Method and device for reporting inference results for artificial intelligence and machine learning-based beam management in wireless communication system

Abstract

The present disclosure relates to artificial intelligence / machine learning (AI / ML)-based beam management in a wireless communication system. This operation method of a terminal may comprise: performing first measurement on at least one reference signal; performing, on the basis of the first measurement, first inference on information related to beams applied during a subsequent time interval; transmitting a beam report generated on the basis of a result of the first inference; performing second measurement on the at least one reference signal during the time interval; and transmitting an additional report generated on the basis of the first inference and the second measurement.

Description

Method and apparatus for reporting inference results for artificial intelligence and machine learning-based beam management in a wireless communication system

[0001] The present disclosure relates to AI / ML (artificial intelligence / machine learning)-based beam management in a wireless communication system, and to a method and apparatus for reporting inference results for AI / ML-based beam management.

[0002] Communication networks (e.g., 5G communication networks, 6G communication networks, etc.) are being developed to provide communication services that are improved over existing communication networks (e.g., LTE (long term evolution), LTE-A (advanced), etc.). 5G communication networks (e.g., NR (new radio) communication networks) can support frequency bands above 6 GHz as well as frequency bands below 6 GHz. That is, 5G communication networks can support the FR1 band and / or FR2 band. 5G communication networks can support a wider variety of communication services and scenarios compared to LTE communication networks. For example, usage scenarios for 5G communication networks may include eMBB (enhanced Mobile BroadBand), URLLC (Ultra Reliable Low Latency Communication), mMTC (massive Machine Type Communication), etc.

[0003] 6G communication networks can support a wider variety of communication services and scenarios compared to 5G communication networks. 6G communication networks can meet the requirements for ultra-performance, ultra-bandwidth, ultra-spatial, ultra-precision, ultra-intelligence, and / or ultra-reliability. 6G communication networks can support a wide range of frequency bands and can be applied to various usage scenarios (e.g., terrestrial communication, non-terrestrial communication, sidelink communication, etc.).

[0004] The use of artificial intelligence in the telecommunications sector is expanding. AI can be utilized in various areas, including network operations monitoring, predictive maintenance, network security and fraud prevention, customer service, and intelligent customer relationship management (CRM) systems.

[0005] Furthermore, AI can be used to reduce the complexity of signaling between terminals and base stations or to improve accuracy. To this end, methods are being researched to replace part or all of the signaling for channel measurement, beam management, and positioning, or to enhance performance, by utilizing artificial intelligence models.

[0006] Meanwhile, the technology forming the background of the invention is written to enhance understanding of the background of the invention and may include content that is not prior art already known to a person with ordinary knowledge in the field to which this technology belongs.

[0007] The present disclosure may provide a method and apparatus for effectively performing AI / ML (artificial intelligence / machine learning)-based beam management.

[0008] The present disclosure may provide a method and apparatus for evaluating a model for AI / ML-based beam management.

[0009] The present disclosure may provide a method and apparatus for reporting inference results for AI / ML-based beam management.

[0010] The technical objectives to be achieved in this disclosure are not limited to those mentioned above, and other unmentioned technical problems may be considered by those skilled in the art to which the technical configuration of this disclosure applies, based on the embodiments of this disclosure described below.

[0011] According to one embodiment of the present disclosure, a method of operation of a terminal in a wireless communication system may include performing a first measurement on at least one reference signal, performing a first inference on information related to beams applied during a subsequent time interval based on the first measurement, transmitting a beam report generated based on the result of the first inference, performing a second measurement on at least one reference signal during the time interval, and transmitting an additional report generated based on the first inference and the second measurement.

[0012] According to one embodiment of the present disclosure, a terminal in a wireless communication system comprises at least one transceiver, at least one processor, and at least one memory connected to the at least one processor to be operable and storing instructions that control the terminal to perform operations when executed by the processor, wherein the operations may include performing a first measurement of at least one reference signal, performing a first inference of information regarding beams applied during a subsequent time interval based on the first measurement, transmitting a beam report generated based on the result of the first inference, performing a second measurement of at least one reference signal during the time interval, and transmitting an additional report generated based on the first inference and the second measurement.

[0013] According to the present disclosure, it is possible to effectively perform beam management.

[0014] The effects obtainable from the embodiments of the present disclosure are not limited to those mentioned above, and other unmentioned effects can be clearly derived and understood by a person skilled in the art to which the technical configuration of the present disclosure applies from the description of the embodiments of the present disclosure below. That is, unintended effects resulting from implementing the configuration described in the present disclosure can also be derived by a person skilled in the art from the embodiments of the present disclosure.

[0015] FIG. 1 illustrates a wireless communication system according to an embodiment of the present disclosure.

[0016] FIG. 2 illustrates a block diagram of a communication node according to an embodiment of the present disclosure.

[0017] FIG. 3 illustrates an apparatus for inferring or learning an AI / ML (artificial intelligence / machine learning) model according to an embodiment of the present disclosure.

[0018] FIGS. 4a and 4b illustrate block diagrams of a transmission path and a reception path of a communication node according to an embodiment of the present disclosure.

[0019] FIG. 5 illustrates the structure of a neural network according to an embodiment of the present disclosure.

[0020] FIG. 6 illustrates an AI / ML framework according to an embodiment of the present disclosure.

[0021] FIG. 7 illustrates the life cycle management (LCM) of a network-side model according to an embodiment of the present disclosure.

[0022] FIG. 8 illustrates an example of a procedure for performing inference for AI / ML-based beam management in a wireless communication system according to one embodiment of the present disclosure.

[0023] FIG. 9 illustrates an example of a procedure for beam reporting according to AI / ML-based beam management in a wireless communication system according to one embodiment of the present disclosure.

[0024] FIG. 10 illustrates an example of a procedure for transmitting additional reports for AI / ML-based beam management in a wireless communication system according to one embodiment of the present disclosure.

[0025] FIG. 11 illustrates an example of a procedure for performing additional inference for AI / ML-based beam management in a wireless communication system according to one embodiment of the present disclosure.

[0026] 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 should be understood to include all modifications, equivalents, and substitutions that fall within the spirit and scope of the present disclosure.

[0027] 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" may mean a combination of a plurality of related described items or any of a plurality of related described items.

[0028] In the present disclosure, "at least one of A and B" may mean "at least one of A or B" or "at least one of one or more combinations of A and B". Additionally, in the present disclosure, "at least one of A and B" may mean "at least one of A or B" or "at least one of one or more combinations of A and B".

[0029] In the present disclosure, (re)transmission may mean "transmission," "retransmission," or "transmission and retransmission"; (re)setting may mean "setting," "resetting," or "setting and resetting"; (re)connection may mean "connection," "reconnection," or "connection and reconnection"; and (re)connection may mean "connection," "reconnection," or "connection and reconnection".

[0030] 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.

[0031] 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.

[0032] 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.

[0033] Hereinafter, preferred embodiments of the present disclosure will be described in more detail with reference to the attached drawings. To facilitate overall understanding in describing the present disclosure, the same reference numerals are used for identical components in the drawings, and redundant descriptions of identical components are omitted. Operations according to combinations of embodiments, extensions of embodiments, and / or modifications of embodiments may be performed, as well as the embodiments explicitly described in the present disclosure. The performance of some operations may be omitted, and the order of operations may be changed.

[0034] In the embodiments, even when a method performed at 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 at the first communication node (e.g., reception or transmission of a signal). That is, when the operation of a UE (user equipment) is described, the corresponding base station may perform an operation corresponding to the operation of the UE. Conversely, when the operation of a base station is described, the corresponding UE may perform an operation corresponding to the operation of the base station.

[0035] A base station may be referred to as Node B, evolved Node B, gNode B (next generation node B), gNB, device, apparatus, node, communication node, BTS (base transceiver station), RRH (radio remote head), TRP (transmission reception point), RU (radio unit), RSU (road side unit), radio transceiver, access point, access node, etc. A UE may be referred to as terminal, device, apparatus, node, communication node, end node, access terminal, mobile terminal, station, subscriber station, mobile station, portable subscriber station, OBU (on-broad unit), etc.

[0036] In the present disclosure, signaling may be at least one of upper-layer signaling, MAC signaling, or PHY (physical) signaling. A message used for upper-layer signaling may be referred to as an "upper-layer message" or an "upper-layer signaling message." A message used for MAC signaling may be referred to as a "MAC message" or a "MAC signaling message." A message used for PHY signaling may be referred to as a "PHY message" or a "PHY signaling message." Upper-layer signaling may refer to the transmission and reception operations of system information (e.g., MIB (master information block), SIB (system information block)) and / or RRC messages. MAC signaling may refer to the transmission and reception operations of MAC CE (control element). PHY signaling may refer to the transmission and reception operations of control information (e.g., DCI (downlink control information), UCI (uplink control information), SCI (sidelink control information)).

[0037] In the present disclosure, "setting an operation (e.g., transmission operation)" may mean that "setting information for said operation (e.g., information element, parameter)" and / or "information directing the performance of said operation" is signaled. "Setting an information element (e.g., parameter)" may mean that said information element is signaled. In the present disclosure, "signal and / or channel" may mean signal, channel, or "signal and channel," and signal may be used to mean "signal and / or channel."

[0038] The communication networks to which the embodiments apply are not limited to those described below, and the embodiments may be applied to various communication networks (e.g., 4G communication networks, 5G communication networks, and / or 6G communication networks). Here, the term "communication network" may be used interchangeably with "communication system."

[0039] FIG. 1 illustrates a wireless communication system according to an embodiment of the present disclosure.

[0040] 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). Additionally, the communication system (100) may further include a core network (e.g., S-GW (serving-gateway), P-GW (PDN (packet data network)-gateway), MME (mobility management entity)). If the communication system (100) is a 5G communication system (e.g., NR (new radio) system), the core network may include an access and mobility management function (AMF), a user plane function (UPF), a session management function (SMF), etc.

[0041] Multiple communication nodes (110 to 130) can support communication protocols defined in 3GPP (3rd generation partnership project) standards (e.g., LTE communication protocol, LTE-A communication protocol, NR communication protocol, etc.). Multiple communication nodes (110 to 130) can support CDMA (code division multiple access) technology, WCDMA (wideband CDMA) technology, TDMA (time division multiple access) technology, FDMA (frequency division multiple access) technology, OFDM (orthogonal frequency division multiplexing) technology, Filtered OFDM technology, CP (cyclic prefix)-OFDM technology, DFT-s-OFDM (discrete Fourier transform-spread-OFDM) technology, OFDMA (orthogonal frequency division multiple access) technology, SC (single carrier)-FDMA technology, NOMA (non-orthogonal multiple access) technology, GFDM (generalized frequency division multiplexing) technology, FBMC (filter bank multi-carrier) technology, UFMC (universal filtered multi-carrier) technology, SDMA (space division multiple access) technology, etc. Each of the multiple communication nodes may have the following structure.

[0042] FIG. 2 illustrates a block diagram of a communication node according to an embodiment of the present disclosure. FIG. 2 illustrates an example of a wireless device (200) in a wireless communication system according to an embodiment of the present disclosure. The wireless device (200) according to an embodiment of the present disclosure may be a mobile terminal such as a smartphone, tablet PC, or wearable device, but may not be limited thereto.

[0043] Referring to FIG. 2, the wireless device (200) may include at least one control unit (210), at least one memory (220), at least one power supply unit (230), at least one transceiver unit (240), at least one input unit (250), at least one output unit (260) and / or at least one antenna (270).

[0044] The control unit (210) can control the memory (220) and / or the transmission / reception unit (240) and may be configured to implement the descriptions, functions, procedures, proposals, methods, and / or operation sequences disclosed in this disclosure. The memory (220) may be connected to the control unit (210) and may store various information related to the operation of the control unit (210). For example, the memory (220) may store software code including instructions for performing some or all of the controls controlled by the control unit (210) or for performing the descriptions, functions, procedures, proposals, methods, and / or operation sequences disclosed in this disclosure. The configuration of the memory is not limited in a particular way. For example, it may be configured as at least one of read-only memory (ROM) and random access memory (RAM).

[0045] At least one control unit (210) may be referred to as a controller, microcontroller, microprocessor, or microcomputer. The descriptions, functions, procedures, proposals, methods, and / or flowcharts of operations disclosed in this disclosure may be implemented using firmware or software in the form of code, instructions, and / or sets of instructions. Here, the firmware or software may execute other programs stored in memory (220), such as an OS. The control unit (210) may be implemented to support differently weighted beamforming or directional routing operations to effectively control the outgoing signal from at least one antenna (270) to a desired direction.

[0046] Additionally, at least one control unit (210) may be coupled with a backhaul or network interface. The wireless device (200) may communicate with other wireless devices through the backhaul or network interface. The control unit (210) may include at least one processor. The processor 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.

[0047] At least one transceiver (240) may be connected to a control unit (210) and may transmit and / or receive a wireless signal through at least one antenna (270). The transceiver (240) may include a transmitter and / or a receiver. At least one transceiver (240) may transmit user data, control information, wireless signals / channels, etc., as described in the methods and / or operation flowcharts of the present disclosure to at least one other device. For example, at least one transceiver (240) may be connected to at least one control unit (210) and may transmit and receive wireless signals. Additionally, at least one control unit (210) may control at least one transceiver (240) to transmit user data, control information, or wireless signals to at least one other device. At least one transmitter (240) may receive a signal transmitted by another wireless device from at least one antenna (270). Additionally, at least one transceiver (240) can down-convert or up-convert the received signal to generate a baseband signal. At least one antenna (270) may be a plurality of physical antennas or a plurality of logical antennas (e.g., antenna ports).

[0048] The input unit (250) can acquire information such as user input, video, and audio, and may include various input means such as various mechanical / electronic input means, cameras, and microphones. The output unit (260) is intended to provide information to a user by generating output related to sight, hearing, or touch, and may include a display, speaker, vibration module, etc. The wireless device (200) supplies power through the power unit (230), and the power unit (230) may include a wired / wireless charging circuit, battery, etc.

[0049] 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). Each of the first base station (110-1), the second base station (110-2), and the third base station (110-3) may form a macro cell. Each of the fourth base station (120-1) and the fifth base station (120-2) may 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).

[0050] Here, each of the multiple base stations (110-1, 110-2, 110-3, 120-1, 120-2) may be referred to as NB (NodeB), eNB (evolved NodeB), gNB, ABS (advanced base station), HR-BS (high reliability-base station), BTS (base transceiver station), radio base station, radio transceiver, access point, access node, RAS (radio access station), MMR-BS (mobile multihop relay-base station), RS (relay station), ARS (advanced relay station), HR-RS (high reliability-relay station), HNB (home NodeB), HeNB (home eNodeB), RSU (road side unit), RRH (radio remote head), TP (transmission point), TRP (transmission and reception point), etc.

[0051] Each of the multiple terminals (130-1, 130-2, 130-3, 130-4, 130-5, 130-6) may be referred to as UE (user equipment), TE (terminal equipment), AMS (advanced mobile station), HR-MS (high reliability-mobile station), terminal, access terminal, mobile terminal, station, subscriber station, mobile station, portable subscriber station, node, device, OBU (on board unit), etc.

[0052] 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.

[0053] 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, carrier aggregation (CA) transmission, transmission in an unlicensed band, sidelink communication (e.g., D2D (device to device communication), ProSe (proximity services)), IoT (Internet of Things) communication, dual connectivity (DC), 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.

[0054] 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 sidelink communication 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 sidelink communication by controlling each of the second base station (110-2) and the third base station (110-3).

[0055]

[0056] FIG. 3 illustrates an apparatus for inferring or learning an AI / ML (artificial intelligence / machine learning) model according to one embodiment of the present disclosure.

[0057] The device for inferring or learning the AI / ML model of FIG. 3 may be an embodiment of the wireless device (200) of FIG. 2. Accordingly, the communication unit, control unit, and storage unit of FIG. 3 may correspond to the transceiver unit (240), control unit (210), and storage unit (220) of FIG. 2. For convenience of explanation, the device for inferring or learning the AI / ML model will be described as a server. In addition, if the server is connected to the outside via wireless communication, the communication unit may include the antenna (270) of FIG. 2. The control unit may perform AI / ML data learning or inference. The AI ​​model may be implemented in various forms. As an example, the AI ​​model may be a model based on a neural network. For the neural network model, models such as a deep neural network (DNN), a recurrent neural network (RNN), or a bidirectional recurrent deep neural network (BRDNN) may be used.

[0058] The communication unit can communicate with external devices, and the server can receive various information through the communication unit. Accordingly, the communication unit can receive monitoring information, user input information, etc., for training AI / ML models and transmit them to the control unit. To this end, the communication unit can perform wired or wireless communication with external devices. The server can transmit and receive signals to and from external devices, such as satellites, mobile devices, and autonomous vehicles, through the communication unit. Wireless communication may include cellular communication, short-range wireless communication, or GNSS (global navigation satellite system) communication.

[0059] The storage unit can store training models, input data, output data, etc. Accordingly, the storage unit can store AI models trained by the control unit or updated AI, and can provide data if necessary according to the commands of the control unit.

[0060] The control unit may include an AL / ML management unit, a model training unit, and a model inference unit. The AL / ML management unit can manage the communication unit, the model inference unit, and the model training unit so that inference or training tasks can be performed efficiently. For example, the management unit can receive monitoring information from the communication unit and output from the model inference unit, and then verify the performance of the inference. The management unit can transmit the verified performance to the model training unit and instruct the model training unit to train an AI / ML model based on that performance. That is, the management unit can transmit performance feedback information to the model inference unit, and the model inference unit can use the performance feedback information to instruct the learning goal or utilize it for reinforcement learning rewards, etc. Additionally, the control unit can instruct the server or external device to use an AI / ML model, or instruct the activation or deactivation of the use of the AI / ML model. For convenience of explanation, FIG. 3 is illustrated as the server including both the model training unit and the model inference unit; however, depending on the purpose of the server, it may include only one of the model training unit or the model inference unit. In addition, the structure of Fig. 3 described above can be applied not only to servers but also to other devices where inference or learning takes place, such as terminals, wireless devices, autonomous vehicles, and mobile devices.

[0061]

[0062] FIGS. 4a and 4b illustrate block diagrams of a transmission path and a reception path of a communication node according to an embodiment of the present disclosure. The transmission path (410) exemplified in FIG. 4a may be implemented in a communication node that transmits a signal, and the reception path (420) exemplified in FIG. 4b may be implemented in a communication node that receives a signal.

[0063] Referring to FIGS. 4a and 4b, the transmission path (410) may include a channel coding and modulation block (411), an S-to-P (serial-to-parallel) block (512), an N IFFT (Inverse Fast Fourier Transform) block (413), a P-to-S (parallel-to-serial) block (414), a CP (cyclic prefix) addition block (415), and an UC (up-converter) (UC) (416). The reception path (420) may include a DC (down-converter) (421), a CP removal block (422), an S-to-P block (423), an N FFT block (424), a P-to-S block (425), and a channel decoding and demodulation block (426). Here, N may be a natural number.

[0064] Information bits in the transmission path (410) can be input to the channel coding and modulation block (411). The channel coding and modulation block (411) can perform coding operations (e.g., LDPC (low-density parity check) coding operations, polar coding operations, etc.) and modulation operations (e.g., QPSK (Quadrature Phase Shift Keying), QAM (Quadrature Amplitude Modulation), etc.) on the information bits. The output of the channel coding and modulation block (411) may be a sequence of modulation symbols.

[0065] The S-to-P block (412) can convert modulated symbols in the frequency domain into parallel symbol streams to generate N parallel symbol streams. N can be the IFFT size or the FFT size. The N IFFT block (413) can generate signals in the time domain by performing an IFFT operation on the N parallel symbol streams. The P-to-S block (414) can convert the output of the N IFFT block (413) (e.g., parallel signals) into a serial signal to generate a serial signal.

[0066] The CP addition block (415) can insert CP into the signal. The UC (416) can up-convert the frequency of the output of the CP addition block (415) to an RF (radio frequency) frequency. Additionally, the output of the CP addition block (415) can be filtered in the baseband before up-conversion.

[0067] A signal transmitted from the transmission path (410) can be input to the reception path (420). The operation in the reception path (420) may be the inverse operation of the operation in the transmission path (410). The DC (421) may down-convert the frequency of the received signal to a baseband frequency. The CP removal block (422) may remove CP from the signal. The output of the CP removal block (422) may be a serial signal. The S-to-P block (423) may convert the serial signal into parallel signals. The N FFT block (424) may generate N parallel signals by performing an FFT algorithm. The P-to-S block (425) may convert the parallel signals into a sequence of modulation symbols. The channel decoding and demodulation block (426) may perform a demodulation operation on the modulation symbols and restore data by performing a decoding operation on the result of the demodulation operation.

[0068] In FIGS. 4a and 4b, Discrete Fourier Transform (DFT) and Inverse DFT (IDFT) may be used instead of FFT and IFFT. In FIGS. 4a and 4b, each of the blocks (e.g., components) may be implemented by at least one of hardware, software, or firmware. For example, in FIGS. 4a and 4b, some blocks may be implemented by software, and the remaining blocks may be implemented by hardware or a "combination of hardware and software." In FIGS. 4a and 4b, one block may be subdivided into multiple blocks, multiple blocks may be integrated into one block, some blocks may be omitted, and blocks supporting other functions may be added.

[0069]

[0070] FIG. 5 illustrates the structure of a neural network according to an embodiment of the present disclosure. In a neural network, the input layer is the first layer that receives external data. The input layer can receive raw data of various forms, feature it, and then transmit it. Here, feature refers to extracting characteristic parts of the raw data. The number of neurons in the input layer can be determined based on the characteristics and requirements of the given data.

[0071] In a neural network, hidden layers can consist of one or more layers located between the input layer and the output layer. Therefore, unlike in Fig. 5, the hidden layer can consist of two or more layers. The neurons in the hidden layer can represent the transformed form of the input data, and the depth and width of the hidden layer can determine the complexity of the AI / ML model. A non-linear relationship between the input layer and the output layer can be implemented using the hidden layer.

[0072] The output layer is the final layer of the neural network and can provide the final result. Based on the final result of the output layer, the error between the neural network's prediction and the actual target value can be calculated.

[0073] Therefore, data input through the input layer can be transmitted to the output layer via a hidden layer consisting of one or more layers. During this process, features of the input data can be extracted. Neural networks can be trained using algorithms such as backpropagation and gradient descent. Training can be provided through feedback by rewarding errors calculated in the output layer. Based on this feedback, the weights of each layer can be updated. This training can be performed iteratively to optimize the inference of AI / ML models.

[0074]

[0075] Terms related to AI used in this disclosure may be defined as follows.

[0076] AI / ML-enabled features refer to specific functions that enable the use of AI / ML. For example, CSI measurements, beam management, and positioning functions may fall under the category of AI / ML-enabled features.

[0077] An AI / ML model refers to an algorithm that applies AI / ML technology to generate a set of outputs based on a set of inputs. In this disclosure, an AI / ML model may refer not to the algorithm itself, but to a set of parameter values ​​for defining the AI / ML model. For example, an AI / ML model may be represented as a set of weights of a neural network.

[0078] AI / ML model delivery refers to the transfer of an AI / ML model from one entity to another. Here, entities can refer to network nodes / functions (e.g., gNBs, LMFs, etc.), UEs, independent servers, etc. Therefore, an AI / ML model trained on a first entity can be transferred to a second entity.

[0079] AI / ML model inference refers to the process of generating a set of outputs based on a set of inputs using a trained AI / ML model within a single entity.

[0080] AI / ML model testing is the process of evaluating the performance of a final AI / ML model using a dataset different from those used for training and validation, and can be included as a sub-process of training. Unlike AI / ML model validation, testing refers to further tuning the model.

[0081] AI / ML model training refers to the process of training an AI / ML model in a data-driven manner to learn input / output relationships and obtain a trained AI / ML model that can be used for inference.

[0082] AI / ML model transfer refers to the transmission of an AI / ML model from a sending side to a receiving side. This transmission can occur wirelessly or via a wired connection, and may include parameters of a model structure known to the receiving side or parameters related to a new model. The transmitted parameters may include those for the entire model or a portion of the model.

[0083] AI / ML model validation is a sub-process of training that evaluates the quality of an AI / ML model using a dataset different from the one used for training. It can be used to determine model parameters for generalized datasets as well as for the dataset used for training.

[0084] Data collection refers to network nodes, management entities, or UEs collecting data for AI / ML model training, data analysis, and inference.

[0085] Federated learning or federated training refers to a machine learning technique that trains AI / ML models by performing local model training using local data at multiple distributed edge nodes (e.g., UEs, gNBs). Federated learning may require various interactions between nodes related to edge AI / ML models. In this process, the exchange of local data is not essential.

[0086] Functionality identification refers to the process of identifying AI / ML functionalities so that they can be commonly recognized between the network and the UE. Information regarding AI / ML functionalities can be shared during the functionality identification process.

[0087] Management instructions refer to the information necessary to ensure appropriate inference operations. Management instructions may include the selection, (de)enable, or switching of AI / ML models or functions, AI / ML operations, as well as other alternative operations.

[0088] Model activation refers to activating an AI / ML model for a specific AI / ML support function, and model deactivation refers to deactivating an AI / ML model for a specific AI / ML support function.

[0089] Model download refers to the transmission of an AI / ML model from the network to the UE, and the UE receiving the AI / ML model. Model identification refers to identifying an AI / ML model based on functional identifications commonly recognized between the network and the UE. Information regarding the AI / ML model may be shared during model identification. Model upload refers to the transmission of a model from the UE to the network.

[0090] Model monitoring refers to the procedure of monitoring the inference performance of an AI / ML model, and model parameter updating refers to the process of updating the model's parameters.

[0091] Model selection refers to choosing one of several models to activate for a single AI / ML support function. Model selection can occur simultaneously with model activation. Model switching refers to deactivating the currently activated AI / ML model for a specific AI / ML support function and activating a different AI / ML model. Model update refers to the process of updating a model's parameters and / or structure.

[0092] AI / ML models can be classified based on the location of inference. A network-side AI / ML model refers to an AI / ML model where inference is performed entirely within the network. A two-sided AI / ML model refers to a pair of AI / ML models where inference is performed jointly by the UE and the network. Joint inference means that some inference is performed first in the UE, and the remainder is performed in the gNB. A UE-side AI / ML model refers to an AI / ML model where inference is performed entirely within the UE.

[0093] The following learning methods can be utilized in relation to AI / ML training. Reinforcement Learning (RL) refers to the process of training an AI / ML model based on inputs (i.e., states) and feedback signals (e.g., rewards) generated by the model's outputs (e.g., actions) in an environment where the model interacts. Semi-supervised learning refers to the process of training a model by mixing labeled and unlabeled data. Supervised learning refers to the process of training a model from inputs and their corresponding labels. Unsupervised learning refers to the process of training a model using unlabeled data.

[0094]

[0095] The following describes the initial connection procedure between a terminal and a base station. When the initial connection procedure with the base station is performed due to reasons such as the terminal's power on / off operation or out of coverage, an identification procedure between the base station and the terminal may be required. First, the terminal may perform an initial cell search operation with the base station. The terminal may perform monitoring to receive a synchronization signal. The synchronization signal may be at least one of a PSS (primary synchronization signal) or a SSS (secondary synchronization signal). The terminal may obtain broadcast information within the cell by receiving a physical broadcast channel (PBCH) signal from the base station. Based on the physical broadcast channel, the terminal may obtain information about the cell using at least one of an MIB or a SIB. A block containing the PSS, SSS, and PBCH may be referred to as an SSB (synchronization signal block).

[0096] The terminal can perform a random access procedure. The terminal can transmit a preamble to the base station and receive a random access response (RAR) from the base station. The RAR message may include a temporary identifier. The terminal can transmit an MSG3 (or RRC connection request message) using the scheduling information within the RAR, and the base station can perform a contention resolution procedure by transmitting an MSG4 (or contention resolution message) to the terminal in response to the MSG3.

[0097] A base station can perform beam management based on RACH occasions used for transmitting preambles in random access procedures. For example, a base station can identify the beam in which a terminal received a synchronization signal based on the RACH occasion where the preamble was transmitted. A synchronization signal may also be included as a reference signal to indicate QCL relationships, and QCL relationships may be established based on the SSB received through the initial access procedure.

[0098] Additionally, a channel measurement procedure may be performed for beam management. The terminal may receive a reference signal from the base station. Based on this, the terminal may report channel state information (CSI) to the base station. The channel state information may include at least one of the following: reference signal received power (RSRP), reference signal received quality (RSRQ), and signal-to-noise ratio (SNR). The base station may use the received channel state information to adjust beamforming for the terminal or to optimize radio resource allocation. For channel measurement, the base station may transmit configuration information for channel measurement to the terminal. The configuration information for channel measurement may include information related to the measurement target, the measurement period, etc.

[0099]

[0100] Enhancing CSI Feedback Using AI

[0101] Terminals and base stations can utilize AI to perform channel state estimation and channel state reporting procedures. When AI is used, the CSI feedback process can be enhanced. Typically, AI models can be used for spatial-frequency domain CSI compression. Bifacial models can be used for spatial-frequency domain CSI compression. Preprocessing, postprocessing, quantization, and dequantization processes can be included in the CSI compression process. AI / ML-based CSI compression can be performed based on existing frameworks, similar to the aforementioned CSI feedback procedure. Additionally, AI models can be used for time-domain CSI prediction. Since CSI measurements are taken at the terminal, UE-side models can be used.

[0102] For CSI compression using two-sided models, three types of AI / ML model training collaboration methods can be considered. The first type of training collaboration refers to a type in which two-sided models are jointly trained on a single side or entity, and can be performed on either the UE side or the network side. Which side to train on can be determined based on at least one of the following: whether the model can be maintained exclusively, privacy protection, support for specific models, device-specific optimization, flexibility for model updates, the possibility of developing / updating each model, whether an integrated CSI reset model for different UEs is built, whether an integrated CSI generation model for different networks is built, scalability, and model performance.

[0103] The second training cooperation type involves jointly training a bilateral model on both the network side and the UE side, respectively. The third training cooperation type involves separate training on the network side and the UE side, where the UE side trains the CSI generation part and the network side trains the CSI reconfiguration part. For joint training, the generative and reconfiguration models must be trained in the same loop using forward and backpropagation, and this can be performed at least one node. Separate training refers to sequential training initiated by either the UE side or the network side.

[0104] In the second training cooperation type, joint training can include both simultaneous training and sequential training. Sequential training in the second training cooperation type can begin with network-side training.

[0105] In the second training cooperation type, the choice between concurrent training or sequential training may be determined based on at least one of the following: whether the model can be maintained exclusively, privacy protection, flexibility regarding support for specific models, device-specific optimization, flexibility for model updates, the possibility of developing / updating each model, whether an integrated CSI reset model for different UEs is established, whether an integrated CSI generation model for different networks is established, scalability, whether the training data distribution can match the inference device, software / hardware compatibility, and model performance. In the third training cooperation type, whether the training is performed first on a network or a UE may be determined based on the possibility of maintaining exclusivity, privacy protection, etc., similar to the conditions considered in the second training cooperation type.

[0106] Data generation for AI models can be implemented in various ways. For example, in a CSI compression use case, training data for model training can be generated by the UE / gNB. For the network portion of bilateral model inference, input data can be generated by the UE and terminated at the base station. For the UE portion of bilateral model inference, input data can be used internally within the UE.

[0107] Additionally, for network-side performance monitoring, calculated performance metrics or data for calculating performance metrics may be generated by the UE and terminated at the base station if necessary. In use cases for CSI compression using a bifacial model, pairing information based on model identification may be configured to select a CSI generation model compatible with the CSI reconstruction model used at the base station.

[0108] CSI prediction using AI can be applied as follows. For model training, training data may be generated by the UE. For UE-side model inference, input data may be used internally by the UE. Performance metrics required for network-side performance monitoring, or data for calculating performance metrics, may be generated by the UE and terminated in the gNB.

[0109]

[0110] Beam management using AI

[0111] The beams of the base station and terminal can manage beam settings through AI. For convenience, Beam Set B refers to the beam set for which measurements are performed as input to an AI / ML model, and Set A refers to the beam set determined based on the inference of the AI / ML model. Beam Set A and Beam Set B may contain beam information regarding the same frequency range.

[0112] AI can be used to infer spatial domain downlink beams for beam set A based on measurements of beam set B. As another example, AI can be used to infer temporal downlink beams for beam set A based on historical measurement results of beam set B. Here, beam set A and beam set B may be different sets, or beam set B may be a subset of beam set A.

[0113] Additionally, the input to the AI ​​model can be formed in various combinations. For example, the input to the AI ​​may include at least one of an L1-RSRP measurement measured based on beam set B, other auxiliary information, a CIR (channel impulse response) based on beam set B, and a downlink Tx / Rx beam identifier (ID) associated with the L1-RSRP measurement of beam set B.

[0114] The aforementioned AI model can be designed to infer a beam including at least one of a downlink receiving beam and a downlink transmitting beam. Additionally, the output of the AI ​​model may include at least one of a transmitting beam, a receiving beam, the L1-RSRP of the transmitting beam, the L1-RSRP of the receiving beam, the angle of the transmitting beam, the angle of the receiving beam, and other information.

[0115] The beam management method using an AI model is not limited to the method described above. The AI ​​model can be configured in various ways by configuring inputs and outputs with various combinations of settings, performance monitoring, data collection, auxiliary information, etc., regarding beam set A and beam set B.

[0116] Furthermore, learning and inference methods can be implemented in various ways. For example, AI can be learned or trained using AI / ML models. Learning and training can be performed by a network or a UE. Additionally, learning and inference can be performed on different devices. For example, learning can be performed on a network and inference on a terminal. Partial learning can be performed in such a way that part of the learning is performed on a first device and another part on a second device. Similarly to learning, partial inference can be performed using multiple devices. Input data for inference can also be generated in various ways. For example, input data can be generated in the UE, and inference can be performed using the input data in a network. Additionally, input data generated in the UE can be processed internally within the UE to enable inference.

[0117]

[0118] AI-enhanced positioning method

[0119] Terminals and base stations can perform positioning procedures to determine the location of the terminal using AI. To improve positioning accuracy, methods for directly determining the location through AI / ML models and methods for determining an auxiliary location may be considered.

[0120] In a method where the location is determined directly through an AI / ML model, the location of the UE is output, and a channel fingerprint based on channel observation can be used as input to the AI / ML model. The fingerprinting method is a positioning technique based on probabilistic modeling that utilizes noise and surrounding environment information as information for location tracking. Therefore, the fingerprinting method utilizes existing devices such as wireless APs (Access Points) to construct a fingerprint map based on signal strength values, and the location of the terminal can be determined based on the channel fingerprint generated by the terminal observing the channel.

[0121] In a method where auxiliary position is determined through an AI / ML model, new measurements and / or enhanced values ​​of existing measurements are output, and LOS / NLOS identification, timing and / or angle of measurement, and measurement feasibility can be input.

[0122] Specifically, the following methods may be considered.

[0123] - UE-based localization, direct AI / ML using UE-side models, or AI / ML-assisted localization

[0124] - UE-assisted / LMF-based positioning, AI / ML-assisted positioning using UE-side models

[0125] - UE-assisted / LMF-based localization, direct AI / ML localization using LMF-side models

[0126] - NG-RAN node-assisted location determination, AI / ML-assisted location determination using gNB-side models

[0127] - NG-RAN node-assisted location determination, direct AI / ML location determination using LMF side models

[0128] The method of generating data for AI models can be implemented in various ways. Specifically, for training AI / ML models for positioning, training data can be generated by the UE, PRU, gNB, or LMF. For model inference on the LMF side, input data can be generated by the UE or gNB and terminated within the LMF. For model inference on the gNB side, input data can be used internally by the gNB. For model inference on the UE side, input data can be used internally by the UE. For performance monitoring on the LMF side, if necessary, calculated performance metrics or data for calculating performance metrics can be generated by the UE or gNB and terminated within the LMF. For performance monitoring on the gNB side, if necessary, calculated performance metrics or data for calculating performance metrics can be generated by at least the gNB.

[0129]

[0130] Among the discussion topics related to AI / ML for NR air interface to be discussed at Rel-19 RAN (radio access network) WG1 (working group 1), the contents of the WID (work item description) document for AI / ML based beam management are as shown in [Table 1] below.

[0131] Provide specification support for the following aspects:....- Beam management - DL Tx beam prediction for both UE-sided model and NW-sided model, encompassing [RAN1 / RAN2]:o Spatial-domain DL Tx beam prediction for Set A of beams based on measurement results of Set B of beams ("BM-Case1")o Temporal DL Tx beam prediction for Set A of beams based on the historic measurement results of Set B of beams ("BM-Case2")o Specify necessary signalling / mechanism(s) to facilitate LCM operations specific to the Beam Management use cases, if anyo Enabling method(s) to ensure consistency between training and inference regarding NW-side additional conditions (if identified) for inference at UENOTE: Strive for common framework design to support both BM-Case1 and BM-Case2

[0132] At 3GPP Rel-18 RAN WG1, discussions were conducted at the SI (study item) stage to develop standards utilizing AI / ML. Specifically, studies were conducted on three major technical categories: CSI feedback, beam management, and positioning accuracy enhancement. Among these, beam management and positioning accuracy enhancement are scheduled to be discussed at the WI (working item) stage in Rel-19. [Table 1] shows the scope of development for AI / ML-based beam management to be discussed in Rel-19. As mentioned above, Rel-19 considers DL transmit (Tx) beam prediction, and UE-sided and NW-sided models are considered as the AI / ML models used here. Here, the UE-side model and the NW-side model may have a structure in which the UE and the NW each perform at least one operation for an AI / ML procedure, rather than utilizing an AI / ML model combined between the UE and the NW. The detailed scope of development includes the following four topics.

[0133] - Spatial Domain Beam Prediction for Set A Beam Based on Set B Beam Measurement Results: BM-Case1

[0134] - Time-domain beam prediction for Set A beam based on Set B beam measurement results: BM-Case2

[0135] - Signaling and mechanisms required for LCM (life cycle management) operation related to beam management

[0136] - A method to ensure continuity between training and inference regarding additional NW conditions for inference at the UE level

[0137] Additionally, standardization aims to develop a common framework that can support both BM-Case1 and BM-Case2 as much as possible.

[0138] Through AI / ML research at 3GPP Rel-18, document TR 38.843 was drafted, and this document covers NR radio interfaces based on AI / ML. Specifically, the standard defines an AI / ML framework to be commonly used in AI / ML for NR radio interfaces, and includes descriptions of three representative use cases (e.g., CSI feedback, positioning accuracy enhancement, beam management, etc.), performance evaluation results, and expected specification changes. Subsequently, Rel-19 RAN WG1 plans to develop standards for positioning accuracy enhancement and beam management, as well as conduct further research on CSI-RS feedback. The aforementioned AI / ML framework is shown in Fig. 6.

[0139] FIG. 6 illustrates an AI / ML framework according to an embodiment of the present disclosure. Depending on the entity where operations for an AI / ML framework such as FIG. 6 are physically performed, AI / ML models are classified into NW-side AI / ML models, UE-side AI / ML models, and two-side AI / ML. Among these, NW-side AI / ML models and UE-side AI / ML models process all AI / ML operations at either the base station or the terminal, while two-side AI / ML models perform AI / ML operations jointly at the base station and the terminal. Consequently, two-side AI / ML models require a much larger amount of data transmission for AI / ML compared to NW-side AI / ML models and UE-side AI / ML models, and may have much greater complexity. Therefore, standardization for NW-side AI / ML models and UE-side AI / ML models is expected in Rel-19.

[0140] In AI / ML-based beam management, beams are classified into multiple sets. Beam set A, beam set B, and beam set C can be defined, and their respective technical meanings can be defined as shown in [Table 2] below.

[0141] Clause 5.2.1 in TR 38.843… The following are selected as representative sub-use cases:- BM-Case1: Spatial-domain Downlink beam prediction for Set A of beams based on measurement results of Set B of beams...- BM-Case2: Temporal Downlink beam prediction for Set A of beams based on the historic measurement results of Set B of beams...Set B is a set of beams whose measurements are taken as inputs of the AI / ML model.… Clause 6.3.2.3 in TR38.843… - (Opt 2D) For the case that Set B of beams (pairs) is a subset of measured beams (pairs) Set C (where Set C is fixed across training and inference), compared to the case with all measurements of measured beam Set C as AI inputs- with Top K=1 / 2 of the measurements of Set C,…

[0142] To summarize, a set of beams A (hereinafter 'Beam Set A') includes a set of beams to be predicted via AI / ML, a set of beams B (hereinafter 'Beam Set B') includes a set of beams to be used as input data for AI / ML to predict beams within Beam Set A, and a set of beams C (hereinafter 'Beam Set C') includes a set of beams for which beam measurements are performed. Here, the beams included in Beam Set C may or may not be used as input data for AI / ML.

[0143] For the aforementioned beam set A, beam set B, and beam set C, beam set B may be a subset of beam set C. The specification does not restrict the inclusion relationship between beam set A, beam set B, and beam set C, but cases in which beam set A includes beam set B and / or beam set C may be considered.

[0144] The present disclosure describes a technique for data collection in beam management based on an NW-side model. In AI / ML-based beam management using an NW-side model, information transfer between the NW and the UE takes place as shown in FIG. 7. FIG. 7 illustrates the life cycle management (LCM) of an NW-side model according to an embodiment of the present disclosure. In FIG. 7, the NW can be understood to include a base station, a gNB, etc.

[0145] Referring to Fig. 7, the NW transmits a reference signal (RS) to the UE, the UE transmits a report of collected data to the NW, and the NW transmits a beam indication. At this time, the NW can perform model training, performance monitoring, and model inference using the collected data. That is, since it is a NW-side model, there are no AI / ML-related operations in the UE, and the UE is responsible for receiving the reference signal transmitted by the NW and performing reporting. On the other hand, the NW performs model training, performance monitoring, and model inference by distributing the data reported from the terminal, and transmits a beam indication to the terminal based on the beam prediction, which is the result of the model inference.

[0146] For data acquisition, the NW can configure reference signal resources (e.g., SSB resources or CSI-RS resources) for L1-RSRP measurements. For example, the NW can configure reference signal resources for Beam Set A and / or Beam Set B. The UE performs measurements based on the configured reference signal resources and reports the collected data under the control of the NW. From the NW's perspective, the collected data can be used for model training, model inference, or performance monitoring. For model training, the collected data can be passed on for offline learning or fine-tuning, or processed. For model inference, the NW can use the collected data as model input and generate outputs, which can be further used for beam directing. The collected data can be used to monitor model performance, and the NW can decide whether to enable, disable, or fallback AI / ML operations based on performance monitoring.

[0147] AI / ML performed on the network operates based on beam measurement reports from the terminal. The network can classify the information reported from the terminal into data for model training, data for model inference, and data for performance monitoring. Data for model training is used to train AI / ML models and is characterized by a relatively large amount of data for reporting beam measurement results and long latency requirements. In contrast, data for model inference requires relatively short latency requirements to generate beam prediction results through AI / ML in near real-time. The amount of data for reporting beam measurement results based on model inference may be less than or equal to the amount of data for model training. Finally, data for performance monitoring may include as little as the data for model inference, as much as the AI / ML model training data, or even more.

[0148] Among the approvals from recent working group meetings (e.g., RAN1 #116bis, #117, #118, #118bis), the approvals [Table 3] regarding the reporting of inference results for UE-sided model BM Case-2 were decided.

[0149] RAN1 #116bisAgreementFor UE-side AI / ML model inference, for BM-Case2, support to report inference results of N(N>=1, FFS on N) future time instance(s) in one report· wherein information of inference results of one time instance is as in one report for BM-Case 1.o Note: overhead reduction is not precluded.· FFS on detailsAgreementFor report content of inference results for UE-sided model for BM-Case 1, for the RSRP of predicted Top K beam(s) in the report of inference results, when applicable, further study the following options:· Option A: Predicted RSRP.· Option B: Predicted RSRP, if the beam is not configured for corresponding measurement, and measured L1-RSRP if the beam is configured for corresponding measurement.· Where the predicted RSRP is based on AI / ML output.· Note: Support both Option A and Option B is not precluded.Working AssumptionFor report content of inference results for UE-sided model for BM-Case 2, the RSRP of predicted beam(s) in the report of inference results, is the predicted RSRP, where the predicted RSRP is based on AI / ML output.AgreementFurther study, for the consistency of NW-side additional condition across training and inference for UE-sided model for BM-Case 1 and BM Case 2, where the NW-side additional condition may at least impact UE assumption on beams of Set A / Set B:· Opt1: Based on associated ID (Referring to AI 9.1.3.3)· FFS on what can be assumed by UE with the same associated ID across training and inference· FFS on how associated ID is introduced, e.g., within CSI framework, or outside of CSI framework· Opt 2: Performance monitoring based· FFS details· Other options are not precluded.RAN1 #117AgreementFollowing Working Assumption is confirmed.Working AssumptionFor report content of inference results for UE-sided model for BM-Case 2, the RSRP of predicted beam(s) in the report of inference results, is the predicted RSRP, where the predicted RSRP is based on AI / ML output.RAN1 #118AgreementFor UE-sided model, for the quantization of a RSRP value at least for the report of inference results, support· Support differential RSRP reporting with legacy quantization step and range for L1-RSRP reportingo For BM-Case 1, support differential RSRP report among multiple beamso For BM-Case 2, support differential RSRP report among multiple beams over multiple time instances♣ FFS detailsAgreementFor UE-sided model for BM-Case 2, for inference results report, support to configure UE with N future time instance(s) for inference by NW when applicable· FFS: how to determinate reference time for the time instance(s).· FFS: duration values of the N time instance(s) that can be predicted.RAN1 #118bisAgreementFor BM-Case 2 of UE-side model, for the reference time of the earliest time instance for the predicted results, consider at least the following alternatives for potential down-selection: · Option 1: Based on the uplink slot for the report · Option 2: Based on the CSI reference resource corresponding to the report · Option 3: Based on the latest transmission occasion of the CSI-RS / SSB resource in Set B for measurement for the report, wherein the transmission occasion is no later than the CSI reference resourceAgreementFor beam management, multiple CSI reports for inference for UE-side model can be configured / activated / triggered, which is up to UE capability.

[0150] The details regarding the inference report for the UE-side model's BM Case-2 in the approval details as shown in [Table 3] are summarized as follows.

[0151] - The terminal reports the inference result of the UE-side model for N future time instances.

[0152] - Among the results inferred from the UE-side model, reporting is performed on the top K beam(s).

[0153] - The report includes the predicted RSRP.

[0154]

[0155] The time instances in the aforementioned 'N future time instances' can be defined in various ways. For example, time instances can be defined as time intervals of 20ms, 40ms, 80ms, 160ms, or longer. As another example, time instances can be defined as non-periodical.

[0156]

[0157] The process between the terminal and the base station when BM Case-2 is performed can be represented as shown in FIG. 8 below. FIG. 8 illustrates an example of a procedure for performing inference for AI / ML-based beam management in a wireless communication system according to one embodiment of the present disclosure.

[0158] Referring to FIG. 8, through a dedicated resource, the terminal reports inference results for N future time instance(s). Subsequently, based on the inference results, communication is initiated between the base station and the terminal. During the time interval corresponding to the N time instance(s), the most recent inference result is applied. Through the next dedicated resource, the terminal reports inference results for the next N future time instance(s). The aforementioned operations are repeated.

[0159] The above-described procedure may be modified according to various embodiments. For example, a dedicated resource for reporting inference results and / or a time interval for applying inference results (e.g., the interval between time point A1 and time point B1 in FIG. 8) may differ from a time interval over N time instance(s), and specifically, the interval between time points for reporting inference results may be greater or smaller than the interval between instances. Additionally, the time point for reporting inference results and the time point for starting to apply inference results may be the same, different, or different but nearly similar.

[0160] As described above, various forms of processes and procedures can be defined for BM Case-2. The present disclosure considers FIG. 8 as a representative example. Multiple bursts containing reference signals (e.g., SSB, CSI-RS, etc.) may be transmitted during N future time instance(s), such as between time point A2 and time point B2 in FIG. 8, or between time point A1 and time point B1. At this time, the terminal may obtain additional inference results based on a measurement performed based on the reference signal received after the inference result report at time point A1.

[0161] As shown in FIG. 8, if the time interval corresponding to N time instances is sufficiently long, the terminal can receive multiple additional reference signals and perform measurements within that time interval. That is, by receiving and measuring additional reference signals, the terminal can perform performance monitoring of inference. Through this, for example, in the case of a UE-side model where AI / ML computation is performed on the terminal, if a measurement of the reference signal is made, the terminal can produce a new inference result for future time interval(s).

[0162] If the UE-side model is sufficiently well trained, the inference result reflecting the measurement of an additional reference signal is expected to be similar to the inference result reported at time point A1. In other words, the inference result for future time instances inferred at time points C, D, and E in Fig. 8 may be similar to the result reported at the existing time point A1. However, if the UE-side model is not sufficiently trained or if there is a significant change in the channel between the transmitter and receiver, the measurement of a new reference signal may result in an inference result with a significant difference. That is, there is a possibility that the inference result for the future time instance estimated at time point D may be significantly different from the result reported at the existing time point A1.

[0163] Additionally, if the UE-side model is sufficiently well trained, the optimal or superior beam(s) identified based on the measurement of an additional reference signal are expected to be included in at least one beam reported as an inference result at time point A1. In other words, the superior beam(s) selected based on new measurements at time points C, D, and E in Fig. 8 may be included in the result reported at the original time point A1. However, if the UE-side model is not sufficiently trained or if there is significant channel variation between the transmitter and receiver, the superior beam(s) selected based on the measurement of the new reference signal may differ from the previously reported inference result. That is, the superior beam(s) selected based on the measurement at time point D may differ from at least one beam included in the result reported at time point A1.

[0164] The present disclosure proposes a response method for cases where additional reporting other than the dedicated inference report is required after the inference report in BM Case-2 of the UE-side model.

[0165]

[0166] FIG. 9 illustrates an example of a procedure for beam reporting according to AI / ML-based beam management in a wireless communication system according to one embodiment of the present disclosure. FIG. 9 illustrates a method performed by a terminal.

[0167] Referring to FIG. 9, in step S901, the terminal performs a first measurement for at least one reference signal. The terminal receives at least one reference signal to obtain a measurement result which is input data for inference, and performs a first measurement for at least one. To do this, the terminal receives configuration information for at least one reference signal from a base station and can receive at least one reference signal from a resource indicated by the configuration information.

[0168] In step S903, the terminal performs inference regarding the beams based on the first measurement. In other words, the terminal can perform inference regarding information related to the beams applied during a subsequent specified or set time interval based on the first measurement. For example, the terminal can predict the quality of the beams (e.g., RSRP) during the subsequent time interval using a terminal-side AI / ML model. Through inference, the terminal can determine at least one of the quality of the beams and the rank of the beams that can be used for scheduling during the subsequent time interval.

[0169] In step S905, the terminal transmits a beam report generated based on inference. The beam report may include at least one of information indicating the quality of the beams, information indicating the optimal beam(s), and information indicating the ranking of the beams. For example, the information indicating the optimal beam(s) may indicate the top-K beam(s) regarding quality. Accordingly, the base station may transmit a signal to the terminal using at least one of the top-K beams during the time interval in which the inference result is applied.

[0170] In step S907, the terminal performs a second measurement on at least one reference signal. The second measurement may be performed during a time interval in which the result of the previously performed inference is applied. That is, the terminal may perform the second measurement to monitor the performance of the result of the inference. Accordingly, the terminal can determine whether the previously selected top-K beam(s) are beams that still provide a relatively high RSRP.

[0171] In step S909, the terminal transmits an additional report generated based on inference and the second measurement. The additional report may include information determined based on a comparison between the result of the inference and the result of the second measurement. That is, the additional report may include information for feedback on the results of an evaluation or monitoring of the performance of the inference. As a specific example, the additional report may include information indicating whether at least one top-M beam(s) identified through the second measurement are included in the beam set included in the result of the inference, i.e., the top-K beam(s). Here, depending on the result of the second measurement, the additional report may be omitted. For example, if the previously selected top-K beam(s) are beam(s) that still provide a relatively high RSRP, the terminal may not transmit the additional report. In other words, if the top-M beam(s) for the quality selected based on the second measurement are included in the previously selected top-K beam(s), the terminal may not transmit the additional report. Alternatively, according to another embodiment, the terminal may transmit additional information including information indicating that the top-M beam(s) for the quality selected based on the second measurement are included in the previously selected top-K beam(s).

[0172]

[0173] The present disclosure below describes various specific embodiments of additional reports following inference.

[0174] [Example #1] Performing additional reporting for updating inference results triggered by the terminal

[0175] For the inference report regarding BM Case-2, the terminal may perform an inference report using not only dedicated resources (e.g., resources for inference reporting set by the base station, resources for reporting at time point A1 and time point B1 in FIG. 8) but also 'additional resources'. Here, dedicated resources refer to resources for beam inference result reporting set in advance by the base station, resources for beam inference result reporting defined in advance by the standard, or resources for reporting similar to these. On the other hand, the aforementioned 'additional resources' (hereinafter referred to as 'additional resources') refer to resources other than the dedicated resources mentioned above, and refer to resources used for reporting to evaluate, correct, or update already reported inference results (hereinafter referred to as 'additional reporting').

[0176] Reporting via the aforementioned additional resources may be made through physical channels such as reporting via PUCCH, dynamic grant (DG) PUSCH, or configured grant (CG) PUSCH. That is, reporting via additional resources may be made periodically, semi-statically, or non-periodically, depending on the configuration type of the additional resources.

[0177]

[0178] FIG. 10 illustrates an example of a procedure for transmitting additional reports for AI / ML-based beam management in a wireless communication system according to one embodiment of the present disclosure. FIG. 10 illustrates a case where the aforementioned additional reports are triggered by a terminal, which can be summarized as follows.

[0179] At time point A1, the terminal performs a report on the inference results for N future time instances. Subsequently, at time point D, an event occurs in the beam inference results of the future time points. In the case of FIG. 10, regarding the time interval of the time instance satisfying the event occurrence condition, the terminal may determine that the existing inference result and the inference result at time point D are not similar. At time point D, an additional report is performed regarding the event that occurred. The time interval for the additional report can be between time point D and the start point of the time instance satisfying the event occurrence condition, as shown in FIG. 10.

[0180] As mentioned above, triggering for additional reporting can be based on events at the terminal. Many events such as the following can be considered as triggering conditions, and similar variations of events can be considered.

[0181] - Event A: When the beam information (e.g., information for identifying the beam) of the upper-1 or upper-L beam(s) changes at any point within the previously reported time instance. That is, when a beam different from the existing report (e.g., the report at time A1 in Fig. 8) becomes the upper-1 or upper-L. Here, L is a natural number greater than 1 and less than or equal to M. For example, M can be 2.

[0182] - Event B: When, at any point within a previously reported time instance, among the beams belonging to the upper-M, there exists a beam where the deviation between the new estimate result and the existing estimate result is large. Here, the threshold value of the deviation referenced as the event occurrence condition may be predefined or set by the base station.

[0183] - Event C: Among the beams that did not belong to the upper-M at any point in time within a previously reported time instance, there exists a beam whose new estimation result belongs to the upper-M, upper-L, or upper-1. Here, L is a natural number greater than 1 and less than or equal to M, and is a value that may be determined by a future standard conference or set by a base station.

[0184] - Event D: For a beam that generated an event, if the event persists for a certain period of time.

[0185] Other events other than the aforementioned Event A, Event B, Event C, and Event D may be considered, and combinations of one or more events may be utilized. For example, “Event A, Event B, or Event C” may be considered and utilized as a single event, and “an event that simultaneously satisfies Event A and Event D” may be utilized by the terminal to trigger additional reporting.

[0186] The contents of the aforementioned additional report may include at least one of the following items.

[0187] - Time instance information (e.g., timestamp)

[0188] - Beam information

[0189] - Predicted RSRP value

[0190] In some cases, the terminal may report an RSRP value or a differential RSRP value. Here, the reference RSRP for the differential RSRP value may be generated based on the existing RSRP value of the corresponding beam of the corresponding time instance in an existing report (e.g., a report at time point A1 in FIG. 10) or a recent additional report (e.g., if there are two or more additional reports). If the beam to be reported is a beam not included in the existing report, the terminal may report an RSRP value or a differential RSRP value generated based on the RSRP value of the beam having the maximum RSRP of the corresponding time instance.

[0191] The aforementioned additional reporting-related operation may be performed up to once between dedicated reporting resources (e.g., between A1 and B1 in FIG. 10), and the number of times it is performed is not limited. In other words, additional reporting may not be performed, may be performed once, or may be performed two or more times.

[0192]

[0193] [Example #2] Performing a new report of inference results triggered by the terminal

[0194] Instead of an additional report resulting from the event occurrence of Example #1, the terminal can report the changed inference result by newly reporting the inference result for N time instances after the time of the event occurrence.

[0195] FIG. 11 illustrates an example of a procedure for performing additional inference for AI / ML-based beam management in a wireless communication system according to one embodiment of the present disclosure. Referring to FIG. 11, if an event occurs at time D after reporting an inference result at time A1, the terminal may report a new inference result for a time corresponding to time instance N from time F, just as the inference result report at time A1 was reported. This report may be performed independently of the existing report at time A1. That is, the report is unrelated to the inference result report at time A1.

[0196] The aforementioned new report may occur prior to the next inference report time through the previously scheduled dedicated resource (e.g., time point B1 in FIG. 11). As a result, the scheduled inference report may be cancelled (the report at time point B1 in FIG. 11 may be cancelled due to the new report at time point F).

[0197] Through a new report (e.g., time point F in Fig. 11), the base station can apply the corresponding inference result for N time instance(s) from time point. (Applying the inference result of the new report for N time instance(s) from time point F in Fig. 11)

[0198] When a new report occurs, the dedicated inference reporting resource to be performed next may be implicitly or explicitly allocated after N time instance(s) or after a time similar to N time instance(s) from the new report.

[0199]

[0200] The operation of the method according to 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 any type of recording device in which information that can be read by a computer system is stored. Additionally, a 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.

[0201] 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.

[0202] 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.

[0203] 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 in this disclosure. A field-programmable gate array may operate with a microprocessor to perform one of the methods described in this disclosure. Generally, it is preferable that the methods be performed by some hardware device.

[0204] 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 a method of operation of a terminal in a wireless communication system, Performing a first measurement for at least one reference signal; Performing a first inference regarding information related to beams applied during a subsequent time interval based on the above first measurement; Transmitting a beam report generated based on the result of the first inference above; Performing a second measurement of at least one reference signal during the above time interval; and A method comprising transmitting an additional report generated based on the first inference and the second measurement.

2. In Claim 1, A method in which the above additional report includes information indicating whether at least one higher quality beam identified through the above second measurement is included in the beam set included in the result of the above first inference.

3. In Claim 1, The above additional report comprises information related to beams generated based on a second inference performed based on the above second measurement.

4. In Claim 3, A method in which the above additional report is transmitted in response to the detection of an event related to the result of the first inference or the result of the second inference.

5. In Claim 4, The above event is, The change in upper beam information at a point within the above time interval, Identifying a beam among the top beams within the above time interval that has a large deviation between the new estimation result and the existing estimation result, At a point in time within the above time interval, among beams other than the upper beams, the identification of a beam whose new estimation result belongs to the upper beams, or A method comprising at least one of the following: for a beam that generated a specified event, the event persists for a certain period of time.

6. In Claim 3, A method comprising at least one of the following: information on a channel quality value included in the result of the second inference, information on a time instance corresponding to the channel quality value, or information on a beam corresponding to the channel quality value.

7. In Claim 3, A method configured such that the above additional report is interpreted in combination with the report including the result of the above first inference.

8. In Claim 3, A method configured such that the above additional report replaces the report containing the result of the above first inference.

9. In Claim 8, The above additional report is a method including the inference result of channel quality for at least one time instance after the above time interval.

10. In claim 8, The above additional report is a method that includes the inference results of channel quality for the same number of time instances as the time instances included in the above time interval.

11. In a terminal of a wireless communication system, At least one transmitter / receiver; At least one processor; and It includes at least one memory connected to the above-mentioned at least one processor to enable operation and storing instructions that control the terminal to perform operations when executed by the processor, and The above operations are, Performing a first measurement for at least one reference signal; Performing a first inference regarding information related to beams applied during a subsequent time interval based on the above first measurement; Transmitting a beam report generated based on the result of the first inference above; Performing a second measurement of at least one reference signal during the above time interval; and A terminal comprising transmitting an additional report generated based on the first inference and the second measurement.

Images