Terminal device, base station device, and communication method executed by terminal device

Abstract

Provided are a terminal device, a base station device, and a communication method performed by the terminal device. The terminal device may include a memory configured to: store at least one artificial intelligence (AI) model or at least one AI function for performing AI-based wireless communication; a communication interface; and a processor configured to: configure capability information of the terminal device, the capability information including information about a number of at least one AI model or a number of at least one AI function that the terminal device can support; a communication interface is controlled to transmit a first message including capability information of the terminal device to a base station device.

Description

[0001] This application claims priority based on and to Korean Patent Application No. 10-2023-0168585, filed with the Korean Intellectual Property Office on November 28, 2023, and Korean Patent Application No. 10-2024-0069520, filed with the Korean Intellectual Property Office on May 28, 2024, the disclosures of which are hereby incorporated by reference in their entireties. Technical Field

[0002] Devices and methods consistent with example embodiments relate to a wireless communication system, and more particularly, to managing an artificial intelligence (AI) / machine learning (ML) model or an AI / ML function based on signaling between devices in a wireless communication system. Background Art

[0003] The fifth-generation (5G) wireless communication technology defines a wide bandwidth to achieve fast transmission speeds and new services, and can be implemented not only in frequency bands below 6 GHz (“Sub 6 GHz”) such as 3.5 GHz, but also in ultra-high frequency bands above 6 GHz (referred to as millimeter waves (mmWave)), such as 28 GHz and 39 GHz. In addition, the sixth-generation (6G) wireless communication technology, which is referred to as a system beyond 5G communication (Beyond 5G), is being considered for implementation in the terahertz band (e.g., a band from 95 GHz to 3 THz) to achieve a transmission speed 50 times faster than 5G wireless communication technology and an ultra-low latency time reduced to 1 / 10.

[0004] In the early days of 5G wireless communication technology, in order to meet the service support and performance requirements of enhanced mobile broadband (eMBB), ultra-reliable low-latency communication (URLLC), and massive machine-type communication (mMTC) as goals, standardization has been carried out for beamforming and large-scale array multiple input / output (large-scale MIMO) to mitigate path loss of radio waves in ultra-high frequency bands and increase the transmission distance of radio waves, various parameter set supports (such as multiple subcarrier spacing operations, etc.) and dynamic operation of time slot formats to efficiently use ultra-high frequency resources, an initial access technology for supporting multi-beam transmission and broadband, the definition and operation of a bandwidth part (BWP), new channel coding methods (such as low-density parity-check (LDPC) codes for big data transmission and polar codes for highly reliable transmission of control information), L2 preprocessing, network slicing to provide a dedicated network specifically for a specific service, etc.

[0005] Currently, discussions are underway to consider the services that 5G wireless communication technology aims to support in order to improve and enhance the initial 5G wireless communication technology, and physical layer standardization is in progress for technologies such as vehicle-to-everything (V2X) technology that helps autonomous vehicles make driving decisions based on their own positions and status information transmitted by the vehicle and improve user convenience, new radio unlicensed (NR-U) technology that aims to operate systems meeting various regulatory requirements in unlicensed bands, NR terminal low-power technology (user equipment (UE) power saving), and non-terrestrial network (NTN) for terminal-satellite direct communication to ensure coverage in areas where communication with the terrestrial network is not possible, as well as position determination (positioning).

[0006] When such a 5G wireless communication system becomes commercially available, the number of connected devices is expected to grow exponentially. Therefore, it may be necessary to enhance the functions and performance of the 5G wireless communication system and the integrated operation of connected devices. To this end, new research based on extended reality (XR), AI / machine learning (ML) will be conducted in areas such as 5G performance improvement and complexity reduction, artificial intelligence (AI) / machine learning (ML) service support, meta-service support, and drone communication to efficiently support augmented reality (AR), virtual reality (VR), and mixed reality (MR).

[0007] The development of a 5G wireless communication system can be the basis for the development of AI / ML-based communication technology, which realizes system optimization by leveraging AI / ML from the design phase and internalizing end-to-end AI / ML support functions.

[0008] Specifically, in 3GPP Release 18, various research projects were conducted to introduce and apply AI / ML-based communication technology to the 5G NR wireless communication system. For example, it is planned to apply AI / ML-based communication technology to three representative scenarios (e.g., channel state information (CSI) reporting, beam management, and positioning). Before applying AI / ML-based communication technology to the above three scenarios, 3GPP constructed a basic framework for AI / ML that represents the entire process of the lifecycle management (LCM) of AI / ML.

[0009] With the development of the wireless communication system as described above, various AI / ML-based services can be provided. Therefore, methods for effectively providing services are needed. In addition, in a wireless communication system, methods for efficiently managing AI / ML models or AI / ML functions based on signaling between a terminal device and a base station device are needed. Summary of the Invention

[0010] One or more embodiments provide a method and apparatus for efficiently managing an AI / ML model or an AI / ML function based on various signaling between a terminal device and a base station device in a wireless communication system using AI / ML-based communication technologies.

[0011] According to an aspect of the present disclosure, a terminal device may include: a memory configured to store at least one AI model or at least one AI function for performing AI-based wireless communication; a communication interface; and a processor configured to configure capability information of the terminal device, the capability information including information about the number of at least one AI model or the number of at least one AI function that the terminal device is capable of supporting; and control the communication interface to send a first message including the capability information of the terminal device to a base station device.

[0012] According to another aspect of the present disclosure, a base station device may include: a memory; a communication interface; and a processor configured to control the communication interface to receive a first message including the capability information of a terminal device from the terminal device, configure control information about at least one AI model or at least one AI function based on the capability information of the terminal device in the first message; and send a second message including the control information to the terminal device, wherein the capability information of the terminal device includes information about the number of at least one AI model or the number of at least one AI function that the terminal device is capable of supporting.

[0013] According to another aspect of the present disclosure, a wireless communication system may include: an electronic device; and an external electronic device, wherein the electronic device is configured to generate a first message including information about the number of at least one AI model or the number of at least one AI function that the electronic device is capable of supporting and send the first message to the external electronic device, and the external electronic device is configured to generate a second message including control information about the at least one AI model or the at least one AI function based on the first message and send the second message to the electronic device. The electronic device is further configured to generate a third message including preference information about the at least one AI model or the at least one AI function set according to the second message and send the third message to the external electronic device, and the external electronic device is further configured to determine whether to update the control information based on the third message.

[0014] According to another aspect of the present disclosure, a communication method performed by a terminal device may include: sending capability information from the terminal device to a base station device, the capability information indicating that the terminal device is capable of reporting an artificial intelligence (AI) function preference of the terminal device to the base station device; receiving a prohibition timer from the base station device based on a communication session established between the terminal device and the base station device for reporting the AI function preference to the base station device, the prohibition timer prohibiting the terminal device from sending an updated AI function preference within a predetermined period of time; and after the prohibition timer expires, sending the updated AI function preference from the terminal device to the base station device. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The above and / or other aspects will become more apparent by describing specific example embodiments with reference to the accompanying drawings.

[0016] Figure 1 FIG. shows a wireless communication system according to one or more embodiments.

[0017] Figure 2 FIG. shows an example of an artificial intelligence (AI) / machine learning (ML) framework according to one or more embodiments.

[0018] Figure 3 FIG. shows a configuration of a terminal device according to one or more embodiments Figure 1 in.

[0019] Figure 4 FIG. shows a configuration of a base station device according to one or more embodiments Figure 1 in.

[0020] Figure 5 FIG. is a flowchart showing a method for managing an AI / ML model or an AI / ML function in a wireless communication system according to one or more embodiments.

[0021] Figure 6 FIG. is a flowchart showing a method for managing an AI / ML model or an AI / ML function in a wireless communication system according to one or more embodiments.

[0022] Figure 7 FIG. is a flowchart showing a method for managing an AI / ML model or an AI / ML function in a wireless communication system according to one or more embodiments.

[0023] Figure 8 FIG. is a flowchart showing a method for managing an AI / ML model or an AI / ML function in a wireless communication system according to one or more embodiments.

[0024] Figure 9A and Figure 9B FIG. is a block diagram showing an electronic device according to one or more embodiments.

[0025] Figure 10 is a diagram showing an electronic device to which AI / ML-based communication techniques are applied according to one or more embodiments. Detailed Description

[0026] Example embodiments will be described in more detail below with reference to the accompanying drawings.

[0027] In the following description, even in different drawings, the same reference numerals are used for the same elements. Matters defined in the description (such as detailed configurations and elements) are provided to assist in a comprehensive understanding of example embodiments. However, it is clear that example embodiments can be practiced without those specific defined matters. In addition, well-known functions or configurations will not be described in detail since they would obscure the description with unnecessary detail.

[0028] It will be understood that each block of the process flow diagrams and combinations of the flow diagrams can be executed by computer program instructions. These computer program instructions can be loaded into a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, and thus, the instructions executed by the processor of the general-purpose computer, special-purpose computer, or other programmable data processing device create a means for performing the functions described in one or more of the flow diagram blocks. These computer program instructions can also be stored in a computer-usable or computer-readable memory, which can direct a computer or other programmable data processing device to implement functions in a particular manner, and thus, it is also feasible to produce a manufacture including an instruction means for performing the functions described in the flow diagram blocks, the instruction means being created by the instructions stored in the computer-usable or computer-readable memory. The computer program instructions can also be loaded onto a computer or other programmable data processing device, and thus, a series of operation steps executed on the computer or other programmable data processing device to create a computer-executed process and to operate the computer or other programmable data processing device can also be provided for performing the steps of the functions described in one or more of the flow diagram blocks.

[0029] In addition, each block can represent a module, segment, or portion of code that includes one or more executable instructions for performing the specified one or more logical functions. In addition, it should be noted that in some alternative execution examples, it is feasible that the functions mentioned in the blocks occur out of order. For example, two consecutively shown blocks can be executed substantially simultaneously, or the two blocks can be executed in the reverse order according to the corresponding functions.

[0030] The term “~ unit” used in the present disclosure represents a software or hardware component (such as an FPGA or an ASIC), and the “~ unit” performs a specific role. However, the “~ unit” is not limited to software or hardware.

[0031] When describing the disclosed embodiments in detail, the new RAN (NR) as the radio access network and the packet core (5G system, 5G core network, or next-generation (NG) core) as the core network based on the 5G mobile communication standard specified by the 3rd Generation Partnership Project Long-Term Evolution (3GPP), which is a mobile communication standard standardization organization, are the main objectives. However, without significantly departing from the scope of the disclosure, the main gist of the disclosure can be slightly modified and applied to other communication systems with a similar technical background, which can be determined by those skilled in the technical field of the present disclosure.

[0032] In the 5G system, in order to support network automation, a network data collection and analysis function (NWDAF) can be defined. NWDAF is a network function that provides the function of analyzing and providing the data collected from the 5G network. NWDAF can collect, store, and analyze information from the 5G network, and provide the results to unspecified network functions (NFs), and the analysis results can be independently used by each NF.

[0033] For the convenience of the following description, some terms and names defined in the 3GPP standard (standards for 5G, NR, LTE, or similar systems) can be used. However, the disclosure is not limited to the terms and names of the present disclosure, and can equally apply to systems that conform to other standards.

[0034] In the following description, for the convenience of explanation, terms representing signals, terms representing channels, terms representing control information, terms representing network entities, terms representing device components, etc. are used as examples. Therefore, the disclosure is not limited to the terms used in the disclosure, and other terms representing objects with equivalent technical meanings can be used.

[0035] Hereinafter, in the disclosure, high-layer signaling refers to the method of sending a signal from a base station to a terminal by using the downlink data channel of the physical layer, or sending a signal from the terminal to the base station by using the uplink data channel of the physical layer. High-layer signaling can be understood as radio resource control (RRC) signaling or media access control (MAC) control element (CE).

[0036] In addition, in the disclosure, the expressions "~ greater than (or more than)" or "~ less than" are used to determine whether a specific condition is satisfied or achieved, but this is only a description for expressing examples, and does not exclude the description of "~ or more" or "~ or less". The condition written as "~ or more" can be replaced by "~ greater than (or more than)", the condition written as "~ or less" can be replaced by "~ less than", and the condition written as "~ or more and ~ less" can be replaced by "~ greater than (or more than) and ~ or less".

[0037] In addition, the disclosure describes embodiments by using terms used in some communication standards (e.g., 3GPP), but this is only an example for illustration. The disclosed embodiments can be easily modified and applied to other communication systems.

[0038] In the disclosure, an artificial intelligence (AI) / machine learning (ML) model can represent a hardware configuration and / or a software configuration for learning (or training) a specific pattern of a data set and inferring / predicting a target function of an AI / ML-based communication technology based on the learned pattern. The AI model can cover the ML model. The AI / ML function can represent a hardware configuration and / or a software configuration (or a set of configurations) for performing a pre-agreed target feature (or a pre-agreed target function) or a pre-agreed target feature group (or a pre-agreed target function group) between the terminal device 100 and the base station device 200 in AI / ML-based communication. The AI function can cover the ML function.

[0039] In the disclosure, a unilateral AI / ML model or a unilateral AI / ML function can represent an AI / ML model or an AI / ML function trained only in the terminal device 100 or the base station device 200.

[0040] In the disclosure, a bilateral AI / ML model or a bilateral AI / ML function can represent an AI / ML model or an AI / ML function trained in both the terminal device 100 and the base station device 200. The bilateral AI / ML model and the bilateral AI / ML function can also be referred to as a distributed AI / ML model and a distributed AI / ML function.

[0041] In the disclosure, an electronic device can represent the terminal device 100, and an external electronic device can represent the base station device 200.

[0042] In the disclosure, embodiments described in terms of the operation of the terminal device 100 can also be performed in the operation of the base station device 200, and embodiments described in terms of the operation of the base station device 200 can also be performed in the operation of the terminal device 100.

[0043] Figure 1 A wireless communication system according to one or more embodiments is shown.

[0044] Figure 1 The terminal (or terminal device) 100, the base station (or base station device) 200, and the terminal (or terminal device) 300 are shown as part of nodes using a wireless channel in the wireless communication system. Figure 1 Only one base station is shown, but other base stations having the same or a similar structure as that of the base station 200 can also be included in the wireless communication system.

[0045] Base station 200 is a network infrastructure that provides wireless access to terminals 100 and 300. The base station 200 has a coverage area defined as a specific geographical area based on the distance of the signal that can be transmitted. The base station 200 can also be referred to as an "access point (AP)", "evolved Node B (eNodeB) (eNB)", "fifth-generation node (5G node)", "next-generation NodeB (gNB)", "radio point", "transmission / reception point (TRP)", or other terms with equivalent technical meanings other than the base station.

[0046] Each of terminals 100 and 300 is a device used by a user and communicates with the base station 200 through a wireless channel. The link from the base station 200 to terminal 100 or terminal 300 is called the downlink (DL), and the link from terminal 100 or terminal 300 to the base station 200 is called the uplink (UL). In addition, terminals 100 and 300 can communicate with each other through a wireless channel. In this case, the link between terminal 100 and terminal 300 is called a sidelink (or side link, auxiliary link, edge link), and the sidelink can also be referred to as the PC5 interface. In some cases, at least one of terminals 100 and 300 can be operated without user participation. That is, at least one of terminals 100 and 300 can be a device that performs machine-type communication (MTC) and may not be carried by a user. Each of terminals 100 and 300 can be referred to as a "user equipment (UE)", "mobile station", "subscriber station", "remote terminal", "wireless terminal", "user device", or other terms with equivalent technical meanings other than the terminal.

[0047] Base station 200, terminal 100, and terminal 300 can transmit and receive wireless signals in the millimeter-wave (mmWave) frequency band (e.g., 28 GHz, 30 GHz, 38 GHz, and 60 GHz). In this case, to improve the channel gain, base station 200, terminal 100, and terminal 300 can perform beamforming. Beamforming can include transmit beamforming and receive beamforming. That is, base station 200, terminal 100, and terminal 300 can impart directionality to the transmitted signal or the received signal. For this purpose, base station 200 and terminals 100 and 300 can select service beams 201, 202, 101, and 301 through a beam search or beam management process. After selecting service beams 201, 202, 101, and 301, subsequent communication can be performed through resources that are in a quasi co-located (QCL) relationship with the resources of transmit service beams 201, 202, 101, and 301.

[0048] If the large-scale characteristics of the channel carrying the symbols on the first antenna port can be inferred from the channel carrying the symbols on the second antenna port, the first antenna port and the second antenna port can be evaluated as being in a QCL relationship. For example, the large-scale characteristics can include at least one of delay spread, Doppler spread, Doppler shift, average gain, average delay, and spatial receiver parameters.

[0049] Figure 1 The terminal 100 shown in is capable of supporting vehicle communication. In the case of vehicle communication, in the LTE system, the standardization work on V2X technology was completed in 3GPP Release 14 and Release 15 based on the device-to-device (D2D) communication structure, and currently efforts are being made to develop V2X technology based on 5G NR. In NR V2X, unicast communication, multicast (or broadcast) communication between terminals can be supported. In addition, different from LTE V2X which aims to transmit and receive the basic safety information required for vehicles driving on the road, NR V2X aims to provide more advanced services (such as platooning, advanced driving, extended sensors, and remote driving).

[0050] Hereinafter, the base station is an entity that performs resource allocation for the terminal, and can be a base station that supports both V2X communication and general cellular communication, or a base station that only supports V2X communication. That is, the base station can represent an NR base station (e.g., gNB), an LTE base station (e.g., eNB), or a roadside unit (RSU). Examples of terminals can include not only general UEs and mobile stations, but also vehicles that support vehicle-to-vehicle (V2V) communication, vehicles that support vehicle-to-pedestrian (V2P) communication, vehicles or pedestrian handsets (e.g., smartphones) that support V2P communication, vehicles that support vehicle-to-network (V2N) communication, vehicles that support vehicle-to-infrastructure (V2I) communication, RSUs equipped with terminal functions, RSUs equipped with base station functions, and RSUs equipped with a part of the base station function and a part of the terminal function.

[0051] The base station and the terminal are connected through the uplink (UL) / downlink (DL) interface. The uplink (UL) represents the radio link through which the terminal sends data or control signals to the base station, and the downlink (DL) represents the radio link through which the base station sends data or control signals to the terminal.

[0052] As the 5G wireless communication system is commercialized, various types of connected devices can be connected to the communication network. Therefore, it is expected that strengthening the functions and performance of the 5G wireless communication system and the integrated operation of the connected devices will be necessary. Specifically, in 3GPP Release 18, various research projects have been conducted to introduce / apply AI / ML-based communication technologies to the 5G NR wireless communication system. For example, in 3GPP Release 18, it is planned to apply AI / ML-based communication technologies to three representative scenarios (e.g., channel state information (CSI) reporting, beam management, and positioning). Before applying AI / ML-based communication technologies to the above three scenarios, 3GPP has constructed a basic framework for AI / ML that represents the entire process of the life cycle management (LCM) of AI / ML. The following will describe this with reference to Figure 2 This is described below.

[0053] Specifically, the LCM of AI / ML can be subdivided into model ID-based LCM and function-based LCM. An AI / ML model can represent the hardware configuration and / or software configuration for learning (or training) a dataset in a specific pattern and inferring / predicting the target function of the AI / ML-based communication technology based on the learned pattern. The model ID represents a unique identifier for each AI / ML model (e.g., logical model ID, serial number, index, etc.). For example, mapping the logical model ID to the physically implemented physical model ID can be performed according to the implementation of each terminal device 100. The AI / ML function can represent the hardware configuration and / or software configuration (or a set of configurations) for performing a pre-agreed target feature (or pre-agreed target function) or a pre-agreed group of target features (or pre-agreed group of target functions) between the terminal device 100 and the base station device 200 in AI / ML-based communication.

[0054] Figure 2 An example of the AI / ML framework 10 according to one or more embodiments is shown. The AI / ML framework 10 can be configured using components included in the terminal device and the base station.

[0055] Referring to Figure 2 According to one or more embodiments, the AI / ML framework 10 can include a data collection module 11, a model training module 12, a management module 13, a model storage device 14, and an inference module 15. Each of the terminal device 100 and the base station device 200 according to one or more embodiments can correspond to (or include) at least one of the data collection module 11, the model training module 12, the management module 13, the model storage device 14, and the inference module 15.

[0056] The data collection module 11 can pre-collect the data required to define at least one AI / ML model or at least one AI / ML function or collect the data in real time. The data collection module 11 can send the collected data as data for training at least one AI / ML model or at least one AI / ML function (e.g., training data) to the model training module 12. The data collection module 11 can send the collected data as data for inferring at least one AI / ML model or at least one AI / ML function (e.g., inference data) to the inference module 15.

[0057] The model training module 12 can train at least one AI / ML model or at least one AI / ML function based on the training data received from the data collection module 11. The model training module 12 can send the trained at least one AI / ML model or at least one AI / ML function to the model storage device 14.

[0058] The model training module 12 can update the previously set at least one AI / ML model or at least one AI / ML function based on the training data received from the data collection module 11 and the feedback of the inference results received from the management module 13. The model training module 12 can send the updated at least one AI / ML model or at least one AI / ML function to the model storage device 14.

[0059] The model storage device 14 can store the trained at least one AI / ML model or at least one AI / ML function received from the model training module 12. The model storage device 14 can store the updated at least one AI / ML model or at least one AI / ML function received from the model training module 12.

[0060] The model storage device 14 can send the stored at least one AI / ML model or at least one AI / ML function to the inference module 15 in response to a model transfer request received from the management module 13. The model can include at least one AI / ML model or at least one AI / ML function.

[0061] The inference module 15 can operate at least one AI / ML model or at least one AI / ML function based on the inference data and perform inference (or estimation) on a specific function (or feature) or a specific function group (or feature group) of the communication technology based on the target AI / ML.

[0062] The inference module 15 can generate monitoring results for the operation performance (e.g., inference performance) of at least one AI / ML model or the operation performance (e.g., inference performance) of at least one AI / ML function and output the monitoring results to the management module 13.

[0063] The management module 13 may maintain or update (or change) the settings of at least one AI / ML model or at least one AI / ML function based on the monitoring results received from the inference module 15. When updating the settings of at least one AI / ML model or at least one AI / ML function, the management module 13 may send control information (settings) regarding the changes of at least one AI / ML model or at least one AI / ML function to the inference module 15. In this case, the control information (settings) may include information indicating any one of switching, selecting, activating, and deactivating at least one AI / ML model or at least one AI / ML function with the previous settings, or a fallback to at least one default AI / ML model or at least one default AI / ML function. For example, when it is necessary to select at least one AI / ML model according to a specific communication environment, the management module 13 may send control information (settings) including a shared model ID (predefined) to the inference module 15. The inference module 15 shares the IDs of at least one AI / ML model that "the inference module 15 can support" with the management module 13.

[0064] The management module 13 may generate feedback on the training direction of at least one AI / ML model or at least one AI / ML function based on the monitoring results and send the generated feedback to the model training module 12. The model training module 12 may retrain at least one AI / ML model or at least one AI / ML function by reflecting the feedback received from the management module 13 during the training process. The management module 13 may store at least one retrained AI / ML model or at least one retrained AI / ML function in the model storage device 14.

[0065] As described above, the management module 13 may require various signaling for communicating with the inference module 15 to efficiently manage at least one AI / ML model or at least one AI / ML function. For example, the management module 13 may update at least one AI / ML model or at least one AI / ML function based on the signaling between the terminal device and the base station.

[0066] According to one or more disclosed embodiments, there may be provided a method and apparatus for efficiently managing at least one AI / ML model or at least one AI / ML function based on signaling (e.g., the capability information, preference information of the inference module 15, and signaling control information between the terminal device 100 and the base station device 200) between the inference module 15 (e.g., the terminal device 100 in Figure 1 and the management module 13 (e.g., the base station device 200 in Figure 1 . A detailed description will be given below with reference to Figures 5 to 8

[0067] Hereinafter, with reference to the accompanying drawings (e.g., Figures 3 to 10), for ease of explanation, the inference module 15 is described as the terminal device 100, and the management module 13 is described as the base station device 200. However, the disclosure is not limited thereto. According to one or more embodiments, the inference module 15 can be the terminal device 100 or the base station device 200 according to the communication environment or scenario, and the management module 13 can be the terminal device 100 or the base station device 200 according to the communication environment or scenario.

[0068] According to various embodiments, the device and method can reduce unnecessary signaling and processing between the terminal device 100 and the base station device 200 by managing the AI / ML model or AI / ML function based on the signaling between the terminal device 100 and the base station device 200 (e.g., the capability information and / or preference information of the terminal device 100).

[0069] In addition, the embodiments of the present disclosure can prevent waste of radio resources and / or processing delay caused by unnecessary signaling / processing. Therefore, the management of the AI / ML model or AI / ML function can be efficiently executed.

[0070] In addition, various disclosed embodiments can prevent excessive signaling overhead by controlling the transmission timing of the signaling between the terminal device 100 and the base station device 200.

[0071] Figure 3 Illustrates the Figure 1 configuration of the terminal device 100 in

[0072] Referring to Figure 3 , the terminal device 100 includes a processor 110, a memory 120, and a communication interface 130. The processor 110, the memory 120, and the communication interface 130 can be implemented as hardware, software, or a combination of hardware and software. For example, the communication interface can be implemented by any one or any combination of a digital modem, a radio frequency (RF) modem, an antenna circuit, a wireless fidelity (WiFi) chip, and related software and / or firmware.

[0073] The communication interface 130 can send and receive signals through a wireless channel. For example, the communication interface 130 performs a conversion function between the baseband signal and the bit string according to the physical layer specification of the system. For example, when sending data, the communication interface 130 generates complex symbols by encoding and modulating the transmission bit string. In addition, when receiving data, the communication interface 130 recovers the received bit string by demodulating and decoding the baseband signal. In addition, the communication interface 130 up-converts the baseband signal to an RF band signal and sends the RF band signal through the antenna, and down-converts the RF band signal received through the antenna to a baseband signal. For example, the communication interface 130 can include a transmit filter, a receive filter, an amplifier, a mixer, an oscillator, a digital-to-analog converter (DAC), an analog-to-digital converter (ADC), etc.

[0074] In addition, the communication interface 130 may include a plurality of transmission paths and reception paths. In addition, the communication interface 130 may include at least one antenna array including a plurality of antenna elements. In terms of hardware, the communication interface 130 may include digital circuits and analog circuits (e.g., radio frequency integrated circuits (RFICs)). In this case, the digital circuits and the analog circuits may be implemented in one package. In addition, the communication interface 130 may include a plurality of RF chains. In addition, the communication interface 130 may perform beamforming.

[0075] As described above, the communication interface 130 transmits and receives signals. Accordingly, all or part of the communication interface 130 may be referred to as a "transmitter", "receiver", or "transceiver". In addition, in the following description, the transmission and reception performed through the wireless channel are used to represent the processing performed by the communication interface 130 as described above.

[0076] The memory 120 stores data (such as a basic program (e.g., an operating system or system software), an application program, and setting information for the operation of the terminal device 100). The memory 120 may include a volatile memory, a non-volatile memory, or a combination of a volatile memory and a non-volatile memory. The memory 120 provides the stored data according to the request of the processor 110. In one or more embodiments, the memory 120 may store at least one AI / ML model or at least one AI / ML function group for AI / ML-based wireless communication between the terminal device 100 and the base station device 200.

[0077] The processor 110 controls the overall operation of the terminal device 100. For example, the processor 110 transmits and receives signals through the communication interface 130. In addition, the processor 110 writes data to the memory 120 and reads data from the memory 120. In addition, the processor 110 may execute protocol stack functions required by communication standards. To this end, the processor 110 may include at least one processor or microprocessor, or may be a part of a processor. In addition, a part of the communication interface 130 and the processor 110 may be referred to as a communication processor (CP). According to an embodiment, the processor 110 may control the terminal device 100 to perform operations according to the embodiments described below ( Figures 5 to 8 operations).

[0078] Figure 4 illustrates the configuration of the base station device 200 in Figure 1 according to one or more embodiments.

[0079] Referring to Figure 4 , the base station device 200 includes a processor 210, a communication interface 220, and a memory 230. The processor 210, the communication interface 220, and the memory 230 may be implemented as hardware, software, or a combination of hardware and software.

[0080] The communication interface 220 performs functions for transmitting and receiving signals through a wireless channel. For example, the communication interface 220 performs a conversion function between a baseband signal and a bit string according to the physical layer specification of the system. For example, when transmitting data, the communication interface 220 generates complex symbols by encoding and modulating the transmitted bit string. In addition, when receiving data, the communication interface 220 recovers the received bit string by demodulating and decoding the baseband signal.

[0081] In addition, the communication interface 220 up-converts the baseband signal to an RF band signal and transmits the RF band signal through an antenna, and down-converts the RF band signal received through the antenna to a baseband signal. To this end, the communication interface 220 may include a transmit filter, a receive filter, an amplifier, a mixer, an oscillator, a DAC, an ADC, etc. In addition, the communication interface 220 may include a plurality of transmit paths and receive paths. In addition, the communication interface 220 may include at least one antenna array including a plurality of antenna elements.

[0082] In terms of hardware, the communication interface 220 may include a digital unit and an analog unit, and the analog unit may include a plurality of sub-units according to the operating power, operating frequency, etc. The digital unit may be implemented by at least one processor (e.g., a digital signal processor (DSP)).

[0083] As described above, the communication interface 220 transmits and receives signals. Therefore, all or part of the communication interface 220 may be referred to as a "transmitter", "receiver", or "transceiver". In addition, in the following description, the transmission and reception performed through the wireless channel are used to represent the processing performed by the communication interface 220 as described above.

[0084] The communication interface 220 provides an interface for performing communication with other nodes in the network (e.g., backhaul communication). That is, the communication interface 220 converts the bit string transmitted from the base station device 200 to another node (e.g., another access node, another base station, an upper node, or a core network) into a physical signal, and converts the physical signal received from another node into a bit string.

[0085] The memory 230 stores data (such as basic programs, application programs, and setting information for the operation of the base station device 200). The memory 230 may include a volatile memory, a non-volatile memory, or a combination of a volatile memory and a non-volatile memory. The memory 230 provides the stored data according to the request of the processor 210.

[0086] The processor 210 controls the overall operation of the base station apparatus 200. For example, the processor 210 transmits signals and receives signals through the communication interface 220. In addition, the processor 210 writes data to the memory 230 and reads data from the memory 230. In addition, the processor 210 may execute the functions of the protocol stack required by the communication standard. According to another exemplary embodiment, the protocol stack may be included in the communication interface 220. To this end, the processor 210 may include at least one processor. According to an embodiment, the processor 210 may control the base station apparatus 200 to perform operations according to the embodiments described below ( Figures 5 to 8 operations).

[0087] Figure 5 is a flowchart showing a method for managing an AI / ML model or AI / ML functions in a wireless communication system according to one or more embodiments.

[0088] Referring to Figure 5 , a method for managing an AI / ML model or AI / ML functions based on signaling between the terminal device 100 and the base station apparatus 200 in a wireless communication system may include operation S110 to operation S150. Figure 5 The signaling between the terminal device 100 and the base station apparatus 200 in

[0089] In the disclosure, the AI / ML model may represent a hardware configuration and / or software configuration for learning (or training) a specific pattern of a data set and inferring / predicting a target function of an AI / ML-based communication technology based on the learned pattern. The AI / ML function may represent a hardware configuration and / or software configuration (or a set of configurations) for performing a pre-agreed target feature (or pre-agreed target function) or a pre-agreed target feature group (or pre-agreed target function group) between the terminal device 100 and the base station apparatus 200 in AI / ML-based communication.

[0090] In the disclosure, a unilateral AI / ML model or unilateral AI / ML function may represent an AI / ML model or AI / ML function trained only in the terminal device 100 or the base station apparatus 200.

[0091] In the disclosure, a bilateral AI / ML model or bilateral AI / ML function may represent an AI / ML model or AI / ML function trained in both the terminal device 100 and the base station apparatus 200.

[0092] In operation S110, the terminal device 100 (e.g., the processor 110 of the terminal device 100) may configure the capability information of the terminal device 100, and the capability information includes information about the number of at least one AI / ML model or the number of at least one AI / ML function that the terminal device 100 can support. For example, different identifiers (IDs) or serial numbers may be assigned to each AI / ML model and / or AI / ML function in a one-to-one correspondence. The terminal device 100 may provide the count of these IDs or serial numbers as the number of AI / ML models and / or AI / ML functions to the base station device 200. In operation S120, the terminal device 100 may send a first message including the capability information of the terminal device 100 to the base station device 200. For example, the processor 110 of the terminal device 100 may control the communication interface 130 to send a first message including the capability information of the terminal device 100 to the base station device 200. In this case, the first message may be sent through various types of signaling including UE capability signaling between the terminal device 100 and the base station device 200.

[0093] In one or more embodiments, the capability information of the terminal device 100 may include information about the maximum number of at least one AI / ML model that the terminal device 100 can support. The capability information of the terminal device 100 may include the information shown in [Table 1]. For example, when the maximum number of at least one AI / ML model that the terminal device 100 can support is "a", the terminal device 100 may enter "a" in NR_AIML_Max_UE_model (e.g., the capability information of the terminal device 100) in [Table 1] and send the NR_AIML_Max_UE_model with the input "a" to the base station device 200 (where a, b, c, d, e, f, and nmax are integers greater than or equal to 0, and nmax may be defined in the 3GPP mobile communication standard). For example, in the following table, ENUMERATED{} may be an enumeration function, and spare or spare~ may represent spare inputs.

[0094] [Table 1]

[0095] In one or more embodiments, the capability information of the terminal device 100 may include information about the maximum number of at least one AI / ML function that the terminal device 100 is capable of supporting. For example, when the maximum number of at least one AI / ML function that the terminal device 100 is capable of supporting is "b", the terminal device 100 may input "b" in NR_AIML_Max_UE_functionality (e.g., the capability information of the terminal device 100) in [Table 2], and send the NR_AIML_Max_UE_functionality with the input "b" to the base station device 200 (where a, b, c, d, e, f, and nmax are integers greater than or equal to 0, and nmax may be defined in the 3GPP mobile communication standard).

[0096] [Table 2]

[0097] In one or more embodiments, the capability information of the terminal device 100 may include at least one of information about the maximum number of one-sided AI / ML models that the terminal device 100 is capable of supporting among at least one AI / ML model and information about the maximum number of two-sided AI / ML models that the terminal device 100 is capable of supporting among at least one AI / ML model. For example, when the maximum number of one-sided AI / ML models that the terminal device 100 is capable of supporting is "c" and the maximum number of two-sided AI / ML models that the terminal device 100 is capable of supporting is "d", the terminal device 100 may input "c" in NR_AIML_Max_UE_model_onesided (e.g., the capability information of the terminal device 100) in [Table 3] and input "d" in NR_AIML_Max_UE_model_twosided (e.g., the capability information of the terminal device 100) in [Table 3], and may send the NR_AIML_Max_UE_model_onesided with the input "c" and the NR_AIML_Max_UE_model_twosided with the input "d" to the base station device 200 (where a, b, c, d, e, f, nmax are integers greater than or equal to 0, and nmax may be defined in the 3GPP mobile communication standard). The information about the maximum number of one-sided AI / ML models that the terminal device 100 is capable of supporting (e.g., NR_AIML_Max_UE_model_onesided) and the information about the maximum number of two-sided AI / ML models that the terminal device 100 is capable of supporting among at least one AI / ML model (e.g., NR_AIML_Max_UE_model_twosided) in [Table 3] may be sent together or separately as needed.

[0098] [Table 3]

[0099] In one or more embodiments, the capability information of the terminal device 100 may include at least one of information about the maximum number of one-sided AI / ML functions that the terminal device 100 can support among at least one AI / ML function and information about the maximum number of two-sided AI / ML functions that the terminal device 100 can support among at least one AI / ML function.

[0100] For example, when the maximum number of one-sided AI / ML functions that the terminal device 100 can support is "e" and the maximum number of two-sided AI / ML functions that the terminal device 100 can support is "f", the terminal device 100 may input "e" in NR_AIML_Max_UE_functionality_onesided of [Table 4] (e.g., the capability information of the terminal device 100) and input "f" in NR_AIML_Max_UE_functionality_twosided of [Table 4] (e.g., the capability information of the terminal device 100), and may send NR_AIML_Max_UE_functionality_onesided with the input "e" and NR_AIML_Max_UE_functionality_twosided with the input "f" to the base station device 200 (where a, b, c, d, e, f, nmax are integers greater than or equal to 0, and nmax may be defined in the 3GPP mobile communication standard). Information about the maximum number of one-sided AI / ML functions that the terminal device 100 can support in [Table 4] (e.g., NR_AIML_Max_UE_functionality_onesided) and information about the maximum number of two-sided AI / ML functions that the terminal device 100 can support among at least one AI / ML function (e.g., NR_AIML_Max_UE_functionality_twosided) may be sent together or separately as needed.

[0101] [Table 4]

[0102] In one or more embodiments, information in [Table 3] regarding the maximum number of single-sided / double-sided AI / ML models that the terminal device 100 can support (e.g., NR_AIML_Max_UE_model_onesided and NR_AIML_Max_UE_model_twosided) and information in [Table 4] regarding the maximum number of single-sided / double-sided AI / ML functionalities that the terminal device 100 can support (e.g., NR_AIML_Max_UE_functionality_onesided and NR_AIML_Max_UE_functionality_twosided) can be sent together or separately as needed.

[0103] In operation S130, the base station device 200 may configure control information regarding at least one AI / ML model or at least one AI / ML functionality in the terminal device 100 based on the first message. In operation S140, the base station device 200 may send a second message including the control information regarding at least one AI / ML model or at least one AI / ML functionality to the terminal device 100. In this case, the control information may include information or instructions for switching, selecting, activating, and / or deactivating at least one previously set AI / ML model or at least one AI / ML functionality. The control information may also include a fallback instruction for reverting to a default AI / ML model or default AI / ML functionality. For example, when it is necessary to select a specific AI / ML model, the base station device 200 may send control information instructing the terminal device 100 to set the AI / ML model according to a pre-shared model ID (pre-defined) to the terminal device 100.

[0104] In operation S150, the terminal device 100 may set at least one AI / ML model or at least one AI / ML functionality according to the second message. For example, the terminal device 100 may set at least one AI / ML model or at least one AI / ML functionality based on the control information of the base station device 200 included in the second message.

[0105] As described above, the base station device 200 may request a specific AI / ML model or function from the terminal device 100 via a second message (e.g., control information). The terminal device 100 may provide feedback to report the AI / ML model or AI / ML function that best matches the request from among the options supported by the terminal device 100 (i.e., at least one AI / ML model or at least one AI / ML function stored in the terminal device 100 and supported by the terminal device 100). Thereafter, if it is found that the operating performance of the AI / ML model or function set according to the previous second message (e.g., control information) is insufficient (e.g., below a predetermined performance threshold), the base station device 200 sends an updated second message (e.g., changed / reconfigured control information) to the terminal device 100. However, if the base station device 200 does not have prior knowledge of the maximum number of AI / ML models or functions that the terminal device 100 can support (i.e., the capability information of the terminal device 100 for AI / ML models or functions), inefficient use of communication resources and communication delays may result due to unnecessary signaling (such as repeated requests for AI / ML model reset or redundant feedback).

[0106] According to one or more embodiments of the present disclosure, these problems can be solved by providing a device and method capable of efficiently managing AI / ML models or functions. By including capability information (specifically, the maximum number of AI / ML models or functions that the terminal device 100 can support) in the signaling, unnecessary signaling and communication delays are minimized. This leads to more efficient management of AI / ML resources, reduced waste, and improved overall system performance.

[0107] In Figure 5 , the operations of the terminal device 100 may also be performed by the base station device 200 or through interaction with the base station device 200, and the operations of the base station device 200 may also be performed by the terminal device 100 or through interaction with the terminal device 100.

[0108] Figure 6 is a flowchart showing a method for managing an AI / ML model or an AI / ML function in a wireless communication system according to one or more embodiments.

[0109] Referring to Figure 6 , a method for managing an AI / ML model or an AI / ML function based on signaling between the terminal device 100 and the base station device 200 in a wireless communication system may include operations S210 to S250. In Figure 6 's description, a description that is substantially the same as the description given with reference to Figure 5 is replaced with Figure 5 's description.

[0110] In the disclosure, an AI / ML model may represent a hardware configuration and / or a software configuration for learning (or training) a specific pattern of a data set and inferring / predicting a target function of an AI / ML-based communication technology based on the learned pattern. The AI / ML function may represent a hardware configuration and / or a software configuration (or a set of configurations) for performing a pre-agreed target feature (or a pre-agreed target function) or a pre-agreed target feature group (or a pre-agreed target function group) between the terminal device 100 and the base station device 200 in AI / ML-based communication.

[0111] During the process of establishing a wireless communication connection with the base station device 200 (e.g., the initial setup process), the terminal device 100 may notify the base station device 200 in advance of the terminal device 100's preference information reporting capability (i.e., the terminal device 100 may send / report opinions on at least one AI / ML model or at least one AI / ML function to the base station device 200, or the terminal device 100 may be able to send / report the terminal device 100's AI / ML model preference or AI / ML function preference to the base station device 200).

[0112] In operation S210, the terminal device 100 may configure the preference information of the terminal device 100 for at least one AI / ML model or at least one AI / ML function, and in operation S220, the terminal device 100 may send a third message including the preference information of the terminal device 100. In this case, the preference information may be included in the UE assistance information (UAI) of 3GPP release 15, but is not limited thereto, and may be sent to the base station device 200 through various forms of higher layer signaling.

[0113] In one or more embodiments, the terminal device 100 may monitor the operational performance of at least one previously set AI / ML model or at least one AI / ML function, and may configure preference information regarding at least one AI / ML model or at least one AI / ML function. In this case, the preference information may include information for the terminal device 100 to request resetting of at least one AI / ML model or at least one AI / ML function (e.g., information for requesting selection, activation, deactivation, switching, or fallback of at least one AI / ML model or at least one AI / ML function). The preference information of the terminal device 100 may include the information shown in [Table 5]. For example, the terminal device 100 may send the preference information of the terminal device 100 regarding each of selection, activation, deactivation, switching, and fallback in AIML-Assistance-r19 of [Table 5] to the base station device 200. In another example, the terminal device 100 may select a preferred one among selection, activation, deactivation, switching, and fallback in AIML-Assistance-r19 of [Table 5] based on the monitoring result and send the selected one to the base station device 200.

[0114] [Table 5]

[0115] In this case, in AIML-Assistance-r19 of [Table 5], each of selection, activation, deactivation, switching, and fallback may be displayed as mapped data. For example, when the terminal device 100 requests the base station device 200 to perform a switching operation on the AI / ML model or AI / ML function, the data “010” mapped to switching may be input into AIML-Assistance-r19 of [Table 5] and sent.

[0116] As described above, an example of a method for displaying the preference information of the terminal device 100 has been described. However, the disclosure is not limited thereto, and the preference information according to one or more embodiments may be displayed in various forms and sent to the base station device 200.

[0117] In operation S230, the base station device 200 may determine whether to reconfigure the control information regarding at least one AI / ML model or at least one AI / ML function based on the third message. The base station device 200 may perform operation S240 when reconfiguring the control information regarding at least one AI / ML model or at least one AI / ML function based on the third message, and may terminate the operation when not reconfiguring the control information regarding at least one AI / ML model or at least one AI / ML function based on the third message (i.e., may maintain the control information regarding at least one existing AI / ML model or at least one existing AI / ML function).

[0118] In operation S240, the base station device 200 may send a fourth message including reconfigured control information regarding at least one AI / ML model or at least one AI / ML function.

[0119] In operation S250, the terminal device 100 may reset at least one AI / ML model or at least one AI / ML function according to the fourth message. For example, the terminal device 100 may perform selection, activation, deactivation, switching, or fallback operations on at least one AI / ML model or at least one AI / ML function according to the fourth message.

[0120] As described above, in accordance with one or more embodiments of the disclosure, when the operation performance of at least one previously set AI / ML model or at least one AI / ML function deteriorates (e.g., when the operation performance becomes lower than a predetermined performance threshold), the terminal device 100 may configure the preference information of the terminal device 100 based on the monitoring result and send the preference information to the base station device 200. The operation performance of the AI / ML model or function may be evaluated using measurements of one or more of the following metrics: accuracy and error rate (e.g., by comparing the model prediction with the ground truth or reference data), response time, throughput (e.g., the number of data items (such as images) processed per minute), resource utilization (e.g., CPU or memory usage), and user feedback. That is, when the operation performance of at least one previously set AI / ML model or at least one AI / ML function deteriorates / becomes unnecessary, the terminal device 100 may send a reset request for at least one AI / ML model or at least one AI / ML function to the base station device 200 as preference information.

[0121] Therefore, according to one or more embodiments of the disclosure, an apparatus and method capable of performing adaptive management of an AI / ML model or an AI / ML function through signaling between the terminal device 100 and the base station device 200 may be provided, the signaling including preference information of the terminal device 100 configured based on the monitoring result of the operation performance of at least one previously set AI / ML model or at least one AI / ML function (i.e., information requesting to reset the AI / ML model or AI / ML function preferred by the terminal device 100).

[0122] In Figure 6 it, the operation of the terminal device 100 may also be performed by the base station device 200 or via interaction with the base station device 200, and the operation of the base station device 200 may also be performed by the terminal device 100 or via interaction with the terminal device 100.

[0123] Figure 7It is a flowchart showing a method for managing an AI / ML model or an AI / ML function in a wireless communication system according to one or more embodiments.

[0124] Specifically, Figure 7 It is a diagram showing a method for managing an AI / ML model or an AI / ML function based on signaling (e.g., signaling of "transmitting control information") between a terminal device 100 and a base station device 200.

[0125] Referring to Figure 7 , a method for managing an AI / ML model or an AI / ML function based on signaling between a terminal device 100 and a base station device 200 in a wireless communication system may include operation S301 to operation S350.

[0126] In the present disclosure, the AI / ML model may represent a hardware configuration and / or a software configuration for learning (or training) a specific pattern of a data set and inferring / predicting a target function of an AI / ML-based communication technology based on the learned pattern. The AI / ML function may represent a hardware configuration and / or a software configuration (or a set of configurations) for performing a pre-agreed target feature (or a pre-agreed target function) or a pre-agreed target feature group (or a pre-agreed target function group) between the terminal device 100 and the base station device 200 in AI / ML-based communication.

[0127] The terminal device 100 may notify the base station device 200 of the preference information reporting ability of the terminal device 100 (i.e., the terminal device 100 may send / report opinions on at least one AI / ML model or at least one AI / ML function) in advance during the process of establishing a wireless communication connection with the base station device 200 (e.g., the initial setup process).

[0128] In operation S301, the base station device 200 may configure "transmission control information" for the terminal device 100, and in operation S303, the base station device 200 may send a control message including the "transmission control information". In this case, the control message may be signaling in the initial setup process or the RRCReconfiguration process of the wireless communication connection between the terminal device 100 and the base station device 200, but is not limited thereto.

[0129] In one or more embodiments, the transmission control information may include specifying "prevent the terminal device 100 from sending Figure 6The prohibition timer (e.g., prohibition timer value) for the "holding" period of the third message (including the preference information of the terminal device 100) shown in [Table 6]. The holding period can be determined in advance by the base station device 200. The transmission control information may include the information shown in [Table 6]. For example, the transmission control information may include information about the transmission time of the third message, the period for pausing the transmission of the third message in the terminal device 100, and the transmission frequency of the third message.

[0130] In one or more embodiments, the base station device 200 may select one of the multiple holding periods (e.g., s0, s0dot5, s1, s2, s5, s10, s20, s30, s60, s90, s120, s300, and s600) included in [Table 6] to configure the transmission control information. For example, when the holding period is "1 second", the base station device 200 may input / declare "s1" in the AIMLPreferenceIndicationProhibitTimer (e.g., transmission control information) in [Table 6] and send the AIMLPreferenceIndicationProhibitTimer with "s1" written / declared to the terminal device 100. The terminal device 100 may need to pause the next third message within the holding period (e.g., 1 second) starting from the time of the first transmission of the third message including the preference information of the terminal device 100 (the initial transmission time of the third message by the terminal device 100 may be scheduled by the base station device 200). After the expiration of the holding period (e.g., 1 second), the terminal device 100 may send the next third message to the base station device 200. In another example, when the holding period is "0.5 second", the base station device 200 may input / declare "s0dot5" in the AIMLPreferenceIndicationProhibitTimer (e.g., transmission control information) in [Table 6] and send it to the terminal device 100. The terminal device 100 may need to pause the next third message within the holding period (e.g., 0.5 second) starting from the time point of the first transmission of the third message including the preference information of the terminal device 100 (the initial transmission time point of the third message by the terminal device 100 may be scheduled by the base station device 200). After the expiration of the holding period (e.g., 0.5 second), the terminal device 100 may send the next third message to the base station device 200.

[0131] [Table 6]

[0132] In the above, it is described that the base station device 200 transmits transmission control information (e.g., AIMLPreferenceIndicationProhibitTimer) configured by selecting one of the pause periods s0, s0dot5, s1, s2, s5, s10, s20, s30, s60, s90, s120, s300, and s600 in [Table 6] to the terminal device 100, but it is not limited thereto. The base station device 200 according to one or more embodiments may configure transmission control information including various pause periods according to the communication environment and scenario and control the transmission of the third message (e.g., the preference information of the terminal device 100) of the terminal device 100.

[0133] Because Figure 7 the description of operations S310 to S350 is substantially the same as Figure 6 the description of operations S210 to S250, so Figure 7 the description of operations S310 to S350 is replaced with Figure 6 the description of operations S210 to S250.

[0134] As described above, in accordance with one or more of the disclosed embodiments, the transmission of the preference information (e.g., Figure 6 the third message) of the terminal device 100 may be paused for a pause period predetermined by the base station device 200 based on the transmission control information.

[0135] Therefore, according to one or more of the disclosed embodiments, a device and method capable of efficiently managing an AI / ML model or AI / ML function by controlling the transmission timing of the preference information of the terminal device 100 via a control message including transmission control information and preventing signaling overhead caused by excessive transmission of the preference information of the terminal device 100 can be provided.

[0136] In Figure 7 the operations of the terminal device 100 may also be performed by the base station device 200 or via interaction with the base station device 200, and the operations of the base station device 200 may also be performed by the terminal device 100 or via interaction with the terminal device 100.

[0137] Figure 8 is a flowchart showing a method for managing an AI / ML model or AI / ML function in a wireless communication system according to one or more embodiments.

[0138] Specifically, Figure 8 is a diagram showing a method for managing an AI / ML model or AI / ML function based on signaling (e.g., signaling of transmission control information) between the terminal device 100 and the base station device 200.

[0139] Referring toFigure 8 In a wireless communication system, a method for managing an AI / ML model or an AI / ML function based on signaling between a terminal device 100 and a base station device 200 may include operation S401 to operation S450. In Figure 8 the description of operation S401 to operation S450, a description substantially the same as Figures 5 to 7 the description of Figures 5 to 7 is replaced with the description of

[0140] In the disclosure, an AI / ML model may represent a hardware configuration and / or a software configuration for learning (or training) a specific pattern of a data set and inferring / predicting a target function of an AI / ML-based communication technology based on the learned pattern. An AI / ML function may represent a hardware configuration or and / or a software configuration (or a set of configurations) for performing a pre-agreed target feature (or a pre-agreed target function) or a pre-agreed target feature group (or a pre-agreed target function group) between the terminal device 100 and the base station device 200 in AI / ML-based communication.

[0141] The terminal device 100 may notify the base station device 200 of the preference information reporting ability of the terminal device 100 (i.e., the terminal device 100 may send / report opinions on at least one AI / ML model or at least one AI / ML function) in advance during a process of establishing a wireless communication connection with the base station device 200 (e.g., an initial setup process).

[0142] In operation S401, the base station device 200 may configure transmission control information of the terminal device 100, and in operation S403, the base station device 200 may send a control message including the transmission control information. For example, the transmission control information may include information on the transmission time of a third message, a period for pausing the transmission of the third message in the terminal device 100, and the transmission frequency of the third message. For example, the base station device 200 may send a control message obtained by inputting / declaring "Sa (i.e., the pause period is 'a seconds')" in the transmission control information (see Figure 7 and [Table 6]). Thereafter, the terminal device 100 may set at least one AI / ML model or at least one AI / ML function based on the signaling with the base station device 200.

[0143] In operation S410, the terminal device 100 may configure preference information (preference information #1) of the terminal device 100 for at least one AI / ML model or at least one AI / ML function, and in operation S420, the terminal device 100 may send a third message (message #1) including the preference information (preference information #1) of the terminal device 100. For example, the terminal device 100 may configure the preference information (preference information #1) of the terminal device 100 based on the monitoring results of at least one previously set AI / ML model or the monitoring results of at least one AI / ML function. The preference information (preference information #1) of the terminal device 100 may include information about the AI / ML model or AI / ML function preferred by the terminal device 100 from among at least one AI / ML model or at least one AI / ML function (e.g., information requesting reset to the AI / ML model or AI / ML function preferred by the terminal device 100) (see Figure 6 and [Table 5]).

[0144] In operation S421, although receiving the third message (message #1) including the preference information (preference information #1) of the terminal device 100, the base station device 200 may determine to maintain the control information for at least one AI / ML model or at least one AI / ML function.

[0145] In operation S423, the terminal device 100 may configure preference information (preference information #2) of the terminal device 100 for at least one AI / ML model or at least one AI / ML function. For example, the terminal device 100 may configure the preference information (preference information #2) of the terminal device 100 based on the monitoring results of at least one previously set AI / ML model or the monitoring results of at least one AI / ML function. The preference information (preference information #2) of the terminal device 100 may include information about the AI / ML model or AI / ML function preferred by the terminal device 100 from among at least one AI / ML model or at least one AI / ML function (e.g., information requesting reset to the AI / ML model or AI / ML function preferred by the terminal device 100) (see Figure 6 and [Table 5]). The preference information (preference information #2) of the terminal device 100 may include the same information as the preference information (preference information #1) of the terminal device 100, or may include information different from the preference information (preference information #1) of the terminal device 100.

[0146] In operation S425, the terminal device 100 may send an updated third message (Message #2) including the preference information (Preference Information #2) of the terminal device 100 after the period indicated in the control message (e.g., the suspension period of the preference information (#2)), that is, after the prohibition timer value expires. For example, according to the transmission control information, the terminal device 100 may suspend the preference information (Preference Information #2) until the suspension period of "a seconds" has elapsed since the time point when the preference information (Preference Information #1) was sent (operation S420), and may send the preference information (Preference Information #2) to the base station device 200 after the suspension period of "a seconds" has elapsed since the time point when the preference information (Preference Information #1) was sent (operation S420).

[0147] In operation S430, the base station device 200 may determine whether to reconfigure the control information regarding at least one AI / ML model or at least one AI / ML function based on the updated third message (Message #2) including the preference information (Preference Information #2) of the terminal device 100. The base station device 200 may perform operation S440 when reconfiguring the control information regarding at least one AI / ML model or at least one AI / ML function based on the updated third message (Message #2), and may terminate the operation when not reconfiguring the control information regarding at least one AI / ML model or at least one AI / ML function based on the updated third message (Message #2) (i.e., may maintain the control information regarding at least one existing AI / ML model or at least one existing AI / ML function).

[0148] In operation S440, the base station device 200 may send a fourth message including the reconfigured control information regarding at least one AI / ML model or at least one AI / ML function.

[0149] In operation S450, the terminal device 100 may reset at least one AI / ML model or at least one AI / ML function according to the fourth message. For example, the terminal device 100 may perform selection, activation, deactivation, switching, or fallback operations on at least one AI / ML model or at least one AI / ML function according to the fourth message.

[0150] Therefore, according to one or more disclosed embodiments, a device and method capable of performing adaptive management of an AI / ML model or an AI / ML function through signaling between the terminal device 100 and the base station device 200 can be provided. The signaling includes preference information of the terminal device 100 configured based on the monitoring results of the operation performance of at least one previously set AI / ML model or at least one AI / ML function (i.e., information requesting to reset the AI / ML model or AI / ML function preferred by the terminal device 100).

[0151] In addition, according to one or more disclosed embodiments, a device and a method capable of efficiently managing an AI / ML model or an AI / ML function can be provided by controlling the transmission timing of preference information of the terminal device 100 via a control message including transmission control information and preventing signaling overhead caused by excessive transmission of the preference information of the terminal device 100.

[0152] In Figure 8 this, the operation of the terminal device 100 can also be performed by the base station device 200 or via interaction with the base station device 200, and the operation of the base station device 200 can also be performed by the terminal device 100 or via interaction with the terminal device 100.

[0153] Figure 9A and Figure 9B are block diagrams showing an electronic device 1500 according to one or more embodiments. Figure 9A The electronic device 1500 of Figures 1 to 8 can correspond to the terminal device 100 of

[0154] Referring to Figure 9A and Figure 9B , the electronic device 1500 may include a memory 1010, a processor unit 1020, an input / output interface (e.g., input / output interface) 1040, a display (e.g., display device or display panel) 1050, an input device 1060, and a communication interface 1090. Here, a plurality of memories 1010 may be provided. Each element will be described below.

[0155] Referring to Figure 9A , the memory 1010 may include a program storage device 1011 that stores a program for controlling the operation of the electronic device 1500 and a data storage device 1012 that stores data generated during program execution. The data storage device 1012 may store data required for the operation of the application program 1013 and the AI / ML program 1014a. The program storage device 1011 may include the application program 1013 and the AI / ML program 1014a. Here, the programs included in the program storage device 1011 may be a set of instructions and may be represented as an instruction set.

[0156] The application program 1013 includes application programs operating in the electronic device 1500. That is, the application program 1013 may include instructions of applications driven by the processor 1022.

[0157] Figure 9A shows a case where the AI / ML program 1014a is included in the memory 1010 (or the program storage device 1011). However, the disclosure is not limited thereto, and the AI / ML program 1014b according to one or more embodiments may be included in various configurations of the electronic device 1500 according to the design. Referring toFigure 9B According to one or more embodiments, the AI / ML program 1014b may be included in the communication interface 1090.

[0158] According to an embodiment, in Figure 9A and Figure 9B , the AI / ML programs 1014a and 1014b may include at least one AI / ML model or at least one AI / ML function for AI / ML-based communication technologies that the electronic device 1500 can support. The at least one AI / ML model may represent a hardware configuration and / or software configuration for learning (or training) a specific pattern of a data set and inferring / predicting a target function of an AI / ML-based communication technology based on the learned pattern. The at least one AI / ML function may represent a hardware configuration and / or software configuration (or a set of configurations) for performing a pre-agreed target feature (or pre-agreed target function) or a pre-agreed group of target features (or pre-agreed group of target functions) between the terminal device 100 and the base station device 200 in AI / ML-based communication.

[0159] According to an embodiment, in Figure 9A and Figure 9B , the AI / ML programs 1014a and 1014b may configure / generate the AI / ML-related capability information of the electronic device 1500. The AI / ML programs 1014a and 1014b may control the communication interface 1090 to send the AI / ML-related capability information of the electronic device 1500 to an external electronic device (e.g., the base station device 200). The AI / ML-related capability information of the electronic device 1500 may include information about the maximum number of at least one AI / ML model that the electronic device 1500 can support or the maximum number of at least one AI / ML function that the electronic device 1500 can support.

[0160] According to an embodiment, in Figure 9A and Figure 9B , the AI / ML programs 1014a and 1014b may configure / generate the AI / ML-related preference information of the electronic device 1500. The AI / ML programs 1014a and 1014b may control the communication interface 1090 to send the AI / ML-related preference information of the electronic device 1500 to an external electronic device (e.g., the base station device 200). Here, the AI / ML-related preference information of the electronic device 1500 may include information about the AI / ML model or AI / ML function preferred by the electronic device 1500 from among at least one AI / ML model or at least one AI / ML function (e.g., information requesting to reset to the AI / ML model or AI / ML function preferred by the electronic device 1500).

[0161] According to an embodiment, in Figure 9A andFigure 9B Among them, the AI / ML programs 1014a and 1014b can control the communication interface 1090 to receive transmission control information from an external electronic device (e.g., Figure 1 the base station device 200). The AI / ML programs 1014a and 1014b can determine the timing for sending preference information regarding AI / ML according to the received transmission control information. The transmission control information may include information for controlling the sending time point of preference information regarding AI / ML of the electronic device 1500 (e.g., the sending time for sending preference information regarding AI / ML, the transmission pause period for preference information regarding AI / ML, and the sending frequency for sending preference information regarding AI / ML).

[0162] The peripheral device interface 1023 can control the connection between the input / output peripheral devices of the base station 200, the processor 1022, and the memory interface 1021.

[0163] The input / output interface 1040 can provide an interface between input / output devices (such as the display 1050 and the input device 1060) and the peripheral device interface 1023. The display 1050 displays status information, input characters, moving pictures, still pictures, etc. For example, the display 1050 can display information about application programs driven by the processor 1022.

[0164] The input device 1060 can provide input data generated by selecting the electronic device 1500 to the processor unit 1020 through the input / output interface 1040. Here, the input device 1060 may include a keyboard including at least one hardware button and a touchpad for sensing touch information. For example, the input device 1060 can provide touch information (such as touch, touch movement, and touch release) sensed by the touchpad to the processor 1022 through the input / output interface 1040. The electronic device 1500 may include a communication interface 1090 that performs communication functions for voice communication and data communication. According to an embodiment, the communication interface 1090 may include a modem, and the modem includes an antenna module 1092 that performs various functions for connecting to a wireless network to perform various wireless communications (e.g., cell search function, cell measurement function, handover function, etc.).

[0165] In Figure 9A and Figure 9B Among them, the operation of the electronic device 1500 can also be performed by an external electronic device or via interaction with an external electronic device, and the operation of the external electronic device can also be performed by the electronic device 1500 or via interaction with the electronic device 1500.

[0166] Figure 10It is a diagram showing an electronic device to which AI / ML-based communication technology is applied according to one or more embodiments.

[0167] In Figure 10 it is assumed that the home appliance 2100, the household appliance 2120, and the entertainment device 2140 are electronic devices capable of performing wireless communication connections according to AI / ML-based communication technology, and it is assumed that the access point 2200 can perform wireless communication connections with at least one of the electronic device (e.g., the terminal device 100), the home appliance 2100, the household appliance 2120, and the entertainment device 2140 according to AI / ML-based communication technology.

[0168] Referring to Figure 10 each of the electronic device (e.g., the terminal device 100), the home appliance 2100, the household appliance 2120, and the entertainment device 2140 may send ability information including the maximum number of at least one AI / ML model or the maximum number of at least one AI / ML function that each device can support to the access point 2200.

[0169] Each of the electronic device (e.g., the terminal device 100), the home appliance 2100, the household appliance 2120, and the entertainment device 2140 may send preference information regarding at least one AI / ML model or at least one AI / ML function to the access point 2200.

[0170] Each of the electronic device (e.g., the terminal device 100), the home appliance 2100, the household appliance 2120, and the entertainment device 2140 may send preference information to the access point 2200 based on the transmission control information configured by the access point 2200.

[0171] The access point 2200 may efficiently manage at least one AI / ML model or at least one AI / ML function stored in each device based on the signaling (e.g., ability information, preference information, and / or transmission control information) among the electronic device (e.g., the terminal device 100), the home appliance 2100, the household appliance 2120, and the entertainment device 2140.

[0172] In some embodiments, the home appliance 2100, the household appliance 2120, the entertainment device 2140, and the access point 2200 may constitute an Internet of Things (IoT) network system. It will be understood that Figure 10 the electronic device shown in Figure 10 is only an example, and the above embodiments may be applied to

[0173] In Figure 10In this case, the operations of the electronic device (e.g., the terminal device 100) can also be performed through the operations of the access point 2200, and the operations of the access point 2200 can be performed by the electronic device (e.g., the terminal device 100).

[0174] The foregoing exemplary embodiments are merely exemplary and should not be construed as restrictive. The present teachings can be readily applied to other types of devices. Further, the description of the exemplary embodiments is intended to be illustrative and not to limit the scope of the claims, and many alternatives, modifications, and variations will be apparent to those skilled in the art.

Claims

1. A terminal device, comprising: A memory configured to: store at least one artificial intelligence model or at least one artificial intelligence function for performing artificial intelligence-based wireless communication; Communication interface; as well as The processor is configured as: configuring capability information of the terminal device, the capability information including information about the number of at least one artificial intelligence model or the number of at least one artificial intelligence function that the terminal device can support; and The communication interface is controlled to send a first message including the capability information of the terminal device to the base station device.

2. The terminal device according to claim 1, wherein: The capability information includes: at least one of information about the maximum number of single-sided artificial intelligence models that the terminal device can support among at least one artificial intelligence model that the terminal device can support and information about the maximum number of bilateral artificial intelligence models that the terminal device can support among the at least one artificial intelligence model, or information about the maximum number of the at least one artificial intelligence model.

3. The terminal device according to claim 1, wherein: The capability information includes: at least one of information about the maximum number of single-sided artificial intelligence functions that the terminal device can support among at least one artificial intelligence function that the terminal device can support and information about the maximum number of bilateral artificial intelligence functions that the terminal device can support among the at least one artificial intelligence function, or information about the maximum number of the at least one artificial intelligence function.

4. The terminal device according to any one of claims 1 to 3, wherein: The processor is further configured to set the at least one artificial intelligence model or the at least one artificial intelligence function based on a second message received from the base station device in response to the first message.

5. The terminal device according to any one of claims 1 to 3, wherein: The processor is also configured to: monitoring the operational performance of the at least one artificial intelligence model or the operational performance of the at least one artificial intelligence function, configuring preference information about the at least one artificial intelligence model or the at least one artificial intelligence function based on the monitoring result, and The communication interface is controlled to send a third message including the preference information to the base station device.

6. The terminal device according to claim 5, wherein: The processor is further configured to receive a control message including transmission control information from the base station device through the communication interface.

7. The terminal device according to claim 6, wherein: To send the third message, the processor is further configured to: determining a transmission timing of the third message based on the transmission control information, and controlling the communication interface to send the third message to the base station device according to the determined sending timing, The transmission control information includes information about the sending time of the third message, the time period for suspending the sending of the third message in the terminal device, and the sending frequency of the third message.

8. The terminal device according to claim 5, wherein: The preference information includes: information for requesting, through the terminal device, to select, switch, activate, deactivate or roll back the at least one artificial intelligence model or the at least one artificial intelligence function.

9. A base station device, comprising: Memory; Communication interface; as well as The processor is configured as: controlling the communication interface to receive a first message including capability information of the terminal device from the terminal device, Based on the capability information of the terminal device in the first message, configure control information about at least one artificial intelligence model or at least one artificial intelligence function; and sending a second message including control information to the terminal device, Among them, the capability information of the terminal device includes information about the number of at least one artificial intelligence model or the number of at least one artificial intelligence function that the terminal device can support.

10. The base station device according to claim 9, wherein: The capability information includes: at least one of information about the maximum number of single-sided artificial intelligence models that the terminal device can support among at least one artificial intelligence model that the terminal device can support and information about the maximum number of double-sided artificial intelligence models that the terminal device can support among the at least one artificial intelligence model, or information about the maximum number of the at least one artificial intelligence model.

11. The base station device according to claim 9, wherein: The capability information includes: at least one of information about the maximum number of single-sided artificial intelligence functions that the terminal device can support among at least one artificial intelligence function that the terminal device can support and information about the maximum number of bilateral artificial intelligence functions that the terminal device can support among the at least one artificial intelligence function, or information about the maximum number of the at least one artificial intelligence function.

12. The base station device according to any one of claims 9 to 11, wherein: The processor is also configured to: Configure transmission control information, and A control message including the transmission control information is sent to the terminal device through the communication interface.

13. The base station device according to claim 12, wherein: The transmission control information includes information indicating a transmission time of the third message predetermined by the base station device, a period of time during which transmission of the third message is suspended in the terminal device, and a transmission frequency of the third message.

14. The base station device according to claim 12, wherein: The processor is also configured to: controlling the communication interface to receive a third message from the terminal device, the third message including preference information about the at least one artificial intelligence model or the at least one artificial intelligence function, determining whether to reconfigure control information about the at least one artificial intelligence model or the at least one artificial intelligence function based on the third message, and When the control information about the at least one artificial intelligence model or the at least one artificial intelligence function is reconfigured, a fourth message including the reconfigured control information is transmitted to the terminal device.

15. The base station device according to claim 14, wherein: The preference information includes: information for requesting, through a terminal device, to select, switch, activate, deactivate or roll back the at least one artificial intelligence model or the at least one artificial intelligence function.

16. The base station device according to claim 14, wherein: The preference information is included in the user equipment auxiliary information.

17. A communication method performed by a terminal device, the communication method comprising: sending capability information from the terminal device to the base station device, the capability information indicating that the terminal device is capable of reporting an artificial intelligence function preference of the terminal device to the base station device; receiving, from the base station device, a prohibit timer value based on a communication session established between the terminal device and the base station device for reporting the artificial intelligence function preference to the base station device, the prohibit timer value indicating that the terminal device is prohibited from sending the updated artificial intelligence function preference for a predetermined period of time; as well as After the inhibit timer value expires, the updated artificial intelligence function preferences are sent from the terminal device to the base station device.

18. The communication method according to claim 17, wherein: The capability information includes information about the number of identifiers assigned to artificial intelligence functions or artificial intelligence models supported by the terminal device, which is the number of artificial intelligence functions or artificial intelligence models supported by the terminal device.

Images