The terms used in this disclosure are used merely to describe specific embodiments and are not intended to limit this disclosure. The singular expression includes the plural expression unless the context clearly indicates otherwise. In this disclosure, terms such as “comprising” or “having” are intended to specify the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0045] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which this disclosure pertains. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this disclosure.

[0046] A communication system to which embodiments according to the present disclosure are applied will be described. The communication system to which embodiments according to the present disclosure are applied is not limited to the details described below, and embodiments according to the present disclosure may be applied to various communication systems. Here, the term "communication system" may be used interchangeably with "communication network."

[0047] Throughout the specification, a network may include, for example, wireless internet such as WiFi (wireless fidelity), mobile internet such as WiBro (wireless broadband internet) or WiMAX (world interoperability for microwave access), 2G mobile communication networks such as GSM (global system for mobile communication) or CDMA (code division multiple access), 3G mobile communication networks such as WCDMA (wideband code division multiple access) or CDMA2000, 3.5G mobile communication networks such as HSDPA (high speed downlink packet access) or HSUPA (high speed uplink packet access), 4G mobile communication networks such as LTE (long term evolution) networks or LTE-Advanced networks, and 5G mobile communication networks.

[0048] Throughout the specification, the term "terminal" may refer to a mobile station, mobile terminal, subscriber station, portable subscriber station, user equipment, access terminal, etc., and may include all or part of the functions of a terminal, mobile station, mobile terminal, subscriber station, portable subscriber station, user equipment, access terminal, etc.

[0049] Here, a desktop computer, laptop computer, tablet PC, wireless phone, mobile phone, smartphone, smart watch, smart glass, e-book reader, PMP (portable multimedia player), portable game console, navigation device, digital camera, DMB (digital multimedia broadcasting) player, digital audio recorder, digital audio player, digital picture recorder, digital picture player, digital video recorder, digital video player, etc., capable of communicating with a terminal can be used.

[0050] Throughout the specification, the term "base station" may refer to an access point, a radio access station, a node B, an evolved node B, a base transceiver station, a mobile multihop relay (MMR)-BS, etc., and may include all or part of the functions of a base station, access point, radio access station, node B, eNodeB, base transceiver station, MMR-BS, etc.

[0051] Hereinafter, preferred embodiments of the present disclosure will be described in more detail with reference to the attached drawings. In order to facilitate an overall understanding of the present disclosure, the same reference numerals are used for identical components in the drawings, and redundant descriptions of identical components are omitted.

[0052] FIG. 1 is a conceptual diagram illustrating an embodiment of a communication system.

[0053] Referring to FIG. 1, the communication system (100) may include a plurality of communication nodes (110-1, 110-2, 110-3, 120-1, 120-2, 130-1, 130-2, 130-3, 130-4, 130-5, 130-6). The plurality of communication nodes may support 4G communication (e.g., LTE (long term evolution), LTE-A (advanced)), 5G communication (e.g., NR (new radio)), etc., as defined in the 3GPP (3rd generation partnership project) standard. 4G communication may be performed in a frequency band of 6 GHz or lower, and 5G communication may be performed not only in a frequency band of 6 GHz or lower but also in a frequency band of 6 GHz or higher.

[0054] For example, for 4G communication and 5G communication, multiple communication nodes can support communication protocols based on CDMA (code division multiple access), WCDMA (wideband CDMA), TDMA (time division multiple access), FDMA (frequency division multiple access), OFDM (orthogonal frequency division multiplexing), Filtered OFDM, CP (cyclic prefix)-OFDM, DFT-s-OFDM (discrete Fourier transform-spread-OFDM), OFDMA (orthogonal frequency division multiple access), SC (single carrier)-FDMA, NOMA (Non-orthogonal Multiple Access), GFDM (generalized frequency division multiplexing), FBMC (filter bank multi-carrier) based communication protocol, UFMC (universal filtered multi-carrier) based communication protocol, SDMA (Space Division Multiple Access) based communication protocol, etc.

[0055] Additionally, the communication system (100) may further include a core network. If the communication system (100) supports 4G communication, the core network may include an S-GW (serving-gateway), a P-GW (PDN (packet data network)-gateway), an MME (mobility management entity), etc. If the communication system (100) supports 5G communication, the core network may include a UPF (user plane function), an SMF (session management function), an AMF (access and mobility management function), etc.

[0056] Meanwhile, each of the plurality of communication nodes (110-1, 110-2, 110-3, 120-1, 120-2, 130-1, 130-2, 130-3, 130-4, 130-5, 130-6) constituting the communication system (100) may have the following structure.

[0057] FIG. 2 is a block diagram illustrating an example of a communication node constituting a communication system.

[0058] Referring to FIG. 2, the communication node (200) may include at least one processor (210), a memory (220), and a transceiver (230) that is connected to a network to perform communication. Additionally, the communication node (200) may further include an input interface device (240), an output interface device (250), a storage device (260), etc. Each component included in the communication node (200) may be connected by a bus (270) to communicate with one another.

[0059] However, each component included in the communication node (200) may be connected via individual interfaces or individual buses centered around the processor (210), rather than via a common bus (270). For example, the processor (210) may be connected via a dedicated interface to at least one of a memory (220), a transmission / reception device (230), an input interface device (240), an output interface device (250), and a storage device (260).

[0060] The processor (210) can execute a program command stored in at least one of the memory (220) and the storage device (260). The processor (210) may mean a central processing unit (CPU), a graphics processing unit (GPU), or a dedicated processor on which the methods according to embodiments of the present disclosure are performed. Each of the memory (220) and the storage device (260) may be composed of at least one of a volatile storage medium and a non-volatile storage medium. For example, the memory (220) may be composed of at least one of read-only memory (ROM) and random access memory (RAM).

[0061] Referring again to FIG. 1, the communication system (100) may include a plurality of base stations (110-1, 110-2, 110-3, 120-1, 120-2) and a plurality of terminals (130-1, 130-2, 130-3, 130-4, 130-5, 130-6). The communication system (100) including the base stations (110-1, 110-2, 110-3, 120-1, 120-2) and terminals (130-1, 130-2, 130-3, 130-4, 130-5, 130-6) may be referred to as an "access network". Each of the first base station (110-1), the second base station (110-2), and the third base station (110-3) can form a macro cell. Each of the fourth base station (120-1) and the fifth base station (120-2) can form a small cell. The fourth base station (120-1), the third terminal (130-3), and the fourth terminal (130-4) may be located within the cell coverage of the first base station (110-1). The second terminal (130-2), the fourth terminal (130-4), and the fifth terminal (130-5) may be located within the cell coverage of the second base station (110-2). The fifth base station (120-2), the fourth terminal (130-4), the fifth terminal (130-5), and the sixth terminal (130-6) may be located within the cell coverage of the third base station (110-3). The first terminal (130-1) may be located within the cell coverage of the fourth base station (120-1). The sixth terminal (130-6) may be located within the cell coverage of the fifth base station (120-2).

[0062] Here, each of the multiple base stations (110-1, 110-2, 110-3, 120-1, 120-2) may be referred to as Node B, evolved Node B, base transceiver station (BTS), radio base station, radio transceiver, access point, access node, road side unit (RSU), radio remote head (RRH), transmission point (TP), transmission and reception point (TRP), eNB, gNB, etc.

[0063] Each of the multiple terminals (130-1, 130-2, 130-3, 130-4, 130-5, 130-6) may be referred to as a UE (user equipment), terminal, access terminal, mobile terminal, station, subscriber station, mobile station, portable subscriber station, node, device, IoT (Internet of Thing) device, mounted module / device / terminal or on board device / terminal, etc.

[0064] Meanwhile, each of the multiple base stations (110-1, 110-2, 110-3, 120-1, 120-2) may operate in different frequency bands or in the same frequency band. Each of the multiple base stations (110-1, 110-2, 110-3, 120-1, 120-2) may be connected to each other via an ideal backhaul link or a non-ideal backhaul link, and may exchange information with each other via an ideal backhaul link or a non-ideal backhaul link. Each of the multiple base stations (110-1, 110-2, 110-3, 120-1, 120-2) may be connected to a core network via an ideal backhaul link or a non-ideal backhaul link. Each of the multiple base stations (110-1, 110-2, 110-3, 120-1, 120-2) can transmit a signal received from the core network to the corresponding terminal (130-1, 130-2, 130-3, 130-4, 130-5, 130-6), and can transmit a signal received from the corresponding terminal (130-1, 130-2, 130-3, 130-4, 130-5, 130-6) to the core network.

[0065] In addition, each of the multiple base stations (110-1, 110-2, 110-3, 120-1, 120-2) can support MIMO transmission (e.g., SU (single user)-MIMO, MU (multi user)-MIMO, massive MIMO, etc.), CoMP (coordinated multipoint) transmission, CA (carrier aggregation) transmission, transmission in an unlicensed band, device-to-device communication (D2D) (or ProSe (proximity services)), etc. Here, each of the plurality of terminals (130-1, 130-2, 130-3, 130-4, 130-5, 130-6) can perform an operation corresponding to the base station (110-1, 110-2, 110-3, 120-1, 120-2) and an operation supported by the base station (110-1, 110-2, 110-3, 120-1, 120-2). For example, the second base station (110-2) can transmit a signal to the fourth terminal (130-4) based on the SU-MIMO method, and the fourth terminal (130-4) can receive a signal from the second base station (110-2) based on the SU-MIMO method. Alternatively, the second base station (110-2) can transmit a signal to the fourth terminal (130-4) and the fifth terminal (130-5) based on the MU-MIMO method, and each of the fourth terminal (130-4) and the fifth terminal (130-5) can receive a signal from the second base station (110-2) by the MU-MIMO method.

[0066] Each of the first base station (110-1), the second base station (110-2), and the third base station (110-3) can transmit a signal to the fourth terminal (130-4) based on the CoMP method, and the fourth terminal (130-4) can receive a signal from the first base station (110-1), the second base station (110-2), and the third base station (110-3) by the CoMP method. Each of the multiple base stations (110-1, 110-2, 110-3, 120-1, 120-2) can transmit and receive signals based on the CA method with terminals (130-1, 130-2, 130-3, 130-4, 130-5, 130-6) within its cell coverage area. Each of the first base station (110-1), the second base station (110-2), and the third base station (110-3) can control D2D between the fourth terminal (130-4) and the fifth terminal (130-5), and each of the fourth terminal (130-4) and the fifth terminal (130-5) can perform D2D by controlling each of the second base station (110-2) and the third base station (110-3).

[0067] Next, methods for configuring and managing wireless interfaces in a communication system will be described. Even when a method performed by a first communication node among the communication nodes (e.g., transmission or reception of a signal) is described, the corresponding second communication node may perform a method corresponding to the method performed by the first communication node (e.g., reception or transmission of a signal). That is, when the operation of a terminal is described, the corresponding base station may perform an operation corresponding to the operation of the terminal. Conversely, when the operation of a base station is described, the corresponding terminal may perform an operation corresponding to the operation of the base station.

[0068] Meanwhile, in a communication system, a base station can perform all functions of the communication protocol (e.g., remote radio transmission and reception functions, baseband processing functions). Alternatively, among all functions of the communication protocol, the remote radio transmission and reception function may be performed by a TRP (transmission reception point) (e.g., f(flexible)-TRP), and among all functions of the communication protocol, the baseband processing function may be performed by a BBU (baseband unit) block. The TRP may be an RRH (remote radio head), RU (radio unit), TP (transmission point), etc. A BBU block may include at least one BBU or at least one DU (digital unit). A BBU block may be referred to as a "BBU pool," "centralized BBU," etc. A TRP may be connected to a BBU block via a wired fronthaul link or a wireless fronthaul link. A communication system composed of backhaul links and fronthaul links may be as follows. When the function split method of the communication protocol is applied, the TRP can selectively perform some functions of the BBU or some functions of MAC (medium access control) / RLC (radio link control).

[0069] In the present disclosure, a phrase containing "~ case (e.g., when ~)" may be expressed as a phrase containing "~ based on (e.g., based on ~)" or a phrase containing "~ in response to (e.g., in response to ~)". In other words, a phrase containing "~ case" may be interpreted as identical or similar to a phrase containing "~ based on" or a phrase containing "~ in response to".

[0070] In the following, a method and apparatus for verifying the performance regarding the accuracy of inference results of artificial intelligence (AI) / machine learning (ML) are described.

[0071] 3GPP is proposing methods for applying AI / ML models to various fields. Among the fields proposed by 3GPP, the following description assumes the use of AI / ML models in beam management (BM) procedures. Although this disclosure assumes that the inference of an AI / ML model is applied to a BM procedure, this is merely an example to aid understanding, and this disclosure should not be understood as being limited to BM procedures.

[0072] AI / ML technology may be considered in terms of single-sided models, where the AI / ML model is deployed on only one side of the network or only on the UE, and two-sided models, where the model is deployed on both the network and the UE. In the disclosure described below, the AI / ML model is assumed to be a model for beam management unless otherwise specifically stated. For convenience of explanation, the network is assumed to be a base station. Furthermore, in the following description, the AI / ML model deployed on the base station will be referred to as the network-side model (NW-side model), and the AI / ML model deployed on the UE will be referred to as the UE-side model (UE-side model).

[0073] FIG. 3 is a conceptual diagram illustrating the configuration in which AI / ML is mounted on a base station and a UE to which the present disclosure applies, and the case in which beamforming is performed.

[0074] The base station (310) may be equipped with an NW side model (311), which is an AI / ML model for beam management, and the UE (320) may be equipped with a UE side model (321), which is an AI / ML model for beam management. The base station (310) may be configured in whole or in part of the configuration of the communication node described above in FIG. 2. If the base station (310) has the configuration of FIG. 2, the NW side model (311), which is an AI / ML model for beam management, may be stored in the memory (220) or storage device (260) of FIG. 2 and may be executed by the processor (210). The base station (310) may also have additional configurations in addition to the configuration described in FIG. 2. For example, the base station (310) may further include an interface for connection with the core network and / or an interface for connection with an adjacent base station. In addition, when the base station (310) is configured in a functionally divided form, the transmission and reception device (230) within the base station (310) may be composed of a plurality of transmission and reception points (TRP) or a plurality of radio units (RU) and / or distributed units (DU).

[0075] The UE (320) may be composed of all or part of the configuration of the communication node described above in FIG. 2. If the UE (320) has the configuration of FIG. 2, the UE side model (321), which is an AI / ML model for beam management, may be stored in the memory (220) or storage device (260) of FIG. 2 and may be executed by the processor (210). The UE (320) may also have additional configurations in addition to the configuration described in FIG. 2. The UE (320) may further include a camera module, a speaker, a microphone and / or various sensors. In addition, the UE (320) may further include various modules for user convenience.

[0076] In FIG. 3, the case where both the base station (310) and the UE (320) have AI / ML models is illustrated, but only the base station (310) may have an AI / ML model, or only the UE (320) may have an AI / ML model. As another example, even if both the base station (310) and the UE (320) have AI / ML models, only the NW side model (311) may perform inference, or only the UE side model (321) may perform inference.

[0077] As illustrated in FIG. 3, the base station (310) can form a plurality of transmission beams (301, 302, 303, 304, 305, 306, 307, 308) and can transmit a signal through the plurality of transmission beams (301-308). For example, the base station (310) can transmit an RS, such as an SSB and / or a CSI-RS, through the plurality of transmission beams (301-308). The following description assumes that the RS is a CSI-RS. When the base station (310) transmits a CSI-RS through the plurality of transmission beams (301-308), the base station (310) can transmit by sweeping each of the transmission beams. When beam sweeping, the base station (310) can transmit CSI-RS through the first transmission beam (301) at the first time point, transmit CSI-RS through the second transmission beam (302) at the second time point, and transmit CSI-RS through the third transmission beam (303) at the third time point. In this way, the base station (310) can transmit CSI-RS through the eighth transmission beam (308) at the eighth time point. In the example of FIG. 3, the case where the number of transmission beams (301-308) of the base station (310) is eight is illustrated, but this is merely an example to aid understanding and should not be interpreted as being limited thereto.

[0078] The UE (320) can receive one or more of the transmission beam(s) among the multiple transmission beams (301-308) transmitted by the base station (310). When the UE (320) receives one or more transmission beam(s), it can receive the transmission beam(s) transmitted by the base station (310) using one or more of the multiple reception beam(s) (331, 332, 333). The UE (320) can receive the transmission beams (301-308) using each of the multiple reception beams (331-333). For example, the UE (320) can receive transmission beams (301-308) transmitted by the base station (310) using the first receiving beam (331), then receive transmission beams (301-308) transmitted by the base station (310) using the second receiving beam (332), and finally receive transmission beams (301-038) transmitted by the base station (310) using the third receiving beam (333). In the example of FIG. 3, the case where the number of receiving beams (331-333) of the UE (320) is three is illustrated, but this is merely an example to aid understanding and should not be interpreted as being limited thereto.

[0079] First, a case in which inference is performed in the NW side model (311) of the base station (310) is described. The UE (320) can receive CSI-RS transmitted through each of the transmission beams (301-308) using each of the receiving beams (331-333), and can measure the layer 1 (L1) reference signal received power (RSRP) (L1-RSRP) for each of the received CSI-RS. The L1-RSRP measured for the CSI-RS is one example and is not limited thereto. For example, the UE (320) may measure the CSI-RS to obtain channel state information (CSI) and report it to the base station (310), or obtain channel quality information (CQI) and report it to the base station (310). At this time, CSI and CQI, etc., may be measurement values ​​for each of the beams formed by the base station (310) to the UE (320). There may be various forms of methods to check the state of the channel, that is, the beam, through RS measurement, in addition to L1-RSRP, CSI, and CQI. It should be noted that the present disclosure may be applicable to all forms capable of measuring and reporting the state of the beam.

[0080] The UE (320) can generate a beam measurement report message based on the beam measurement report settings received from the base station (310) and can transmit the generated beam measurement report message to the base station (310). The NW side model (311) of the base station (310) can predict a high-quality beam that the base station (310) can use when transmitting a signal to the UE (320) at a specific point in time (e.g., the next RS transmission time, the data transmission time) using the beam measurement report message received from the UE (320). For example, if the base station (310) performs a beam prediction for the next CSI-RS transmission time, the base station may possess a beam-specific quality value that can be obtained when the UE (320) measures each beam through reception at the predicted CSI-RS transmission time. The base station (310) can transmit CSI-RS to the UE (320) at the predicted CSI-RS transmission time. The UE (320) can receive the CSI-RS received from the base station (310), measure it, and then send a beam measurement report message to the base station (310). The base station (310) can monitor the inference results of the NW side model (311) of the base station (310) by comparing the previously predicted value with the value included in the measurement report message reported by the UE (320).

[0081] Next, the case where inference is performed in the UE side model (321) of the UE (320) is described. When the UE (320) uses an AI / ML model, the base station (310) may configure and / or indicate that measurements for a specific beam set be performed so that the inference results of the UE side model (321) are reported to the base station (310). The UE (320) may generate an inference report message containing the inference results based on the method configured (or indicated) by the base station (310) and may transmit the generated inference report message to the base station (310). The inference results of the UE side model (321) may be predictive information regarding the beam(s) at a specific future point in time(s) that are of good quality within another beam set or available between the base station (310) and the UE (320). The predictive information regarding the beam(s) may be predicted L1-RSRP values.

[0082] Predictive information for the beam(s) may vary depending on beam management case 1 (BM case 1) and beam management case 2 (BM case 2).

[0083] The UE side model (321) used in BM case 1 may be an AI / ML model that takes the measurement results of beam set B as input and predicts the top K beam(s) with good quality in the spatial domain for a single time instance, and the quality corresponding to those beam(s). In other words, the UE side model (321) used in BM case 1 may take the measurement results of beam set B as input, perform inference, and output an inference result. The inference result of the UE side model (321) used in BM case 1 may be information on the top K beam(s) selected based on quality in beam set A corresponding to the beams to be predicted. Here, K may be a natural number greater than or equal to 1.

[0084] The UE side model (321) used in BM case 2 may be an AI / ML model that takes past measurement results for the beams to be measured in beam set B as input and predicts the top M beam(s) with good quality and the quality corresponding to those beam(s) for multiple future time instances. In other words, the UE side model (321) used in BM case 2 may infer using past measurement results of beam set B as input and output an inference result. The inference result of the UE side model (321) used in BM case 2 may be information on the top M beam(s) selected based on quality in beam set A corresponding to the beams to be predicted. Here, M may be a natural number greater than or equal to 1.

[0085] In the present disclosure described below, the beam set measured to be used as an input value in BM case 1 and BM case 2 is referred to as "beam set B" in the present disclosure, and the beam set used for the inference result is referred to as "beam set A".

[0086] The base station (310) can set configuration information for a reference signal (RS) resource or RS resource set for the transmission of beam set A and beam set B to the UE (320) through an information element (IE) of the upper layer signaling message "CSI-ReportConfig". Additionally, the base station (310) can set configuration information for an RS resource or RS resource set for monitoring to the UE (320) through CSI-ReportConfig. The RS resource set can be set by the RS resource setting IE. RS resource configuration information elements may be, for example, "CSI-RS-ConfigIE" and "SSB-ConfigIE". Within CSI-ReportConfig, RS resources (e.g., CSI-RS resources) or RS resource sets (e.g., CSI-RS resource sets) for beam set A and beam set B, and RS resources (e.g., CSI-RS resources for monitoring) or RS resource sets (e.g., CSI-RS resource sets for monitoring) for monitoring may be configured in association with specific associated ID values.

[0087] CSI-RS can be configured and operated for the transmission of Beam Set B, and CSI-RS can be configured and operated as periodic, semi-persistent, aperioditic, or event-triggered. The configuration of CSI-RS for the transmission of Beam Set B can be set in CSI-ReportConfig, which is a higher-level signaling message (e.g., a radio resource control (RRC) signaling message).

[0088] The RS resources for Beam Set A and Beam Set B, and the CSI-RS resources for monitoring, may be configured by the same RRC signaling message or by separate RRC signaling messages. If the CSI-RS resources for Beam Set A and Beam Set B, and the CSI-RS resources for monitoring, are configured by the same RRC signaling message, they may be configured in the same IE. If the CSI-RS resources for Set A and Beam Set B, and the CSI-RS resources for monitoring, are configured by different RRC signaling messages, they may be configured in different IEs.

[0089] In the present disclosure, the CSI-ReportConfigIE for monitoring may include one or more of the following information.

[0090] 1) Types and amounts of information to be reported as a result of monitoring,

[0091] 2) RS resource information for monitoring, or

[0092] 3) Uplink (UL) resource information to be used for monitoring result reporting

[0093] In the configuration information, periodic, semi-permanent, or non-periodic RS resources for monitoring can be configured.

[0094] In the case of beam set A, since the UE (320) is not measuring the beam for inference, for reporting inference results, each beam in beam set A may be operated by setting only the beam identifier (ID) information for identifying each beam. For example, if the number of beams in beam set A is 8, the beam identifier for identifying each beam in beam set A inferred through measurement of beam set B may be indicated as 3 bits. Therefore, when setting the RRC related to beam set A, the number of beams in beam set A may be set and operated. As another example, the CSI-RS for beam set A may be set, and the UE may indicate specific beams in beam set A by indicating the CSI-RS resource (e.g., CSI-RS Resource Indicator, CRI) when reporting inference results based on the resource information.

[0095] For convenience of explanation, in this disclosure, RSs transmitted for inference (e.g., CSI-RS) are referred to as "RS for inference," "CSI-RS for inference," "RS for beam measurement," or "CSI-RS for beam measurement." Information or signaling reported by the UE to the base station based on the measurement results of the CSI-RS for inference is referred to as an "inference report message." RSs transmitted for monitoring (e.g., CSI-RS) are referred to as "monitoring RS" or "monitoring CSI-RS." Information or signaling reported by the UE to the base station based on the measurement results of the monitoring CSI-RS is referred to as a "monitoring report message." Additionally, for convenience of explanation, this disclosure will describe monitoring RS based on CSI-RS. However, this is merely to aid in understanding this disclosure and should not be understood as limiting monitoring RS to CSI-RS. For monitoring RS, other RSs besides CSI-RS, such as a synchronization signal block (SSB), may be used.

[0096] First, the inference through the AI / ML model, the reporting of inference results, the monitoring operation, and the monitoring RS mentioned in the first to third embodiments described below will be explained.

[0097] The base station (310) can transmit RSs for beam measurement to the UE (320) through each beam. The RSs may be CSI-RSs transmitted through beams included in beam set B. The UE (320) can receive CSI-RSs from the base station (310) through each of the multiple beams, and can obtain L1-RSRP values ​​for each of the CSI-RSs by measuring each of the CSI-RSs received through each beam. The UE (320) can infer the L1-RSRP values ​​using them as input to the UE side model (321). The inference result may be the L1-RSRP value at a future point in time for the beam(s) belonging to beam set A. The UE (320) can transmit an inference report message containing the inference result to the base station (310).

[0098] The base station (310) may receive an inference report message. Based on the inference report message, the base station (310) may communicate with the UE (320) through the beams included in beam set A. In this situation, a procedure may be required to verify whether the beams determined based on the inference of the UE side model (321) are appropriate for communication. Alternatively, a procedure may be required to verify whether the performance of the AI / ML model is appropriate. In the present disclosure, monitoring RS may refer to CSI-RS transmitted by the base station (310) to the UE (320) through the selected beam(s) to verify whether the beam(s) selected by the UE side model (321) through inference among the beams included in beam set A are the correct beam(s).

[0099] The NW side model (311) included in the base station (310) can also be understood in the same form as above.

[0100] [First Embodiment: Monitoring method based on periodic monitoring RS]

[0101] FIG. 4 is a flowchart illustrating a first embodiment in which a UE transmits a monitoring report to a base station when periodic RS monitoring is set by upper layer setting information.

[0102] The base station and UE exemplified in FIG. 4 may be the base station and UE described earlier in FIG. 3. In FIG. 4, the base station (310) and UE (320) may both have AI / ML models. Additionally, the base station and UE exemplified in FIG. 4 may include all or part of the configuration of the communication node described earlier in FIG. 2.

[0103] In step S400, the base station (310) can transmit high layer configuration information to the UE (320). The high layer configuration information may be transmitted, for example, as an RRC signaling message. In the following description, it is assumed that the high layer configuration information is included in the RRC signaling message. The RRC signaling message may include configuration information for a monitoring RS (e.g., CSI-RS) to be monitored by the UE (320) and monitoring report configuration information. The configuration information for the monitoring RS and the monitoring report configuration information may be transmitted via the CSI-ReportConfigIE of the RRC signaling message.

[0104] The configuration information for the monitoring CSI-RS may be configured such that the method of transmission of the CSI-RS is, for example, periodic, semi-permanent, or non-periodic. Since FIG. 4 assumes that the CSI-RS is transmitted periodically, the configuration information for the monitoring CSI-RS to be monitored by the UE (320) may be configured to be periodic. Additionally, the configuration information for the monitoring CSI-RS may further include resource mapping information for the transmission of the CSI-RS (e.g., time and frequency density, subcarrier location, etc.) and transmission period information of the CSI-RS (e.g., repetition period and transmission time information).

[0105] Monitoring report configuration information may include a report type (e.g., periodic report, semi-permanent report, non-periodic report, trigger-based report) and one or more of the three pieces of information described above for CSI-ReportConfigIE. In the embodiment of FIG. 4, the report type included in the monitoring report configuration information may be set to periodic. When the report type included in the monitoring report configuration information is set to periodic, the monitoring report configuration information may further include report resource information. The report resource information may refer, for example, to uplink channel resources.

[0106] In step S400, the UE (320) can receive an RRC signaling message. Based on the received RRC signaling message, the UE (320) can check the configuration information for the RS to be monitored and the monitoring report configuration information.

[0107] In step S410, the base station (310) can transmit the monitoring CSI-RS to the UE (320) at the time of transmission of the periodic monitoring CSI-RS. The periodic monitoring CSI-RS can be transmitted, for example, through each of the beams included in beam set A set by the RRC signaling message. Or it can be transmitted through the beams associated with beam set A set by the RRC signaling message. The following description will be based on beam set A. The time of transmission and transmission resources (e.g., time resources, frequency resources, and time / frequency density, etc.) of the periodic monitoring CSI-RS can be transmitted based on the configuration information for the monitoring CSI-RS included in the RRC signaling message. The configuration information for the monitoring CSI-RS may indicate specific resources in other IEs of the RRC signaling message, for example, CSI-RS-ResourceConfigIE or CSI-RS-ResourceSetConfigIE.

[0108] In step S410, the UE (320) can receive monitoring CSI-RS transmitted by the base station (310) based on the RRC signaling message received in step S400.

[0109] In step S420, the UE (320) can measure the monitoring CSI-RS transmitted by the base station (310) and generate a monitoring report message. The monitoring report message can be generated based on monitoring report setting information. The generation of the monitoring report message is described in more detail in the reporting method of BM case 1 and the reporting method of BM case 2, which will be described below.

[0110] In step S430, the UE (320) can transmit the generated monitoring report message to the base station via the uplink channel (e.g., physical uplink control channel (PUCCH) or physical uplink shared channel (PUSCH)) assigned from the monitoring report setting information received in step S400. In step S430, the base station (310) can receive the monitoring report message from the UE (320) via the uplink channel. The base station (310) can verify the performance of the AI / ML model based on the monitoring report message received from the UE (320).

[0111] After the time of the RS transmission cycle (401) has elapsed following the monitoring CSI-RS transmission in step S410, the base station (310) can transmit the monitoring CSI-RS to the UE (320) in step S440. In step S410, the UE (320) can receive the monitoring CSI-RS from the base station (310).

[0112] Steps S450 and S460 correspond to the previously described Steps S420 and S430, respectively, so a redundant explanation is omitted.

[0113] As described above, the base station (310) can cause the UE (320) to receive periodic monitoring CSI-RS transmitted by the base station (310) even without giving special monitoring instructions to the UE (320) based on upper layer configuration information. Additionally, the base station (310) can cause the UE (320) to report monitoring report messages regarding the periodic monitoring CSI-RS transmitted by the base station (310) to the base station (310) even without giving special reporting instructions to the UE (320) based on upper layer configuration information.

[0114] FIG. 5 is a flowchart illustrating a second embodiment in which a UE receives a monitoring report instruction message from a base station when periodic RS monitoring is set by upper layer setting information.

[0115] The base station and UE exemplified in FIG. 5 may be the base station and UE described earlier in FIG. 3. In FIG. 5, the base station (310) and the UE (320) may both have AI / ML models. The base station and the UE exemplified in FIG. 5 may include all or part of the configuration of the communication node described earlier in FIG. 2.

[0116] In step S500, the base station (310) can transmit upper layer configuration information to the UE (320). The upper layer configuration information can be configured and transmitted as an RRC signaling message as previously described. The RRC signaling message may include configuration information for the monitoring CSI-RS to be monitored by the UE (320) and monitoring report configuration information. The configuration information for the monitoring CSI-RS and the monitoring report configuration information can be transmitted via the CSI-ReportConfigIE of the RRC signaling message.

[0117] The configuration information for the monitoring CSI-RS may be configured such that the method of transmission of the CSI-RS is, for example, periodic, semi-permanent, or non-periodic. Since FIG. 5 assumes that the CSI-RS is transmitted periodically, the configuration information for the monitoring CSI-RS to be monitored by the UE (320) may be configured to be periodic. Additionally, the configuration information for the monitoring CSI-RS may further include resource mapping information for the transmission of the CSI-RS (e.g., time and frequency density, subcarrier location, etc.) and transmission period information of the CSI-RS (e.g., repetition period and transmission time information).

[0118] Monitoring report configuration information may include a report type (e.g., periodic report, semi-permanent report, non-permanent report, trigger-based report) and one or more of the three pieces of information described above for CSI-ReportConfigIE. An example of FIG. 5 may be a case where the monitoring report configuration information is set to semi-permanent report or non-permanent report.

[0119] In step S500, the UE (320) can receive an RRC signaling message. Based on the received RRC signaling message, the UE (320) can check the configuration information for the monitoring CSI-RS and the monitoring report configuration information.

[0120] In step S510, the base station (310) can transmit the monitoring CSI-RS to the UE (320) at the time of transmission of the periodic monitoring CSI-RS. The periodic monitoring CSI-RS can be transmitted, for example, through each of the beams included in beam set A set by the RRC signaling message. The time of transmission and the transmission resources (e.g., time resources, frequency resources, and time / frequency density, etc.) of the periodic monitoring CSI-RS can be transmitted based on the configuration information for the monitoring RS included in the RRC signaling message. As described in FIG. 4, the configuration information for the monitoring RS can indicate specific resources of other IEs of the RRC signaling message, for example, CSI-RS-ResourceConfigIE or CSI-RS-ResourceSetConfigIE.

[0121] In step S510, the UE (320) can receive the monitoring CSI-RS transmitted by the base station (310) based on the RRC signaling message received in step S500. However, the UE (320) may not measure the received monitoring CSI-RS or generate a monitoring report message. As another example, the UE (320) may measure the received monitoring CSI-RS but not generate a monitoring report message. This is because the monitoring report is not set as a periodic report in the monitoring report setting information received in step S500.

[0122] In step S520, the base station (310) can transmit a monitoring report instruction message to the UE (320). The report instruction message may consist only of a MAC-CE signaling message, may consist only of a DCI, or may use both MAC-CE and DCI.

[0123] A case in which the report instruction message consists solely of MAC-CE signaling messages is described first. In the case where the report instruction message consists solely of MAC-CE signaling messages, for example, when the monitoring report setting information has the monitoring CSI-RS report semi-permanently set, the base station (310) may generate a MAC-CE signaling message containing information instructing the activation of the monitoring report message. If the monitoring report setting information has an uplink channel set (or allocated) for the transmission of the monitoring report message, the MAC-CE signaling message may not include uplink channel information for the transmission of the monitoring report message. On the other hand, if the monitoring report setting information has no uplink channel set (or allocated) for the transmission of the monitoring report message, the MAC-CE signaling message may include uplink channel information for the transmission of the monitoring report message.

[0124] Next, a case in which the report instruction message consists solely of a DCI is described. A case in which the report instruction message consists solely of a DCI may be, for example, when the monitoring report setting information is configured to have the monitoring CSI-RS report non-periodically. When the monitoring CSI-RS report is configured to have the monitoring CSI-RS report non-periodically, the base station (310) may instruct the UE (320) to transmit the monitoring CSI-RS report to the DCI whenever the monitoring CSI-RS report is needed. Additionally, the DCI may instruct the uplink resources for transmitting the monitoring report message.

[0125] Next, the report instruction message may be transmitted by the MAC-CE signaling message and the DCI. This may be the case where one or more beam sets A for monitoring CSI-RS are configured in the RRC signaling message. The MAC-CE signaling message may include information instructing the activation of one beam set A. Alternatively, the MAC-CE signaling message may include information instructing the beam set A for monitoring among two or more beam sets A. Additionally, the MAC-CE signaling message may instruct the activation of monitoring reports. Furthermore, the DCI may include uplink channel allocation information to which the monitoring report message will be transmitted. If the RRC signaling message includes uplink channel allocation information to which the monitoring report message will be transmitted, and the DCI includes different uplink channel allocation information, the UE (320) may use the uplink channel with the higher priority of the two. For example, if the uplink channel allocation information included in the DCI has a higher priority, the UE (320) may transmit the monitoring report message based on the uplink channel allocation information included in the DCI when transmitting the monitoring report message thereafter.

[0126] The method of using MAC-CE and DCI described above is one example to aid in understanding the present disclosure. Accordingly, in addition to the method described above, information indicated through MAC-CE and DCI may be indicated and operated through various combinations of methods.

[0127] In step S520, the UE (320) may receive a monitoring report instruction message. As described above, the monitoring report instruction message may consist only of MAC-CE, may consist of MAC-CE and DCI, or may consist only of DCI. It should be noted that in FIG. 5, these are collectively referred to as monitoring report instruction messages.

[0128] In step S530, the base station (310) can transmit monitoring CSI-RS to the UE (320) based on the monitoring CSI-RS transmission cycle (501). In step S530, the UE (320) can receive the monitoring CSI-RS transmitted by the base station.

[0129] In step S540, the UE (320) can measure the received monitoring CSI-RS. The reason the UE (320) measures the received monitoring CSI-RS in step S540 may be because it received a monitoring report instruction message from the base station (310) in step S520.

[0130] In step S540, the UE (320) can generate a monitoring report message to be transmitted to the base station (310) based on the specific CSI-RS transmitted in the step prior to S540 and the specific CSI-RS measurement information to be transmitted in the step after S540. Here, the specific CSI-RS refers to valid CSI-RSs to be used for monitoring reporting. The generation of the monitoring report message is described in more detail in the reporting method of BM case 1 and the reporting method of BM case 2, which will be described below.

[0131] In step S550, the UE (320) can transmit the generated monitoring report message to the base station (310) via an uplink channel (e.g., PUCCH or PUSCH) assigned by the monitoring report setting information received in step S500, or assigned by the MAC-CE signaling message, or assigned by the DCI. In step S550, the base station (310) can receive the monitoring report message from the UE (320) via the uplink channel. The base station (310) can verify the performance of the AI / ML model based on the monitoring report message received from the UE (320).

[0132] After the time of the RS transmission cycle (501) has elapsed following the monitoring CSI-RS transmission in step S530, the base station (310) can transmit the monitoring CSI-RS to the UE (320) in step S560. In step S560, the UE (320) can receive the monitoring CSI-RS from the base station (310).

[0133] Meanwhile, in FIG. 5, the UE can store the measurement value for the periodic monitoring CSI-RS, which is transmitted prior to receiving the monitoring report instruction message, for a preset time. The preset time can be set by the upper layer signaling.

[0134] As another example, the UE may store only the most recent CSI-RS measurement value. Explaining this with reference to FIG. 5, the UE may measure the periodic monitoring CSI-RS received from the base station in step S510 and store the measurement value of the monitoring CSI-RS. At the time the UE receives the monitoring report instruction message from the base station in step S520, the stored monitoring CSI-RS measurement value may be the CSI-RS received in step S510. After the UE receives the monitoring report instruction message in step S520, if it receives the monitoring CSI-RS transmitted from the base station in step S530, it may measure the received monitoring CSI-RS. Then, the UE may store the measurement value of the CSI-RS. At this time, the measurement value of the CSI-RS measured in step S510 may be deleted or updated with the value received and measured in step S530. In other words, the CSI-RS measurement value may be stored only for the period (501) of the periodic monitoring CSI-RS.

[0135] As another example, the UE can store measurement(s) for the monitored CSI-RS that are within a specific time offset relative to the CSI reference resource time associated with a specific inference reporting time. The specific time offset can be set in units of symbols or slots. The specific time offset can be determined based on a value set by the upper-level signaling.

[0136] As another example, in the case of BM case 2, the UE can store measurements for the monitored CSI-RS within a specific time offset relative to the time of the monitored instance among multiple future time instances that are the subject of the prediction.

[0137] FIG. 6 is a flowchart illustrating a third embodiment in which a UE transmits a monitoring report to a base station when periodic RS monitoring is set by upper layer setting information.

[0138] The base station and UE exemplified in FIG. 6 may be the base station and UE described earlier in FIG. 3. In FIG. 6, the base station (310) and UE (320) may both have AI / ML models. The base station and UE exemplified in FIG. 6 may include all or part of the configuration of the communication node described earlier in FIG. 2.

[0139] In step S600, the base station (310) can transmit upper layer configuration information to the UE (320). The upper layer configuration information can be transmitted as an RRC signaling message as previously described. The RRC signaling message may include configuration information for a monitoring RS (e.g., CSI-RS) to be monitored by the UE (320) and monitoring report configuration information. The configuration information for the monitoring CSI-RS and the monitoring report configuration information can be transmitted via the CSI-ReportConfigIE of the RRC signaling message.

[0140] The configuration information for the monitoring CSI-RS may be configured such that the method of transmission of the CSI-RS is, for example, periodic, semi-permanent, or non-periodic. Since FIG. 6 assumes that the CSI-RS is transmitted periodically, the configuration information for the monitoring CSI-RS to be monitored by the UE (320) may be configured to be periodic. Additionally, the configuration information for the monitoring CSI-RS may further include resource mapping information for the transmission of the CSI-RS (e.g., time and frequency density, subcarrier location, etc.) and transmission period information of the CSI-RS (e.g., repetition period and transmission time information).

[0141] Monitoring report configuration information may include a report type (e.g., periodic report, semi-permanent report, non-periodic report, or trigger-based report), information on the uplink resource to which the monitoring report message will be sent, and one or more of the three pieces of information described above for CSI-ReportConfigIE. An example of the embodiment in FIG. 6 may be a case where the monitoring report configuration information is set to trigger-based reporting. Thus, the monitoring report configuration information may further include trigger conditions for sending the monitoring report message.

[0142] In step S600, the UE (320) can receive an RRC signaling message. Based on the received RRC signaling message, the UE (320) can check the configuration information for the CSI-RS to be monitored and the monitoring report configuration information. In the embodiment of FIG. 6, the monitoring report configuration information may include monitoring report configuration information based on an event trigger. Additionally, configuration information for the event trigger condition may be included.

[0143] In step S610, the base station (310) can transmit the monitoring CSI-RS to the UE (320) at the time of transmission of the periodic monitoring CSI-RS. The periodic monitoring CSI-RS can be transmitted, for example, through each of the beams included in beam set A set by the RRC signaling message. The time of transmission and the transmission resources (e.g., time resources, frequency resources, and time / frequency density, etc.) of the periodic monitoring CSI-RS can be transmitted based on the configuration information for the monitoring CSI-RS included in the RRC signaling message. As described in FIG. 4, the configuration information for the monitoring CSI-RS can indicate specific resources of other IEs of the RRC signaling message, for example, CSI-RS-ResourceConfigIE or CSI-RS-ResourceSetConfigIE.

[0144] In step S610, the UE (320) can receive monitoring CSI-RS transmitted by the base station based on the RRC signaling message received in step S600.

[0145] In step S615, the UE (320) can measure the received monitoring CSI-RS. The UE (320) can then determine whether the trigger condition for sending a monitoring report message is met by comparing the measured CSI-RS value with the trigger condition. The trigger condition may be determined, for example, based on whether the inference result is the same as or similar to a previously predicted value. If the trigger condition is not met as a result of the inspection in step S615, the UE (320) can determine that the transmission of the event-based monitoring report message was not triggered. Therefore, even if the UE (320) measures the monitoring CSI-RS in step S615, it may not perform a monitoring report.

[0146] In step S620, the base station (310) can transmit monitoring CSI-RS to the UE (320) based on the monitoring RS transmission cycle (601). In step S620, the UE (320) can receive the monitoring CSI-RS transmitted by the base station.

[0147] In step S630, the UE (320) can measure the received monitoring CSI-RS. Then, the UE (320) can determine whether the trigger condition for sending a monitoring report message is met by comparing the measured CSI-RS value with the trigger condition. The trigger condition may be, for example, the following conditions.

[0148] The trigger condition may be whether the predicted value, which is the result inferred from the AI / ML model previously, is the same as the measured value of CSI-RS. For example, let's assume that beam set A includes beam #A1, beam #A2, …, beam #A(p-1), and beam #Ap, and that when the three beams with the highest quality inferred by the AI / ML model are listed in order of quality, the result is "beam #A2, beam #A1, beam #A3". Under this assumption, the UE (320) may determine that the trigger condition is not satisfied if the order of the predicted beams is the same as the order of the beams with the highest quality among the beams with measured CSI-RS when listed in order of beam quality. On the other hand, if the order of the beams with the highest quality among the beams with measured CSI-RS differs from the order of the predicted beams, the UE (320) may determine that the trigger condition is satisfied.

[0149] As another example, if the difference between the previously inferred quality value for a specific beam (e.g., beam #A2) (or multiple beams) and the CSI-RS measurement value received through that beam (beam #A2) (or multiple beams) exceeds a preset threshold, the UE (320) may determine that the trigger condition is satisfied. On the other hand, if the difference between the previously inferred quality value for a specific beam (e.g., beam #A2) and the CSI-RS measurement value received through that beam (beam #A2) is within a preset threshold, the UE (320) may determine that the trigger condition is not satisfied. In this embodiment, the specific beam may be the beam of the best quality. As another example, the specific beam may be one of the inferred beams. According to one embodiment of the present disclosure, event triggering conditions can be set and operated based on the specific beam.

[0150] Since the present disclosure utilizes the inference results of an AI model, cases satisfying the trigger condition may include instances where the inference results of the AI ​​model differ from the actual measurement results, or where the difference between the inference results of the AI ​​model and the actual measurement results exceeds a preset threshold. The examples above are embodiments to aid in understanding the present disclosure, and various other forms using inference results (e.g., predicted values) and actual measurement values ​​can be set as trigger conditions.

[0151] In FIG. 6, the operation of comparing with an event trigger condition through a monitoring CSI-RS measurement may mean comparing with an event trigger condition based on measurement(s) for the monitoring CSI-RS that are within a specific time offset relative to a CSI reference resource time associated with a specific inference reporting time. The specific time offset may be set in units of symbols or slots. The specific time offset may be determined based on a value set by upper-level signaling.

[0152] As another example, if the operation of comparing with event trigger conditions through monitoring CSI-RS measurements in Fig. 6 is applied to BM case 2, the UE can perform the following operation. The UE can compare with event trigger conditions based on measurements for monitoring CSI-RS that are within a specific time offset relative to the time of the instance being monitored among multiple future time instances that are the subject of prediction.

[0153] In the embodiment of FIG. 6, step S630 may be an embodiment in which the trigger condition is satisfied. The UE (320) may perform step S640 when the trigger condition is satisfied.

[0154] In step S640, the UE (320) can generate a monitoring report message to be transmitted to the base station (310) using the monitoring CSI-RS measured in step S630. The generation of the monitoring report message is described in more detail in the reporting method of BM case 1 and the reporting method of BM case 2, which will be described below.

[0155] In step S650, the UE (320) can transmit the generated monitoring report message to the base station (310) via a PUCCH or PUSCH that is pre-set (or assigned) in the monitoring report setting information received in step S600. In step S650, the base station (310) can receive the monitoring report message from the UE (320) via an uplink channel. The base station (310) can verify the performance of the AI / ML model based on the monitoring report message received from the UE (320).

[0156] After the time of the RS transmission cycle (601) has elapsed following the monitoring CSI-RS transmission in step S620, the base station (310) can transmit the monitoring CSI-RS to the UE (320) in step S660. In step S660, the UE (320) can receive the monitoring CSI-RS from the base station (310).

[0157] If the trigger condition is satisfied, the continuous time (or number of durations) for transmitting the monitoring report message is set. In this case, the UE (320) can measure the monitoring CSI-RS for the duration (or number of durations) based on the trigger condition and transmit the monitoring report message based on the measured result to the base station (310).

[0158] If the inference result of the AI ​​model continues to satisfy the event trigger condition during this trigger duration (or number of durations), the base station (310) may disable the AI ​​model so that it is not used or update it.

[0159] Meanwhile, FIG. 6 describes an example where uplink resources are pre-allocated (or configured) in the monitoring report configuration information of an RRC signaling message. If uplink resources are not pre-allocated (or configured) in the monitoring report configuration information of an RRC signaling message, the UE (320) may request uplink channel resources from the base station (310). When the base station (310) receives a request for uplink channel resources from the UE (320), it may allocate an uplink channel (e.g., PUCCH or PUSCH) capable of transmitting monitoring report messages to the UE (320). If the uplink channel is PUCCH, it may be a case where the amount of monitoring report messages that the UE (320) needs to transmit to the base station (310) is small. On the other hand, if the amount of monitoring report messages is large, the base station (310) may allocate PUSCH as the uplink channel to the UE (320). The UE (320) can transmit a monitoring report message to the base station (310) via an uplink channel allocated from the base station (310) at the request of the UE (320).

[0160] In the embodiments of FIGS. 4 to 6 described above, the periodic monitoring RS (e.g., periodic monitoring CSI-RS) set by the RRC can be transmitted based on beam set A. Additionally, the periodic monitoring CSI-RS can be operated in different ways depending on the BM case. Below, the method of operating the periodic monitoring CSI-RS for each BM case is described. Also, the method of generating the monitoring report message described in steps S420 and S450 of FIG. 4, step S540 of FIG. 5, and step S640 of FIG. 6 is also described.

[0161] <BM case 1의 빔 세트 운용 방식>

[0162] The UE (320) can measure a beam using the RS (e.g., CSI-RS for beam measurement) received from the base station (310) and infer the measured CSI-RS as input to the UE side model (321). The UE (320) can predict the beam(s) within beam set A through the inference of the UE side model (321). At this time, the CSI-RS may be transmitted through the beams included in beam set B. It should be noted that in the embodiments of the present disclosure described in FIGS. 4 to 6, the procedure in which the UE (320) receives the CSI-RS through the beams of beam set B from the base station (310), measures the received CSI-RS, and infers the measured value of the CSI-RS as input to the UE side model (321) is omitted.

[0163] When performing monitoring, the UE (320) may measure periodic monitoring CSI-RS through each of the beams and generate a monitoring report message containing the results of comparing the measured value and the inferred value of the periodic monitoring CSI-RS. In this case, the base station may transmit monitoring CSI-RS using only all or some of the beams within the predicted beam set A obtained through the inference of the UE side model (321) and the inference report of the UE. Some of the beams within the predicted beam set A may be the top K beams predicted to be of the highest quality among all the beams within the predicted beam set A. As another example, the base station may configure the beam(s) for monitoring CSI-RS transmission using some of the beam(s) within the predicted beam set A obtained through the inference of the UE side model (321) and the inference report of the UE, and some of the beam(s) within the beam set A that are not included in the prediction. Through this, the base station can efficiently configure resources for monitoring CSI-RS transmission.

[0164] When the UE generates a monitoring report message based on the base station operation described above, the UE (320) may not be configured to include comparison results for all beams having predicted values ​​within beam set A. For example, the UE (320) may configure the monitoring report message to include only the comparison results for the top K beams predicted by the inference result. As another example, even if the beams transmitting the monitoring CSI-RS consist of some of the beams predicted by the inference result and other beams, the monitoring report message may be configured to include only the results for the top N beams based on the beams transmitting the monitoring CSI-RS. In other words, the UE (320) may generate a monitoring report message containing only a predetermined number (K ​​or N) of beams and transmit the generated monitoring report message to the base station (310). By configuring the monitoring report message to include only the predicted values ​​for some of the beams within beam set A, the UE (320) can efficiently configure the resources of the monitoring report message transmitted to the base station (310).

[0165] Additionally, when the UE (320) reports measurement results for the monitoring CSI-RS, it may be configured in the same form as an inference report message containing inference results. For example, the CSI-RS measurement values ​​(for inference) for beam measurements transmitted using beam set B may be L1-RSRP values ​​for the received CSI-RS. If the measurement value for the CSI-RS is an L1-RSRP value, the predicted value of the UE side model may also be the predicted L1-RSRP value. And subsequently, the measurement value for the monitoring CSI-RS may also be an L1-RSRP. Thus, the monitoring report message may be configured based on beam quality values ​​in the same form as the inference report message. Additionally, the monitoring report message may include identifiers such as CRI for the monitoring CSI-RS for each beam included in the report message. In other words, the monitoring report message may include additional information regarding which beam the quality value is for.

[0166] The UE (320) may report the L1-RSRP value, which is the measured value for CSI-RS, in a modified form rather than as is. For example, instead of the measured value for CSI-RS, the UE (320) may report the difference between the previously predicted L1-RSRP value for the beam and the L1-RSRP value actually measured for CSI-RS through the beam.

[0167] As another example, the UE (320) may configure the measurement report message to use the L1-RSRP value for the best quality beam as is, but to include the difference value between the L1-RSRP value for the best beam and the L1-RSRP value for the next best beams.

[0168] In addition, there may be various forms in which the L1-RSRP value, a measurement of CSI-RS, is reported in a modified form rather than directly. Furthermore, while the above assumed that the CSI-RS measurement was the L1-RSRP value for ease of understanding, other values ​​may also be used. For example, other values ​​that indicate beam quality, such as L1-SINR for the measured monitoring CSI-RS, may be used.

[0169] As previously described, the UE (320) may measure periodic monitoring CSI-RS and, based on the measurement results, include information about the predicted beam, as well as other beams included in beam set A, or all beams included in beam set A, or beams associated with the predicted beam, in a monitoring report message and transmit it to the base station. In this case, the number of beams included in the monitoring report message may be limited to N. Limiting the number of beams included in the monitoring report message to N means that only information about N beams is included in the monitoring report message. In this case, the N beams may be the top N (Top N) beams with the best beam quality among the beams that measured periodic monitoring CSI-RS. Here, the value of N may be a natural number greater than or equal to 1. The value of N may be set to the UE (320) through one or more messages such as RRC signaling messages, MAC-CE signaling messages, or DCI. As another example, the value of N may be set to the same value as the value of K.

[0170] Meanwhile, the monitoring CSI-RS may be all or part of the beams of beam set A, or one or more beams associated with beam set A. The time at which the monitoring CSI-RS is transmitted may be within a specific time offset relative to the time at which the CSI-RS is transmitted through the beams of beam set B. Alternatively, as described in FIGS. 4 to 6, if the monitoring CSI-RS is transmitted periodically according to upper layer configuration information, the inference performance evaluation may be performed through the monitoring CSI-RS transmitted within a specific time offset relative to the time at which the CSI-RS is transmitted through the beams of beam set B.

[0171] For example, the monitoring CSI-RS(s) used for performance evaluation of a specific inference reporting may be CSI-RSs that exist within a predefined maximum time offset relative to the CSI reference resource configured for that inference reporting. Here, the monitoring CSI-RS may be transmitted through a separate resource distinct from the CSI-RS for beam measurements, and the UE may evaluate inference performance using the measurement results of one or more monitoring CSI-RSs that are temporally closest to the configured CSI reference resource. Additionally, the time offset may be set in slots or in symbol units.

[0172] If the time at which the monitoring CSI-RS is transmitted deviates from a specific time offset relative to the time at which the CSI-RS is transmitted through the beams of beam set B (if the offset time has elapsed), the UE measures the monitoring CSI-RS but may not perform a monitoring report.

[0173] <BM case 2의 빔 세트 운용 방식>

[0174] The UE (320) can measure a beam using the RS (e.g., CSI-RS for beam measurement) received from the base station (310) and infer the measured CSI-RS as input to the UE side model (321). The UE (320) can predict the beam(s) within beam set A through the inference of the UE side model (321). At this time, the CSI-RS may be transmitted through the beams included in beam set B. It should be noted that in the embodiments of the present disclosure described in FIGS. 4 to 6, the procedure in which the UE (320) receives the CSI-RS through the beams of beam set B from the base station (310), measures the received CSI-RS, and infers the measured value of the CSI-RS as input to the UE side model (321) is omitted.

[0175] The base station (310) can predict high-quality beams that can be used in specific future instances as a result of inference. And to evaluate the inference performance for the predicted values ​​in specific future instances, the base station can transmit monitoring CSI-RS to the UE. The UE can evaluate the inference performance through measurements of the monitoring CSI-RS. And the UE can report the results of the evaluation of the inference performance to the base station. The operation thereof will be examined in more detail with reference to the attached drawings.

[0176] Figure 7 is a conceptual diagram illustrating a case where monitoring RS is transmitted and measured from an instance that is the prediction target through a UE side model.

[0177] Referring to FIG. 7, inference and monitoring instance #1 (710) and inference and monitoring instance #2 (720) may be time instances in which RS (e.g., CSI-RS) is transmitted through specific beam(s). The UE (320) may measure the CSI-RS received from the base station (310) prior to inference and monitoring instance #1 (710) and perform inference using the measured CSI-RS as input to the UE side model (321). The UE side model (321) may have at least the information inferred from inference and monitoring instance #1 (710) (e.g., predicted L1-RSRP value). Additionally, the UE (320) may have more information inferred from inference and monitoring instance #2 (720) (e.g., predicted L1-RSRP value).

[0178] The base station (310) can transmit monitoring CSI-RS through the beam(s) predicted by the UE side model (321) in the inference and monitoring instance #1 (710). The UE (320) can measure the monitoring CSI-RS received from the base station (310) in the inference and monitoring instance #1 (710) and evaluate the performance of the UE side model (321) by comparing the measured result with the previously predicted value. The result of the performance evaluation can be included in a monitoring report message and transmitted to the base station (310).

[0179] Additionally, the base station (310) can transmit CSI-RS for inference to the UE (320) through the beams belonging to beam set B in the inference and monitoring instance #1 (710). The UE (320) can measure the CSI-RS received through the beams belonging to beam set B and perform inference using the measured CSI-RS value as input to the UE side model (321). The result of the inference performed by the UE side model (321) using the CSI-RS value measured in the inference and monitoring instance #1 (710) may be a prediction for the beams belonging to beam set A in the inference and monitoring instance #2 (720) and / or the inference and monitoring instance #3 (not illustrated in FIG. 7).

[0180] Subsequently, the base station (310) can transmit monitoring CSI-RS to the UE (320) from the inference and monitoring instance #2 (720) through the beam(s) predicted by the UE side model (321). Additionally, the base station (310) can transmit CSI-RS for inference of the UE side model (321) from the inference and monitoring instance #2 (720) to the UE (320) through the beams belonging to beam set B. Since the operation of the base station (310) and the UE (320) in the inference and monitoring instance #2 (720) is the same as the operation of the base station (310) and the UE (320) in the inference and monitoring instance #1 (710), a redundant description is omitted.

[0181] Inference and monitoring instance #1 (710) and inference and monitoring instance #2 (720), each exemplified in FIG. 7, may be in the form of symbol(s), slot(s), or subframe(s). Additionally, the monitoring CSI-RS transmitted via beams in each of the inference and monitoring instance #1 (710) and inference and monitoring instance #2 (720) may be transmitted based on configuration information for the monitoring CSI-RS as previously described in FIG. 4 through 6. In other words, the monitoring CSI-RS transmitted to the UE (320) via beams in each of the inference and monitoring instance #1 (710) and inference and monitoring instance #2 (720) may be transmitted based on resources (e.g., time, frequency, and density) configured by the configuration information for the monitoring CSI-RS.

[0182] Assuming that the inference and monitoring instances (710, 720) exemplified in FIG. 7 consist of two or more time intervals, the inference and monitoring instance #1 (710) and the inference and monitoring instance #2 (720) can be understood as follows:

[0183] The base station can transmit CSI-RS in each time interval of an inference and monitoring instance composed of multiple time intervals. The UE can receive monitoring CSI-RS in each time interval of the inference and monitoring instance and evaluate the performance of the inference result using the received CSI-RS. In this case, the inference instance may refer to multiple future time instances in BM case 2. When evaluating inference performance, the UE may use the CSI-RS(s) received in the time interval closest to the inference instance being predicted for the inference performance evaluation.

[0184] Additionally, only one monitoring CSI-RS can be used for evaluating inference performance in a specific inference instance, and in this case, only the monitoring CSI-RS received in the time interval closest to the inference instance can be used for evaluation. At this time, only monitoring CSI-RS within a specific time offset relative to the inference instance being evaluated can be set as valid monitoring CSI-RS for inference performance evaluation. In other words, among the monitoring CSI-RS within a specific time offset (time threshold) from the inference instance, the single closest CSI-RS can be used for inference performance evaluation. When one or more monitoring CSI-RS are configured to be used, X CSI-RS can be used for inference performance evaluation in order of proximity among the monitoring CSI-RS within a specific time offset (time threshold). The specific time offset and X can be configured and operated by upper-layer signaling, and X can be a natural number greater than or equal to 1.

[0185] Additionally, the monitoring CSI-RS may be valid for evaluating inference performance for multiple inference instances. For example, assume a time sequence with the same time intervals as t1, t2, t3, t4, t5… Let us assume that the UE (320) and / or base station (310) have values ​​inferred from an inference instance at time t1 and values ​​inferred at time t3, and that the monitoring CSI-RS is transmitted at times t2 and t5. In this case, the monitoring CSI-RS at time t2 may be a valid monitoring CSI-RS used for evaluating inference performance at times t1 and t3.

[0186] If the system is operated to evaluate inference performance using multiple monitoring CSI-RSs, if the monitoring CSI-RS at time t5 is within a time offset (time threshold) set relative to time t3, the monitoring CSI-RS at time t5 can be a valid monitoring CSI-RS used for evaluating inference performance at time t3. On the other hand, if the monitoring CSI-RS at time t5 is outside the time offset set relative to time t3, the monitoring CSI-RS at time t5 can be operated so that it cannot be used for evaluating inference performance at time t3. In other words, the UE (320) does not transmit a monitoring report message using the monitoring CSI-RS at time t5 to the base station (310).

[0187] FIG. 7 may be a diagram assuming a configuration in which an inference instance and a monitoring instance to which a monitoring CSI-RS is transmitted exist within a single time interval. That is, as an example of a configuration in which the inference instance and the monitoring CSI-RS exist in adjacent intervals, the temporal sequence of the inference instance and the monitoring instance for transmitting the monitoring CSI-RS can be configured in various ways. FIGS. 8a to 8c show examples of various temporal sequences of the inference instance and the monitoring instance. That is, the instance for inference (or the instance to be predicted) and the monitoring instance to which the monitoring CSI-RS is transmitted may be configured separately.

[0188] FIG. 8a is a conceptual diagram of a first embodiment of a transmission method of a monitoring instance in which an inference instance and a monitoring RS are transmitted.

[0189] Referring to FIG. 8a, inference instance #1 (811) and inference instance #2 (812) can be set at adjacent times, and monitoring instance #1 (813) and monitoring instance #2 (814) can also be set at adjacent times. In other words, the time interval in which the inference instances (811, 812) are set and the time interval in which the monitoring instances (813, 814) are set can be distinguished.

[0190] The base station (310) may have obtained an inference result from the UE (320) using the RS (e.g., CSI-RS for inference) transmitted to the UE (320) via beams included in beam set B at each of inference instance #1 (811) and inference instance #2 (812). Accordingly, the UE (320) can measure the CSI-RS received via the beams at a previous time point of inference instance #1 (811) and / or inference instance #2 (811), and obtain an inference value for inference instance #1 (811) and inference instance #2 (812) by using the measured CSI-RS value as an input to the UE side model (321). At this time, the CSI-RS used for inference may be the CSI-RS of the previous time point closest to each of the inference instances (811, 812), or the CSI-RS(s) received within a preset interval for the time prior to inference and the time closest to the time of inference.

[0191] The UE side model (321) can compare monitoring CSI-RS measurements received from the base station (310) at monitoring instance #1 (813) using inference values ​​inferred for inference instance #1 (811) or for inference instance #1 (811) and inference instance #2 (812). In other words, it can compare values ​​predicted through inference (e.g., predicted L1-RSRP) with values ​​measured by receiving monitoring CSI-RS (e.g., measured L1-RSRP). At this time, the UE (320) can perform an evaluation of the inference performance of inference instance #2 (812), which is the closest inference instance, through monitoring instance #1 (813). As another example, if pre-configured by upper layer signaling, etc., the UE (320) may use the inference results of inference instance #1 (811) within a time offset (810), rather than inference instance #2 (812), which is the closest inference instance to monitoring instance #1 (813). The UE (320) may report the comparison results to the base station (310). In this case, the UE (320) may compare the measured value (e.g., measured L1-RSRP) for the CSI-RS in monitoring instance #2 (814) using the inferred value (e.g., predicted L1-RSRP) inferred from inferred instance #2 (812). The UE (320) may report the comparison results to the base station (310).

[0192] Meanwhile, the base station (310) can transmit CSI-RS for monitoring from the monitoring instance #1 (813) to the UE (320) through beams included in beam set A based on the results of inference by the UE (320) for inference instance #1 (811) and the results of inference by the UE (320) for inference instance #2 (812). In this case, the UE (320) can receive CSI-RS from the base station (310) at the monitoring instance #1 (813) through beams included in beam set A. The UE (320) can measure the received CSI-RS and compare the measured value of the CSI-RS (e.g., the measured L1-RSRP value) with the predicted value (e.g., the predicted L1-RSRP value) from inference instance #1 (811) and inference instance #2 (812). Based on the difference between the measured L1-RSRP value and the predicted L1-RSRP value, the UE (320) can evaluate the performance of the UE side model (321). The results of the performance evaluation can be included in a monitoring report message and transmitted to the base station (310). Thus, the base station (310) can also evaluate the performance of the UE side model (321) based on the monitoring report message received from the UE (320).

[0193] In the examples described above with respect to FIG. 8a, the inference instance and the monitoring instance corresponding to the inference instance must exist within a certain time interval to be valid. In other words, the time interval between the inference instances and the monitoring instance evaluating the performance of the instance may exist within a preset time offset. For example, if the inference performance evaluation for inference instance #2 (812) is performed through monitoring instance #1 (813), the time interval between the two instances may exist within a preset time offset (810). This is because the performance evaluation for the actual inference is possible at a point temporally close to the inference instance, which is the prediction target time for the beam by the UE side model (321). Therefore, the time offset (810) between the inference instance and the monitoring instance corresponding to the inference instance may be determined based on the capability of the UE side model (321) and / or based on UE capability, etc.

[0194] Since the base station (310) can know in advance the capabilities of the UE side model (321) and / or the UE capabilities, it can determine the time offset (810) based on the UE side model (321) and / or the UE capabilities. The time offset (810) can be set to the UE (320) through one or more of the RRC signaling messages, MAC-CE, or DCI.

[0195] Meanwhile, the CSI-RS(s) used for inference described in FIG. 8 may be the nearest CSI-RS(s) among the previous times of each inference instance as described in FIG. 7, and / or CSI-RS(s) received within a preset interval among the previous times of each inference instance.

[0196] FIG. 8b is a conceptual diagram of a second embodiment of a transmission method of a monitoring instance in which an inference instance and a monitoring RS are transmitted.

[0197] Referring to FIG. 8b, the inference instance #1 (821) and the monitoring instance #1 (822) may be set in temporally adjacent, and the inference instance #2 (823) and the monitoring instance #2 (824) may be set in temporally adjacent.

[0198] The base station (310) can perform inference for inference instance #1 (821) using RS for inference (e.g., CSI-RS for inference) received through the beams included in beam set B. The CSI-RS used for inference for inference instance #1 (821) may be the CSI-RS closest to the time of inference and / or CSI-RS(s) received within a preset interval for a time prior to inference. Additionally, the UE (320) can obtain inference results for one or more beams among the beams of beam set A. The inference results may be an identifier (index, which may be CRI if CSI-RS is used) value for one or more high-quality beams among the beams included in beam set A, and a predicted beam quality value corresponding to that beam (e.g., predicted L1-RSRP). The UE (320) can report the predicted value from the UE side model (321) to the base station (310).

[0199] The base station (310) can transmit monitoring CSI-RS(s) to the UE (320) through beam(s) included in beam set A, which are reported based on the results inferred by the UE (320) for inference instance #1 (821) in monitoring instance #1 (822). Thus, the UE (320) can receive monitoring CSI-RS(s) from the base station (310) in monitoring instance #1 (822) through beams included in beam set A. The UE (320) can measure the received monitoring CSI-RS(s) and compare the measured value of the CSI-RS(s) (e.g., measured L1-RSRP value) with the predicted value for inference instance #1 (821) (e.g., predicted L1-RSRP value). Based on the difference between the measured L1-RSRP value and the predicted L1-RSRP value, the UE (320) can evaluate the performance of the UE side model (321). The results of the performance evaluation can be included in a monitoring report message and transmitted to the base station (310). In this case, the beams used for CSI-RS transmission in monitoring instance #1 (822) may be the beams predicted in inference instance #1 (821).

[0200] In addition, the same operation and management as above are possible in subsequent inference instance #2 (823) and monitoring instance #2 (824). As exemplified in FIG. 8b, inference instances (821, 823) for two or more prediction time points in BM case 2 can perform performance evaluation using the closest monitoring CSI-RS(s) among the monitoring CSI-RS(s) within a specific time offset relative to each inference instance. The monitoring CSI-RS(s) can be configured to evaluate inference performance using the closest monitoring CSI-RS(s) among the time points after the inference instance. That is, the closest CSI-RS among the monitoring CSI-RS(s) within a specific time offset (time threshold) relative to the time points after the inference instance can be used for inference performance evaluation. When one or more monitoring CSI-RS(s) are configured to be used, X number of CSI-RS(s) among the monitoring CSI-RS(s) within a specific time offset (time threshold) can be used for inference performance evaluation in order of proximity. A specific time offset and X can be set and operated by upper-layer signaling, and X can be a natural number greater than or equal to 1.

[0201] As another example, as shown in FIG. 8c, the inference performance evaluation can be configured and operated using the nearest monitoring CSI-RS(s) among the time points prior to the inference instances (832, 834). As yet another example, as described in FIG. 7, the inference performance evaluation can be configured and operated using the nearest monitoring CSI-RS(s) regardless of the time points prior to or after the inference instance time point.

[0202] In FIG. 8b, the presence of monitoring instance #1 (822) between inference instance #1 (821) and inference instance #2 (823) may be due to a time offset (820) based on the UE side model (321) and / or UE capabilities, etc., as previously described. In other words, if the interval between inference instances is set to be greater than the time offset (820), it can be operated as exemplified in FIG. 8b.

[0203] Since the base station (310) can know in advance the capabilities of the UE side model (321) and / or the UE capabilities, it can determine the time offset (820) based on the capabilities of the UE side model (321) and / or the UE capabilities. The time offset (820) can be set to the UE (320) through one or more of an RRC signaling message, MAC-CE, or DCI.

[0204] FIG. 8c is a conceptual diagram of a third embodiment of a transmission method of a monitoring instance in which an inference instance and a monitoring RS are transmitted.

[0205] FIG. 8c may be an example where the monitoring instances (831, 833) and inference instances (832, 834) are configured in a different form than the first and second embodiments described above. That is, FIG. 8c is an example where, unlike FIG. 8b, the closest monitoring instances are used for inference performance evaluation based on a point in time prior to the inference instances. It should be noted that the monitoring instances (831, 833), inference instances (832, 834), and time offset (830) can be understood as being the same as in the first and second embodiments described above, so a redundant description is omitted.

[0206] The CSI-RS transmission settings and inference instances for beam set B described in FIGS. 7, FIGS. 8a, FIGS. 8b, and FIGS. 8c, and the monitoring CSI-RS transmission settings (monitoring instances) can be configured in conjunction. For example, the monitoring CSI-RS transmission settings can be configured to be x times the CSI-RS transmission settings for beam set B (where x is a natural number greater than or equal to 1). If x has a value of 1, the monitoring CSI-RS transmission settings and the CSI-RS transmission settings for beam set B may have the same transmission period.

[0207] In addition, the monitoring CSI-RS transmission time can be set to be the same as the CSI-RS transmission time for beam set B, or to be transmitted within a time offset from the CSI-RS transmission time for beam set B.

[0208] Additionally, the base station (310) can determine the beams to transmit CSI-RS in the monitoring instance based on the inference result value received from the UE (320) via the inference report message. The base station (3100) can transmit the monitoring CSI-RS to the UE (320) through the determined beams. At this time, the base station (310) may transmit CSI-RS using additional beams in addition to the beams based on the inference result value. For example, the base station (310) may transmit CSI-RS through all beams included in the entire beam set A. Alternatively, the base station (310) may transmit CSI-RS through the beams included in the inference report message of the UE (320) and some beams included in the entire beam set A that are not included in the inference report message. If the base station (310) transmits CSI-RS only through the beams included in the inference report message received from the UE (320), the efficiency of resource utilization may be increased.

[0209] Additionally, if the UE (320) configures the performance evaluation results by including them in a monitoring report message, the performance evaluation results may be configured as values ​​in the same form as the predicted value (or inference value) of the CSI-RS included in the inference report message. As another example, if the UE (320) configures the performance evaluation results by including them in a monitoring report message, the performance evaluation results may be configured as the difference between the predicted value (or inference value) of the CSI-RS included in the inference report message and the measured value of the monitoring CSI-RS.

[0210] When transmitting periodic monitoring CSI-RS, the UE (320) can be configured to include information on the top B (Top B) beams of the best quality among the beams measured through the monitoring CSI-RS in the monitoring report message, in the case where the beams predicted by the inference result (i.e., beams included in beam set A) are included and other beams associated with a different beam set A are transmitted. Then the UE (320) can transmit the monitoring report message to the base station (310). In this case, the value of B is a natural number greater than or equal to 1 and can be set and operated as one or more of RRC, MAC-CE, and DCI.

[0211] As another example, the UE (320) may be configured to include the L1-RSRP value for the best quality beam and the difference between that L1-RSRP and the L1-RSRP values ​​for the remaining beams in a monitoring report message and transmit it to the base station (310). In addition, reporting by various combinations is possible, and it can be operated in a form that reports other beam quality values, such as L1-SINR, as measured monitoring results.

[0212] Meanwhile, the method described above with reference to FIGS. 7, FIGS. 8a, FIGS. 8b, and FIGS. 8c may be applied to the following description.

[0213] The UE (320) can evaluate performance by associating the measurement results of the monitoring CSI-RS received from each monitoring instance with the prediction target time instance (inference instance) that is temporally closest to the monitoring instance. In this case, the time offset between the monitoring instance and the prediction target time instance associated with it may be considered valid only if it is less than or equal to a preset maximum time offset value. A monitoring instance whose time offset exceeds the preset maximum time offset value may not be associated with any prediction target time instance.

[0214] Additionally, the base station (310) may calculate performance metrics based on measurement results from the latest N monitoring CSI-RS (monitoring instances) existing within a preset time offset from the predicted time instance to evaluate the performance of the AI / ML model at a specific predicted time instance. Here, N may be a natural number greater than or equal to 1. The value of N may be indicated to the UE (320) through a combination of one or more of RRC signaling, MAC-CE, and DCI.

[0215] <BM case 별로 서로 다른 모니터링 보고를 운용하는 방식>

[0216] Below, cases where different monitoring report messages are configured for each BM case are explained.

[0217] In the case of BM case 2, the base station (310) may need to transmit monitoring RS (e.g., CSI-RS) to the UE (320) from multiple instances. On the other hand, in the case of BM case 1, the base station (310) may transmit monitoring CSI-RS to the UE (320) from a single instance. For this reason, the configuration for transmitting monitoring CSI-RS may differ depending on BM case 1 and BM case 2. The base station (310) may configure one or more monitoring CSI-RS settings in monitoring report configuration information (e.g., CSI-ReportConfigIE). If two or more monitoring CSI-RS settings are included in the monitoring report configuration information, each of the monitoring CSI-RS settings may include a CSI-RS resource configuration transmitted from a single instance and a CSI-RS resource configuration transmitted from multiple instances, as in BM case 1.

[0218] The monitoring report transmitted by the UE (320) to the base station (310) can also be operated in different forms depending on BM case 1 and BM case 2. For example, in the case of BM case 1, the UE (320) may be configured to transmit measurement information values ​​for one or more beam(s) to the base station (310) based on results measured in one monitoring instance. On the other hand, in the case of BM case 2, the UE (320) may be configured to transmit measurement values ​​for beams for each instance to the base station (310) based on results measured in multiple monitoring instances.

[0219] The beam quality reported for each BM case may be an actual measured value, such as an L1-RSRP value. If the predicted L1-RSRP value from the UE side model (321) needs to be transmitted, the beam quality may be the predicted L1-RSRP value. Accordingly, the base station (310) can operate by setting a method for a specific monitoring report in conjunction with the settings of the monitoring CSI-RS within the monitoring report setting information.

[0220] When the base station (310) instructs the UE (320) on a monitoring reporting method, the monitoring reporting method instruction may be given via a MAC-CE signaling message or DCI, as previously described in FIG. 4. If the monitoring reporting method is instructed by the base station (310) via a MAC-CE signaling message or DCI, the UE (320) may generate a monitoring report message containing measurement results for the monitoring CSI-RSs based on the instructed monitoring reporting method. The UE (320) may transmit the generated monitoring report message to the base station (310).

[0221] In addition, monitoring configurations may be configured separately for BM case 1 and BS case 2, respectively. For example, a higher-level signaling message (e.g., an RRC signaling message) may include one or more monitoring configurations for BM case 1 and one or more monitoring configurations for BM case 2. In other words, the RRC signaling message can configure the monitoring configuration(s) for BM case 1 and the monitoring configuration(s) for BM case 2 separately. Examples of the above cases may include the following.

[0222] If the monitoring settings for BM case 1 are, for example, monitoring setting #1 and monitoring setting #2, the monitoring settings for BM case 2 may be different settings, such as monitoring setting #3 and monitoring setting #4. As such, the monitoring settings for BM case 1 and the monitoring settings for BM case 2 can be completely different from each other.

[0223] As another example, if the monitoring settings for BM case 1 are, for instance, monitoring setting #1 and monitoring setting #2, the monitoring settings for BM case 2 can be set to monitoring setting #1 and monitoring setting #2. In this way, the monitoring settings for BM case 1 and the monitoring settings for BM case 2 are set separately in the RRC signaling message, but they may be the same settings.

[0224] As another example, if the monitoring settings for BM case 1 are, for instance, monitoring setting #1 and monitoring setting #2, the monitoring settings for BM case 2 may be set to monitoring setting #1 and monitoring setting #3, or to monitoring setting #2 and monitoring setting #3. In this way, parts of the monitoring settings for BM case 1 and BM case 2, which are configured separately in the RRC signaling message, may overlap.

[0225] In addition, when performing monitoring reporting based on RRC event settings as described in Fig. 6 above, the UE (320) can generate monitoring report messages based on BM case-specific event settings. In other words, the setting information for monitoring CSI-RS can be operated by associating it with specific events.

[0226] Meanwhile, the time offset described above may be set in the UE (320) based on a combination of one or more of the RRC signaling message, the MAC-CE signaling message, or the DCI. If the time offset is set in the UE (320) using only the RRC signaling message, the time offset may be included in the monitoring report setting information.

[0227] [Second Embodiment: Monitoring method based on semi-permanent monitoring RS]

[0228] The cases of FIGS. 4 through 6 described above in the first embodiment were described assuming that periodic monitoring is set. However, monitoring RSs (e.g., CSI-RS) can be operated in a semi-permanent form.

[0229] In the case of semi-permanent monitoring CSI-RS, monitoring CSI-RS transmission can be initiated via low-layer triggering. For example, this may occur when the base station has the CSI-RS monitoring method set to semi-permanent in the configuration information for monitoring CSI-RS within the RRC signaling message, which is a higher-layer signaling message.

[0230] The base station (310) may transmit (or instruct) a MAC-CE signaling message to the UE (320) instructing the activation of a semi-permanent monitoring CSI-RS to instruct the start of transmission and monitoring of the monitoring CSI-RS. Additionally, the base station (310) may transmit to the UE (320) resources to transmit monitoring report messages using DCI. If resources to transmit monitoring report messages are pre-configured (or allocated) in the configuration information for the monitoring CSI-RS, DCI may not be transmitted.

[0231] When the activation of the semi-permanent monitoring CSI-RS is instructed as described above and the resource to transmit the monitoring report message is allocated to the UE (320), the base station (310) can receive the monitoring report message from the UE (320) without instructing the UE (320) to send an additional monitoring report, as described in FIG. 4 above.

[0232] The base station (310) can stop the transmission of monitoring report messages from the UE (320) by instructing the deactivation of the semi-permanent monitoring CSI-RS via a MAC-CE signaling message when the reception of monitoring report messages is no longer needed.

[0233] As another example, as described in FIG. 5 above, if the base station (310) has instructions to the UE (320) for monitoring reporting, the UE (320) can perform monitoring reporting in the same way for the semi-permanent monitoring CSI-RS as described in FIG. 5 above. However, the monitoring report instruction information may be included in the activation information for the monitoring semi-permanent CSI-RS and used.

[0234] Instructions for monitoring reports transmitted by the base station (310) to the UE (320) may be in the form of activation instructions. The UE (320) may generate monitoring report messages whenever it continuously receives a monitoring semi-permanent CSI-RS in the absence of deactivation instructions from the base station (310). The UE (320) may then transmit the generated monitoring report messages to the base station (310) at each reporting time corresponding to the transmission cycle of the semi-permanent CSI-RS.

[0235] The report message of the monitoring semi-permanent CSI-RS may be interrupted when the transmission of the monitoring semi-permanent CSI-RS is interrupted, that is, when the base station instructs the deactivation of the monitoring semi-permanent CSI-RS. The deactivation of the monitoring semi-permanent CSI-RS may be instructed to the UE (320) by the base station (310) using deactivation via MAC-CE signaling messages, as previously described. As another example, the base station (310) may use DCI to instruct the interruption of the monitoring CSI-RS transmission or to instruct the interruption of the transmission of the monitoring report message for the monitoring CSI-RS.

[0236] As another example, the monitoring semi-permanent CSI-RS may be operated to report only once when instructed to report from the base station (310).

[0237] Next, when an event-based monitoring report is performed, the base station (310) can activate the event by lower-level triggering. For example, the base station may have the CSI-RS monitoring method set to event-based monitoring in the configuration information for the monitoring CSI-RS of the RRC signaling message.

[0238] The base station (310) can use a MAC-CE signaling message or DCI to instruct the semi-permanent monitoring CSI-RS transmission and event-based CSI-RS monitoring method to be activated. When the UE (320) receives an instruction from the base station (310) to activate the event-based CSI-RS monitoring method, it can activate the event included in the RRC signaling message and check the trigger condition corresponding to the event.

[0239] Subsequently, when the UE (320) receives a semi-permanent monitoring CSI-RS, it can check whether the trigger condition is satisfied and generate a monitoring report message as described in FIG. 6. Then, the UE (320) can transmit the generated monitoring report message to the base station (310).

[0240] [Third Embodiment: Monitoring Method Based on Non-Periodic Monitoring RS]

[0241] Figure 9 is a flowchart illustrating the case where a UE transmits a monitoring report to a base station when non-periodic RS monitoring is set by upper layer setting information.

[0242] The base station and UE exemplified in FIG. 9 may be the base station and UE described earlier in FIG. 3. In FIG. 9, the base station (310) and UE (320) may both have AI / ML models. Additionally, the base station and UE exemplified in FIG. 9 may include all or part of the configuration of the communication node described earlier in FIG. 2.

[0243] In step S900, the base station (310) can transmit upper layer configuration information to the UE (320). The upper layer configuration information may be configured and transmitted as an RRC signaling message as previously described. The RRC signaling message may include configuration information for a monitoring RS (e.g., monitoring CSI-RS) to be monitored by the UE (320) and monitoring report configuration information. The configuration information for the monitoring CSI-RS and the monitoring report configuration information may be transmitted via the CSI-ReportConfigIE of the RRC signaling message.

[0244] The configuration information for the monitoring CSI-RS may be configured such that the method of transmission of the CSI-RS is, for example, periodic, semi-permanent, or non-periodic. Since FIG. 9 assumes that the CSI-RS is transmitted non-periodically, the configuration information for the monitoring CSI-RS to be monitored by the UE (320) may be configured for non-periodic transmission. Additionally, the configuration information for the monitoring CSI-RS to be monitored by the UE (320) may further include resource mapping information (e.g., time and frequency density, subcarrier location, etc.) in which the CSI-RS is transmitted.

[0245] Monitoring report configuration information may include a report type (e.g., periodic report, semi-permanent report, non-periodic report, or trigger-based report) and one or more of the three pieces of information described above for CSI-ReportConfigIE. In the embodiment of FIG. 9, the report type included in the monitoring configuration information may be set to non-periodic. Even when the report type included in the monitoring configuration information is set to non-periodic, the monitoring configuration information may further include report resource information.

[0246] In step S900, the UE (320) can receive an RRC signaling message. Based on the received RRC signaling message, the UE (320) can check the configuration information for the CSI-RS to be monitored and the monitoring report configuration information.

[0247] In step S910, the base station (310) can transmit a monitoring report instruction message to the UE (320). The report instruction message may consist only of a MAC-CE signaling message, may consist only of a DCI, or may use both a MAC-CE signaling message and a DCI.

[0248] First, we will explain the case where a report instruction message is configured using only MAC-CE signaling messages.

[0249] The embodiment of FIG. 9 may be a case where the reporting of the monitoring CSI-RS is set to be non-periodical. In other words, it may be a case where the reporting of the monitoring CSI-RS is set to be non-periodical in the monitoring report setting information transmitted via the RRC signaling message. Additionally, the information regarding the transmission resources of the CSI-RS transmitted non-periodically in the setting information for the monitoring CSI-RS transmitted via the RRC signaling message may be in a state obtained at step S900.

[0250] A MAC-CE signaling message may include information instructing the activation of transmission for aperiodic monitoring CSI-RS and the activation of transmission for monitoring report instruction messages. Each of the activation of transmission for aperiodic monitoring CSI-RS and the activation of transmission for monitoring report instruction messages may consist of 1 bit of information. Here, the activation of transmission for aperiodic monitoring CSI-RS may mean that the aperiodic monitoring CSI-RS is activated once. The activation instruction for the transmission of monitoring report instruction messages may also mean that the transmission of monitoring report instruction messages is activated once.

[0251] If a resource for transmitting monitoring report messages is configured (or not allocated) in the RRC signaling message, the MAC-CE signaling message may further include information about the resource to which the monitoring report messages will be transmitted. Here, the resource information may be, for example, uplink channel information.

[0252] The base station (310) can transmit a MAC-CE signaling message containing information as described above to the UE (320). The UE (320) can receive the MAC-CE signaling message at step S910. Based on the received MAC-CE signaling message, the UE (320) can confirm that the transmission of the non-periodic monitoring CSI-RS is enabled and can identify the resource to which the monitoring report message is to be transmitted. The UE (320) can also identify the resource to which the received non-periodic monitoring CSI-RS is transmitted at step S900.

[0253] Next, we will explain the case where the report instruction message consists only of DCI.

[0254] The base station (310) may generate a DCI that includes one or more of an aperiodic monitoring CSI-RS activation instruction and / or an aperiodic monitoring report instruction. The DCI may further include uplink channel information for transmitting an aperiodic monitoring report message. The base station (310) may transmit the DCI to the UE (320). The UE (320) may receive the DCI from the base station (310). The UE (320) may identify the resource to which the aperiodic monitoring CSI-RS is transmitted based on the received DCI and the RRC signaling message received in step S900. The UE (320) may also know that a monitoring report message must be transmitted based on the received DCI and may identify the uplink channel information to which the monitoring report message is transmitted.

[0255] Next, the case where MAC-CE messages and DCI are used together is described. If MAC-CE and DCI are used together, step S910 of FIG. 9 may actually include MAC-CE transmission and DCI transmission.

[0256] A MAC-CE signaling message may include information instructing the transmission enable instruction of an aperiodic monitoring CSI-RS and the transmission enable instruction of a monitoring report instruction message. As another example, a MAC-CE signaling message may include only the transmission enable instruction of an aperiodic monitoring CSI-RS. If it includes both the transmission enable instruction of an aperiodic monitoring CSI-RS and the transmission enable instruction of a monitoring report instruction message, each of the transmission enable instruction of an aperiodic monitoring CSI-RS and the transmission enable instruction of a monitoring report instruction message may consist of 1 bit of information. Here, the transmission enable instruction of an aperiodic monitoring CSI-RS may mean that the aperiodic monitoring CSI-RS is activated once. The transmission enable instruction of a monitoring report instruction message may also mean that the transmission of a monitoring report instruction message is activated once.

[0257] The base station (310) can first transmit a MAC-CE signaling message containing the above information to the UE (320). Then, the base station (310) can generate a DCI and transmit it to the UE (320).

[0258] The DCI may include uplink channel information for transmitting non-periodic monitoring report messages. If the MAC-CE signaling message does not include an instruction to enable the transmission of monitoring report instruction messages, the DCI may further instruct the transmission of monitoring report instruction messages.

[0259] In step S920, the base station (310) can transmit aperiodic monitoring CSI-RS to the UE (320). In step S920, the UE (320) can receive the aperiodic monitoring CSI-RS through a resource to which the aperiodic monitoring CSI-RS is transmitted based on an RRC signaling message.

[0260] In step S930, the UE (320) can measure the monitoring CSI-RS. Based on the measurement of the monitoring CSI-RS, the UE (320) can generate a monitoring report message. The monitoring report message can be configured in the form described in BM case 1 and BM case 2 above.

[0261] In step S940, the UE (320) can transmit a monitoring report message to the base station (310) via an uplink channel allocated by lower-level triggering for non-periodic monitoring reporting. The base station (310) can receive the non-periodic monitoring report message from the UE (320) in step S940.

[0262] Meanwhile, the transmission type of the monitoring CSI-RS described above (e.g., periodic, semi-permanent, or non-periodic) may be the same as the transmission type of the CSI-RS for AI model inference. In other words, the transmission type of the monitoring CSI-RS can be configured in association with the transmission type of the CSI-RS for AI model inference.

[0263] The operation of the method according to an embodiment of the present disclosure can be implemented as a computer-readable program or code on a computer-readable recording medium. A computer-readable recording medium includes all types of recording devices in which information that can be read by a computer system is stored. Additionally, the computer-readable recording medium may be distributed across networked computer systems, allowing the computer-readable program or code to be stored and executed in a distributed manner.

[0264] In addition, computer-readable recording media may include hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory. Program instructions may include machine code, such as that generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc.

[0265] Some aspects of the present disclosure have been described in the context of a device, but may also be described according to a corresponding method, wherein a block or device corresponds to a method step or a feature of a method step. Similarly, aspects described in the context of a method may also be described according to a corresponding block or item or a feature of a corresponding device. Some or all of the method steps may be performed by (or using) a hardware device, such as, for example, a microprocessor, a programmable computer, or an electronic circuit. In some embodiments, at least one of the most important method steps may be performed by such a device.

[0266] In the embodiments, a programmable logic device (e.g., a field-programmable gate array) may be used to perform some or all of the functions of the methods described herein. In the embodiments, the field-programmable gate array may operate with a microprocessor to perform one of the methods described herein. Generally, it is preferable that the methods be performed by some hardware device.

[0267] Although the present disclosure has been described with reference to preferred embodiments, those skilled in the art will understand that various modifications and changes can be made to the present disclosure without departing from the spirit and scope of the present disclosure as set forth in the following claims.

Claims

1. In the method of user equipment (UE), A step of receiving a first message from a base station that includes one or more monitoring reference signal (monitoring RS) transmission information and monitoring report setting information; A step of measuring one or more monitoring RS received from the base station based on the above RS transmission information; A step of selecting a monitoring RS for evaluating the performance of the AI ​​model from among one or more monitoring RSs having a temporal correlation with inference-related information by the above-mentioned UE-side artificial intelligence (AI) model; A step of generating a monitoring report message using the measurement value of the selected monitoring RS and the inference result of the AI ​​model; and The above monitoring report message includes the step of transmitting the above monitoring report message to a base station via an uplink channel, UE's method.

2. In Claim 1, The selection of the above monitoring RS involves selecting the monitoring RS closest to the inference RS time point within preset time conditions and a specific RS transmission time point set in association with the inference RS or inference report used for inference, UE's method.

3. In Claim 1, The selected monitoring RS is the monitoring RS closest to the inference target instance among the monitoring RSs existing within preset time conditions and the future inference target time instance determined based on the inference of the AI ​​model, UE's method.

4. In Claim 1, The above temporal association includes one or more of temporal proximity, whether a time offset is satisfied, or time conditions set at a base station, UE's method.

5. In Claim 1, The above monitoring RS is transmitted through beams based on the inference results of the above AI model, UE's method.

6. In Claim 1, The transmission type of the above monitoring RS is associated with the transmission type of the RS for the inference of the above AI model, and The transmission type of the above monitoring RS is set to one of periodic, semi-permanent, or non-periodic, UE's method.

7. In Claim 1, The first message further includes first information configured such that the monitoring RS set to a periodic transmission type and the RS for inference of the AI ​​model set to a periodic type are interconnected. UE's method.

8. In Claim 1, The above monitoring report message includes performance evaluation indicators of the AI ​​model configured based on monitoring RS values ​​measured by the same criteria as the quality indicators of the inference results of the AI ​​model, UE's method.

9. In Claim 1, The above monitoring report message includes measurement results of monitoring RS corresponding to the top K beams based on the quality of the beams among a plurality of beams included in the inference result of the AI ​​model, wherein K is a natural number greater than or equal to 1. UE's method.

10. In Claim 1, The above inference-related information includes one or more of the following: a predicted value of beam quality in a future time instance, a beam quality metric in a future time instance, or an identifier for the top X beams available in a future time instance, wherein X is a natural number greater than or equal to 1. UE's method.

11. In user equipment (UE), The above UE includes at least one processor, and the at least one processor is the UE: Receiving a first message from a base station that includes one or more monitoring reference signal (monitoring RS) transmission information and monitoring report setting information; Based on the above RS transmission information, one or more monitoring RSs received from the base station are measured; Select a monitoring RS for evaluating the performance of the AI ​​model from among one or more monitoring RSs that have a temporal correlation with inference-related information by the above UE-side artificial intelligence (AI) model; Generate a monitoring report message using the measurement value of the selected monitoring RS and the inference result of the AI ​​model; and Causing the above monitoring report message to be transmitted to the base station via the uplink channel, UE.

12. In Claim 11, The selection of the above monitoring RS involves selecting the monitoring RS closest to the inference RS time point within preset time conditions and a specific RS transmission time point set in association with the inference RS or inference report used for inference, UE.

13. In Claim 11, The selected monitoring RS is the monitoring RS closest to the inference target instance among the monitoring RSs existing within preset time conditions and the future inference target time instance determined based on the inference of the AI ​​model, UE.

14. In Claim 11, The above temporal association includes one or more of temporal proximity, whether a time offset is satisfied, or time conditions set at a base station, UE.

15. In Claim 11, The above monitoring RS is transmitted through beams based on the inference results of the above AI model, UE.

16. In Claim 11, The transmission type of the above monitoring RS is associated with the transmission type of the RS for the inference of the above AI model, and The transmission type of the above monitoring RS is set to one of periodic, semi-permanent, or non-periodic, UE.

17. In Claim 11, The first message further includes first information configured such that the monitoring RS set to a periodic transmission type and the RS for inference of the AI ​​model set to a periodic type are interconnected. UE.

18. In Claim 11, The above monitoring report message includes performance evaluation indicators of the AI ​​model configured based on monitoring RS values ​​measured by the same criteria as the quality indicators of the inference results of the AI ​​model, UE.

19. In Claim 11, The above monitoring report message includes measurement results of monitoring RS corresponding to the top K beams based on the quality of the beams among a plurality of beams included in the inference result of the AI ​​model, wherein K is a natural number greater than or equal to 1. UE.

20. In Claim 11, The above inference-related information includes one or more of the following: a predicted value of beam quality in a future time instance, a beam quality metric in a future time instance, or an identifier for the top X beams available in a future time instance, wherein X is a natural number greater than or equal to 1. UE.

Images