Method for changing artificial intelligence model in wireless communication system, and device therefor

Abstract

The present disclosure relates to a 5G or 6G communication system for providing a higher data transmission rate than the previous 4G communication system, such as LTE. The operation method of a terminal in a wireless communication system includes the operations of receiving, from a base station, configuration information that includes criteria information for changing an AI model according to a channel state, and identifying whether the criteria is satisfied on the basis of the measurement information of the terminal and criteria information. The AI model is used to obtain DCI prediction information for at least one prediction unit, and the DCI prediction information indicates whether DCI is present in the corresponding prediction unit.

Description

Method for changing an artificial intelligence model in a wireless communication system and device therefor

[0001] The present disclosure relates to a method for changing an artificial intelligence (AI) model in a wireless communication system and a device therefor.

[0002] Looking back at the evolution of wireless communication over successive generations, technologies have primarily been developed for human-facing services such as voice, multimedia, and data. With the commercialization of 5G (5th-generation) communication systems, an explosive increase in connected devices is expected to be connected to communication networks. Examples of networked objects include vehicles, robots, drones, home appliances, displays, smart sensors installed in various infrastructures, construction equipment, and factory equipment. Mobile devices are expected to evolve into diverse form factors, including augmented reality glasses, virtual reality headsets, and holographic devices. In the 6th-generation (6G) era, efforts are being made to develop improved 6G communication systems to connect hundreds of billions of devices and objects and provide diverse services. For this reason, 6G communication systems are often referred to as "beyond 5G."

[0003] The 6G communication system, expected to be realized around 2030, will have a maximum transmission speed of terabytes per second (i.e., 1,000 gigabits per second) and a wireless latency of 100 microseconds (μsec). In other words, compared to 5G, the transmission speed in a 6G communication system will be 50 times faster, while the wireless latency will be reduced to one-tenth.

[0004] To achieve these high data rates and ultra-low latency, 6G communication systems are being considered for implementation in the terahertz band (e.g., from 95 gigahertz (GHz) to 3 terahertz (THz)). Compared to the millimeter wave (mmWave) band introduced in 5G, the terahertz band is expected to experience more severe path loss and atmospheric absorption, making it more crucial to ensure signal reach, or coverage, in this band. Key technologies to ensure coverage include radio frequency (RF) components, antennas, new waveforms that offer better coverage than OFDM (orthogonal frequency division multiplexing), beamforming, and multiple antenna transmission technologies such as massive multiple-input and multiple-output (MIMO), full-dimensional MIMO (FD-MIMO), array antennas, and large-scale antennas. In addition, new technologies such as metamaterial-based lenses and antennas, high-dimensional spatial multiplexing using orbital angular momentum (OAM), and reconfigurable intelligent surfaces (RIS) are being discussed to improve the coverage of terahertz band signals.

[0005] In addition, in order to improve frequency efficiency and system network, 6G communication systems are developing full duplex technology that utilizes the same frequency resources at the same time for uplink and downlink; network technology that integrates satellites and high-altitude platform stations (HAPS); network structure innovation technology that supports mobile base stations and enables optimization and automation of network operation; dynamic spectrum sharing technology through collision avoidance based on spectrum usage prediction; AI-based communication technology that utilizes artificial intelligence (AI) from the design stage and internalizes end-to-end AI support functions to realize system optimization; and next-generation distributed computing technology that realizes services with complexity that exceeds the limits of terminal computing capabilities by utilizing ultra-high-performance communication and computing resources (mobile edge computing (MEC), cloud, etc.). In addition, efforts are being made to further strengthen connectivity between devices, further optimize networks, promote softwareization of network entities, and increase the openness of wireless communications through the design of new protocols to be used in 6G communication systems, the implementation of hardware-based security environments, the development of mechanisms for the safe use of data, and the development of technologies for maintaining privacy.

[0006] Research and development of these 6G communication systems are expected to enable a new level of hyper-connected experience through the hyper-connectivity of 6G communication systems, which encompass not only connections between things but also connections between people and things. Specifically, 6G communication systems are expected to enable services such as truly immersive extended reality (Truly Immersive XR), high-fidelity mobile holograms, and digital replicas. Furthermore, services such as remote surgery, industrial automation, and emergency response, which are provided through enhanced security and reliability, will find application in diverse fields such as industry, healthcare, automotive, and home appliances.

[0007] According to one embodiment of the present disclosure, a method for changing an AI model set in a terminal and a base station in a wireless communication system and a device therefor are provided.

[0008] According to one embodiment of the present disclosure, a method of operating a terminal in a wireless communication system includes an operation of receiving, from a base station, configuration information including information on a criterion for changing an AI model according to a channel state, wherein the AI ​​model is used to obtain downlink control information (DCI) prediction information for at least one prediction unit, and the DCI prediction information indicates whether DCI exists in the corresponding prediction unit; and an operation of identifying whether the criterion is satisfied based on the information on the criterion and measurement information of the terminal, wherein the changing of the AI ​​model corresponds to a change from one AI model to another AI model within a plurality of preset AI models, and the number of prediction units per inference for each of the plurality of AI models can be set differently.

[0009] According to one embodiment of the present disclosure, a method of operating a base station in a wireless communication system includes an operation of generating configuration information including information on a criterion for changing an AI model according to a channel state, wherein the AI ​​model is used to obtain DCI prediction information for at least one prediction unit, and the DCI prediction information indicates whether a DCI exists in the corresponding prediction unit; and an operation of transmitting the configuration information to a terminal, wherein the changing of the AI ​​model corresponds to a change from one AI model to another AI model among a plurality of preset AI models, and the number of prediction units per inference for each of the plurality of AI models can be set differently.

[0010] In accordance with one embodiment of the present disclosure, in a wireless communication system, a terminal comprises: a transceiver; and at least one processor connected to the transceiver and including a processing circuit, wherein the at least one processor is configured to: receive, from a base station, configuration information including information on a criterion for changing an AI model according to a channel state, wherein the AI ​​model is used to obtain DCI prediction information for at least one prediction unit, and the DCI prediction information indicates whether a DCI exists in the corresponding prediction unit; and identify, based on the information on the criterion and measurement information of the terminal, whether the criterion is satisfied, wherein the changing of the AI ​​model corresponds to a change from one AI model to another AI model within a plurality of preset AI models, and the number of prediction units per inference for each of the plurality of AI models can be set differently.

[0011] In one embodiment of the present disclosure, in a wireless communication system, a base station comprises a transceiver; and at least one processor connected to the transceiver and including a processing circuit, wherein the at least one processor is configured to: generate configuration information including information on a criterion for changing an AI model according to a channel condition, wherein the AI ​​model is used to obtain DCI prediction information for at least one prediction unit, and the DCI prediction information indicates whether a DCI exists in the corresponding prediction unit; and transmit the configuration information to a terminal, wherein the changing of the AI ​​model corresponds to a change from one AI model to another AI model within a plurality of preset AI models, and the number of prediction units per inference for each of the plurality of AI models can be set differently.

[0012] In a wireless communication system according to one embodiment of the present disclosure, power consumption can be reduced and / or prediction accuracy can be increased by changing an AI model set in a terminal and a base station according to the state of a channel.

[0013] The effects that can be obtained from the exemplary embodiments of this specification are not limited to the effects mentioned above, and other effects not mentioned can be clearly derived and understood by a person having ordinary skill in the technical field to which the exemplary embodiments of this specification belong, from the description below. In other words, unintended effects that come from practicing the exemplary embodiments of this specification can also be derived by a person having ordinary skill in the technical field from the exemplary embodiments of this specification.

[0014] FIG. 1 illustrates an exemplary wireless communication system according to one embodiment of the present disclosure.

[0015] FIG. 2 illustrates an exemplary structure of a wireless communication system for supporting wireless communication according to one embodiment of the present disclosure.

[0016] FIG. 3 illustrates an example of a signal transmission method of a wireless communication system according to one embodiment of the present disclosure.

[0017] FIG. 4 illustrates an exemplary wireless frame structure according to one embodiment of the present disclosure.

[0018] FIG. 5 illustrates an example of a physical downlink control channel (PDCCH) signaling procedure according to one embodiment of the present disclosure.

[0019] FIG. 6 illustrates an example of an AI-based PDCCH signaling procedure according to one embodiment of the present disclosure.

[0020] FIG. 7 illustrates an example of an AI model used in an AI-based PDCCH signaling procedure according to one embodiment of the present disclosure.

[0021] FIG. 8 is a flowchart of an AI-based PDCCH signaling procedure according to one embodiment of the present disclosure.

[0022] FIG. 9 illustrates an example of an AI model for an AI-based PDCCH signaling procedure according to one embodiment of the present disclosure.

[0023] FIG. 10 is a diagram comparing a PDCCH monitoring ratio when AI-based PDCCH monitoring is performed and a PDCCH monitoring ratio when AI-based PDCCH monitoring is not performed, according to one embodiment of the present disclosure.

[0024] FIG. 11 illustrates an example of an AI model set in a base station and a terminal according to one embodiment of the present disclosure.

[0025] FIG. 12a illustrates an example of a method in which a base station and a terminal perform PDCCH prediction using an AI model according to one embodiment of the present disclosure.

[0026] FIG. 12b illustrates an example of a method in which a base station and a terminal perform PDCCH prediction using an AI model according to one embodiment of the present disclosure.

[0027] FIG. 12c illustrates an example of a method in which a base station and a terminal perform PDCCH prediction using an AI model according to one embodiment of the present disclosure.

[0028] FIG. 13 illustrates an example of output mismatch due to input mismatch of an AI model according to one embodiment of the present disclosure.

[0029] FIG. 14 illustrates the PDCCH failure probability according to the number of prediction units per inference, according to one embodiment of the present disclosure.

[0030] FIG. 15 is a flowchart of a procedure for a base station to set criteria for changing an AI model according to one embodiment of the present disclosure.

[0031] FIG. 16A is a flowchart illustrating an example of a method by which a terminal initiates a procedure for changing an AI model, according to one embodiment of the present disclosure.

[0032] FIG. 16b is a flowchart illustrating an example of a method by which a terminal initiates a procedure for changing an AI model, according to one embodiment of the present disclosure.

[0033] FIG. 17 is a flowchart illustrating a method by which a base station initiates a procedure for changing an AI model, according to one embodiment of the present disclosure.

[0034] FIG. 18 is a flowchart of a procedure for changing and synchronizing an AI model according to one embodiment of the present disclosure.

[0035] FIG. 19 is a flowchart of a procedure for changing and synchronizing an AI model according to one embodiment of the present disclosure.

[0036] FIG. 20 illustrates an example of a synchronization adjustment method according to one embodiment of the present disclosure.

[0037] FIG. 21 is a flowchart of an AI model change procedure according to one embodiment of the present disclosure.

[0038] FIG. 22 is a flowchart of an AI model change procedure according to one embodiment of the present disclosure.

[0039] FIG. 23 is a flowchart of an AI model change procedure according to one embodiment of the present disclosure.

[0040] FIG. 24 is a flowchart of an AI model change procedure according to one embodiment of the present disclosure.

[0041] FIG. 25 is a flowchart of a method of operating a terminal according to one embodiment of the present disclosure.

[0042] FIG. 26 is a flowchart of a method of operating a base station according to one embodiment of the present disclosure.

[0043] FIG. 27 is a flowchart of a method of operating a terminal according to one embodiment of the present disclosure.

[0044] FIG. 28 is a flowchart of a method of operating a base station according to one embodiment of the present disclosure.

[0045] FIG. 29 is a flowchart of a method of operating a terminal according to one embodiment of the present disclosure.

[0046] FIG. 30 is a flowchart of a method of operating a base station according to one embodiment of the present disclosure.

[0047] FIG. 31 illustrates a configuration of a terminal according to one embodiment of the present disclosure.

[0048] FIG. 32 illustrates a configuration of a base station according to one embodiment of the present disclosure.

[0049] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings so that those skilled in the art can easily implement the present disclosure. However, the present disclosure may be implemented in various different forms and is not limited to the embodiments described herein. In connection with the description of the drawings, the same or similar reference numerals may be used for identical or similar components. Furthermore, in the drawings and related descriptions, descriptions of well-known functions and configurations may be omitted for clarity and conciseness.

[0050] FIG. 1 illustrates an exemplary wireless communication system according to one embodiment of the present disclosure.

[0051] Referring to FIG. 1, as an example, a wireless communication system (100) may include at least one base station (110) and at least one terminal (e.g., at least one of a plurality of terminals (120a, 120b, 120c, 120d)). The base station (110) and / or the terminals (120a, 120b, 120c, 120d) may be wireless communication devices or end devices that perform wireless communication.

[0052] A plurality of terminals (120a, 120b, 120c, 120d) included in a wireless communication system (100) may be in the same or different operating states (e.g., active state, inactive state, or idle state). The plurality of terminals (120a, 120b, 120c, 120d) may perform operations according to the operating states. For example, the plurality of terminals (120a, 120b, 120c, 120d) may include one or more terminals in the same operating state. The reference numeral '120' below may be used to refer to all of the plurality of terminals (120a, 120b, 120c, 120d) or to refer to at least one specific terminal. At least one specific terminal may be a terminal having substantially the same or similar operating state.

[0053] The base station (110) is a network infrastructure that provides wireless access to the terminal (120). The base station (110) has coverage defined as a certain geographical area based on the distance at which a signal can be transmitted. The coverage may be a service area corresponding to the cell. In addition to the base station, the base station (110) may be referred to as an 'access point (AP)', 'eNodeB (eNB)', '5th generation node', 'next Generation NodeB (gNB)', '5G NodeB (5gNB)', 'wireless point', 'transmission / reception point (TRP)', 'digital unit (DU)', 'radio unit (RU)', 'remote radio head (RRH)', or other terms having equivalent technical meanings thereto.

[0054] The terminal (120) is a device used by a user and performs communication with the base station (110) via a wireless channel. In some cases, the terminal (120) may be operated without the user's involvement. That is, the terminal (120) is a device that performs machine type communication (MTC) and may not be carried by the user. The terminal (120) may be referred to as a 'user equipment (UE)', a 'mobile station', a 'subscriber station', a 'remote terminal', a 'wireless terminal', an 'electronic device', a 'user device', or other terms having an equivalent technical meaning. The terminal (120) may include, for example, at least one of a cellular phone, a smart phone, a computer, a tablet PC, a mobile phone, a video phone, an e-book reader, a desktop PC, a laptop PC, a netbook computer, a workstation, a server, a PDA, a portable multimedia player (PMP), an MP3 player, a medical device, a camera, a wearable device, or a multimedia system capable of performing a communication function. Furthermore, the types of terminals are not limited to the above examples.

[0055] To improve channel gain, the base station (110) or the terminal (120) may perform beamforming. Beamforming may include transmit beamforming or receive beamforming. For example, the base station (110) may perform transmit beamforming for signal transmission via downlink and receive beamforming for signal reception via uplink. For example, the terminal (120) may perform receive beamforming for signal reception via downlink and transmit beamforming for signal transmission via uplink. That is, the base station (110) or the terminal (120) may provide directionality to the transmit signal or the receive signal. For this purpose, the base station (110) or the terminal (120) may select serving beams through a beam search or beam management procedure. After serving beams are selected, the base station (110) or terminal (120) can perform communication through resources that are in a quasi-co-located (QCL) relationship with the resources that transmitted the serving beams. For example, the base station (110) or terminal (120) can transmit and receive wireless signals in the millimeter wave (mmWave) band (e.g., 28 GHz, 30 GHz, 38 GHz, 60 GHz).

[0056] The terminal (120) can perform a synchronization process and / or a cell search process through a primary synchronization signal (PSS), a secondary synchronization signal (SSS), and a physical broadcast channel (PBCH). The terminal (120) can perform an access process to complete connection to a network through a base station (110). The terminal (120) can transmit a preamble through a physical random access channel (PRACH) and receive a response message to the preamble. The terminal that has performed the above-described process can then receive a downlink control channel (e.g., a physical downlink control channel (PDCCH)) or a downlink data channel (e.g., a physical downlink shared channel (PDSCH)) as an uplink / downlink signal transmission process, or transmit an uplink control channel (e.g., a physical uplink control channel (PUCCH)) or an uplink data channel (e.g., a physical uplink shared channel (PUSCH)). These access procedures can be used for various purposes, such as initial connection, uplink synchronization coordination, resource allocation, or handover.

[0057] An AI model in a wireless communication system (100) can be trained using information input from a base station (110) and / or information input from terminals (120). The AI ​​model can evolve through learning through training. The AI ​​model can be defined for each function implemented in the wireless communication system (100) and / or for each terminal.

[0058] According to one embodiment, a wireless communication system (100) may implement an AI model for acquiring downlink control information (DCI). The AI ​​model for DCI acquisition may be implemented, for example, for each terminal, or as a single model for multiple terminals (e.g., all terminals). In the present disclosure, the AI ​​model for DCI acquisition may also be referred to as an AI model for PDCCH prediction.

[0059] According to one embodiment, an AI model for acquiring DCI prediction information may use channel quality indication (CQI) information that may be acquired or provided by a terminal (120) and scheduling information that may be acquired or provided by a base station (110) as input information. The AI ​​model may evolve to provide a corresponding function through learning based on the CQI information and scheduling information corresponding to the input information. Evolution means that the accuracy of the DCI prediction information can be gradually increased through repetitive learning through training.

[0060] According to one embodiment, an artificial intelligence model for obtaining DCI may be applied (or set) to each of a base station (110) and a terminal (120). The base station (110) and the terminal (120) must be able to share input information for the corresponding AI model with each other. For example, the terminal (120) may measure channel quality and report CQI information corresponding to the measured channel quality to the base station (110). The PF (proportional fair) scheduler of the base station (110) may obtain scheduling information using the CQI information. The base station (110) may transmit the obtained scheduling information to the terminal (120). The terminal (120) may obtain scheduled information, which is scheduling information transmitted from the base station (110).

[0061] The base station (110) can input CQI information and scheduling information of each terminal (120) into an AI model activated for the terminal to obtain DCI prediction information for the terminal. The terminal (120) can input its own CQI information and scheduling information into an AI model activated to perform the corresponding function to obtain DCI prediction information for itself. If the AI ​​models in the base station (110) and the terminal (120) have been trained through the same training process and the input information is synchronized, the base station (110) and the terminal (120) will obtain substantially the same DCI prediction information. For example, the DCI prediction information from the AI ​​model can indicate whether a DCI for the terminal exists in a specific prediction unit (e.g., slot) (e.g., '1' if a DCI exists or '0' if a DCI does not exist).

[0062] FIG. 2 illustrates an exemplary structure of a wireless communication system for supporting wireless communication according to one embodiment of the present disclosure.

[0063] Referring to FIG. 2, as an example, a wireless communication system (200) may include one or more base stations (210a, 210b) (e.g., base station (110) of FIG. 1) or a core network (CN) (220) (e.g., evolved packet core (EPC) or New Generation Core (NGC)). One or more base stations (210a, 210b) may be connected to the CN (220). For example, the base stations (210a, 210b) may include eNBs corresponding to base stations in an LTE network or gNBs corresponding to base stations in an NR network, and may include base stations with new names that may be introduced in a 6G network. For example, the base stations (210a, 210b) may also be devices unrelated to a radio access technology (RAT), such as APs in a WiFi network.

[0064] One or more base stations (210a, 210b) may be configured as a single unit or may be divided into various units. The various units constituting a single base station may support mobile communication functions corresponding to the base station's role. The mobile communication functions may include, for example, functions for each PDCP / RLC / MAC / PHY / RF layer. One unit may support one or more functions. Multiple units may support one or more functions in a distributed manner.

[0065] One or more base stations (210a, 210b) may be interconnected with other base stations via an interface such as X2 or Xn. For example, the base stations (210a, 210b) may be interconnected with each other via an X2 or Xn interface. In one example, the base stations (210a, 210b) may be interconnected with a CN (220) (e.g., NGC) via an interface such as S1 or NG. The interface such as S1 or NG may be an interface between the base station and the core network. For example, the one or more base stations (210a, 210b) may be interconnected with an Access and Mobility Management Function (AMF) via an N2 interface. The one or more base stations (210a, 210b) may be interconnected with a User Plane Function (UPF) via an N3 interface.

[0066] FIG. 3 illustrates an example of a signal transmission method of a wireless communication system according to one embodiment of the present disclosure.

[0067] Referring to FIG. 3, in a wireless communication system (100), a terminal (UE) (120) (e.g., one of the terminals 120a, 120b, 120c, and 120d of FIG. 1) can receive information from a base station (BS) (110) (e.g., the base station (110) of FIG. 1) via a downlink. The terminal (120) can transmit information to the base station (110) via an uplink. The information transmitted and received by the base station (110) and / or the terminal (120) may include data and / or various control information. The physical channels in the base station (110) and / or the terminal (120) may exist in various ways depending on the type and / or purpose of the information transmitted and received.

[0068] In operation 301, the terminal (120) may perform an initial cell search to synchronize with the base station (110). The terminal (120) may perform the initial cell search, for example, in response to being powered on or entering a new cell. During the initial cell search process, the terminal (120) may receive a synchronization signal (e.g., a primary synchronization signal (PSS) and / or a secondary synchronization signal (SSS)) from the base station (110) to synchronize with the base station (110). During the initial cell search process, the terminal (120) may obtain information such as a cell identifier that may identify the base station (110). The terminal (120) may receive a physical broadcast channel (PBCH) from the base station (110) identified based on the synchronization signal. The terminal (120) may obtain cell broadcast information (e.g., a master information block (MIB)) transmitted by the base station (110) through the PBCH. During initial cell search, the terminal (120) can receive a downlink reference signal (DL RS) to identify the downlink channel status. As a result, the terminal (120) can complete downlink synchronization with the base station (110).

[0069] In operation 302, the terminal (120) may receive system information (SI) via a downlink physical channel (e.g., PDCCH or PDSCH). For example, the terminal (120) may identify a radio resource on which a PDSCH will be transmitted via the PDCCH, and receive system information via the PDSCH that uses the identified radio resource. The system information may include, for example, a lot of information that the terminal (120) requires for wireless communication with the base station (110).

[0070] When the terminal (120) acquires system information through synchronization for the downlink, it can perform a random access procedure for the base station (110) using the system information. The random access procedure may include multiple operations (e.g., operations 303, 304, 305, and 306) for implementing random access.

[0071] Specifically, in operation 303, the terminal (120) can transmit a preamble corresponding to a specific sequence via a physical random access channel (PRACH) (msg1). The terminal (120) can obtain resources for the PRACH from a first system information block (SIB1: system information block 1) received via a PDSCH. The preamble is a unique signal used for the initial connection between the terminal (120) and the base station (110) in wireless communication. Based on the preamble, the terminal (120) can notify the base station (110) of its presence and the start of wireless communication.

[0072] In operation 304, the terminal (120) may receive a random access response (RAR) for the preamble signal via the PDCCH or its corresponding PDSCH (msg2). The RAR may be a response that notifies the terminal (120) that the base station (110) is ready to start communication in response to recognizing the preamble of the terminal (120).

[0073] In operation 305, the terminal (120) may transmit an RRC connection request message (msg3) to the base station (110) using scheduling information in the RAR. The terminal (120) may receive a response message (msg4) corresponding to the RRC connection request message (msg3) from the base station (110). The terminal (120) may transmit a PUSCH to the base station (110) using information obtained from the response message (msg4). The scheduling information may include, for example, information indicating when or which frequency band the base station (110) will use to transmit data to the terminal (120). As a result, the terminal (120) may complete uplink synchronization with the base station (110).

[0074] In the case of a contention-based RACH, the terminal (120) may additionally perform a contention resolution procedure at operation 306. Specifically, the terminal (120) may perform a contention resolution procedure, such as receiving a PDCCH and its corresponding PDSCH. This may be a process for determining whether to grant priority to a specific terminal when multiple terminals attempt to access the base station (110) simultaneously using the same preamble. This process is necessary for efficient use of resources in wireless communications.

[0075] The terminal (120) that has performed at least some of operations 301 to 306 may perform a procedure for receiving or transmitting a signal via uplink / downlink. For example, the terminal (120) may perform a downlink signal reception procedure for receiving a signal via PDCCH and / or PDSCH in operation 307. For example, the terminal (120) may perform an uplink signal transmission procedure for transmitting a signal via PUSCH and / or PUCCH in operation 308.

[0076] According to an example, the terminal (120) can receive downlink control information (DCI) via a PDCCH. The terminal (120) can obtain control information such as resource allocation information from the DCI. The DCI can be configured in different formats depending on the intended use. The intended use may include, but is not limited to, PUSCH scheduling within a cell, one or multiple PUSCH scheduling within a cell, or indication of cell group downlink feedback information for the terminal (120), PUSCH scheduling within a cell, PDSCH scheduling within a downlink cell, and / or PDCSH scheduling within a cell. In addition, the terminal (120) may also transmit uplink control information (UCI) to the base station (110). The UCI may include at least one of HARQ-ACK / NACK (hybrid automatic repeat and request acknowledgement / negative-ACK), SR (scheduling request), CQI (channel quality indication), PMI (precoding matrix indication), RI (rank indication), and / or BI (beam indication).

[0077] In the signal transmission method of the wireless communication system (100) described above, an AI model may be used. The AI ​​model may be used in at least a portion of the signal transmission procedure. At least one function of the base station (110) and / or terminal (120) of the wireless communication system (100) may be performed by or supported by the AI ​​model.

[0078] FIG. 4 illustrates an exemplary wireless frame structure according to one embodiment of the present disclosure.

[0079] Referring to FIG. 4, each radio frame may have a length of, for example, 10 ms. The radio frame may be divided into two half frames. Each half frame may have a length of, for example, 5 ms. The half frame may be divided into five subframes. Each subframe may have a length of, for example, 1 ms.

[0080] A subframe can be divided into one or more slots. The number of slots within a subframe is called the numerology ( ) (or, subcarrier spacing (SCS)). For example, = 0 (or, if SCS=15khz), a subframe can contain 1 slot. For example, = 1 (or, if SCS=30khz), a subframe can contain two slots. For example, = 2 (or, if SCS=60khz), a subframe can contain 4 slots. For example, = 3 (or, if SCS=120khz), a subframe can contain 8 slots. For example, = 4 (or, SCS = 240 kHz), a subframe may include 16 slots. Meanwhile, depending on the embodiment, an additional numerology value may be further set, and in this case, the number of slots in a subframe may be determined based on the numerology value. For example, = 5 (or, if SCS=480khz), a subframe can contain 32 slots.

[0081] A slot can contain 12 or 14 OFDM symbols, depending on the cyclic prefix (CP). For example, if a normal CP is used, each slot can contain 14 OFDM symbols. For example, if an extended CP is used, each slot can contain 12 OFDM symbols.

[0082] Meanwhile, the structure of the wireless frame of FIG. 4 is merely an example, and the number of subframes, the number of slots, the number of symbols, and the length of frames / subframes / slots / symbols within the frame may be varied.

[0083] Below, various examples of the PDCCH signaling procedure are described. In the present disclosure, the PDCCH signaling procedure may also be referred to as the PDCCH monitoring procedure.

[0084] Before this, an example of PDCCH / DCI is explained as follows.

[0085] The PDCCH can carry DCI. Scheduling information for uplink data (or PUSCH) or downlink data (or PDSCH) can be included in the DCI and transmitted from the base station to the terminal. The terminal can monitor the fallback DCI format and the non-fallback DCI format for the PUSCH or PDSCH. The fallback DCI format can be composed of fixed fields predefined between the base station and the terminal, and the non-fallback DCI format can include configurable fields.

[0086] DCI can be transmitted via PDCCH after going through channel coding and modulation process. A cyclic redundancy check (CRC) is attached to the DCI message payload, and the CRC can be scrambled with a radio network temporary identifier (RNTI) corresponding to the identity of the UE. Different RNTIs can be used depending on the purpose of the DCI message, such as UE-specific data transmission, power control command, or random access response. That is, the RNTI is not transmitted explicitly, but is scrambled in the CRC after calculating it, or is included in the CRC after being operated on a bit-by-bit basis (e.g., OR, AND, XOR, etc.) and transmitted. When the UE receives a DCI message transmitted on the PDCCH, it verifies the CRC using the assigned RNTI. If the CRC verification result is correct, the UE can determine that the message has been transmitted to the UE.

[0087] For example, a DCI scheduling a PDSCH for system information (SI) may be scrambled with SI-RNTI. A DCI scheduling a PDSCH for a random access response (RAR) message may be scrambled with RA-RNTI. A DCI scheduling a PDSCH for a paging message may be scrambled with P-RNTI. A DCI notifying a slot format indicator (SFI) may be scrambled with SFI-RNTI. A DCI notifying a transmit power control (TPC) may be scrambled with TPC-RNTI. A DCI scheduling a UE-specific PDSCH or PUSCH may be scrambled with C-RNTI (cell RNTI).

[0088] A PDCCH may be composed of, for example, 1, 2, 4, 8, or 16 control channel elements (CCEs) depending on the aggregation level (AL). A CCE is a logical allocation unit used to provide a PDCCH with a preset code rate depending on the wireless channel conditions. A CCE may be composed of, for example, six resource element groups (REGs). An REG may be defined as, for example, one OFDM symbol and one (P)RB.

[0089] The PDCCH can be transmitted via a CORESET (control resource set). A CORESET can be defined, for example, as a set of REGs having a given pneumologue. Multiple CORESETs for a single UE can overlap in the time / frequency domain. A CORESET can be configured via system information (e.g., MIB) or UE-specific higher layer (e.g., RRC layer) signaling. For example, the number of RBs and the number of OFDM symbols (up to 3) constituting a CORESET can be configured via higher layer signaling.

[0090] For PDCCH reception / detection, the UE may monitor PDCCH candidates. The PDCCH candidates may, for example, indicate CCE(s) that the UE should monitor for PDCCH detection. For example, each PDCCH candidate may be defined as 1, 2, 4, 8, or 16 CCEs depending on the AL. Monitoring may include (blind) decoding the PDCCH candidates. The set of PDCCH candidates that the UE monitors is defined as a PDCCH search space. The search space includes a common search space (CSS) or a UE-specific search space (USS). The UE may acquire DCI by monitoring PDCCH candidates in one or more search spaces configured by the MIB or higher layer signaling. Each CORESET is associated with one or more search spaces, and each search space may be associated with one CORESET.

[0091] FIG. 5 illustrates an example of a PDCCH signaling procedure according to one embodiment of the present disclosure.

[0092] The PDCCH signaling procedure of the embodiment of FIG. 5 may be a non-AI-based PDCCH signaling procedure that does not utilize an AI model. In the present disclosure, the non-AI-based PDCCH signaling procedure may be referred to as a PDCCH monitoring procedure or a non-AI-based PDCCH monitoring procedure.

[0093] The non-AI PDCCH signaling procedure may be, for example, a signaling procedure using an always awake scheme or a signaling procedure using a discontinuous reception (DRX) scheme (e.g., C(connected mode)-DRX).

[0094] In the always awake mode, the terminal (120) always monitors the PDCCH in an awake state, regardless of whether the PDCCH carrying DCI is transmitted. For example, as illustrated in FIG. 5 , in the always awake mode, the terminal (120) monitors the PDCCH in every slot, regardless of whether the PDCCH carrying DCI is transmitted. For example, the terminal (120) monitors all possible PDCCH candidates in every slot, regardless of whether DCI is present (blind decoding). This results in significant power consumption.

[0095] In the DRX method (e.g., C-DRX), the base station (110) sets a DRX cycle to the terminal (120), and the terminal (120) can wake up periodically and monitor the PDCCH for the period set by the on-duration timer (onDurationTimer). If the PDCCH is successfully received, the terminal (120) can extend the active time based on the period set by the inactive timer (InactiveTimer). Even when the DRX method is used, while the terminal (120) is awake, the PDCCH is monitored in every slot regardless of whether the PDCCH carrying the DCI is transmitted. That is, the terminal (120) cannot skip PDCCH monitoring for slots in which the PDCCH is not transmitted. This may cause unnecessary power consumption of the terminal (120).

[0096] Therefore, it is necessary to consider a method for reducing power consumption of a terminal (120) by predicting whether DCI (or PDCCH carrying DCI) is transmitted (or exists).

[0097] FIG. 6 illustrates an example of an AI-based PDCCH signaling procedure according to an embodiment of the present disclosure. FIG. 7 illustrates an example of an AI model used in an AI-based PDCCH signaling procedure according to an embodiment of the present disclosure.

[0098] The PDCCH signaling procedure of the embodiment of FIG. 6 may be an AI-based PDCCH signaling procedure utilizing an AI model. In the present disclosure, the AI-based PDCCH signaling procedure may also be referred to as an AI-based PDCCH monitoring procedure, an AI-driven PDCCH signaling procedure, or an AI-driven PDCCH monitoring procedure.

[0099] Referring to FIGS. 6 and 7, in the AI-based PDCCH signaling procedure, the terminal (120) and the base station (110) can perform PDCCH prediction (or inference) using a preset AI model.

[0100] According to one embodiment, the AI ​​model may be initially set upon cell connection or may be additionally set when performance deteriorates during operation. According to one embodiment, the AI ​​model may be set from the base station (110) to the terminal (120), for example, via upper layer signaling (e.g., RRC message). According to one embodiment, activation of the set AI model may be performed, for example, via PHY layer signaling (e.g., DCI). For example, multiple AI models may be set by the base station (110) to the terminal (120) via RRC message, and at least one of the multiple AI models set via DCI may be activated in the terminal (120). However, this is merely an example, and there may be various methods for setting and / or activating AI models.

[0101] For example, the terminal (120) may perform PDCCH prediction (or inference) using the first AI model (710). In this case, information related to channel conditions and / or PDCCH may be used as inputs of the first AI model (710), and first DCI prediction information (711) for at least one prediction unit may be generated as an output of the first AI model (710). In the present disclosure, the first DCI prediction information (711) may be referred to as first DCI output information or first DCI presence information.

[0102] For example, the base station (110) may perform PDCCH prediction (or inference) using a second AI model (720) associated with a first AI model (710). In this case, information related to a channel state and / or a PDCCH may be used as input to the second AI model (720), and second DCI prediction information (721) for at least one prediction unit may be generated as an output of the second AI model (720). In the present disclosure, the second DCI prediction information (721) may be referred to as second DCI output information or second DCI presence information.

[0103] A prediction unit corresponds to a unit that performs PDCCH prediction using an AI model. The length of the prediction unit may be, for example, the same as the length of one slot. That is, each prediction unit may correspond to each slot. However, this is merely an example, and the length of the prediction unit may vary. For example, the length of the prediction unit may correspond to the length of a preset time interval or a predefined scheduling unit (e.g., a scheduling unit for non-slot-based scheduling (e.g., a scheduling unit for scheduling one or more symbol units, a scheduling unit corresponding to a mini-slot, a scheduling unit newly defined in a communication system after 5G (e.g., a communication system after 6G)) or the length of a unit (or time interval) scheduled by PDCCH / DCI.

[0104] Below, the AI-based PDCCH prediction operations of terminals and base stations are explained with examples.

[0105] <Terminal AI-based PDCCH prediction / inference operation>

[0106] According to one embodiment, the terminal (120) may perform inference to obtain first DCI prediction information (711) for at least one prediction unit using the first AI model (710). For example, if the first AI model (710) is an AI model for generating first DCI prediction information (711) for one prediction unit with one inference (an AI model configured to have one prediction unit per inference), the terminal (120) may repeatedly (or continuously) perform inference to obtain first DCI prediction information (711) for each of a plurality of prediction units. For example, if the first AI model (710) is an AI model set to have a number of prediction units per AI model inference to generate first DCI prediction information (711) for multiple prediction units with a single inference, the terminal (120) can perform a single inference to obtain first DCI prediction information (711) for each of the multiple prediction units.

[0107] According to one embodiment, the first DCI prediction information (711) may indicate whether DCI (or a PDCCH carrying DCI) is transmitted within the corresponding prediction unit (or whether DCI exists within the corresponding prediction unit). The first DCI prediction information (711) may be set to one of, for example, a first value (e.g., 0 (e.g., DCI=0)) indicating that DCI (or a PDCCH carrying DCI) is not transmitted within the corresponding prediction unit or a second value (e.g., 1 (e.g., DCI=1)) indicating that DCI (or a PDCCH carrying DCI) is transmitted within the corresponding prediction unit.

[0108] According to one embodiment, when the first DCI prediction information (711) is set to a first value indicating that DCI (or a PDCCH carrying DCI) is not transmitted within the corresponding prediction unit, the terminal (120) may be in a sleep state and may not monitor the PDCCH to acquire DCI. For example, as illustrated in FIG. 6, the terminal (120) may enter a sleep state in the second, fourth, sixth, and eighth prediction units having the first DCI prediction information (711) set to a value of '0' and may not monitor the PDCCH. This may save power of the terminal (120).

[0109] According to one embodiment, when the first DCI prediction information (711) is set to a second value indicating that DCI (or PDCCH carrying DCI) is not transmitted within the corresponding prediction unit, the terminal (120) may be in an awake state and may monitor the PDCCH to acquire DCI. In this case, the terminal (120) may receive the PDSCH using the acquired DCI. For example, as illustrated in FIG. 6, the terminal (120) may enter the awake state in the 1st, 3rd, 5th, and 7th prediction units having the first DCI prediction information (711) set to a value of '1' and may monitor the PDCCH.

[0110] AI-based PDCCH prediction / inference operation at the base station

[0111] According to one embodiment, the base station (110) can perform inference to obtain second DCI prediction information (721) for at least one prediction unit using a second AI model (720) associated with the first AI model (710). For example, if the second AI model (720) is an AI model for generating second DCI prediction information (721) for one prediction unit with one inference (an AI model configured to have one prediction unit per inference), the base station (110) can perform inference iteratively (or continuously) to obtain second DCI prediction information (721) for each of the plurality of prediction units. For example, if the second AI model (720) is an AI model for generating second DCI prediction information (721) for multiple prediction units with a single inference (an AI model set to have multiple prediction units per inference), the base station (110) can perform a single inference to obtain second DCI prediction information (721) for each of the multiple prediction units.

[0112] In one embodiment, the second AI model (720) may be the same AI model as the associated first AI model (710). The AI ​​model being the same may mean that when the same input information is input, the same output information (or substantially the same) is output.

[0113] For example, the first AI model (710) may be a neural network model that is substantially identical or similar to the second AI model (720). In this case, the first AI model (710) and the second AI model (720) may be trained to output substantially identical or similar results based on the same input being applied.

[0114] For example, since the first AI model (710) and the second AI model (720) are trained from the same AI model, the first AI model (710) and the second AI model (720) may have substantially the same or similar neural network structures. Specifically, the first AI model (710) and the second AI model (720) may have substantially the same or similar input layers, hidden layers(s), and / or output layers, respectively. Nodes of each layer constituting the first AI model (710) or the second AI model (720) and / or correlations between nodes may be substantially the same or similar in both AI models.

[0115] For example, since the first AI model (710) and the second AI model (720) have substantially the same or similar neural network structures, the first AI model (710) and the second AI model (720) can produce substantially the same or similar outputs. When the first AI model (710) and the second AI model (720) produce substantially the same or similar outputs, it can be said that synchronization has been achieved for the first AI model (710) and the second AI model (720).

[0116] According to one embodiment, the second DCI prediction information (721) may indicate whether DCI (or a PDCCH carrying DCI) is transmitted within the corresponding prediction unit (or whether DCI exists within the corresponding prediction unit). The second DCI prediction information (721) may be set to, for example, one of a first value (e.g., 0 (e.g., DCI=0)) indicating that DCI (or a PDCCH carrying DCI) is not transmitted within the corresponding prediction unit or a second value (e.g., 1 (e.g., DCI=1)) indicating that DCI (or a PDCCH carrying DCI) is transmitted within the corresponding prediction unit.

[0117] According to one embodiment, when the second DCI prediction information (721) is set to a first value indicating that the DCI (or the PDCCH carrying the DCI) is not transmitted within the corresponding prediction unit, the base station (110) may not transmit (or schedule) the PDCCH carrying the DCI within the corresponding prediction unit. For example, as illustrated in FIG. 6, the base station (110) may not transmit the PDCCH in the second, fourth, sixth, and eighth prediction units having the second DCI prediction information (721) set to a value of '0'.

[0118] According to one embodiment, when the second DCI prediction information (721) is set to a second value indicating that the DCI (or the PDCCH carrying the DCI) is not transmitted within the corresponding prediction unit, the base station (110) may transmit (or schedule) the PDCCH carrying the DCI within the corresponding prediction unit. For example, as illustrated in FIG. 6, the base station (110) may transmit the PDCCH in the 1st, 3rd, 5th, and 7th prediction units having the second DCI prediction information (721) set to a value of '1'.

[0119] In this way, in the embodiments of FIGS. 6 and 7, the terminal (120) monitors the PDCCH only when it is predicted that the PDCCH carrying the DCI within the corresponding prediction unit will be transmitted using the DCI prediction information obtained using the AI ​​model, thereby reducing the power of the terminal (120) required for PDCCH monitoring compared to the embodiment of FIG. 5.

[0120] FIG. 8 is a flowchart of an AI-based PDCCH signaling procedure according to an embodiment of the present disclosure. FIG. 9 illustrates an example of an AI model for an AI-based PDCCH signaling procedure according to an embodiment of the present disclosure.

[0121] Referring to FIGS. 8 and 9, the AI-based PDCCH signaling procedure may include an AI model setup step / operation (810), a PDCCH prediction step / operation (820), and / or a PDCCH monitoring step / operation (830).

[0122] <AI 모델 설정 단계 / 동작(810)>

[0123] According to one embodiment, the AI ​​model setup step / operation (810) may include an operation (811) of training an AI model for PDCCH prediction and / or an operation (812) of delivering (e.g., distributing) the AI ​​model.

[0124] According to one embodiment, the operation (811) of training an AI model may be performed by the base station (110) (or core network). In one example, the base station (110) may train the AI ​​model (910) by training the AI ​​model (910) using training data. According to one embodiment, the training data may be collected from, for example, all terminals (120) belonging to the base station (110).

[0125] According to one embodiment, the base station (110) can train at least one AI model (910). The at least one AI model (910) can include, but is not limited to, at least one of an AI model (Model A) for generating DCI prediction information for one prediction unit with a single inference, an AI model (Model B) for generating DCI prediction information for each of n (e.g., 3) prediction units with a single inference, or an AI model (Model C) for generating DCI prediction information for each of m (e.g., 5) prediction units greater than n with a single inference.

[0126] According to one embodiment, the base station (110) can train the AI ​​model (910) on a per-terminal and / or per-function basis.

[0127] According to one embodiment, input data for training or learning may include scheduling information (911) and / or CQI information (912).

[0128] According to one embodiment, scheduling information (911) may be acquired or provided by the base station (110). The scheduling information (911) may include, for example, information for scheduling at least one terminal (120a, 120b) associated with the base station (110). For example, the scheduling information (911) may include information for scheduling all terminals (120) belonging to the base station (110).

[0129] According to one embodiment, CQI information (912) may be acquired or provided by at least one terminal (120a, 120b). The CQI information (912) may include, for example, a CQI value acquired by at least one terminal (120a, 120b) associated with a base station. For example, the CQI information may include CQI values ​​acquired by all terminals (120) belonging to the base station.

[0130] According to one embodiment, the PF scheduler of the base station (110) may generate scheduling information using channel state information (e.g., CQI) reported from the terminal. The base station (110) may transmit the generated scheduling information to the terminal (120), and the terminal (120) may receive the scheduling information transmitted from the base station (110).

[0131] According to one embodiment, the operation (812) of transmitting the AI ​​model may include an operation in which the base station (110) transmits (or sets) data related to the learned AI model to at least one terminal (120a, 120b).

[0132] According to one embodiment, the base station (110) may transmit at least one learned AI model or parameter(s) specifying at least one learned AI model to the terminal (120).

[0133] According to one embodiment, the terminal (120a, 120b) may store at least one received AI model or update at least one pre-stored AI model based on parameter(s) specifying at least one AI model.

[0134] <PDCCH 예측 단계 / 동작(820)>

[0135] According to one embodiment, the PDCCH prediction step / operation (720) may include an inference operation for generating DCI prediction information for at least one prediction unit using a stored AI model by the terminal (120) and the base station (110).

[0136] According to one embodiment, the terminal (120a, 120b) can generate first DCI prediction information (911a-1, 911a-2) for at least one prediction unit using the first AI model (910a-1, 910a-2). For example, the first terminal (120a) can input, as input of the first AI model (910a-1), scheduled information received by the first terminal (120a) and CQI information measured by the first terminal (120a), and can obtain, as output of the first AI model (910a-1), first DCI prediction information (911a-1) for at least one prediction unit. For example, the second terminal (120b) can input, for example, scheduled information received from the second terminal (120b) and CQI information measured from the second terminal (120b) as inputs of the first AI model (910a-2), and can obtain, for example, first DCI prediction information (911a-2) for at least one prediction unit as outputs of the first AI model (910a-2).

[0137] According to one embodiment, the terminals (120a, 120b) may use the first DCI prediction information (911a-1, 911a-2) to determine whether to monitor a PDCCH carrying DCI in the corresponding prediction unit. For example, the first terminal (120a) may use the first DCI prediction information (911a-1) to determine whether to monitor a PDCCH carrying DCI in the corresponding prediction unit. For example, the second terminal (120b) may use the first DCI prediction information (911a-2) to determine whether to monitor a PDCCH carrying DCI in the corresponding prediction unit.

[0138] According to one embodiment, the base station (110) can generate second DCI prediction information (911b-1, 911b-2) for at least one prediction unit using a second AI model (910b-1, 910b-2) associated with the first AI model (910a-1, 910a-2). In one example, the base station (110) can input, for example, scheduling information generated by the base station (110) and CQI information measured by the first terminal (120a) and reported to the base station (110) as inputs of the second AI model (910b-1) for the first terminal (120a), and can obtain, for example, second DCI prediction information (911b-1) for at least one prediction unit as outputs of the second AI model (910b-1). For example, the base station (110) may input scheduling information generated by the base station (110) and CQI information measured by the second terminal (120b) and reported to the base station (110) as inputs of the second AI model (910b-2) for the second terminal (120b), and may obtain second DCI prediction information (911b-2) for at least one prediction unit as outputs of the second AI model (910b-2).

[0139] According to one embodiment, the base station (110) may use the second DCI prediction information (911b-1, 911b-2) to determine whether to transmit a PDCCH carrying DCI in the corresponding prediction unit. According to one example, the base station (110) may use the second DCI prediction information (911b-1) for the first terminal (120a) to determine whether to transmit a PDCCH carrying DCI associated with the first terminal (120a) in the corresponding prediction unit. According to one example, the base station (110) may use the second DCI prediction information (911b-2) for the second terminal (120b) to determine whether to transmit a PDCCH carrying DCI associated with the second terminal (120b) in the corresponding prediction unit.

[0140] In one embodiment, the first AI model (910a-1, 910a-2) may be the same AI model as the associated second AI model (910b-1, 910b-2). In one example, the first AI model (910a-1) of the first terminal (120a) used for PDCCH prediction in the same prediction unit and the second AI model (910b-1) of the base station (110) may be the same AI model. In one example, the first AI model (910a-2) of the second terminal (120b) used for PDCCH prediction in the same prediction unit and the second AI model (910b-2) of the base station (110) may be the same AI model. Accordingly, when the input of the first AI model (910a-1, 910a-2) and the input of the second AI model (910b-1, 910b-2) for the corresponding prediction unit are the same, the first AI model (910a-1, 910a-2) and the second AI model (910b-1, 910b-2) can each generate DCI prediction information set to the same value as output.

[0141] <PDCCH 모니터링 단계 / 동작(830)>

[0142] According to one embodiment, the PDCCH monitoring step / operation (830) may include an operation of the base station (110) transmitting or not transmitting a PDCCH in a corresponding prediction slot based on PDCCH prediction (e.g., DCI prediction information) using the second AI model (910b-1, 910b-2).

[0143] According to one embodiment, the PDCCH monitoring step / operation (830) may include an operation in which the terminal (120a, 120b) monitors or does not monitor the PDCCH in the corresponding prediction slot based on the PDCCH prediction (e.g., DCI prediction information) using the first AI model (910a-1, 910a-2).

[0144] According to one embodiment, when a terminal (120a, 120b) performs PDCCH monitoring, the terminal (120a, 120b) may transmit an ACK message to the base station (110) in response to receiving (or acquiring) a PDCCH in a corresponding prediction unit, and may transmit a NACK message to the base station (110) in response to not receiving (or acquiring) a PDCCH in a corresponding prediction unit. Through this, the base station (110) may check whether the PDCCH prediction was performed accurately. If the PDCCH prediction is determined to be inaccurate, the base station (110) may initiate a procedure for changing the AI ​​model. In one example, the ACK / NACK for PDCCH reception may be used as an input for learning / training the AI ​​model.

[0145] FIG. 10 is a diagram comparing a PDCCH monitoring ratio when AI-based PDCCH monitoring is performed and a PDCCH monitoring ratio when AI-based PDCCH monitoring is not performed, according to one embodiment of the present disclosure.

[0146] In the embodiment of Fig. 10, the PDCCH monitoring ratio when AI-based PDCCH monitoring is not performed corresponds to the PDCCH monitoring ratio when PDCCH monitoring is performed using the C-DRX method. Referring to Fig. 10, it can be confirmed that when AI-based PDCCH monitoring is performed, the average PDCCH monitoring ratio is reduced by approximately 35% compared to when AI-based PDCCH monitoring is not performed.

[0147] FIG. 11 illustrates an example of an AI model set in a base station and a terminal according to one embodiment of the present disclosure.

[0148] Referring to FIG. 11, the base station (110) and the terminal (120) may include at least one AI model for AI-based PDCCH prediction.

[0149] For example, the terminal (120) may include at least one first AI model. The at least one first AI model may include, but is not limited to, at least one of a first AI model (first AI model A) (1120a) for generating DCI prediction information for one prediction unit with a single inference, a first AI model (first AI model B) (1120b) for generating DCI prediction information for each of n (e.g., 3) prediction units with a single inference, or a first AI model (first AI model C) (1120c) for generating DCI prediction information for each of m (e.g., 5) prediction units greater than n with a single inference.

[0150] For example, the base station (110) may include at least one second AI model associated with at least one first AI model. The at least one second AI model may include, but is not limited to, at least one of a second AI model (second AI model A) (1110a) for generating DCI prediction information for one prediction unit in a single inference, a second AI model (second AI model B) (1110b) for generating DCI prediction information for each of n (e.g., 3) prediction units in a single inference, or a second AI model (second AI model C) (1110c) for generating DCI prediction information for each of m (e.g., 5) prediction units greater than n in a single inference.

[0151] For example, the first AI model A (1120a), the first AI model B (1120b), and the first AI model C (1120c) of the terminal (120) may be associated with the second AI model A (1110a), the second AI model B (1110b), and the second AI model C (1110c) of the base station (110), respectively. For example, the first AI model A (1120a), the first AI model B (1120b), and the first AI model C (1120c) may be the same AI models as the second AI model A (1110a), the second AI model B (1110b), and the second AI model C (1110c), respectively.

[0152] According to one embodiment, each AI model may be identified by an AI model identifier (ID). For example, a first AI model A (1120a), a first AI model B (1120b), and a first AI model C (1120c) may each have a unique AI model ID. For example, a second AI model A (1110a), a second AI model B (1110b), and a second AI model C (1110c) may each have a unique AI model ID.

[0153] In one embodiment, the second AI models associated with the first AI model may have the same AI model ID or different AI model IDs. For example, the first AI model A (1120a), the first AI model B (1120b), and the first AI model C (1120c) may have the same AI model ID or different AI model IDs from the associated second AI model A (1110a), the second AI model B (1110b), and the second AI model C (1110c), respectively.

[0154] FIG. 12a illustrates an example of a method for a base station and a terminal to perform PDCCH prediction using an AI model, according to an embodiment of the present disclosure. FIG. 12b illustrates an example of a method for a base station and a terminal to perform PDCCH prediction using an AI model, according to an embodiment of the present disclosure. FIG. 12c illustrates an example of a method for a base station and a terminal to perform PDCCH prediction using an AI model, according to an embodiment of the present disclosure.

[0155] In the embodiment of FIG. 12a, a first AI model A (e.g., the first AI model A (1120a) of FIG. 11) and an associated second AI model A (e.g., the second AI model A (1110a) of FIG. 11) may be used for PDCCH prediction. As described above, the first AI model A (1120a) and the second AI model A (1110a) may be the same AI model, which may be a model that generates DCI prediction information for one prediction unit (one prediction unit per inference) through one inference.

[0156] Referring to FIG. 12a, the terminal (120) can input first input information (1221a) into the first AI model A (1120a).

[0157] According to one embodiment, the first input information (1221a) may include, but is not limited to, measured CQI information, scheduling probability information, average throughput information, and / or last scheduled information.

[0158] The measured CQI information may include, for example, a CQI value measured at the terminal based on DL RS (e.g., CSI (channel state information)-RS).

[0159] The scheduling probability information may, for example, indicate a scheduling probability for a preset period (e.g., a period corresponding to n slots prior to the current slot). Here, the current slot may be a slot in which inference is performed, and may be a slot (e.g., slot i-1) preceding a slot (e.g., slot i) in which PDCCH prediction is performed.

[0160] The average transmission rate information may, for example, represent an average of the transmission rates over a preset period (e.g., a period corresponding to the previous m slots from the current slot).

[0161] The last scheduled information may be, for example, the last scheduled information acquired by the terminal.

[0162] The above-described information is merely an example, and various information related to channel status and / or PDCCH may be used as input information. For example, information such as ACK / NACK information for PDCCH / DCI reception, H-ARQ (hybrid automatic repeat request) feedback, channel status conditions, DCI grant size, frequency range, numerology value, BWP setting, CORESET information, search space information, AL information, and / or duplex mode information (e.g., TDD / FDD) may be used as first input information (1221a).

[0163] The terminal (120) can generate first output information (1222a) as an output of the first AI model A (1120a).

[0164] According to one embodiment, the first output information (1222a) may include first DCI prediction information for prediction unit i (e.g., slot i). The first DCI prediction information for prediction unit i may be set to either a first value (e.g., DCI=0) indicating that DCI (or a PDCCH carrying DCI) is not transmitted within prediction unit i or a second value (e.g., DCI=1) indicating that DCI (or a PDCCH carrying DCI) is transmitted within prediction unit i.

[0165] The terminal (120) can generate first DCI prediction information for each prediction unit by continuously performing inference using the first AI model A (1120a).

[0166] The base station (110) can input the second input information (1211a) into the second AI model A (1110a).

[0167] According to one embodiment, the second input information (1211a) may include, but is not limited to, reported CQI information, scheduling probability information, average throughput information, and / or last scheduled information.

[0168] The reported CQI information may include, for example, a CQI value measured at the terminal and reported to the base station based on DL RS (e.g., CSI-RS).

[0169] The scheduling probability information may, for example, indicate a scheduling probability for a preset period (e.g., a period corresponding to n slots prior to the current slot). Here, the current slot may be a slot in which inference is performed, and may be a slot (e.g., slot i-1) preceding a slot (e.g., slot i) in which PDCCH prediction is performed.

[0170] The average transmission rate information may, for example, represent an average of the transmission rates over a preset period (e.g., a period corresponding to the previous m slots from the current slot).

[0171] The last scheduled information may be, for example, the last scheduled information transmitted by the base station.

[0172] The above-described information is only an example, and various information related to channel status and / or PDCCH may be used as input information. For example, information such as ACK / NACK information for PDCCH / DCI reception, H-ARQ (hybrid automatic repeat request) feedback, channel status conditions, DCI grant size, frequency range, numerology value, BWP setting, CORESET information, search space information, AL information, and / or duplex mode information (e.g., TDD / FDD) may be used as second input information (1211a).

[0173] The base station (110) can generate second output information (1212a) as an output of the second AI model A (1110a).

[0174] According to one embodiment, the second output information (1212a) may include second DCI prediction information for prediction unit i (e.g., slot i). The second DCI prediction information for prediction unit i may be set to either a first value (e.g., DCI=0) indicating that no DCI (or a PDCCH carrying the DCI) is transmitted within prediction unit i or a second value (e.g., DCI=1) indicating that the DCI (or a PDCCH carrying the DCI) is transmitted within prediction unit i.

[0175] The base station (110) can generate second DCI prediction information for each prediction unit by continuously performing inference using the second AI model A (1110a).

[0176] Meanwhile, since the first AI model A (1120a) and the second AI model A (1110a) correspond to the same AI model, if the first input information (1221a) and the second input information (1211a) are the same, the first output information (1222a) and the second output information (1212a) may be the same.

[0177] In the embodiment of FIG. 12b, a first AI model B (e.g., the first AI model B (1120b) of FIG. 11) and an associated second AI model A (e.g., the second AI model B (1110b) of FIG. 11) may be used for PDCCH prediction. As described above, the first AI model B (1120b) and the second AI model B (1110b) may be the same AI model that generates DCI prediction information for n (e.g., 3) prediction units (n prediction units per inference) through a single inference.

[0178] The embodiment of Fig. 12b can perform inference / prediction on the same principle as the embodiment of Fig. 12a, except that DCI prediction information for multiple prediction units is generated through a single inference. Therefore, redundant descriptions are omitted.

[0179] Referring to FIG. 12b, the terminal (120) can input first input information (1221b) as an input of the first AI model B (1120b) and generate first output information (1222b) as an output of the first AI model B (1120b). The first output information (1222b) can include, for example, first DCI prediction information for prediction unit i (e.g., slot i), first DCI prediction information for prediction unit i+1 (e.g., slot i+1), and first DCI prediction information for prediction unit i+2 (e.g., slot i+2). The terminal (120) can generate first DCI prediction information for each prediction unit by continuously performing inference using the first AI model B (1120b).

[0180] The base station (110) can input second input information (1211b) as an input of the second AI model B (1110b) and generate second output information (1212b) as an output of the second AI model B (1110b). The second output information (1212b) can include, for example, second DCI prediction information for prediction unit i (e.g., slot i), second DCI prediction information for prediction unit i+1 (e.g., slot i+1), and second DCI prediction information for prediction unit i+2 (e.g., slot i+2). The base station (110) can generate second DCI prediction information for each prediction unit by continuously performing inference using the second AI model B (1110b).

[0181] In the embodiment of FIG. 12c, a first AI model C (e.g., the first AI model C (1120c) of FIG. 11) and an associated second AI model C (e.g., the second AI model C (1110c) of FIG. 11) may be used for PDCCH prediction. As described above, the first AI model C (1120c) and the second AI model C (1110c) may be the same AI model, which may be a model that generates DCI prediction information for m (e.g., 5) prediction units through a single inference (m prediction units per inference).

[0182] The embodiment of Fig. 12c can perform inference / prediction on the same principle as the embodiment of Fig. 12b, except that DCI prediction information for a larger number of prediction units is generated with a single inference. Therefore, redundant descriptions are omitted.

[0183] Referring to FIG. 12c, the terminal (120) may apply first input information (1221c) as an input of the first AI model C (1120c) and generate first output information (1222c) as an output of the first AI model C (1120c). The first output information (1222c) may include, for example, first DCI prediction information for prediction unit i (e.g., slot i), first DCI prediction information for prediction unit i+1 (e.g., slot i+1), first DCI prediction information for prediction unit i+2 (e.g., slot i+2), first DCI prediction information for prediction unit i+3 (e.g., slot i+3), and first DCI prediction information for prediction unit i+4 (e.g., slot i+4). The terminal (120) can generate first DCI prediction information for each prediction unit by continuously performing inference using the first AI model C (1120c).

[0184] The base station (110) can input second input information (1211c) as an input of the second AI model C (1110c) and generate second output information (1212c) as an output of the second AI model C (1110c). The second output information (1212c) can include, for example, second DCI prediction information for prediction unit i (e.g., slot i), second DCI prediction information for prediction unit i+1 (e.g., slot i+1), second DCI prediction information for prediction unit i+2 (e.g., slot i+2), second DCI prediction information for prediction unit i+3 (e.g., slot i+3), and second DCI prediction information for prediction unit i+4 (e.g., slot i+4). The base station (110) can generate second DCI prediction information for each prediction unit by continuously performing inference using the second AI model C (1110c).

[0185] FIG. 13 illustrates an example of output mismatch due to input mismatch of an AI model according to one embodiment of the present disclosure.

[0186] Referring to FIG. 13, depending on the channel status (e.g., DL channel status), the first input information (1321) input to the first AI model of the terminal (e.g., the first AI model A (1120a) and the first AI model C (1120c) of FIG. 11) and the second input information (1311) input to the associated second AI model of the base station (e.g., the second AI model A (1110a) and the second AI model C (1110c) of FIG. 11) may be different. That is, an input mismatch may occur. For example, when the channel status deteriorates, the probability of failure in DCI decoding and CSI report decoding increases, and thus, the input information of each AI model of the base station and the terminal may be different. For example, the measured CQI value of the first input information (1321) and the reported CQI value of the second input information (1311) may be different. For example, the last scheduled information of the first input information (1321) and the last scheduled information of the second input information (1311) may be different.

[0187] Such inconsistency in input information may cause inconsistency in output information in the AI ​​model. For example, as illustrated in FIG. 13, the first output information (1322c) of the first AI model C (1120c) and the second output information (1312c) of the second AI model C (1110c) may be inconsistent. For example, as illustrated, the values ​​of the first DCI prediction information for the second and fifth prediction units according to the first inference of the first AI model C (1120c) may be different from the values ​​of the first DCI prediction information for the second and fifth prediction units according to the first inference of the second AI model C (1110c).

[0188] Meanwhile, for a specific AI model, even if the input information is different, the output information may not be inconsistent. For example, as illustrated in FIG. 13, the first output information (1322a) of the first AI model A (1120a) and the second output information (1312a) of the second AI model A (1110a) may be consistent despite the inconsistency in the input information. In this way, the first AI model A (1120a) and the second AI model A (1110a), which perform PDCCH prediction for one prediction unit with a single inference, may have prediction accuracy that is less dependent on the channel state than the first AI model C (1120c) and the second AI model C (1110c) (or the first AI model B (1120b) and the second AI model B (1110b)), which perform PDCCH prediction for multiple prediction units with a single inference.

[0189] FIG. 14 illustrates the PDCCH failure probability according to the number of prediction units per inference, according to one embodiment of the present disclosure.

[0190] Referring to FIG. 14, it can be seen that as the number of prediction units (e.g., slots) per inference increases, the probability of PDCCH decoding failure increases, thereby increasing the probability of error occurrence.

[0191] As described above, when performing inference using an AI model, additional power consumption occurs due to inference. Therefore, for example, performing inference for each prediction unit, as in the embodiment of FIG. 12a, may increase the accuracy of inference, but may cause unnecessary power consumption. On the other hand, when performing prediction for multiple prediction units at once with a single inference, as in the embodiments of FIGS. 12b and 12c, power consumption due to inference can be reduced, but prediction accuracy may decrease depending on the channel condition. For example, as illustrated in FIG. 13, when the channel condition is poor, an input mismatch of the AI ​​model may occur in the terminal and the base station, which may cause an output mismatch of the AI ​​model, thereby lowering the prediction accuracy. This leads to performance degradation in the terminal and the base station. As such, each AI model has different characteristics.

[0192] Therefore, a method needs to be considered to reduce power consumption due to unnecessary inference without performance degradation by using an appropriate AI model depending on the channel condition (e.g., DL channel condition). For example, when the channel condition (or quality) is good, an AI model having a relatively large number of prediction units per inference among the configured AI models (e.g., AI model B (1110b, 1120b) or AI model C (1110c, 1120c) of FIG. 11) may be used to reduce the number of inferences, thereby reducing power consumption of the terminal and reducing scheduling overhead of the base station. For example, when the channel condition (or quality) is bad, an AI model having a relatively small number of prediction units per inference among the configured AI models (e.g., AI model A (1110a, 1120a) of FIG. 11) may be used to prevent performance degradation. Hereinafter, a method of changing an AI model to use an appropriate AI model depending on such channel condition will be described.

[0193] FIG. 15 is a flowchart of a procedure for a base station to set criteria for changing an AI model, according to one embodiment of the present disclosure.

[0194] Referring to FIG. 15, in operation 15010, a base station (e.g., the base station (110) of FIG. 1) may generate configuration information (AI model change criteria configuration information) for setting criteria used for changing an AI model according to a channel state (e.g., a DL channel state). In the present disclosure, the criteria used for changing an AI model may be referred to as a change criterion or a measurement criterion. In the present disclosure, the AI ​​model change criterion configuration information may be referred to as configuration information, change criterion configuration information, or measurement criterion configuration information.

[0195] According to one embodiment, the AI ​​model change criteria setting information may include, but is not limited to, information on at least one of the following change criteria (measurement criteria):

[0196] - When the RSRP (reference signals received power) threshold is satisfied for a certain period of time (TTT) (change criterion 1 / measurement criterion 1)

[0197] - When the average of RSRP time series data satisfies the threshold (change criterion 2 / measurement criterion 2)

[0198] - When the SINR (signal-to-interference-plus-noise ratio) threshold is satisfied for a certain period of time (TTT) or longer (change criterion 3 / measurement criterion 3)

[0199] - When the average of the SINR time series data satisfies the threshold (change criterion 4 / measurement criterion 4)

[0200] For example, the AI ​​model change criteria setting information may include information about change criteria 1, information about change criteria 2, information about change criteria 3, and / or information about change criteria 4.

[0201] Information about change criterion 1 may include, but is not limited to, information for identifying change criterion 1, information indicating an RSRP threshold of change criterion 1, and / or information indicating a TTT.

[0202] Information about change criterion 2 may include, but is not limited to, information for identifying change criterion 2 and / or information indicating a threshold of change criterion 2.

[0203] Information about change criterion 3 may include, but is not limited to, information for identifying change criterion 3, information indicating an SINR threshold of change criterion 3, and / or information indicating a TTT.

[0204] Information about change criterion 4 may include, but is not limited to, information for identifying change criterion 4 and / or information indicating a threshold for change criterion 4.

[0205] In one embodiment, the change criterion (measurement criterion) can also be used as a criterion for determining (or selecting) the AI ​​model to be changed. For example, if the change criterion (measurement criterion) is set to multiple thresholds, a separate AI model can be assigned to each range defined by each threshold. For example, if the measurement result falls within a first range of the change criterion (measurement criterion) (e.g., a range smaller than the first threshold), AI model A (e.g., model A (1110a, 1120a) of FIG. 11) may be selected as the AI ​​model to be changed, if the measurement result falls within a second range of the change criterion (measurement criterion) (e.g., a range between the first threshold and the second threshold), AI model B (e.g., model B (1110b, 1120b) of FIG. 11) may be selected as the AI ​​model to be changed, and if the measurement result falls within a third range of the change criterion (measurement criterion) (e.g., a range larger than the second threshold), AI model C (e.g., model C (1110c, 1120c) of FIG. 11) may be selected as the AI ​​model to be changed.

[0206] In operation 15020, the base station can transmit AI model change criterion setting information to a terminal (e.g., terminal (120) of FIG. 1).

[0207] According to one embodiment, the base station may transmit AI model change criterion setting information to the terminal using RRC layer signaling (e.g., RRC message), MAC layer signaling, or PHY layer signaling (e.g., DCI message).

[0208] According to one embodiment, the base station may transmit to the terminal a report configuration (e.g., ReportConfig or CSI-ReportConfig IE (information element)) including AI model change reference setting information. For example, when the terminal and the base station use DL RS (e.g., CSI-RS) as a reference signal for checking and sharing channel conditions (e.g., DL channel conditions), the base station may include the AI ​​model change reference setting information in a report configuration for a measurement report (e.g., a measurement report for CSI-RS) and transmit it to the terminal. According to one example, the report configuration may be transmitted by being included in an RRC message (e.g., an RRC configuration message or an RRC reconfiguration message).

[0209] Meanwhile, when the terminal and the base station use UL RS (e.g., SRS (sounding reference signal)) as a reference signal to check and share channel conditions (e.g., DL channel conditions) (e.g., when the DL channel and UL channel have reciprocity), the base station may not transmit AI model change reference setting information to the terminal. That is, operation 15020 may be omitted.

[0210] FIG. 16a is a flowchart illustrating an example of a method by which a terminal initiates an AI model change procedure according to an embodiment of the present disclosure. FIG. 16b is a flowchart illustrating an example of a method by which a terminal initiates an AI model change procedure according to an embodiment of the present disclosure.

[0211] Referring to FIGS. 16a and 16b, in operations 16010a and 16010b, the terminal may acquire (or receive) a DL RS. The DL RS may be, for example, a CSI-RS, but is not limited thereto.

[0212] According to one embodiment, the terminal may perform measurements (e.g., measured signal strength, measured signal quality, measured signal-to-noise ratio, measured signal-to-interference-noise ratio, etc.) on the received DL RS. For example, the terminal may measure (or obtain) RSRP, RSRQ (reference signal received quality) and / or SINR for the CSI-RS based on the received CSI-RS, but is not limited thereto. For example, various measurements that can indicate channel conditions (or quality) may also be used for this.

[0213] In operations 16020a and 16020b, the terminal can determine whether the AI ​​model change criteria are satisfied based on the AI ​​model change criteria setting information. The AI ​​model change criteria setting information can be set in the terminal by a procedure for setting criteria for changing the AI ​​model (e.g., the procedure for setting criteria for changing the AI ​​model of FIG. 15).

[0214] According to one embodiment, the terminal can determine whether the change criteria of the AI ​​model are satisfied by comparing the measurement result of the received DL RS (e.g., CSI-RS) with the change criteria of the AI ​​model based on the AI ​​model change criteria setting information.

[0215] For example, the terminal may determine that the change criterion of the AI ​​model is satisfied when the RSRP for the measured CSI-RS satisfies the RSRP threshold TTT or more (e.g., when the change criterion 1 of FIG. 15 is satisfied).

[0216] For example, the terminal may determine that the change criterion of the AI ​​model is satisfied when the average of the time series data of RSRP for the measured CSI-RS satisfies the threshold (e.g., when change criterion 2 of FIG. 15 is satisfied).

[0217] For example, the terminal may determine that the change criterion of the AI ​​model is satisfied when the SINR for the measured CSI-RS satisfies the SINR threshold for a certain period of time, TTT (e.g., when the change criterion 3 of FIG. 15 is satisfied).

[0218] For example, the terminal may determine that the change criterion of the AI ​​model is satisfied when the average of the time series data of the SINR for the measured CSI-RS satisfies the threshold (e.g., when the change criterion 4 of FIG. 15 is satisfied).

[0219] If the change criteria are satisfied, the terminal can initiate a procedure for changing the AI ​​model.

[0220] For example, if the change criteria are satisfied, the terminal may transmit a message requesting a change of the AI ​​model (e.g., a model change request (MCR) message) to the base station, as in operation 16030a.

[0221] When using DL RSs, such as CSI-RS, to monitor channel conditions, directly changing the AI ​​model at the terminal may grant the terminal too much authority. Therefore, if the change criteria are met, instead of directly changing the AI ​​model, the terminal can send an MCR message to the base station requesting an AI model change.

[0222] According to one embodiment, the MCR message may include information about a measurement result for the CSI-RS and / or a comparison result of a change criterion of the AI ​​model based on the measurement result and AI model change criterion setting information.

[0223] According to one embodiment, the terminal may transmit an MCR message to the base station using RRC layer signaling (e.g., an RRC message), MAC layer signaling, and / or PHY layer signaling (e.g., a UCI message).

[0224] In one embodiment, the base station may determine whether to change the AI ​​model (e.g., change the AI ​​model to an AI model with a larger number of prediction units per inference) based on the MCR message (or information included in the MCR message) and / or traffic information of the base station (e.g., current traffic).

[0225] For example, if the change criteria are satisfied, the terminal may determine whether to change the AI ​​model, as in operation 16030b. If necessary, the base station may authorize the terminal to directly change the AI ​​model. In this case, the terminal may directly determine whether to change the AI ​​model (e.g., change the AI ​​model to an AI model with a larger number of prediction units per inference or an AI model with a smaller number of prediction units) by using, for example, the comparison result of the AI ​​model change criteria based on the measurement result for the CSI-RS and the AI ​​model change criteria setting information. If it is determined to change the AI ​​model, the terminal may perform a procedure for changing the AI ​​model (e.g., the procedure for changing and synchronizing the AI ​​model of FIG. 19). If it is determined not to change the AI ​​model, the terminal may perform operation 16020b again.

[0226] FIG. 17 is a flowchart illustrating a method by which a base station initiates a procedure for changing an AI model according to one embodiment of the present disclosure.

[0227] Referring to FIG. 17, in operation 17010, the base station may acquire (or receive) a UL RS. The UL RS may be, for example, an SRS, but is not limited thereto.

[0228] According to one embodiment, the base station may perform measurements (e.g., measured signal strength, measured signal quality, measured signal-to-noise ratio, measured signal-to-interference-noise ratio, etc.) on the received UL RS. For example, the terminal may measure (or obtain) RSRP and / or SINR for the SRS based on the received SRS, but is not limited thereto. For example, various measurements that can indicate channel conditions (or quality) may also be used for this.

[0229] In operation 17020, the base station may determine whether the AI ​​model change criteria are satisfied based on the AI ​​model change criteria setting information. The AI ​​model change criteria setting information may be generated at the base station by a procedure for setting criteria for changing the AI ​​model (e.g., the procedure for setting criteria for changing the AI ​​model of FIG. 15 ).

[0230] According to one embodiment, the base station can determine whether the change criteria of the AI ​​model are satisfied by comparing the measurement result for the received UL RS (e.g., SRS) with the change criteria of the AI ​​model based on the AI ​​model change criteria setting information.

[0231] For example, the base station may determine that the change criterion of the AI ​​model is satisfied when the RSRP for the measured SRS satisfies the RSRP threshold TTT or more (e.g., when change criterion 1 of FIG. 15 is satisfied).

[0232] For example, the base station may determine that the change criterion of the AI ​​model is satisfied when the average of the time series data of the RSRP for the measured SRS satisfies the threshold (e.g., when the change criterion 2 of FIG. 15 is satisfied).

[0233] For example, the base station may determine that the change criterion of the AI ​​model is satisfied when the SINR for the measured SRS satisfies the SINR threshold for a certain period of time, TTT (e.g., when change criterion 3 of FIG. 15 is satisfied).

[0234] For example, the base station may determine that the change criterion of the AI ​​model is satisfied when the average of the time series data of the SINR for the measured SRS satisfies the threshold (e.g., when the change criterion 4 of FIG. 15 is satisfied).

[0235] If the change criteria are satisfied, the base station can initiate a procedure for changing the AI ​​model.

[0236] For example, if the change criteria are satisfied, the base station may determine whether to change the AI ​​model, as in operation 17030. If it is determined to change the AI ​​model, the base station may perform a procedure for changing the AI ​​model (e.g., the procedure for changing and synchronizing the AI ​​model of FIG. 18). If it is determined not to change the AI ​​model, the base station may perform operation 17020b again.

[0237] According to one embodiment, the base station may determine whether to change the AI ​​model based on a comparison result of a measurement result for a UL RS (e.g., SRS) and a change criterion of the AI ​​model based on AI model change criterion setting information.

[0238] In one embodiment, the base station may determine whether to change the AI ​​model (e.g., change the AI ​​model to an AI model having a larger number of prediction units per inference or to an AI model having a smaller number of prediction units) based on a comparison result of a measurement result for an UL RS (e.g., SRS) and a change criterion for an AI model based on AI model change criterion setting information and traffic information of the base station (e.g., current traffic). Meanwhile, if it is determined to change the AI ​​model, a synchronization procedure (e.g., time synchronization) is required so that the terminal and the base station can change the AI ​​model to the same AI model at the same time and perform inference for PDCCH prediction based on the changed AI model. The following describes a procedure for changing and synchronizing the AI ​​model.

[0239] FIG. 18 is a flowchart of a procedure for modifying and synchronizing an AI model according to one embodiment of the present disclosure. FIG. 20 illustrates an example of a synchronization adjustment method according to one embodiment of the present disclosure.

[0240] Referring to FIG. 18, in operation 18010, the base station may transmit a message for model change and synchronization (e.g., a model change and synchronous alignment (MCSA) message) to the terminal. For example, the base station may transmit the MCSA message to the terminal in response to the base station determining to change an AI model (e.g., determining to change an AI model according to operation 17030 of FIG. 17).

[0241] In one embodiment, the MCSA message may include synchronization information for changing the AI ​​model. This synchronization information may be used, for example, to allow the terminal and base station to change the AI ​​model at the same point in time (or timing). Based on this synchronization information, the terminal and base station can synchronize and change the AI ​​model to the same AI model.

[0242] According to one embodiment, the synchronization information may be set in preset units. For example, the synchronization information may be set in units of slots. For example, the synchronization information may indicate after which slot the AI ​​model is to be changed based on the slot in which the MCSA message is transmitted (and / or received). However, the present invention is not limited thereto, and the units in which the synchronization information is set may be variously changed. For example, the synchronization information may be set in non-slot-based scheduling units (e.g., scheduling units for scheduling one or more symbol units, scheduling units corresponding to mini-slots, scheduling units newly defined in communication systems after 5G (e.g., communication systems after 6G or later)).

[0243] In one embodiment, the MCSA message may include information about the AI ​​model. For example, the MCSA message may include information for identifying the AI ​​model to be changed (e.g., an AI model ID).

[0244] According to one embodiment, the MCSA message may include information about the difference in the number of prediction units per inference between the current AI model and the AI ​​model to be changed (hereinafter referred to as difference information in the number of prediction units). For example, if the current AI model is an AI model that obtains DCI prediction information for one prediction unit per inference (e.g., model A (1110a, 1120a) of FIG. 11) and the AI ​​model to be changed is an AI model that obtains DCI prediction information for three prediction units per inference (e.g., model B (1110b, 1120b) of FIG. 11), the MCSA message may include difference information in the number of prediction units indicating the difference (e.g., 2) between two values ​​(e.g., 3 and 1). For example, if the current AI model is an AI model that obtains DCI prediction information for three prediction units per inference (e.g., model B (1110b, 1120b) of FIG. 11) and the AI ​​model to be changed is an AI model that obtains DCI prediction information for one prediction unit per inference (e.g., model A (1110a, 1120a) of FIG. 11), the MCSA message may include prediction unit number difference information indicating the difference (e.g., -2) between two values ​​(e.g., 1 and 3).

[0245] According to one embodiment, the terminal can identify the AI ​​model to be changed based on the difference information in the number of prediction units in the MCSA message. For example, if the current AI model is an AI model that obtains DCI prediction information for one prediction unit per inference (e.g., the first AI model A (1120a) of FIG. 11) and the difference information in the number of prediction units indicates 2, the terminal can identify that the AI ​​model to be changed is an AI model that obtains DCI prediction information for three prediction units per inference (e.g., the first AI model B (1120b) of FIG. 11).

[0246] According to one embodiment, the base station may transmit an MCSA message to the terminal using RRC layer signaling (e.g., an RRC message), MAC layer signaling, and / or PHY layer signaling (e.g., a DCI message).

[0247] According to one embodiment, the base station may transmit the MCSA message according to a preset transmission method.

[0248] For example, a base station can transmit (or unicast) an MCSA message to a terminal that requires an AI model change. For example, the base station can transmit an MCSA message to a terminal that previously transmitted an MCR message.

[0249] For example, when a base station determines that an AI model change is necessary (e.g., when it determines that an AI model change is necessary), it may transmit (or broadcast) AI model information (e.g., a table including AI model information (e.g., AI model ID) for all terminals belonging to the base station) to all terminals belonging to the base station. In this case, terminals belonging to the base station can check their own AI models included in the AI ​​model information. Through this, each terminal can identify whether an AI model change is necessary. For example, each terminal can identify whether an AI model change is necessary by comparing its current AI model with its AI model included in the received AI model information.

[0250] For example, a base station may periodically transmit (or broadcast) AI model information for all terminals belonging to the base station (e.g., a table including AI model information for all terminals belonging to the base station). For example, the base station may periodically transmit (or broadcast) AI model information for all terminals belonging to the base station (e.g., a table including AI model information for all terminals belonging to the base station) regardless of whether an AI model change is required. In this case, terminals belonging to the base station may periodically check their AI models included in the AI ​​model information. Through this, each terminal may identify whether an AI model change is required. For example, each terminal may compare its current AI model with the AI ​​model included in the received AI model information to identify whether an AI model change is required.

[0251] In operation 18020, the terminal may perform synchronization coordination based on the received MCSA message. For example, the terminal may perform synchronization coordination using synchronization information included in the received MCSA message. For example, as illustrated in FIG. 20, if the synchronization information instructs to change the AI ​​model after x slots based on the slot (e.g., slot index n) in which the MCSA message is transmitted (and / or received), the terminal may maintain the current AI model (e.g., model A) for x slots (e.g., slot index n+x) from the slot in which the MCSA message is received, and change the AI ​​model (e.g., to model B) in the slot after x slots (e.g., slot index n+x+1).

[0252] In operation 18030, the terminal and the base station can change the AI ​​model at the same time. In one embodiment, the terminal and the base station can change the AI ​​model at the start time of the slot specified by the synchronization information (or at a time point that is a preset offset from the start time).

[0253] FIG. 19 is a flowchart of a procedure for changing and synchronizing an AI model according to one embodiment of the present disclosure.

[0254] Referring to FIG. 19, in operation 19010, the terminal may transmit an MCSA message to the base station. For example, the terminal may transmit the MCSA message to the base station in response to identifying that a change in the AI ​​model has been determined by the terminal (e.g., an AI model change decision according to operation 16030b of FIG. 16b).

[0255] In one embodiment, the MCSA message may include synchronization information for changing the AI ​​model. This synchronization information may be used, for example, to allow the terminal and base station to change the AI ​​model at the same point in time (or timing). Based on this synchronization information, the terminal and base station can synchronize and change the AI ​​model to the same AI model.

[0256] According to one embodiment, the synchronization information may be set in preset units. For example, the synchronization information may be set in units of slots. For example, the synchronization information may indicate after which slots the AI ​​model will be changed based on the slot in which the MCSA message is transmitted (and / or received). However, this is merely an example, and the synchronization information may be set in various units. For example, the synchronization information may be set in a non-slot-based scheduling unit (e.g., a scheduling unit for scheduling one or more symbol units, a scheduling unit corresponding to a mini-slot, or a scheduling unit newly defined in a communication system after 5G (e.g., a communication system after 6G or later)).

[0257] In one embodiment, the MCSA message may include information about the AI ​​model. For example, the MCSA message may include information for identifying the AI ​​model to be changed (e.g., an AI model ID).

[0258] According to one embodiment, the MCSA message may include information about the difference in the number of prediction units per inference between the current AI model and the AI ​​model to be changed (hereinafter referred to as difference information in the number of prediction units). For example, if the current AI model is an AI model that obtains DCI prediction information for one prediction unit per inference (e.g., model A (1110a, 1120a) of FIG. 11) and the AI ​​model to be changed is an AI model that obtains DCI prediction information for three prediction units per inference (e.g., model B (1110b, 1120b) of FIG. 11), the MCSA message may include difference information in the number of prediction units indicating the difference (e.g., 2) between two values ​​(e.g., 3 and 1). For example, if the current AI model is an AI model that obtains DCI prediction information for three prediction units per inference (e.g., model B (1110b, 1120b) of FIG. 11) and the AI ​​model to be changed is an AI model that obtains DCI prediction information for one prediction unit per inference (e.g., model A (1110a, 1120a) of FIG. 11), the MCSA message may include prediction unit number difference information indicating the difference (e.g., -2) between two values ​​(e.g., 1 and 3).

[0259] According to one embodiment, the base station can identify the AI ​​model to be changed based on the difference information in the number of prediction units in the MCSA message. For example, if the current AI model is an AI model that obtains DCI prediction information for one prediction unit per inference (e.g., the second AI model A (1110a) of FIG. 11) and the difference information in the number of prediction units indicates 2, the base station can identify that the AI ​​model to be changed is an AI model that obtains DCI prediction information for three prediction units per inference (e.g., the second AI model B (1110b) of FIG. 11).

[0260] According to one embodiment, the terminal may transmit an MCSA message to the base station using RRC layer signaling (e.g., an RRC message), MAC layer signaling, and / or PHY layer signaling (e.g., a UCI message).

[0261] In operation 19020, the base station may perform synchronization adjustment based on the received MCSA message. For example, the base station may perform synchronization adjustment using synchronization information included in the received MCSA message. For example, if the synchronization information instructs to change the AI ​​model after x slots based on the slot (e.g., slot index n) in which the MCSA message is transmitted (or received), the base station may maintain the current AI model for x slots (e.g., slot index n+x) from the slot in which the MCSA message is received, and change the AI ​​model in the slot after x slots (e.g., slot index n+x+1).

[0262] In operation 19030, the terminal and the base station can change the AI ​​model at the same time. In one embodiment, the terminal and the base station can change the AI ​​model at the start time of the slot specified by the synchronization information (or at a time point that is a preset offset from the start time).

[0263] FIG. 21 is a flowchart of an AI model change procedure according to one embodiment of the present disclosure.

[0264] For example, the embodiment of FIG. 21 may be an example of an AI model change procedure when DL RS (e.g., CSI-RS) is used for measuring channel conditions and the authority to change the AI ​​model is granted to the base station.

[0265] Referring to FIG. 21, in operation 21010, the base station may transmit a report setting including a change criterion (or measurement criterion) to the terminal. For a description of the generation and transmission of the measurement criterion, see, for example, the description of FIG. 15. For example, operation 21010 may include operations 15010 and / or 15020 of FIG. 15.

[0266] In operation 21020, the base station and the terminal may perform an AI-based PDCCH monitoring procedure. For a description of the AI-based PDCCH monitoring procedure, see, for example, the descriptions of FIGS. 6 to 12C . For example, operation 21020 may include operations 8010, 8020, and / or 8030 of FIG. 8 .

[0267] In operations 21030 and 21040, the terminal may identify that the measurement criteria are satisfied and transmit an MCR message to the base station. For a description of determining whether the measurement criteria are satisfied and transmitting the MCR message, see, for example, the description of FIG. 16A . For example, operation 21030 may include operations 16010a and / or 16020a of FIG. 16A , and operation 21040 may include operation 16030a of FIG. 16A .

[0268] In steps 21050 and 21060, the base station may identify that a model change has been determined and transmit an MCSA message to the terminal. For a description of the determination of whether to change the model and the transmission of the MCSA message, see, for example, the description of FIG. 18 . For example, steps 21050 and 21060 may include step 18010 of FIG. 18 .

[0269] In operations 21070 and 21080, the terminal performs synchronization adjustment, and the terminal and base station may change the AI ​​model. For a description of synchronization adjustment and AI model change, see, for example, the description of FIG. 18 . For example, operation 21070 may include operation 18020 of FIG. 18 , and operation 21080 may include operation 18030 of FIG. 18 .

[0270] Figure 22 is a flowchart of an AI model change procedure according to one embodiment of the present disclosure.

[0271] For example, the embodiment of FIG. 22 may be an example of an AI model change procedure when DL RS (e.g., CSI-RS) is used for measuring channel status and the terminal is granted permission to change the AI ​​model.

[0272] Referring to FIG. 22, in operation 22010, the base station may transmit a report configuration including a change criterion (or measurement criterion) to the terminal. For a description of the generation and transmission of the measurement criterion, see, for example, the description of FIG. 15. For example, operation 22010 may include operations 15010 and / or 15020 of FIG. 15.

[0273] In operation 22020, the base station and the terminal may perform an AI-based PDCCH monitoring procedure. For a description of the AI-based PDCCH monitoring procedure, see, for example, the descriptions of FIGS. 6 to 12C . For example, operation 22020 may include operations 8010, 8020, and / or 8030 of FIG. 8 .

[0274] In operations 22030 and 22040, the terminal can identify that the measurement criteria are satisfied and that a model change is determined. For a description of the determination of whether the measurement criteria are satisfied and the determination of whether to change the model, see, for example, the description of FIG. 16B . For example, operation 22030 may include operations 16010b and / or 16020b of FIG. 16B , and operation 22040 may include operation 16030b of FIG. 16B .

[0275] In operation 22050, the terminal may transmit an MCSA message to the base station. For a description of the transmission of the MCSA message, see, for example, the description of FIG. 19 . For example, operation 22050 may include operation 19010 of FIG. 19 .

[0276] In operations 22060 and 22070, the base station performs synchronization adjustment, and the terminal and the base station may change the AI ​​model. For a description of synchronization adjustment and AI model change, see, for example, the description of FIG. 19 . For example, operation 22070 may include operation 19020 of FIG. 19 , and operation 22080 may include operation 19030 of FIG. 19 .

[0277] FIG. 23 is a flowchart of an AI model change procedure according to one embodiment of the present disclosure.

[0278] For example, the embodiment of FIG. 23 may be an example of an AI model change procedure when UL RS (e.g., SRS) is used for measuring channel conditions and authority for AI model change is granted to the base station.

[0279] Referring to FIG. 23, in operation 23010, the base station and the terminal may perform an AI-based PDCCH monitoring procedure. For a description of the AI-based PDCCH monitoring procedure, see, for example, the descriptions of FIGS. 6 to 12C . For example, operation 23020 may include operations 8010, 8020, and / or 8030 of FIG. 8 .

[0280] In operations 23020 and 23030, the base station may determine that a measurement criterion is satisfied and that a model change is determined. For a description of the determination of whether a measurement criterion is satisfied and the determination of whether to change the model, see, for example, the description of FIG. 17 . For example, operation 23020 may include operations 17010 and / or 17020 of FIG. 17 , and operation 23030 may include operation 17030 of FIG. 17 .

[0281] In operation 23040, the base station may transmit an MCSA message to the terminal. For a description of the transmission of the MCSA message, see, for example, the description of FIG. 18 . For example, operation 23040 may include operation 18010 of FIG. 18 .

[0282] In operations 23050 and 23060, the terminal performs synchronization adjustment, and the terminal and base station may change the AI ​​model. For a description of synchronization adjustment and AI model change, see, for example, the description of FIG. 18 . For example, operation 23050 may include operation 18020 of FIG. 18 , and operation 23060 may include operation 18030 of FIG. 18 .

[0283] Figure 24 is a flowchart of an AI model change procedure according to one embodiment of the present disclosure.

[0284] For example, the embodiment of FIG. 24 may be an example of an AI model change procedure when UL RS (e.g., SRS) is used for measuring channel conditions and authority for AI model change is granted to the base station.

[0285] Referring to FIG. 24, in operation 24010, the base station and the terminal may perform an AI-based PDCCH monitoring procedure. For a description of the AI-based PDCCH monitoring procedure, see, for example, the descriptions of FIGS. 6 to 12C . For example, operation 24010 may include operations 8010, 8020, and / or 8030 of FIG. 8 .

[0286] In operation 24020, the base station may periodically broadcast an MCSA message. For a description of the periodic broadcasting of the MCSA message, see, for example, the description of FIG. 18 . For example, operation 24020 may include operation 18010 of FIG. 18 .

[0287] In operations 24030 and 24040, the base station may determine that a measurement criterion is satisfied and that a model change is determined. For a description of the determination of whether a measurement criterion is satisfied and the determination of whether to change the model, see, for example, the description of FIG. 17 . For example, operation 24030 may include operations 17010 and / or 17020 of FIG. 17 , and operation 24040 may include operation 17030 of FIG. 17 .

[0288] In operations 24050 and 24060, the terminal performs synchronization adjustment, and the terminal and base station may change the AI ​​model. For a description of synchronization adjustment and AI model change, see, for example, the description of FIG. 18 . For example, operation 24050 may include operation 18020 of FIG. 18 , and operation 24060 may include operation 18030 of FIG. 18 .

[0289] FIG. 25 is a flowchart of a method of operating a terminal according to one embodiment of the present disclosure.

[0290] The embodiment of FIG. 25 may be, for example, an example of the operation of the terminal in the embodiment of FIG. 21.

[0291] Referring to FIG. 25, the terminal can receive a configuration message from the base station (25010). For a description of the generation and transmission / reception of the configuration message (information), see, for example, the description of FIG. 15.

[0292] According to one embodiment, the configuration message may include configuration information for PDCCH prediction using an AI model and / or configuration information related to a change in an AI model used for PDCCH prediction.

[0293] According to one embodiment, the configuration information related to the change of the AI ​​model may include information on change criteria for changing the AI ​​model.

[0294] The terminal can determine whether the change criteria are satisfied (25020). For a description of whether the change criteria are satisfied, see, for example, the description in FIG. 16a.

[0295] According to one embodiment, the terminal may determine whether a change criterion is satisfied based on measurement results (e.g., RSRP, SINR) based on received DL RS (e.g., CSI-RS) and information about the change criterion included in the configuration information.

[0296] If the change criteria are met, the terminal can transmit an MCR message to the base station (25030). For a description of the transmission of the MCR message, see, for example, the description in FIG. 16a.

[0297] According to one embodiment, the MCR message may include information about a measurement result for a CSI-RS and / or a comparison result between the measurement result and a change criterion.

[0298] The terminal can receive an MCSA message from the base station (25040). For a description of the transmission and reception of the MCSA message, see, for example, the description in FIG. 18.

[0299] In one embodiment, when the base station determines to change the AI ​​model based on the MCR message, the base station may transmit an MCSA message to the terminal, and the terminal may receive the MCSA message.

[0300] In one embodiment, the MCSA message may include synchronization information for changing the AI ​​model. This synchronization information may be used, for example, to allow the terminal and base station to change the AI ​​model at the same point in time (or timing). Based on this synchronization information, the terminal and base station can synchronize and change the AI ​​model to the same AI model.

[0301] In one embodiment, synchronization information may be set in preset units. For example, synchronization information may be set in units of slots. For example, synchronization information may indicate the number of slots after which the AI ​​model will be changed, relative to the slot in which the MCSA message is transmitted (and / or received).

[0302] In one embodiment, the MCSA message may include information about the AI ​​model. For example, the MCSA message may include information for identifying the AI ​​model to be changed (e.g., an AI model ID).

[0303] The terminal can change the AI ​​model based on the MCSA message (25050). For a description of the AI ​​model change, see, for example, the description in FIG. 18.

[0304] FIG. 26 is a flowchart of an operation method of a base station according to one embodiment of the present disclosure.

[0305] The embodiment of FIG. 26 may be, for example, an example of the operation of the base station in the embodiment of FIG. 21.

[0306] Referring to FIG. 26, the base station can transmit a configuration message to the terminal (26010). For a description of the generation and transmission / reception of the configuration message (information), see, for example, the description of FIG. 15.

[0307] According to one embodiment, the configuration message may include configuration information for PDCCH prediction using an AI model and / or configuration information related to a change in an AI model used for PDCCH prediction.

[0308] According to one embodiment, the configuration information related to the change of the AI ​​model may include information on change criteria for changing the AI ​​model.

[0309] The base station can receive an MCR message from the terminal (26020). For a description of the transmission of the MCR message, see, for example, the description in FIG. 16a.

[0310] In one embodiment, the terminal may determine whether a change criterion is satisfied based on a configuration message, and if the change criterion is satisfied, transmit an MCR message to the base station. The base station may receive the MCR message transmitted from the terminal.

[0311] According to one embodiment, the terminal may determine whether a change criterion is satisfied based on measurement results (e.g., RSRP, SINR) based on received DL RS (e.g., CSI-RS) and information about the change criterion included in the configuration information.

[0312] According to one embodiment, the MCR message may include information about a measurement result for a CSI-RS and / or a comparison result between the measurement result and a change criterion.

[0313] The base station can decide whether to change the AI ​​model based on the MCR message (26030).

[0314] If it is decided to change the AI ​​model, the base station may transmit an MCSA message to the terminal (26040). For a description of the transmission and reception of the MCSA message, see, for example, the description in FIG. 18.

[0315] In one embodiment, the MCSA message may include synchronization information for changing the AI ​​model. This synchronization information may be used, for example, to allow the terminal and base station to change the AI ​​model at the same point in time (or timing). Based on this synchronization information, the terminal and base station can synchronize and change the AI ​​model to the same AI model.

[0316] In one embodiment, synchronization information may be set in preset units. For example, synchronization information may be set in units of slots. For example, synchronization information may indicate the number of slots after which the AI ​​model will be changed, relative to the slot in which the MCSA message is transmitted (and / or received).

[0317] In one embodiment, the MCSA message may include information about the AI ​​model. For example, the MCSA message may include information for identifying the AI ​​model to be changed (e.g., an AI model ID).

[0318] The base station can change the AI ​​model (25050). For a description of the AI ​​model change, see, for example, the description in FIG. 18.

[0319] Figure 27 is a flowchart of a method of operating a terminal according to one embodiment of the present disclosure.

[0320] The embodiment of FIG. 27 may be, for example, an example of the operation of the terminal in the embodiment of FIG. 22.

[0321] Referring to FIG. 27, the terminal can receive a configuration message from the base station (27010). For a description of the generation and transmission / reception of the configuration message (information), see, for example, the description of FIG. 15.

[0322] According to one embodiment, the configuration message may include configuration information for PDCCH prediction using an AI model and / or configuration information related to a change in an AI model used for PDCCH prediction.

[0323] According to one embodiment, the configuration information related to the change of the AI ​​model may include information on change criteria for changing the AI ​​model.

[0324] The terminal can determine whether the change criteria are satisfied (27020). For a description of whether the change criteria are satisfied, see, for example, the description in FIG. 16b.

[0325] According to one embodiment, the terminal may determine whether a change criterion is satisfied based on measurement results (e.g., RSRP, SINR) based on received DL RS (e.g., CSI-RS) and information about the change criterion included in the configuration information.

[0326] If the change criteria are met, the terminal can decide whether to change the AI ​​model (27030). For a description of the decision on whether to change the model, see, for example, the description in Figure 16b.

[0327] If it is decided to change the AI ​​model, the terminal may transmit an MCSA message to the base station (27040). For a description of the transmission and reception of the MCSA message, see, for example, the description in FIG. 19.

[0328] In one embodiment, the MCSA message may include synchronization information for changing the AI ​​model. This synchronization information may be used, for example, to allow the terminal and base station to change the AI ​​model at the same point in time (or timing). Based on this synchronization information, the terminal and base station can synchronize and change the AI ​​model to the same AI model.

[0329] In one embodiment, synchronization information may be set in preset units. For example, synchronization information may be set in units of slots. For example, synchronization information may indicate the number of slots after which the AI ​​model will be changed, relative to the slot in which the MCSA message is transmitted (and / or received).

[0330] In one embodiment, the MCSA message may include information about the AI ​​model. For example, the MCSA message may include information for identifying the AI ​​model to be changed (e.g., an AI model ID).

[0331] The terminal can change the AI ​​model (27050). For a description of changing the AI ​​model, see, for example, the description in Fig. 19.

[0332] FIG. 28 is a flowchart of an operation method of a base station according to one embodiment of the present disclosure.

[0333] The embodiment of FIG. 28 may be, for example, an example of the operation of the base station in the embodiment of FIG. 22.

[0334] Referring to FIG. 28, the base station can transmit a configuration message to the terminal (28010). For a description of the generation and transmission / reception of the configuration message (information), see, for example, the description of FIG. 15.

[0335] According to one embodiment, the configuration message may include configuration information for PDCCH prediction using an AI model and / or configuration information related to a change in an AI model used for PDCCH prediction.

[0336] According to one embodiment, the configuration information related to the change of the AI ​​model may include information on change criteria for changing the AI ​​model.

[0337] The base station can receive an MCSA message from the terminal (28020). For a description of the transmission and reception of the MCSA message, see, for example, the description in FIG. 19.

[0338] In one embodiment, the MCSA message may include synchronization information for changing the AI ​​model. This synchronization information may be used, for example, to allow the terminal and base station to change the AI ​​model at the same point in time (or timing). Based on this synchronization information, the terminal and base station can synchronize and change the AI ​​model to the same AI model.

[0339] In one embodiment, synchronization information may be set in preset units. For example, synchronization information may be set in units of slots. For example, synchronization information may indicate the number of slots after which the AI ​​model will be changed, relative to the slot in which the MCSA message is transmitted (and / or received).

[0340] In one embodiment, the MCSA message may include information about the AI ​​model. For example, the MCSA message may include information for identifying the AI ​​model to be changed (e.g., an AI model ID).

[0341] The base station can change the AI ​​model based on the MCSA message (28030). For a description of the AI ​​model change, see, for example, the description in FIG. 19.

[0342] Figure 29 is a flowchart of a method of operating a terminal according to one embodiment of the present disclosure.

[0343] The embodiment of FIG. 29 may be an example of the operation of the terminal in the embodiment of FIG. 23 or 24, for example.

[0344] Referring to FIG. 29, a terminal can receive an MCSA message from a base station. For a description of the transmission of the MCSA message, see, for example, the description of FIG. 18.

[0345] In one embodiment, the MCSA message may include synchronization information for changing the AI ​​model. This synchronization information may be used, for example, to allow the terminal and base station to change the AI ​​model at the same point in time (or timing). Based on this synchronization information, the terminal and base station can synchronize and change the AI ​​model to the same AI model.

[0346] In one embodiment, synchronization information may be set in preset units. For example, synchronization information may be set in units of slots. For example, synchronization information may indicate the number of slots after which the AI ​​model will be changed, relative to the slot in which the MCSA message is transmitted (and / or received).

[0347] In one embodiment, the MCSA message may include information about the AI ​​model. For example, the MCSA message may include information for identifying the AI ​​model to be changed (e.g., an AI model ID).

[0348] The terminal can change the AI ​​model based on the MCSA message (29020). For a description of the AI ​​model change, see, for example, the description in FIG. 18.

[0349] FIG. 30 is a flowchart of an operation method of a base station according to one embodiment of the present disclosure.

[0350] The embodiment of FIG. 30 may be an example of the operation of the base station in the embodiment of FIG. 23 or 24, for example.

[0351] Referring to FIG. 30, the base station can determine whether a change criterion is satisfied (30010). For a description of determining whether a change criterion is satisfied, see, for example, the description of FIG. 17.

[0352] According to one embodiment, the base station can determine whether a change criterion is satisfied based on measurement results (e.g., RSRP, SINR) based on received UL RS (e.g., SRS) and information about a change criterion generated by the base station.

[0353] If the change criteria are met, the base station can decide whether to change the AI ​​model (30020). For a description of the decision on whether to change the model, see, for example, the description in FIG. 17.

[0354] If it is decided to change the AI ​​model, the base station may transmit an MCSA message to the terminal (30030). For a description of the transmission and reception of the MCSA message, see, for example, the description in FIG. 18.

[0355] In one embodiment, the MCSA message may include synchronization information for changing the AI ​​model. This synchronization information may be used, for example, to allow the terminal and base station to change the AI ​​model at the same point in time (or timing). Based on this synchronization information, the terminal and base station can synchronize and change the AI ​​model to the same AI model.

[0356] In one embodiment, synchronization information may be set in preset units. For example, synchronization information may be set in units of slots. For example, synchronization information may indicate the number of slots after which the AI ​​model will be changed, relative to the slot in which the MCSA message is transmitted (and / or received).

[0357] In one embodiment, the MCSA message may include information about the AI ​​model. For example, the MCSA message may include information for identifying the AI ​​model to be changed (e.g., an AI model ID).

[0358] The base station can change the AI ​​model (30040). For a description of the AI ​​model change, see, for example, the description in FIG. 18.

[0359] Figure 31 shows the configuration of a terminal according to one embodiment of the present disclosure.

[0360] The terminal of FIG. 31 may be, for example, the terminal (120) of FIG. 1, or may include the terminal (120) of FIG. 1, or may be an electronic device included in the terminal (120) of FIG. 1.

[0361] Referring to FIG. 31, the terminal may include a transceiver (3110), a control unit (3120), and a storage unit (3130). In the present disclosure, the control unit (3110) may be defined as a circuit or an application-specific integrated circuit or at least one processor.

[0362] The transceiver (3110) can transmit and receive signals with other entities. For example, the transceiver (3110) can transmit and receive data for modifying an AI model for PDCCH prediction.

[0363] The control unit (3120) can control the overall operation of the terminal according to the embodiment proposed in this disclosure. For example, the control unit (3120) can control the signal flow between each block to perform operations according to the flowchart described above. Specifically, the control unit (3120) can control the operations of the terminal described with reference to FIGS. 1 to 30, for example.

[0364] The storage unit (3130) can store at least one of the information transmitted and received through the transceiver unit (3110) and the information generated through the control unit (3120). For example, the storage unit (3130) can store data and information necessary for changing the AI ​​model for PDCCH prediction described with reference to FIGS. 1 to 30.

[0365] Figure 32 shows the configuration of a base station according to one embodiment of the present disclosure.

[0366] The base station of FIG. 32 may be, for example, the base station (110) of FIG. 1, or may include the base station (110) of FIG. 1, or may be an electronic device included in the base station (110) of FIG. 1.

[0367] Referring to FIG. 32, the base station may include a transceiver (3210), a control unit (3220), and a storage unit (3230). In the present disclosure, the control unit (3210) may be defined as a circuit or an application-specific integrated circuit or at least one processor.

[0368] The transceiver (3210) can transmit and receive signals with other entities. For example, the transceiver (3210) can transmit and receive data for modifying an AI model for PDCCH prediction.

[0369] The control unit (3220) can control the overall operation of the base station according to the embodiment proposed in the present disclosure. For example, the control unit (3220) can control the signal flow between each block to perform operations according to the flowchart described above. Specifically, the control unit (3220) can control the operations of the base station described with reference to FIGS. 1 to 30, for example.

[0370] The storage unit (3230) can store at least one of the information transmitted and received through the transceiver unit (3210) and the information generated through the control unit (3220). For example, the storage unit (3230) can store data and information necessary for changing the AI ​​model for PDCCH prediction described with reference to FIGS. 1 to 30.

[0371] Electronic devices according to various embodiments disclosed herein may take various forms. Electronic devices may include, for example, display devices, portable communication devices (e.g., smartphones), computer devices, portable multimedia devices, portable medical devices, cameras, wearable devices, or home appliances. Electronic devices according to embodiments disclosed herein are not limited to the aforementioned devices.

[0372] The various embodiments of this specification and the terminology used therein are not intended to limit the technical features described in this specification to specific embodiments, but should be understood to include various modifications, equivalents, or substitutes of the embodiments. For example, a component expressed in the singular should be understood to include plural components unless the context clearly indicates only the singular. It should be understood that the term "and / or" used herein encompasses any and all possible combinations of one or more of the listed items. The terms "comprise," "have," "consist of," and the like used herein are intended only to specify the presence of a feature, component, part, or combination thereof described in this specification, and the use of such terms does not exclude the presence or addition of one or more other features, components, parts, or combinations thereof. In this specification, phrases such as "A or B", "at least one of A and B", "at least one of A or B", "A, B, or C", "at least one of A, B, and C", and "at least one of A, B, or C" can each include any one of the items listed together in that phrase, or all possible combinations thereof. Terms such as "first", "second", or "first" or "second" may be used merely to distinguish the corresponding element from other corresponding elements and do not limit the corresponding elements in any other respect (e.g., importance or order).

[0373] The term "part" or "module" used in various embodiments of the present specification may include a unit implemented by hardware, software, or firmware, and may be used interchangeably with terms such as logic, logic block, component, or circuit, for example. The "part" or "module" may be an integrally configured component or a minimum unit of the component or a part thereof that performs one or more functions. For example, according to one embodiment, the "part" or "module" may be implemented in the form of an application-specific integrated circuit (ASIC).

[0374] The term "if" as used in various embodiments of this specification may be interpreted to mean "when", or "when", or "in response to determining", or "in response to detecting", depending on the context. Similarly, "if it is determined that", or "if ~ is detected" may be interpreted to mean "upon determining", or "in response to determining", or "upon detecting", or "in response to detecting", depending on the context.

[0375] The program executed by the electronic device described herein may be implemented as hardware components, software components, and / or a combination of hardware components and software components. The program may be executed by any system capable of executing computer-readable instructions.

[0376] Software may include a computer program, code, instructions, or a combination of one or more of these, which can configure a processing device to perform a desired operation or command the processing device, either independently or collectively. Software may be implemented as a computer program including instructions stored on a computer-readable storage medium. Examples of the computer-readable storage medium include magnetic storage media (e.g., read-only memory (ROM), random-access memory (RAM), floppy disks, hard disks, etc.) and optical reading media (e.g., CD-ROMs, digital versatile discs (DVDs)). The computer-readable storage medium may be distributed across network-connected computer systems so that the computer-readable code can be stored and executed in a distributed manner. Computer programs can be distributed online (e.g., by download or upload) through an application store (e.g., Play Store™) or directly between two user devices (e.g., smartphones). In the case of online distribution, at least a portion of the computer program product may be temporarily stored or temporarily created on a device-readable storage medium, such as the memory of a manufacturer's server, an application store's server, or an intermediary server.

[0377] According to various embodiments, each component (e.g., a module or a program) of the above-described components may include one or more entities, and some of the entities may be separated and placed in other components. According to various embodiments, one or more components or operations of the aforementioned components may be omitted, or one or more other components or operations may be added. Alternatively or additionally, a plurality of components (e.g., a module or a program) may be integrated into a single component. In such a case, the integrated component may perform one or more functions of each of the plurality of components identically or similarly to those performed by the corresponding component among the plurality of components prior to the integration. According to various embodiments, the operations performed by a module, program, or other component may be executed sequentially, in parallel, iteratively, or heuristically, or one or more of the operations may be executed in a different order, omitted, or one or more other operations may be added.

Claims

1. In a method of operating a terminal in a wireless communication system, An operation of receiving, from a base station, configuration information including information on criteria for changing an AI (artificial intelligence) model according to a channel condition, wherein the AI ​​model is used to obtain DCI (downlink control information) prediction information for at least one prediction unit, and the DCI prediction information indicates whether DCI exists in the corresponding prediction unit; and An operation for identifying whether the criterion is satisfied based on information about the above criteria and measurement information of the terminal, A method wherein the change of the above AI model corresponds to a change from one AI model to another AI model within a plurality of preset AI models, and the number of prediction units per inference for each of the plurality of AI models is set differently.

2. In paragraph 1, the method: If it is identified that the above criteria are satisfied, the method further includes an action of transmitting a model change request (MCR) message requesting a change in the AI ​​model to the base station. A method wherein the MCR message includes at least one of measurement information of the terminal or information on a comparison result between the measurement information of the terminal and information on the criterion.

3. In the second paragraph, the method: An operation of receiving a model change and synchronous alignment (MCSA) message for changing and synchronizing the AI ​​model from the base station; and Further comprising an action of changing the AI ​​model based on the MCSA message; A method wherein the MCSA message includes at least one of identification information of the AI ​​model to be changed, information about the difference in the number of prediction units per inference between the current AI model and the AI ​​model to be changed, or synchronization information for changing the AI ​​model, wherein the synchronization information is used by the terminal and the base station to change the AI ​​model at the same time.

4. In paragraph 3, the method: If it is identified that the above criteria are satisfied, an action for identifying whether to change the AI ​​model; and A method further comprising the action of transmitting an MCSA message to the base station when it is identified that the AI ​​model is changed.

5. In paragraph 1, A method wherein the measurement information of the terminal includes at least one of RSRP (reference signal received power), RSRQ (reference signal received quality), or SINR (signal to interference plus noise ratio) measured based on CSI-RS (channel state information-reference signal) received from the base station.

6. In paragraph 1, A method wherein the input of the AI ​​model includes channel quality indication (CQI) information measured by the terminal and scheduling information provided by the base station.

7. In paragraph 1, A method in which the length of one prediction unit corresponds to the length of one slot.

8. A method in accordance with claim 1, wherein multiple identical AI models are set in the terminal and the base station.

9. In paragraph 1, If the DCI prediction information is set to a first value indicating that the DCI does not exist in the prediction unit, the terminal does not perform PDCCH (physical downlink control channel) monitoring for acquiring the DCI in the prediction unit. A method wherein, when the DCI prediction information is set to a second value indicating that the DCI exists in the corresponding prediction unit, the terminal performs PDCCH monitoring for acquiring the DCI in the corresponding prediction unit.

10. In a method of operating a base station in a wireless communication system, An operation of generating configuration information including information on criteria for changing an AI (artificial intelligence) model according to channel conditions, wherein the AI ​​model is used to obtain DCI (downlink control information) prediction information for at least one prediction unit, and the DCI prediction information indicates whether DCI exists in the corresponding prediction unit; and Includes an action of transmitting the above setting information to the terminal, A method wherein the change of the above AI model corresponds to a change from one AI model to another AI model within a plurality of preset AI models, and the number of prediction units per inference for each of the plurality of AI models is set differently.

11. In paragraph 10, the method: An operation of receiving a model change request (MCR) message requesting a change to the AI ​​model from the terminal; and Further comprising an action for identifying whether to change the AI ​​model based on the MCS message; A method wherein the MCR message includes at least one of measurement information of the terminal or information on a comparison result between the measurement information of the terminal and information on the criterion.

12. In paragraph 11, the method: When it is identified that the above AI model is changed, an action of transmitting an MCSA message to the terminal; An operation of receiving an MCSA message from the terminal; and Further comprising an action of changing the AI ​​model based on the MCSA message; A method wherein the MCSA message includes at least one of identification information of the AI ​​model to be changed, information about the difference in the number of prediction units per inference between the current AI model and the AI ​​model to be changed, or synchronization information for changing the AI ​​model, wherein the synchronization information is used by the terminal and the base station to change the AI ​​model at the same time.

13. In paragraph 10, If the DCI prediction information is set to a first value indicating that the DCI does not exist in the corresponding prediction unit, the base station does not transmit a PDCCH (physical downlink control channel) carrying the DCI in the corresponding prediction unit. A method wherein, when the DCI prediction information is set to a second value indicating that the DCI exists in the corresponding prediction unit, the base station transmits a PDCCH carrying the DCI in the corresponding prediction unit.

14. In a wireless communication system, at a terminal, Transmitter and receiver; and At least one processor coupled to the transceiver, wherein the at least one processor comprises: Receive from a base station configuration information including information on criteria for changing an AI (artificial intelligence) model according to channel conditions, wherein the AI ​​model is used to obtain DCI (downlink control information) prediction information for at least one prediction unit, and the DCI prediction information indicates whether DCI exists in the corresponding prediction unit; Based on the information on the above criteria and the measurement information of the terminal, it is set to identify whether the above criteria are satisfied. A terminal in which the change of the above AI model corresponds to a change from one AI model to another AI model within a plurality of preset AI models, and the number of prediction units per inference for each of the plurality of AI models is set differently.

15. In a base station in a wireless communication system, Transmitter and receiver; and At least one processor coupled to the transceiver, wherein the at least one processor comprises: Generating configuration information including information on criteria for changing an AI (artificial intelligence) model according to channel conditions, wherein the AI ​​model is used to obtain DCI (downlink control information) prediction information for at least one prediction unit, and the DCI prediction information indicates whether DCI exists in the corresponding prediction unit; It is set to transmit the above setting information to the terminal, A base station wherein the change of the above AI model corresponds to a change from one AI model to another AI model within a plurality of preset AI models, and the number of prediction units per inference for each of the plurality of AI models is set differently.

Images