Base station, terminal and communication method

Abstract

A base station according to the present invention is provided with: a control circuit that determines whether or not to notify a terminal of settings relating to an artificial intelligence model on the terminal side for a wireless interface, in accordance with whether or not there is a request from the terminal; and a transmission circuit that transmits a signal relating to the setting when it is determined that the setting is to be notified.

Description

Technical Field

[0001] The present disclosure relates to a base station, a terminal, and a communication method. Background Art

[0002] In recent years, the rapid development of the Internet of Things (IoT) has been anticipated, driven by the expansion and diversification of wireless services. Mobile communications are expanding beyond information devices like smartphones to encompass all areas, including vehicles, homes, home appliances, and industrial equipment. To support this diverse range of services, significant improvements in the performance and functionality of mobile communication systems are required, in addition to increasing system capacity to meet various requirements, such as the increase in the number of connected devices and low latency. Fifth-generation mobile communication systems (5G) offer features such as large capacity and ultra-high speeds (enhanced Mobile Broadband (eMBB)), multi-device connectivity (massive Machine Type Communication (mMTC)), and ultra-reliable and low-latency communication (URLLC), enabling flexible wireless communication to meet diverse needs.

[0003] The 3rd Generation Partnership Project (3GPP), an international standards organization, is developing specifications for New Radio (NR), one of the 5G wireless interfaces.

[0004] Prior art literature

[0005] Non-patent literature

[0006] Non-Patent Literature 1: RP-213599, “New SI: Study on Artificial Intelligence (AI) / Machine Learning (ML) for NR Air Interface,” Qualcomm (Moderator), December 2021.

[0007] Non-Patent Literature 2: R1-2212966, “Summary #6 for CSI evaluation of [111-R18-AI / ML],” Moderator (Huawei), November 2022.

[0008] Non-Patent Document 3: 3GPP TSG RAN WG1 #110bis-e, “RAN1 Chair's Notes” Summary of the Invention

[0009] However, there is still room for research on methods to improve the efficiency of wireless communications.

[0010] The non-limiting embodiments of the present disclosure contribute to providing a base station, a terminal, and a communication method that can improve the efficiency of wireless communications.

[0011] A base station of one embodiment of the present disclosure comprises: a control circuit that determines whether to notify the terminal of settings related to an artificial intelligence model on the terminal side of a wireless interface based on whether there is a request from the terminal; and a sending circuit that sends a signal related to the settings when it is decided to notify the settings.

[0012] It should be noted that these general or specific aspects may be implemented by a system, device, method, integrated circuit, computer program or recording medium, or by any combination of systems, devices, methods, integrated circuits, computer programs and recording media.

[0013] According to one embodiment of the present disclosure, the efficiency of wireless communication can be improved.

[0014] Further advantages and effects of an embodiment of the present disclosure will be clarified through the description and drawings. These advantages and / or effects are provided by several embodiments and the features described in the description and drawings, but not all of them need to be provided in order to obtain one or more of the same features. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This figure shows an example of channel state information (CSI) compression using machine learning (ML) / artificial intelligence (AI) technology.

[0016] Figure 2 This is a diagram showing an example of training method 1 for CSI compression.

[0017] Figure 3This is a diagram showing an example of training method 2 for CSI compression.

[0018] Figure 4 This is a diagram showing an example of training method 3 for CSI compression.

[0019] Figure 5 This is a block diagram showing a configuration example of a portion of a base station.

[0020] Figure 6 This is a block diagram showing a configuration example of a part of a terminal.

[0021] Figure 7 This is a diagram showing an example of mapping between Cell ID and AI / ML model ID.

[0022] Figure 8 A diagram showing an example of operations of a terminal and a base station.

[0023] Figure 9 This is a diagram showing an example of mapping between Cell ID, AI / ML model ID, and AI / ML model version ID.

[0024] Figure 10 A diagram showing an example of operations of a terminal and a base station.

[0025] Figure 11 A diagram showing an example of operations of a terminal and a base station.

[0026] Figure 12 A diagram showing an operation example of a terminal and a base station.

[0027] Figure 13 This is a diagram showing an example of architecture.

[0028] Figure 14 This is a block diagram showing a configuration example of a base station.

[0029] Figure 15 This is a block diagram showing a structural example of a terminal.

[0030] Figure 16 This is a diagram of an exemplary architecture of a 3GPP NR (3rd generation partnership project new radio) system.

[0031] Figure 17 This diagram shows the functional separation between NG-RAN (Next Generation-Radio Access Network) and 5GC (5th Generation Core).

[0032] Figure 18 This is a sequence diagram of the procedure for setting up / resetting a radio resource control (RRC) connection.

[0033] Figure 19 This is a schematic diagram showing the utilization scenarios of large-capacity high-speed communications (eMBB: enhanced Mobile Broadband), multi-simultaneous machine-type communications (mMTC: massive Machine Type Communications), and highly reliable and low-latency communications (URLLC: Ultra Reliable and Low Latency Communications).

[0034] Figure 20 is a block diagram representing an exemplary 5G system architecture for a non-roaming scenario. DETAILED DESCRIPTION

[0035] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings.

[0036] In NR, for example, an access method based on orthogonal frequency division multiplexing (OFDM) is adopted in downlink transmission. In addition, in order to increase the data rate, multiple-input multiple output (MIMO) is adopted. In order to effectively utilize the performance of MIMO, closed-loop control is introduced to feedback channel state information (CSI) from the terminal (for example, also called user equipment (UE)) to the base station (for example, also called gNB). In the feedback of CSI, high performance is required to be achieved with a small amount of control information.

[0037] In 3GPP Release 18 (e.g., also known as Rel. 18), the application of artificial intelligence (AI) technologies (hereinafter also referred to as "AI / ML"), such as machine learning (ML), to the NR radio interface is being discussed (e.g., see Non-Patent Document 1). Research is underway on use cases utilizing AI / ML technologies, such as CSI feedback, beam steering, and position estimation.

[0038] For example, for CSI feedback, methods are being studied to compress the amount of CSI information in the spatial and frequency domains using AI / ML technology (hereinafter also referred to as "CSI compression"). For example, in machine learning, by training AI / ML models such as neural networks (e.g., artificial intelligence models) using a large amount of training data, its features can be automatically extracted. In CSI compression, for example, an algorithm called an autoencoder can be used, which is mainly used for dimensional compression and reconstruction of image data. For example, Figure 1 As shown, the CSI matrix represented by the two-dimensional area of ​​the spatial domain and the frequency domain is regarded as an image, the encoding part of the autoencoder for compression is applied to the CSI compression processing on the terminal side, and the decoding part of the autoencoder for reconstruction is applied to the CSI reconstruction processing on the base station or network (for example, also expressed as base station / network) side.

[0039] In CSI compression, the following three methods have been studied regarding the training methods of the encoding unit (or CSI generation unit) on the terminal side and the decoding unit (or CSI reconstruction unit) on the base station / network side (for example, non-patent document 2).

[0040] The first training method for CSI compression (hereinafter referred to as "training method 1") is as follows Figure 2 As shown in FIG, it is a joint training of a two-sided model in a single side / entity on the terminal side or the network side. In training method 1, for example, Figure 2 As shown, on the network side, the AI / ML models (also referred to as "models") of both the CSI generation unit and the CSI reconstruction unit are jointly trained. The trained CSI generation unit model on the network side is delivered / transferred to the terminal.

[0041] The second CSI compression training method (hereinafter referred to as "training method 2") is as follows: Figure 3 As shown in FIG, the joint training of the two-side models is performed on the network side and the terminal side. In training method 2, for example, Figure 3 As shown, by exchanging information such as forward propagation (FP) and backpropagation (BP) (for example, gradient) between the terminal and the base station, the terminal side and the network side jointly train the respective models of the CSI generation unit and the CSI reconstruction unit.

[0042] The third CSI compression training method (hereinafter referred to as "training method 3") is as follows: Figure 4As shown in FIG, the method is as follows: the terminal side and the network side respectively train the model of the CSI generation unit on the terminal side and the model of the CSI reconstruction unit on the network side. In training method 3, for example, Figure 4 As shown, on the network side, the models of the CSI generation unit and the CSI reconstruction unit are trained jointly. After the training on the network side is completed, the network side sends the training data set (for example, the input data used in the training (for example, V in ) and output data (e.g., V q The terminal uses the data set shared from the network side to train the model of the CSI generation unit on the terminal side.

[0043] An example of the training method in CSI compression has been described above.

[0044] Here, the characteristics of the wireless interface or wireless propagation path may vary depending on the scenario, configuration or site. When applying AI / ML to wireless interfaces, a unified-model method can be cited, in which the AI / ML model is trained by assuming all environments. However, when a unified model is applied, it is possible that a large amount of time and data will be spent on training, or the scale of the model will increase, or the amount of computational processing or latency consumed by the inference processing of the model will increase. Therefore, in contrast to the unified model, the use of models dedicated to specific environments, such as models dedicated to scenarios / configurations or sites, has been studied (for example, refer to non-patent document 3).

[0045] As one of the cases where a model is deployed that is specific to a scenario / configuration or site, a cell-specific model is sometimes deployed on a per-cell basis. For example, in the use case of configuring an AI / ML model on the terminal side such as the above-mentioned CSI compression, the following situation is envisioned: if the terminal moves from one cell to another, the AI / ML model on the terminal side is updated or switched to match the model deployed in the new cell. In addition, in addition to movement between cells, the AI / ML model on the terminal side may also be updated or switched to match the new environment due to movement between transmission and reception points (TRP), switching of beams, or movement of the terminal location within the cell.

[0046] For example, in the aforementioned training method 1, the model trained on the network side may be distributed / transmitted to the terminal side as the cell moves. Furthermore, for example, in training method 2, at least when online training (online learning or online training) is required, information such as FP, BP, and gradients may be exchanged between the network and the terminal side as the cell moves. Furthermore, for example, in training method 3, at least when online training is required, information such as datasets may be shared between the network and the terminal side as the cell moves.

[0047] When implementing scenario / configuration or site-specific AI / ML model deployment, for example, it is desirable to account for the overhead and latency required to update the AI / ML model on the terminal side. For example, without properly defining control methods for updating or switching the AI / ML model on the terminal side in response to terminal mobility, it is possible that the AI / ML model or information used for training and updating the AI / ML model must be distributed / transmitted and shared every time mobility occurs, making it impossible to efficiently apply AI / ML technology to wireless interfaces and achieving improvements in communication efficiency or quality.

[0048] In one non-limiting embodiment of the present disclosure, a method for appropriately and efficiently updating an AI / ML model on a terminal side in a system where a scenario / configuration or site-specific AI / ML model is deployed will be described.

[0049] According to a non-limiting embodiment of the present disclosure, for example, in a system where a scenario / configuration or site-specific AI / ML model is deployed, AI / ML technology can be efficiently applied to a wireless interface, thereby improving the communication efficiency and quality of the wireless system.

[0050] Hereinafter, non-limiting embodiments of the present disclosure will be described.

[0051] (Implementation Method 1)

[0052] [Overview of the Communication System]

[0053] The communication system of this embodiment includes, for example, at least one base station and at least one terminal.

[0054] Figure 5 is a block diagram showing a configuration example of a portion of a base station 100 according to an embodiment of the present disclosure. Figure 6 2 is a block diagram showing a configuration example of a portion of the terminal 200 according to an embodiment of the present disclosure.

[0055] exist Figure 5In the illustrated base station 100, a control unit (e.g., corresponding to a control circuit) determines whether to notify (e.g., distribute, transmit, or share) terminal-side AI / ML model settings for the wireless interface to terminal 200, based on whether or not a request has been received from terminal 200. If notification of terminal-side AI / ML model settings is determined, a transmission unit (e.g., a transmission circuit) transmits a signal related to the settings.

[0056] exist Figure 6 In the illustrated terminal 200, a control unit (e.g., corresponding to a control circuit) determines whether to request notification of terminal-side AI / ML model settings from the network based on stored information related to the AI / ML model for the wireless interface (e.g., mapping information described later). If notification of terminal-side AI / ML model settings is requested, a receiving unit (e.g., corresponding to a receiving circuit) receives a signal related to the settings.

[0057] In this embodiment, as an example, a use case is assumed in which an AI / ML model is configured (deployed) on the terminal side, such as the above-mentioned CSI compression.

[0058] In this case, the terminal 200 may store multiple AI / ML models. In the Mobility state, if the terminal 200 stores an AI / ML model for the cell deployed at the terminal 200's destination (e.g., the target cell), the TRP of the destination, the beam of the handover destination, and the location of the destination, the network (e.g., the base station 100) may not need to distribute / transmit or share a new AI / ML model (or information used for training and updating the new AI / ML model) with the terminal 200 for model updates or handovers.

[0059] If the terminal 200 can update and switch the AI / ML model on the terminal side without distributing / transmitting and sharing the new AI / ML model (or information for training and updating the new AI / ML model) used for updating or switching the AI / ML model to the terminal 200, the overhead during mobility can be reduced. In addition, there is room for reducing the delay in updating and switching the model.

[0060] In this embodiment, regarding the AI / ML model on the terminal side, whether to distribute / transmit or share a new AI / ML model (or information for training and updating the AI / ML model) for model updating or switching to the terminal 200 is determined based on a request made by the terminal 200.

[0061] This can prevent unnecessary distribution, transmission, or sharing of a new AI / ML model (or information used for training and updating a new AI / ML model) to the terminal 200 when the model is updated or switched.

[0062] In this embodiment, in order for the terminal 200 to appropriately determine whether to request the terminal 200 to distribute / transmit or share a new AI / ML model (or information for training and updating the new AI / ML model) for updating or switching the model in the terminal 200, the terminal 200 may be notified of information related to the mapping between the following two (hereinafter also referred to as "mapping information"), which are information related to the wireless environment such as cells, sites, TRPs, beams and locations (for example, IDs), and information on the AI / ML models deployed corresponding to the cells, sites, TRPs, beams and locations (for example, model IDs or functionality IDs).

[0063] As a method of notifying the mapping information, for example, it may be broadcasted using cell-specific higher layer signaling (eg, Radio Resource Control (RRC) signal), or may be notified using an RRC signal dedicated to the terminal 200 .

[0064] Figure 7 This shows an example of mapping information related to the mapping between the cell ID and the AI / ML model ID notified to the terminal 200. It can be, for example Figure 7 As shown, terminal 200 stores mapping information related to a plurality of AI / ML models (model IDs: #A, #B, #C, ...) corresponding to a plurality of cells (cell IDs: 0, 1, 2, ...).

[0065] in addition, Figure 7Examples of mappings between cells and AI / ML models are provided, but are not limited thereto. For example, the mapping between TRP and the AI / ML model may be a mapping between a transmission configuration indication (TCI) and information about the AI / ML model (e.g., model ID, function ID). In addition, for example, the mapping between the beam and the AI / ML model may be a mapping between a synchronization signal block (SSB: Synchronization Signal Block) index or a channel state information reference signal (CSI-RS: Channel State Information-Reference Signal) index and information about the AI / ML model (e.g., model ID, function ID). In addition, for example, the mapping between location and the AI / ML model may be a mapping between an area ID and information about the AI / ML model (e.g., model ID, function ID).

[0066] For example, after acquiring the above-mentioned mapping information during mobility, when an AI / ML model (e.g., also referred to as a "target AI / ML model") of a cell deployed at a mobile destination (e.g., a target cell), a TRP of a mobile destination, a beam of a switching destination, or a location of a mobile destination is not stored, the terminal 200 requests the network (e.g., the base station 100) to distribute / transmit or share the target AI / ML model (or information for training or updating the target AI / ML model) to the terminal 200 for updating or switching the AI / ML model on the terminal side.

[0067] During the random access process during mobility, a request from terminal 200 to the network or a message indicating whether to distribute / transmit or share the target AI / ML model (or information used for training or updating the target AI / ML model) to terminal 200 can be sent from terminal 200.

[0068] For example, in the above-mentioned use case of CSI compression, when the terminal 200 requests the network to distribute / transmit or share the target AI / ML model (or information for training and updating the target AI / ML model) to the terminal 200 for the purpose of updating or switching the AI / ML model on the terminal side, in the case of training method 1, the AI / ML model itself is distributed / transmitted or shared; in the case of training method 2, information related to FP, BP or gradient is distributed / transmitted or shared; in the case of training method 3, the data set is distributed / transmitted or shared. In addition, the information distributed / transmitted or shared can be distributed / transmitted or shared via, for example, an RRC reset message or a user plane message. In addition, the distribution / transmission or sharing of this information can be performed after the connection with the cell of the mobile destination (target cell) or the TRP of the mobile destination, after switching to the beam of the switching destination, or after mobility to the location of the mobile destination.

[0069] Terminal 200 may, for example, enable the target AI / ML model delivered / transmitted to the destination cell (target cell), the destination TRP, the handover destination beam, or the destination location to be used. Alternatively, terminal 200 may train the target AI / ML model based on information shared with terminal 200 for training and updating the target AI / ML model, and enable the target AI / ML model to be used.

[0070] On the other hand, for example, if the terminal 200 stores an AI / ML model of a cell deployed at the mobile destination (target cell), a TRP at the mobile destination, a beam at the handover destination, or a location at the mobile destination (target AI / ML model), the terminal 200 may not send the above request message to the network. Alternatively, in this case, the terminal 200 may send a request message containing the following information to the network, notifying that the target AI / ML model (or information used for training and updating the target AI / ML model) does not need to be distributed / transmitted or shared.

[0071] The terminal 200 can be set to the cell of the mobile destination (target cell), the TRP of the mobile destination, the beam of the switching destination, or the location of the mobile destination, and the saved target AI / ML model can be used.

[0072] Figure 8 FIG. 2 shows an example of operations of the terminal 200 and the network (for example, the base station 100) according to this embodiment. Figure 8 As an example, an action example related to the AI / ML model deployed for each cell is described, but not limited to this, the AI / ML model can also correspond to TRP, location, beam or position.

[0073] exist Figure 8In the process, the network notifies the terminal 200 of mapping information related to the mapping between the cell and the AI / ML model (e.g., Figure 7 )(S101). The terminal 200 acquires (or receives) mapping information notified from the network (S102).

[0074] The terminal 200 determines whether an AI / ML model (target AI / ML model) corresponding to the cell of the terminal 200's movement destination (target cell) is stored based on the mapping information, for example ( S103 ).

[0075] If the target AI / ML model is not stored ( S103 : No), the terminal 200 requests the network to distribute / transmit or share the target AI / ML model (or information for training or updating the target AI / ML model) ( S104 ).

[0076] When there is a request from terminal 200 for distribution / transmission or sharing of a target AI / ML model (or information used for training or updating the target AI / ML model), the network distributes / transmits or shares the target AI / ML model (or information used for training or updating the target AI / ML model) to terminal 200 (S105).

[0077] The terminal 200 obtains the target AI / ML model (or information for training or updating the target AI / ML model) notified from the network (S106). For example, the terminal 200 may also update the stored mapping information based on the target AI / ML model notified from the network.

[0078] When the target AI / ML model is stored ( S103 : Yes) or when the target AI / ML model is acquired from the network ( S106 ), the terminal 200 makes the target AI / ML model available for use ( S107 ).

[0079] As described above, the terminal 200 determines whether to request notification (e.g., distribution / transmission or sharing) of settings related to the terminal-side AI / ML model (e.g., target AI / ML model) from the network based on the stored mapping information related to the AI / ML model for the wireless interface. Furthermore, the network (e.g., base station 100) determines whether to notify (e.g., distribution / transmission or sharing) settings related to the terminal-side AI / ML model for the wireless interface (e.g., target AI / ML model) to the terminal 200 based on whether there is a request from the terminal 200. Then, when the base station 100 has decided to distribute / transmit or share the settings to the terminal 200, it transmits a signal related to the settings of the target AI / ML model, and the terminal 200 receives the signal related to the settings of the target AI / ML model.

[0080] As a result, the network does not distribute / transmit or share information related to the target AI / ML model, for example, when there is no request from the terminal 200 (or when there is a message indicating that it is not necessary). In this way, it is possible to suppress the unnecessary distribution / transmission or sharing of a new AI / ML model (or information used for training and updating the new AI / ML model) to the terminal 200 when the model is updated or switched for the terminal 200 after mobility. Therefore, it is possible to reduce the overhead and latency for updating or switching the AI / ML model on the terminal side. Therefore, according to this embodiment, the efficiency of wireless communication can be improved.

[0081] (Implementation Method 2)

[0082] The configurations of base station 100 and terminal 200 in this embodiment may be the same as those in the first embodiment.

[0083] In this embodiment, as an example, a use case is assumed in which an AI / ML model is configured (deployed) on the terminal side, such as the above-mentioned CSI compression.

[0084] In this case, the terminal 200 may store multiple AI / ML models. In the Mobility state, if the terminal 200 stores an AI / ML model for the cell deployed at the terminal 200's mobile destination (e.g., target cell), the TRP of the mobile destination, the beam of the handover destination, and the location of the mobile destination, the network (e.g., the base station 100) may not need to distribute / transmit or share a new AI / ML model (or information used for training and updating the new AI / ML model) with the terminal 200 for model updates or handovers.

[0085] If the terminal 200 can update and switch the AI / ML model on the terminal side without distributing / transmitting and sharing the new AI / ML model (or information for training and updating the new AI / ML model) used for updating or switching the AI / ML model to the terminal 200, the overhead during mobility can be reduced. In addition, there is room for reducing the delay in updating and switching the model.

[0086] In this embodiment, similar to embodiment 1, regarding the AI / ML model on the terminal side, a decision is made based on a request made by terminal 200 whether to distribute / transmit or share a new AI / ML model (or information for training and updating the new AI / ML model) to terminal 200 for model updating or switching.

[0087] This can prevent unnecessary distribution, transmission, or sharing of a new AI / ML model (or information used for training and updating a new AI / ML model) to the terminal 200 when the model is updated or switched.

[0088] In this embodiment, in order for the terminal 200 to appropriately determine whether to request the terminal 200 to distribute / transmit or share a new AI / ML model (or information for training and updating the new AI / ML model) for updating or switching the model in the terminal 200, the terminal 200 may be notified of mapping information related to the mapping between the following three: information related to wireless environments such as cells, sites, TRPs, beams, and locations (for example, IDs), information on AI / ML models deployed corresponding to cells, sites, TRPs, beams, and locations (for example, model IDs or function IDs), and information related to the versions of each AI / ML model (for example, version IDs or timestamps of updates to the current model).

[0089] As a method of notifying the mapping information, for example, it may be broadcasted using cell-specific higher layer signaling (eg, RRC signal), or it may be notified using an RRC signal dedicated to the terminal 200 .

[0090] Figure 9 This represents an example of mapping information related to the mapping between the cell ID, AI / ML model ID, and version ID notified to the terminal 200. It can be, for example Figure 9 As shown, terminal 200 stores mapping information related to multiple AI / ML models (model IDs: #A, #B, #C, ...) corresponding to multiple cells (cell IDs: 0, 1, 2, ...) and the version (version ID) of each AI / ML model.

[0091] in addition, Figure 9Examples of mappings between cells, AI / ML models, and versions of AI / ML models are shown, but are not limited thereto. For example, the mapping between TRP and AI / ML model can be a mapping between TCI, information about the AI / ML model (e.g., model ID, function ID), and information about the version of the AI / ML model. In addition, for example, the mapping between beam and AI / ML model can be a mapping between SSB index or CSI-RS index, information about the AI / ML model (e.g., model ID, function ID), and information about the version of the AI / ML model. In addition, for example, the mapping between location and AI / ML model can be a mapping between area ID, information about the AI / ML model (e.g., model ID, function ID), and information about the version of the AI / ML model.

[0092] For example, after obtaining the above-mentioned mapping information during mobility, when the current version (target AI / ML model) of the AI / ML model of the cell deployed at the mobile destination (e.g., the target cell), the TRP of the mobile destination, the beam of the switching destination, or the location of the mobile destination is not saved, the terminal 200 requests the network (e.g., the base station 100) to distribute / transmit or share the target AI / ML model (or information for training or updating the target AI / ML model) to the terminal 200 for the purpose of updating or switching the AI / ML model on the terminal side.

[0093] During the random access process during mobility, a request from terminal 200 to the network or a message indicating whether to distribute / transmit or share the target AI / ML model (or information used for training or updating the target AI / ML model) to terminal 200 can be sent from terminal 200.

[0094] For example, in the above-mentioned use case of CSI compression, when the terminal 200 requests the network to distribute / transmit or share the target AI / ML model (or information for training and updating the target AI / ML model) to the terminal 200 for the purpose of updating or switching the AI / ML model on the terminal side, in the case of training method 1, the AI / ML model itself is distributed / transmitted or shared; in the case of training method 2, information related to FP, BP or gradient is distributed / transmitted or shared; in the case of training method 3, the data set is distributed / transmitted or shared. In addition, the information distributed / transmitted or shared can be distributed / transmitted or shared via, for example, an RRC reset message or a user plane message. In addition, the distribution / transmission or sharing of this information can be performed after the connection with the cell of the mobile destination (target cell) or the TRP of the mobile destination, after switching to the beam of the switching destination, or after mobility to the location of the mobile destination.

[0095] Terminal 200 may, for example, enable the target AI / ML model delivered / transmitted to the destination cell (target cell), the destination TRP, the handover destination beam, or the destination location to be used. Alternatively, terminal 200 may train the target AI / ML model based on information shared with terminal 200 for training and updating the target AI / ML model, and enable the target AI / ML model to be used.

[0096] On the other hand, for example, if the terminal 200 stores the current version (target AI / ML model) of the AI / ML model deployed at the mobile destination cell (target cell), the mobile destination TRP, the handover destination beam, or the mobile destination location, the terminal 200 may not send the above request message to the network. Alternatively, in this case, the terminal 200 may send a request message containing the following information to the network, which is information notifying that the target AI / ML model (or information used for training and updating the target AI / ML model) does not need to be distributed / transmitted or shared.

[0097] The terminal 200 can be set to the cell of the mobile destination (target cell), the TRP of the mobile destination, the beam of the switching destination, or the location of the mobile destination, and the saved target AI / ML model can be used.

[0098] In addition, the operation of the terminal 200 and the network (for example, the base station 100) of this embodiment can be the same as Figure 8 This embodiment uses mapping information (e.g., Figure 9 ) This point is different from implementation mode 1.

[0099] As described above, the terminal 200 determines whether to request notification (e.g., distribution / transmission or sharing) of settings related to the terminal-side AI / ML model (e.g., target AI / ML model) from the network based on the stored mapping information related to the AI / ML model for the wireless interface. Furthermore, the network (e.g., base station 100) determines whether to notify (e.g., distribution / transmission or sharing) settings related to the terminal-side AI / ML model for the wireless interface (e.g., target AI / ML model) to the terminal 200 based on whether or not there is a request from the terminal 200. Then, if the base station 100 decides to distribute / transmit or share the settings to the terminal 200, it transmits a signal related to the settings of the target AI / ML model, and the terminal 200 receives the signal related to the settings of the target AI / ML model.

[0100] As a result, the network does not distribute / transmit or share information related to the target AI / ML model, for example, when there is no request from the terminal 200 (or when there is a message indicating that it is not necessary). In this way, it is possible to suppress the unnecessary distribution / transmission or sharing of a new AI / ML model (or information used for training and updating the new AI / ML model) to the terminal 200 when the model is updated or switched for the terminal 200 after mobility. Therefore, it is possible to reduce the overhead and latency for updating or switching the AI / ML model on the terminal side. Therefore, according to this embodiment, the efficiency of wireless communication can be improved.

[0101] Furthermore, according to this embodiment, AI / ML models corresponding to a cell, site, TRP, beam, or location can be managed using not only the AI / ML model ID but also the version ID of each AI / ML model. This can help prevent the proliferation of model IDs assigned to AI / ML models.

[0102] (Variation 1 of Embodiment 1 and Embodiment 2)

[0103] The timing when the terminal 200 requests the network side to distribute / transmit or share the target AI / ML model (or information used for training and updating the target AI / ML model) is not limited to mobility.

[0104] For example, before the mobility, the terminal 200 may request the network to distribute / transmit or share the target AI / ML model (or information used for training and updating the target AI / ML model). In this case, the terminal 200 may request the network to distribute / transmit or share the AI / ML model (or information used for training and updating the AI / ML model) corresponding to the cell of the mobile destination (target cell), the TRP of the mobile destination, the beam of the handover destination, or the location of the mobile destination estimated before the mobility.

[0105] Furthermore, for example, after Mobility, the terminal 200 may request the network to distribute / transmit or share an AI / ML model (or information used for training and updating the AI / ML model) corresponding to the destination cell (target cell), destination TRP, destination beam, and destination location to the terminal 200. For example, even if the terminal 200 does not store an AI / ML model corresponding to the destination cell (target cell), destination TRP, destination beam, and destination location after Mobility, the request related to the AI / ML model may be made as needed.

[0106] The request from the terminal 200 to the network or the message indicating whether the AI / ML model (or information used for training and updating the AI / ML model) needs to be distributed / transmitted or shared to the terminal 200 can be sent using a control plane message such as uplink control information (UCI: Uplink Control Information) or a medium access control control element (MAC-CE: Medium Access Control-Control Element), or can be sent using a user plane message.

[0107] For example, in the above-mentioned CSI compression use case, when the terminal 200 requests the network to distribute / transmit or share the target AI / ML model (or information used for training and updating the target AI / ML model) to the terminal 200 for updating or switching the AI / ML model on the terminal side, the AI / ML model itself is distributed / transmitted or shared in the case of training method 1, information related to FP, BP or gradient is distributed / transmitted or shared in the case of training method 2, and the dataset is distributed / transmitted or shared in the case of training method 3. In addition, the distributed / transmitted or shared information can be distributed / transmitted or shared via user plane messages or via RRC signals.

[0108] In addition, when distribution / transmission or sharing is performed via user-plane messages, the distribution / transmission or sharing of such information may be performed before, during, or after mobility.

[0109] For example, when delivering / transmitting or sharing during mobility, part of the information may be delivered / transmitted or shared from the source cell (source cell), the source TRP, the beam before switching, or the position before moving, and the remaining part of the information may be delivered / transmitted or shared after connecting to the destination cell (target cell) or the destination TRP, after switching to the beam of the switching destination, or from the location of the destination. This enables seamless delivery / transmission or sharing of AI / ML models (or information used for training and updating AI / ML models).

[0110] (Variation 2 of Embodiment 1 and Embodiment 2)

[0111] In embodiments 1 and 2, the function for processing AI / ML (e.g., access and mobility management function (AMF) or a network node dedicated to AI / ML) may not be on the base station 100 (e.g., gNB).

[0112] As an example of a situation where the AI / ML processing function (e.g., AI / ML function) is not located on base station 100 (e.g., gNB), it is possible to configure the AI / ML processing function within the core network or outside the 3GPP network via a server. In this case, the AI / ML processing function may not recognize information (e.g., ID) corresponding to the cell, site, TRP, beam, or location of the serving cell. In such a hypothetical network architecture, the following options may also be applied.

[0113] <Option 1>

[0114] Figure 10 This section shows an example of the operation of the terminal 200, base station 100 (e.g., gNB), and the function for processing AI / ML (AI / ML function) in Option 1. Figure 10 In the embodiment, the same operations as those in the first and second embodiments are denoted by the same reference numerals.

[0115] In Option 1, terminal 200 requests base station 100 (gNB) to distribute / transmit or share (or not distribute / transmit or share) a new AI / ML model (or information used for training and updating the AI / ML model) for model update or switching to terminal 200 (S104). A signal containing this request (e.g., request information) may be sent in a control plane message (e.g., UCI or MAC-CE). Furthermore, the signal containing this request may include, for example, information related to the AI / ML model requested by terminal 200 (e.g., model ID, function ID).

[0116] The base station 100 (e.g., gNB) transmits the request information from the terminal 200 to the function for processing AI / ML (S301).

[0117] When a request is received from terminal 200 to distribute / transmit or share information related to a target AI / ML model, the function for processing AI / ML distributes / transmits or shares the target AI / ML model (or information used for training and updating the target AI / ML model) with terminal 200 (S302). Terminal 200 obtains the target AI / ML model (or information used for training and updating the target AI / ML model) from the function for processing AI / ML (S303).

[0118] In this way, by sending the request information from the terminal 200 to the function of processing AI / ML via the base station 100, the base station 100 (gNB) serving as the service cell for the terminal 200 can recognize the new AI / ML model that can be used in the terminal 200.

[0119] <Option 2>

[0120] Figure 11 This section shows an example of the operation of the terminal 200, base station 100 (e.g., gNB), and the function for processing AI / ML (AI / ML function) in Option 2. Figure 11 In the embodiment, the same operations as those in the first and second embodiments are denoted by the same reference numerals.

[0121] In Option 2, the terminal 200 requests the AI / ML processing function to distribute / transmit or share (or not distribute / transmit or share) a new AI / ML model (or information used for training and updating the AI / ML model) for model update or switching to the terminal 200 (S401). The signal containing this request (e.g., request information) can be included in a user plane message and sent.

[0122] When a request is received from terminal 200 to distribute / transmit or share information related to a target AI / ML model, the function for processing AI / ML distributes / transmits or shares the target AI / ML model (or information used for training and updating the target AI / ML model) with terminal 200 (S402). Terminal 200 obtains the target AI / ML model (or information used for training and updating the target AI / ML model) from the function for processing AI / ML (S403).

[0123] Terminal 200 notifies the network (e.g., base station 100 (gNB)) of information related to the available new AI / ML model (S404). The signal containing this notification can be sent via a control plane message (e.g., UCI or MAC-CE) or a user plane message.

[0124] In this way, by notifying the base station 100 of information related to the new AI / ML model that can be used, the base station 100 (gNB) serving as the service cell for the terminal 200 can recognize the new AI / ML model that can be used in the terminal 200 even when the request information is sent from the terminal 200 to the function for processing AI / ML without passing through the base station 100.

[0125] In addition, Option 2 describes a case in which the terminal 200 notifies the network (e.g., the base station 100 (gNB)) of information related to a new AI / ML model that can be used by the terminal 200, but is not limited to this. For example, notification can also be made to the network (e.g., the base station 100 (gNB)) from the function of processing AI / ML.

[0126] (Implementation 3)

[0127] The configurations of base station 100 and terminal 200 in this embodiment may be the same as those in Embodiment 1 or 2.

[0128] For example, in base station 100, a control unit (e.g., corresponding to a control circuit) determines whether to notify (e.g., distribute, transmit, or share) terminal-side AI / ML model-related settings to terminal 200 based on information related to the terminal's stored AI / ML model. If a decision is made to notify terminal-side AI / ML model-related settings, a transmission unit (e.g., a transmission circuit) transmits a signal related to the settings.

[0129] In addition, for example, in terminal 200, the transmitting unit (e.g., corresponding to the transmitting circuit) transmits information related to the stored AI / ML model, and the receiving unit (e.g., the receiving circuit) receives settings related to the AI / ML model notified based on the information.

[0130] In this embodiment, as an example, a use case is assumed in which an AI / ML model is configured (deployed) on the terminal side, such as the above-mentioned CSI compression.

[0131] In this case, the terminal 200 may store multiple AI / ML models. In the Mobility state, if the terminal 200 stores an AI / ML model for the cell deployed at the terminal 200's mobile destination (e.g., target cell), the TRP of the mobile destination, the beam of the handover destination, and the location of the mobile destination, the network (e.g., the base station 100) may not need to distribute / transmit or share a new AI / ML model (or information used for training and updating the new AI / ML model) with the terminal 200 for model updates or handovers.

[0132] If the terminal 200 can update and switch the AI / ML model on the terminal side without distributing / transmitting and sharing the new AI / ML model (or information used for training and updating the AI / ML model) for updating or switching the AI / ML model to the terminal 200, the overhead during mobility can be reduced. In addition, there is room for reducing the delay in updating and switching the model.

[0133] In this embodiment, regarding the AI / ML model on the terminal side, it is determined according to the instructions of the network whether to distribute / transmit or share a new AI / ML model (or information for training and updating the new AI / ML model) to the terminal 200 for model updating or switching.

[0134] This can prevent unnecessary distribution, transmission, or sharing of a new AI / ML model (or information used for training and updating a new AI / ML model) to the terminal 200 when the model is updated or switched.

[0135] In this embodiment, the terminal 200 reports the AI / ML model stored in the terminal 200 to the network (eg, the base station 100 ), for example, through a terminal capability report.

[0136] In addition, the report on the AI / ML model stored in the terminal 200 can be included in the existing terminal capability report (UE capability signaling) or in a newly introduced function report such as the AI / ML function report. In addition, when the capabilities of the terminal 200 are the same among multiple terminals (for example, when all terminals 200 can store the same model), the reporting step can be omitted.

[0137] Regarding the AI / ML model on the terminal side, the network (e.g., base station 100) determines whether to distribute / transmit or share a new AI / ML model (or information used for training and updating the new AI / ML model) with terminal 200 for model updates or handovers based on the reported capabilities of terminal 200 (e.g., the AI / ML model stored by terminal 200). For example, the above event may occur during the mobility period of terminal 200.

[0138] An indication of whether a new AI / ML model (or information for training and updating a new AI / ML model) is distributed / transmitted or shared with the terminal 200 for model update or switching can be notified together with an RRC signal such as a handover notification.

[0139] In addition, the network in this embodiment can also be a base station (gNB), an access and mobility management function (AMF), or an AI / ML-dedicated network node.

[0140] For example, in the above-mentioned CSI compression use case, when the network decides to distribute / transmit or share the AI / ML model (or information used for training and updating the AI / ML model) with the terminal 200 for updating or switching the AI / ML model on the terminal side, the AI / ML model itself may be distributed / transmitted or shared in the case of training method 1; information related to FP, BP, or gradient may be distributed / transmitted or shared in the case of training method 2; and the dataset may be distributed / transmitted or shared in the case of training method 3. Furthermore, the distributed / transmitted or shared information may be distributed / transmitted or shared, for example, via an RRC reconfiguration message or a user plane message.

[0141] In addition, the distribution / transmission or sharing of this information can be performed before, during, or after mobility. For example, when distribution / transmission or sharing is performed during mobility, part of the information may be distributed / transmitted or shared from the cell of the mobile source (source cell), the TRP of the mobile source, the beam before switching, or the position before moving, and the remaining part of the information may be distributed / transmitted or shared after connecting with the cell of the mobile destination (target cell) or the TRP of the mobile destination, after switching to the beam of the switching destination, or from the position of the mobile destination. In this way, seamless distribution / transmission or sharing of AI / ML models (or information used for training and updating of AI / ML models) can be achieved.

[0142] Figure 12 FIG. 2 shows an example of operations of the terminal 200 and the network (for example, the base station 100) according to this embodiment. Figure 12 As an example, an action example related to the AI / ML model deployed for each cell is described, but not limited to this, the AI / ML model can also correspond to TRP, location, beam or position.

[0143] exist Figure 12 In the example, the terminal 200 reports information related to the AI / ML model stored in the terminal 200 to the network (S201).

[0144] For example, based on information reported from terminal 200 , the network determines whether terminal 200 stores an AI / ML model (target AI / ML model) corresponding to the cell (target cell) of terminal 200 's movement destination ( S202 ).

[0145] If terminal 200 does not store the target AI / ML model (S202: No), the network distributes / transmits or shares the target AI / ML model (or information used for training or updating the target AI / ML model) with terminal 200 (S203). Terminal 200 obtains the target AI / ML model (or information used for training or updating the target AI / ML model) notified from the network (S204). If terminal 200 stores the target AI / ML model (S202: Yes), the network may not perform S203.

[0146] The network determines an AI / ML model (eg, a target AI / ML model) to be used by the terminal 200 ( S205 ).

[0147] The network activates the AI / ML model to be used by the terminal 200 for the terminal 200 ( S206 ).

[0148] The terminal 200 makes the AI / ML model activated by the network available for use ( S207 ).

[0149] As described above, the network (e.g., base station 100) receives information related to the AI / ML model for the wireless interface stored by terminal 200, and decides whether to notify (e.g., distribute / transmit or share) the settings related to the AI / ML model (e.g., target AI / ML model) to terminal 200 based on the received information.

[0150] As a result, the network does not distribute / transmit or share information related to the target AI / ML model, for example, when there is no need to distribute / transmit or share information related to the target AI / ML model to the terminal 200. In this way, it is possible to suppress the unnecessary distribution / transmission or sharing of a new AI / ML model (or information used for training and updating a new AI / ML model) to the terminal 200 when the model is updated or switched for the terminal 200. Therefore, it is possible to reduce the overhead and latency for updating or switching the AI / ML model on the terminal side. Therefore, according to this embodiment, the efficiency of wireless communication can be improved.

[0151] Furthermore, according to this embodiment, it is not necessary for the network to notify terminal 200 of mapping information, as in embodiments 1 and 2. This reduces downlink overhead. Furthermore, according to this embodiment, for example, it is not necessary to disclose the AI / ML model to be used by terminal 200 for a specific cell, TRP, location, beam, or position through mapping information, thereby providing advantages in network deployment.

[0152] In the above, various embodiments of one non-limiting example of the present disclosure have been described.

[0153] Furthermore, for example, Embodiment 1 or 2 may be combined with Embodiment 3. For example, in some cases, the distribution / transmission or sharing of information related to the target AI / ML model to terminal 200 may be determined based on a request from terminal 200, as in Embodiment 1 or 2, while in other cases, the distribution / transmission or sharing of information related to the target AI / ML model to terminal 200 may be determined based on an instruction from the network (e.g., base station 100), as in Embodiment 3.

[0154] [About Architecture]

[0155] Figure 13 An example of an architecture that includes functions for processing AI / ML. Alternatively, a portion of the network architecture may be Figure 13 The structure shown. In addition, the AI / ML processing function can also be included in the AMF or RAN (gNB).

[0156] [Structure of base station]

[0157] Figure 14 is a block diagram showing a configuration example of the base station 100. Figure 14 In FIG, the base station 100 includes a control unit 101 , a signal generation unit 102 , a transmission unit 103 , a reception unit 104 , an extraction unit 105 , a demodulation unit 106 , and a decoding unit 107 .

[0158] Alternatively, Figure 14 At least one of the control unit 101, the signal generating unit 102, the extracting unit 105, the demodulating unit 106 and the decoding unit 107 is included in Figure 5 In addition, Figure 14 The receiving unit 104 shown may also be included in Figure 5 The receiving part is shown.

[0159] For example, the control unit 101 determines control information related to the AI / ML model of the terminal 200 and outputs the determined information to the signal generator 102. Furthermore, when receiving control information from a function that processes AI / ML (for example, when input from the decoding unit 107), the control unit 101 may also output control information related to the AI / ML model of the terminal 200 to the signal generator 102.

[0160] In addition, the control unit 101 determines, for example, information used by the terminal 200 to receive downlink signals, and outputs the determined information to the signal generation unit 102. The information used by the terminal 200 to receive downlink signals may include, for example, information related to resource allocation of a downlink data channel (e.g., PDSCH: Physical Downlink Shared Channel) or a downlink control channel (e.g., PDCCH: Physical Downlink Control Channel), and information related to a coding / modulation scheme (e.g., MCS: Modulation and Coding Scheme).

[0161] Furthermore, the control unit 101 determines, for example, information used by the terminal 200 to transmit an uplink signal, and outputs the determined information to the signal generation unit 102, the extraction unit 105, the demodulation unit 106, and the decoding unit 107. The information used by the terminal 200 to transmit an uplink signal may include, for example, information related to resource allocation of an uplink data channel (e.g., PUSCH: Physical Uplink Shared Channel) or an uplink control channel (e.g., PUCCH: Physical Uplink Control Channel), information related to a coding / modulation scheme (e.g., MCS), and information related to CSI reporting.

[0162] The signal generator 102 generates a data signal or a control signal bit sequence, for example, using information input from the control unit 101, and performs encoding as needed. Furthermore, the signal generator 102 modulates the encoded bit sequence to generate a modulated signal (e.g., a symbol sequence), and maps it to the radio resource indicated by the control unit 101. The signal generator 102 outputs the mapped signal to the transmitter 103.

[0163] Transmitter 103 performs transmission waveform generation processing, such as orthogonal frequency division multiplexing (OFDM), on the signal input from signal generator 102. Furthermore, in the case of OFDM transmission with a cyclic prefix (CP), transmitter 103 performs an inverse fast Fourier transform (IFFT) on the signal and adds the CP to the IFFT signal. Transmitter 103 also performs RF (radio frequency) processing, such as D / A (digital / analog) conversion or upconversion, on the signal and transmits the wireless signal to terminal 200 via an antenna.

[0164] The receiving unit 104 performs RF processing, such as down-conversion or A / D (Analog / Digital) conversion, on the uplink signal received from the terminal 200 via the antenna. Furthermore, in the case of OFDM transmission, the receiving unit 104 performs, for example, Fast Fourier Transform (FFT) processing on the received signal and outputs the obtained frequency domain signal to the extraction unit 105.

[0165] The extraction unit 105 extracts the radio resource portion where the uplink signal (eg, PUSCH or PUCCH) is transmitted from the reception signal input from the reception unit 104 , for example, based on information input from the control unit 101 , and outputs the extracted radio resource portion to the demodulation unit 106 .

[0166] The demodulation unit 106 demodulates the uplink signal (for example, PUSCH or PUCCH) input from the extraction unit 105 based on, for example, the information input from the control unit 101. The demodulation unit 106 outputs the demodulation result to the decoding unit 107, for example.

[0167] Decoding unit 107 performs error correction decoding on an uplink signal (e.g., PUSCH or PUCCH) based on, for example, information input from control unit 101 and a demodulation result input from demodulation unit 106, to obtain a decoded received bit sequence. For example, if the decoded received bit sequence includes control information related to the AI / ML model of terminal 200, decoding unit 107 outputs the information related to the AI / ML model to control unit 101.

[0168] [Structure of the terminal]

[0169] Figure 15 is a block diagram showing a configuration example of a terminal 200 according to an embodiment of the present disclosure. Figure 15 In the embodiment, the terminal 200 includes a receiving unit 201, an extracting unit 202, a demodulating unit 203, a decoding unit 204, a controlling unit 205, a signal generating unit 206, and a transmitting unit 207.

[0170] Alternatively, Figure 15 At least one of the extraction unit 202, the demodulation unit 203, the decoding unit 204, the control unit 205 and the signal generation unit 206 is included in Figure 6 In addition, Figure 15 The sending unit 207 shown may also be included in Figure 6 The sending part is shown.

[0171] The receiving unit 201 receives downlink signals (e.g., downlink data signals or downlink control signals) from the base station 100 via an antenna, for example, and performs RF processing such as downconversion or A / D conversion on the wireless received signals to obtain received signals (baseband signals). Furthermore, when receiving OFDM signals, the receiving unit 201 performs FFT processing on the received signals to convert them into the frequency domain. The receiving unit 201 outputs the received signals to the extraction unit 202.

[0172] For example, the extraction unit 202 extracts a radio resource portion that may contain a downlink control signal from the received signal input from the reception unit 201 based on information related to the radio resource of the downlink control signal input from the control unit 205, and outputs the extracted radio resource portion to the demodulation unit 203. Furthermore, the extraction unit 202 extracts a radio resource portion that contains a downlink data signal based on information related to the radio resource of the data signal input from the control unit 205, and outputs the extracted radio resource portion to the demodulation unit 203.

[0173] The demodulation unit 203 demodulates the signal (for example, PDCCH or PDSCH) input from the extraction unit 202 based on information input from the control unit 205 , for example, and outputs the demodulation result to the decoding unit 204 .

[0174] The decoding unit 204 performs error correction decoding on the PDCCH or PDSCH, for example, using the demodulation result input from the demodulation unit 203 , and obtains, for example, a control signal or a downlink data signal. The decoding unit 204 outputs the control signal to the control unit 205 .

[0175] The control unit 205 determines information related to downlink transmission based on, for example, the information obtained from the control signal input from the decoding unit 204, and outputs it to the extraction unit 202 and the demodulation unit 203. Furthermore, the control unit 205 determines information related to uplink transmission based on, for example, the information obtained from the control signal input from the decoding unit 204, and outputs it to the signal generation unit 206. Furthermore, the control unit 205 generates control information related to the AI / ML model using a terminal method and outputs it to the signal generation unit 206.

[0176] Based on the control information related to the terminal AI / ML model and information related to uplink transmission input from the control unit 205, the signal generation unit 206 generates an uplink data signal or an uplink control signal, encodes and modulates the bit stream of the generated signal, and maps it to the radio resources. Furthermore, the control information related to the terminal AI / ML model can be included in a user plane message or a control plane message, for example. The signal generation unit 206 outputs the uplink signal to which the signal is mapped, for example, to the transmission unit 207.

[0177] Transmitter 207 generates a transmit signal waveform, such as OFDM, on the signal input from signal generator 206. Furthermore, for example, when performing OFDM transmission or DFT-s-OFDM transmission using a CP, transmitter 207 performs IFFT processing on the signal and adds a CP to the IFFT-processed signal. Alternatively, when generating a single-carrier waveform such as DFT-s-OFDM, transmitter 207 may, for example, add a DFT unit (not shown) upstream of signal generator 206. Furthermore, transmitter 207 performs RF processing, such as D / A conversion and up-conversion, on the transmit signal and transmits the wireless signal to base station 100 via an antenna.

[0178] (Other embodiments)

[0179] In addition, the use cases for applying the above-mentioned embodiments are not limited to CSI compression, but can be applied to any use cases in which an AI / ML model is configured on the terminal side. Examples of use cases in which an AI / ML model is configured on the terminal side include CSI prediction based on the AI / ML model on the terminal side, beam prediction in the spatial or temporal domain based on the AI / ML model on the terminal side, and positioning accuracy enhancement based on the AI / ML model on the terminal side.

[0180] In addition, the AI / ML model is not limited to mapping to a cell, site, TRP, beam, or location. For example, different AI / ML models may be used depending on other specific parameters such as the terminal's moving speed, wireless channels such as multipath, cell congestion, and transmitted traffic. For example, in Implementation 1 or Implementation 2, the relationship between these specific parameters and the information of the AI / ML model (e.g., mapping information) may also be notified to the terminal 200. In addition, in Implementation 3, the relationship between these specific parameters and the information of the AI / ML model (e.g., UE capabilities) may also be notified to the network.

[0181] In addition, the "Mobility" used in the above embodiment may include switching in RRC connected state (CONNETED) mode, reselection in RRC idle state (IDLE) or RRC inactive state (INACTIVE) mode. In addition, "Mobility" may also include movement between TRPs, switching of beams, and movement of location, etc.

[0182] In addition, in the present disclosure, the signal / message / signaling used for notification can be a control plane message (for example, UCI or MAC-CE), or an RRC signal, or a notification using DCI as physical layer signaling.

[0183] (Replenish)

[0184] The information indicating whether the terminal 200 supports the functions, actions or processes described in the above embodiments and supplements may be sent (or notified) by the terminal 200 to the base station 100 as capability information or capability parameters of the terminal 200 .

[0185] The capability information may also include the following information element (IE: Information Element), which individually indicates whether terminal 200 supports at least one of the functions, actions, and processes described in each of the above embodiments, variations, and supplements. Alternatively, the capability information may also include the following information element, which indicates whether terminal 200 supports a combination of two or more of the functions, actions, and processes described in each of the above embodiments, variations, and supplements.

[0186] For example, the base station 100 may determine (or determine or assume) the functions, actions, or processes supported (or not supported) by the terminal 200 that sent the capability information based on the capability information received from the terminal 200. The base station 100 may perform actions, processes, or controls corresponding to the determination result based on the capability information. For example, the base station 100 may control processes related to the AI / ML model based on the capability information received from the terminal 200.

[0187] It should be noted that the fact that the terminal 200 does not support some of the functions, actions, or processes described in the above-mentioned embodiments, modifications, and supplements may alternatively mean that such some functions, actions, or processes are restricted in the terminal 200. For example, information or a request related to such restrictions may be notified to the base station 100.

[0188] Information related to the capabilities or limitations of the terminal 200 may be defined in a standard, for example, or may be implicitly notified to the base station 100 by being associated with information known to the base station 100 or information transmitted to the base station 100 .

[0189] The embodiments, modifications, and supplements of a non-limiting embodiment of the present disclosure have been described above.

[0190] (Control Signal)

[0191] In the present disclosure, downlink control signals (information) associated with the present disclosure may be signals (information) sent in the PDCCH of the physical layer, or signals (information) sent in the MAC (Medium Access Control) CE (Control Element) or RRC of the higher layer. Furthermore, predefined signals (information) may be used as downlink control signals.

[0192] The uplink control signal (information) associated with the present disclosure may be a signal (information) transmitted in the PUCCH of the physical layer, or a signal (information) transmitted in the MAC CE or RRC of a higher layer. Furthermore, the uplink control signal may be a predefined signal (information). Furthermore, the uplink control signal may be converted to UCI (uplink control information), first stage SCI (sidelink control information), or second stage SCI.

[0193] (Base Station)

[0194] In this disclosure, a base station can be a TRP (Transmission Reception Point), cluster head, access point, RRH (Remote Radio Head), eNodeB (eNB), gNodeB (gNB), BS (Base Station), BTS (Base Transceiver Station), motherboard, gateway, or the like. Furthermore, in sidelink communications, a base station can also be a terminal. A base station can also be a relay device that relays communications between a higher-level node and a terminal. Furthermore, a base station can also be a roadside device.

[0195] (Uplink / Downlink / Sidelink)

[0196] The present disclosure can be applied to any link in the uplink, downlink, and sidelink. For example, the present disclosure can be applied to the PUSCH, PUCCH, and PRACH (Physical Random Access Channel) of the uplink, the PDSCH, PDCCH, and PBCH (Physical Broadcast Channel) of the downlink, and the PSSCH (Physical Sidelink Shared Channel), PSCCH (Physical Sidelink Control Channel), and PSBCH (Physical Sidelink Broadcast Channel) of the sidelink.

[0197] It should be noted that PDCCH, PDSCH, PUSCH, and PUCCH are examples of downlink control channels, downlink data channels, uplink data channels, and uplink control channels. PSCCH and PSSCH are examples of sidelink control channels and sidelink data channels. PBCH and PSBCH are examples of broadcast channels, and PRACH is an example of a random access channel.

[0198] (Data channel / Control channel)

[0199] The present disclosure can be applied to any channel among data channels and control channels. For example, the channels of the present disclosure can be replaced with PDSCH, PUSCH, PSSCH of data channels and PDCCH, PUCCH, PBCH, PSCCH, PSBCH of control channels.

[0200] (Reference signal)

[0201] In this disclosure, a reference signal is a signal known to both the base station and the terminal, and is sometimes also referred to as an "RS (Reference Signal)" or a "pilot signal." A reference signal can also be one of the following: DMRS (Demodulation Reference Signal), CSI-RS (Channel State Information-Reference Signal), TRS (Tracking Reference Signal), PTRS (Phase Tracking Reference Signal), CRS (Cell-specific Reference Signal), or SRS (Sounding Reference Signal).

[0202] (Time interval)

[0203] In the present disclosure, the unit of time resources is not limited to one of a time slot and a symbol, or a combination thereof. For example, it may be a time resource unit such as a frame, a superframe, a subframe, a time slot, a subslot, a microslot, or a symbol, an OFDM (Orthogonal Frequency Division Multiplexing Access) symbol, or an SC-FDMA (Single Carrier-Frequency Division Multiple Access) symbol, or other time resource units. In addition, the number of symbols contained in a time slot is not limited to the number of symbols exemplified in the above embodiment, and may be other numbers of symbols.

[0204] (frequency band)

[0205] The present disclosure can be applied to any band including the authorized band and the unlicensed band.

[0206] (communication)

[0207] The present disclosure can be applied to any communication between a base station and a terminal (Uu link communication), between terminals (sidelink communication), or in V2X (Vehicle to Everything) communication. For example, the channels of the present disclosure can be replaced with PSCCH, PSSCH, PSFCH (Physical Sidelink Feedback Channel), PSBCH, PDCCH, PUCCH, PDSCH, PUSCH, or PBCH.

[0208] Furthermore, the present disclosure can be applied to any network, including terrestrial networks and non-terrestrial networks (NTNs) using satellites or high-altitude pseudo-satellites (HAPS). Furthermore, the present disclosure can also be applied to terrestrial networks where the transmission delay is greater than the symbol length or slot length, such as networks with large cell sizes and ultra-wideband transmission networks.

[0209] (Antenna port)

[0210] An antenna port refers to a logical antenna (antenna group) consisting of one or more physical antennas. That is, an antenna port does not necessarily refer to a single physical antenna; it can also refer to an array antenna consisting of multiple antennas. For example, an antenna port is not defined as consisting of a number of physical antennas, but rather as the minimum unit by which a terminal can transmit a reference signal. Furthermore, an antenna port is sometimes defined as the minimum unit by which a precoding vector weight is multiplied.

[0211] 5G NR System Architecture and Protocol Stack

[0212] 3GPP continues to work on the next version of fifth-generation mobile technology (also known as "5G"), including the development of new radio access technology (NR) operating in the frequency range up to 100 GHz. The first version of the 5G standard was completed in late 2017, enabling the transition to trial production and commercial deployment of devices (e.g., smartphones) compliant with the 5G NR standard.

[0213] For example, the overall system architecture envisions an NG-RAN (Next Generation Radio Access Network) consisting of gNBs. The gNBs provide UE-side termination of the user plane (SDAP (Service Data Adaptation Protocol) / PDCP (Packet Data Convergence Protocol) / RLC (Radio Link Control) / MAC / PHY (Physical Layer)) and control plane (RRC) protocols of the NG radio access. gNBs are connected to each other via the Xn interface. Furthermore, gNBs are connected to the Next Generation Core (NGC) via the Next Generation (NG) interface, more specifically, to the Access and Mobility Management Function (AMF) (e.g., a core entity that implements the AMF) via the NG-C interface, and to the User Plane Function (UPF) (e.g., a core entity that implements the UPF) via the NG-U interface. Figure 16 Represents the NG-RAN architecture (for example, refer to 3GPP TS 38.300 v15.6.0, section 4).

[0214] The NR user plane protocol stack (e.g., see 3GPP TS 38.300, Section 4.4.1) includes the PDCP (Packet Data Convergence Protocol) sublayer (see Section 6.4 of TS 38.300), the RLC (Radio Link Control) sublayer (see Section 6.3 of TS 38.300), and the MAC (Media Access Control) sublayer (see Section 6.2 of TS 38.300), which terminates on the network side in the gNB. Furthermore, a new access stratum (AS) sublayer (SDAP: Service Data Adaptation Protocol) has been introduced on top of PDCP (e.g., see Section 6.5 of 3GPP TS 38.300). Furthermore, a control plane protocol stack has been defined for NR (e.g., see Section 4.4.2 of TS 38.300). An overview of Layer 2 functionality is provided in Section 6 of TS 38.300. The functions of the PDCP sublayer, RLC sublayer, and MAC sublayer are listed in Sections 6.4, 6.3, and 6.2 of TS 38.300, respectively. The functions of the RRC layer are listed in Section 7 of TS 38.300.

[0215] For example, the media access control layer handles multiplexing of logical channels, scheduling including processing of various parameter sets, and various functions associated with the scheduling.

[0216] For example, the physical layer (PHY) is responsible for coding, PHY HARQ (Physical Layer Hybrid Automatic Repeat Request) processing, modulation, multi-antenna processing, and mapping signals to appropriate physical time-frequency resources. In addition, the physical layer handles the mapping of transport channels to physical channels. The physical layer provides services to the MAC layer in the form of transport channels. A physical channel corresponds to a set of time-frequency resources used to send a specific transport channel, and each transport channel is mapped to a corresponding physical channel. For example, among physical channels, uplink physical channels include PRACH (Physical Random Access Channel), PUSCH (Physical Uplink Shared Channel), and PUCCH (Physical Uplink Control Channel), and downlink physical channels include PDSCH (Physical Downlink Shared Channel), PDCCH (Physical Downlink Control Channel), and PBCH (Physical Broadcast Channel).

[0217] The use cases / extended scenarios of NR may include enhanced mobile broadband (eMBB), ultra-reliable low-latency communications (URLLC), and massive machine type communications (mMTC), which have various necessary conditions in terms of data rate, latency, and coverage. For example, eMBB is expected to support a peak data rate (20 Gbps in the downlink and 10 Gbps in the uplink) and an effective (user-experienced) data rate that is about three times the data rate provided by IMT-Advanced (International Mobile Telecommunications-Advanced). On the other hand, in the case of URLLC, ultra-low latency (the latency of the user plane is 0.5 ms in UL and DL, respectively) and high reliability (1-10 -5 Finally, in mMTC, a high connection density (1,000,000 devices / km in urban environments) is preferably required. 2 ), wide coverage in harsh environments and extremely long battery life (15 years) for an inexpensive device.

[0218] Therefore, sometimes the OFDM parameter set (e.g., subcarrier spacing (SCS), OFDM symbol length, cyclic prefix (CP) length, number of symbols per scheduling interval) that is suitable for one use case is not valid for other use cases. For example, in low-latency services, it is preferred that the symbol length is shorter than that of mMTC services (therefore, the subcarrier spacing is larger) and / or the number of symbols per scheduling interval (also known as "TTI (Transmission Time Interval)") is smaller. Moreover, in extended scenarios where the channel delay spread is large, it is preferred that the CP length is longer than that in scenarios where the delay spread is short. The subcarrier spacing can also be optimized according to the situation to maintain the same CP overhead. The subcarrier spacing supported by NR can have more than one value. Correspondingly, subcarrier spacings of 15kHz, 30kHz, 60kHz... are currently considered. The symbol length Tu and the subcarrier spacing Δf are directly related according to the formula Δf=1 / Tu. Similarly, in the LTE system, the term “resource element” can be used to represent the smallest resource unit consisting of one subcarrier with a length of one OFDM / SC-FDMA (Single-Carrier Frequency Division Multiple Access) symbol.

[0219] In the new wireless system 5G-NR, a resource grid of subcarriers and OFDM symbols is defined for each numerology set and each carrier in both the uplink and downlink. Each element of the resource grid is called a "resource element" and is identified by a frequency index in the frequency domain and a symbol position in the time domain (see 3GPP TS 38.211 v15.6.0).

[0220] Functional separation between NG-RAN and 5GC in 5G NR

[0221] Figure 17 This indicates the functional separation between NG-RAN and 5GC. The logical node of NG-RAN is the gNB or ng-eNB. The 5GC has the logical nodes AMF, UPF, and SMF (Session Management Function).

[0222] For example, gNB and ng-eNB host the following key functions:

[0223] - Radio bearer control (RBC), radio admission control (RAC), connection mobility control (CMC), and radio resource management (RRM) functions, such as dynamically allocating (scheduling) resources to UEs in both uplink and downlink;

[0224] - IP (Internet Protocol) header compression, encryption, and integrity protection of data;

[0225] - Selection of the AMF upon attaching the UE in case the routing towards the AMF cannot be decided based on the information provided by the UE;

[0226] - Routing of user plane data towards the UPF;

[0227] - Routing of control plane information towards the AMF;

[0228] -Setting up and disconnecting connections;

[0229] - Scheduling and sending of paging messages;

[0230] - Scheduling and sending of system broadcast information (AMF or Operation, Admission, Maintenance (OAM) function as the originator);

[0231] - Settings for measurements and measurement reports for mobility and scheduling;

[0232] - Packet marking of transmission class in uplink;

[0233] -Session management;

[0234] - Network slicing support;

[0235] -QoS (Quality of Service) flow management and mapping to data radio bearers;

[0236] - Support for UEs in RRC_INACTIVE (RRC inactive) state;

[0237] -NAS (Non Access Stratum) message distribution function;

[0238] -Sharing of wireless access networks;

[0239] -Dual connection;

[0240] -Close collaboration between NR and E-UTRA (Evolved Universal Terrestrial Radio Access).

[0241] The Access and Mobility Management Function (AMF) hosts the following main functions:

[0242] - Functionality to terminate Non-Access Stratum (NAS) signaling;

[0243] -Security of NAS signaling;

[0244] -Security control of the access layer (AS);

[0245] -Core Network (CN) inter-node signaling for mobility between 3GPP access networks;

[0246] - the possibility of reaching the UE in idle mode (including the control and execution of paging retransmission);

[0247] - Management of the registration area;

[0248] -Support for intra-system and inter-system mobility;

[0249] -Access authentication;

[0250] - Access permission including roaming permission check;

[0251] -Mobility management control (subscription and policy);

[0252] - Network slicing support;

[0253] - Selection of Session Management Function (SMF).

[0254] In addition, the User Plane Function (UPF) hosts the following main functions:

[0255] - Anchor point for intra-RAT (Radio Access Technology) mobility / inter-RAT (Inter-RAT) mobility (where applicable);

[0256] - External PDU (Protocol Data Unit) session point for interconnection with data networks;

[0257] -Packet routing and forwarding;

[0258] - Packet inspection and policy rule enforcement in the user plane;

[0259] - Business usage reports;

[0260] - Uplink classifier for supporting routing of traffic towards the data network;

[0261] -BranchingPoint used to support multi-homed PDU session;

[0262] - QoS processing for the user plane (e.g., packet filtering, gating, UL / DL rate enforcement);

[0263] - Verification of uplink services (mapping of SDF (Service Data Flow) to QoS flows);

[0264] - Downlink packet buffering and downlink data notification triggering function.

[0265] Finally, the Session Management Function (SMF) hosts the following main functions:

[0266] -Session management;

[0267] -Allocation and management of UE IP addresses;

[0268] - UPF selection and control;

[0269] - Traffic steering configuration function in the user plane function (UPF) for directing traffic to the appropriate destination;

[0270] - Enforcement of policies and QoS in the control part;

[0271] -Notification of downlink data.

[0272] <RRC connection setup and re-setup process>

[0273] Figure 18 Represents several interactions between the UE, gNB and AMF (5GC entity) when the UE transitions from RRC_IDLE (RRC Idle) to RRC_CONNECTED (RRC Connected) in the NAS part (refer to TS 38.300 v15.6.0).

[0274] RRC is a high-layer signaling protocol used for the configuration of the UE and gNB. Through this transition, the AMF prepares UE context data (including, for example, PDU session context, security keys, UE radio capabilities, UE security capabilities, etc.) and sends it to the gNB along with the Initial Context Setup Request (INITIAL CONTEXT SETUP REQUEST). The gNB then activates AS security together with the UE. This is done by the gNB sending a Security Mode Command message to the UE, and the UE responding to the gNB with a Security Mode Complete message. The gNB then sends an RRC Reconfiguration message to the UE, and the gNB receives an RRC Reconfiguration Complete message from the UE in response to the RRC Reconfiguration message, thereby reconfiguring the Signaling Radio Bearer 2 (SRB2) and Data Radio Bearer (DRB). For signaling-only connections, since SRB2 and DRB are not configured, the steps related to RRC reconfiguration can be omitted. Finally, the gNB notifies the AMF of the completion of the setup process using the Initial Context Setup Response (INITIAL CONTEXT SETUP RESPONSE).

[0275] Therefore, the present disclosure provides the following fifth-generation core network (5GC) entity (e.g., AMF, SMF, etc.), which includes: a control circuit that, when operating, establishes a next-generation (NG) connection with a gNodeB (gNodeB); and a transmitter that, when operating, sends an initial context setup message to the gNodeB via the NG connection to set up a signaling radio bearer between the gNodeB and a user equipment (UE). Specifically, the gNodeB sends radio resource control (RRC) signaling including a resource allocation setup information element (IE) to the UE via the signaling radio bearer. The UE then transmits in the uplink or receives in the downlink based on the resource allocation setup.

[0276] IMT Utilization Scenarios After 2020

[0277] Figure 19Indicates several use cases for 5G NR. In the 3rd Generation Partnership Project New Radio (3GPP NR), three use cases supporting a wide range of services and applications, as envisioned by IMT-2020, have been studied. The first phase of specification development for high-capacity, high-speed communications (eMBB: enhanced mobile broadband) has been completed. Current and future work includes the gradual expansion of support for eMBB and the standardization of ultra-reliable and low-latency communications (URLLC: ultra-reliable and low-latency communications) and multi-machine-type communications (mMTC: massive machine-type communications). Figure 19 Several examples of conceptual use cases for IMT after 2020 (for example, see ITU-R M.2083) Figure 2 ).

[0278] The use cases of URLLC have strict requirements related to performance such as throughput, latency (delay) and availability. The use cases of URLLC are conceived as an element technology for realizing future applications such as wireless control of industrial production processes or manufacturing processes, remote medical surgery, automation of power transmission and distribution in smart grids, and traffic safety. The ultra-high reliability of URLLC is supported by determining technologies that meet the requirements set by TR38.913. In the NR URLLC version 15, as an important requirement, the requirement that the target user plane latency is 0.5ms in UL (uplink) and 0.5ms in DL (downlink) is included. The overall URLLC requirement for a packet transmission is a block error rate (BLER) of 1E-5 for a packet size of 32 bytes with a user plane latency of 1ms.

[0279] Considering the physical layer, there are a number of possible approaches to improve reliability. Current scope for improving reliability includes defining additional CQI (Channel Quality Indicator) tables for URLLC, more compact DCI formats, PDCCH iteration, etc. However, as NR (an important prerequisite for NR URLLC) becomes more stable and receives further development, this scope can be expanded to achieve ultra-high reliability. Specific use cases for NR URLLC in Release 15 include augmented reality / virtual reality (AR / VR), e-health, e-safety, and critical applications.

[0280] In addition, the technical enhancements to the goals of NR URLLC are aimed at improving latency and increasing reliability. Technical enhancements for improving latency include configurable parameter sets, non-slot-based scheduling with flexible mapping, unauthorized (authorized) uplinks, slot-level repetitions in data channels, and preemption in downlinks. Preemption means stopping the transmission of allocated resources and using the allocated resources for other transmissions that are requested later and must meet the necessary conditions for lower latency / higher priority. Therefore, the transmission that has been allowed will be replaced by the subsequent transmission. Preemption can be applied regardless of the specific service type. For example, the transmission of service type A (URLLC) can also be replaced by the transmission of service type B (eMBB, etc.). Technical enhancements related to improving reliability include a dedicated CQI / MCS table for a target BLER of 1E-5.

[0281] mMTC (Massive Machine Type Communications) use cases typically involve a large number of connected devices transmitting relatively small amounts of data that are less susceptible to latency. These devices are required to be low-priced and have very long battery life. From the perspective of NR, utilizing very narrow bandwidth segments is one approach to conserving UE power and extending battery life.

[0282] As mentioned above, it is predicted that the room for reliability improvement in NR will be further expanded. It is one of the important requirements for all situations. For example, the important requirement related to URLLC and mMTC is high reliability or ultra-high reliability. From the perspective of wireless and the perspective of the network, reliability can be improved in several mechanisms. In general, there are two to three important areas that may help improve reliability. These areas include compact control channel information, repetition of data channels / control channels, and diversity related to the frequency domain, time domain and / or spatial domain. These areas can be used universally to improve reliability regardless of the specific communication scenario.

[0283] Regarding NR URLLC, further use cases with more stringent requirements are envisioned, such as factory automation, transportation, and power transmission. Strict requirements include high reliability (up to 10 -6 level of reliability), high availability, packet sizes up to 256 bytes, time synchronization up to a few microseconds (μs) (capable of being set to 1 μs or a few μs depending on the use case, depending on the frequency range and short latency of around 0.5 ms to 1 ms (e.g., 0.5 ms latency in the user plane as the target).

[0284] Moreover, regarding NR URLLC, from the perspective of the physical layer, there may be several technical enhancements. These technical enhancements include the enhancement of PDCCH (Physical Downlink Control Channel) related to compact DCI, repetition of PDCCH, and increased monitoring of PDCCH. In addition, the enhancement of UCI (Uplink Control Information) is related to the enhancement of enhanced HARQ (Hybrid Automatic Repeat Request) and CSI feedback. In addition, there may be enhancement of PUSCH related to frequency hopping at the mini-slot level and enhancement of retransmission / repetition. The term "mini-slot" refers to a transmission time interval (TTI) that contains fewer code elements than a time slot (a time slot has 14 code elements).

[0285] QoS Control

[0286] The 5G QoS (Quality of Service) model is based on QoS flows, supporting both QoS flows that require guaranteed bit rates (GBR: Guaranteed Bit Rate QoS flows) and QoS flows that do not (non-GBR QoS flows). Therefore, at the NAS level, a QoS flow is the most granular QoS division within a PDU session. A QoS flow is identified within a PDU session by the QoS Flow ID (QFI: QoS Flow ID) transmitted in the encapsulation header via the NG-U interface.

[0287] For each UE, 5GC establishes one or more PDU sessions. For each UE, in conjunction with the PDU session, NG-RAN, for example, as described in the previous reference Figure 18 As described, at least one data radio bearer (DRB) is established. Additionally, DRBs added to the QoS flows of this PDU session may be configured later (when this is configured depends on the NG-RAN). The NG-RAN maps packets belonging to various PDU sessions to various DRBs. NAS-level packet filters in the UE and 5GC are used to associate UL and DL packets with QoS flows, and AS-level mapping rules in the UE and NG-RAN associate UL and DL QoS flows with DRBs.

[0288] Figure 20 Indicates the non-roaming reference architecture of 5G NR (refer to TS 23.501 v16.1.0, section 4.23). Application Function (AF) (e.g., host Figure 19The 5G service is an external application server for the 5G service illustrated in the example, which interacts with the 3GPP core network to provide services. For example, the network exposure function (NEF) is accessed to support applications that affect the routing of the service, or the policy framework is interacted with for policy control (for example, QoS control) (see Policy Control Function (PCF)). Based on the operator's deployment, the operator considers that the application function that is trusted can interact directly with the associated network function (Network Function). Application functions that are not allowed by the operator to directly access the network function interact with the associated network function via the NEF using an open framework to the outside world.

[0289] Figure 20 It also represents further functional units of the 5G architecture, namely, the Network Slice Selection Function (NSSF), the Network Repository Function (NRF), the Unified Data Management (UDM), the Authentication Server Function (AUSF), the Access and Mobility Management Function (AMF), the Session Management Function (SMF), and the Data Network (DN, e.g., services provided by an operator, Internet access, or services provided by a third party). All or part of the core network functions and application services may also be deployed and operated in a cloud computing environment.

[0290] Therefore, the present disclosure provides the following application server (for example, AF of 5G architecture), which includes: a sending unit, which, in action, sends a request containing QoS requirements for at least one of URLLC service, eMMB service and mMTC service to at least one of the functions of 5GC (for example, NEF, AMF, SMF, PCF, UPF, etc.) in order to establish a PDU session of wireless bearer between gNodeB and UE corresponding to QoS requirements; and a control circuit, which, in action, uses the established PDU session to provide service.

[0291] The expression "...part" used in the present disclosure may also be replaced with other expressions such as "...circuitry", "...device", "...unit" or "...module".

[0292] The present disclosure can be implemented by software, hardware, or software in collaboration with hardware. The functional blocks used in the description of the above embodiments are partially or entirely implemented as LSIs (Large Scale Integration) as integrated circuits, and the processes described in the above embodiments may also be partially or entirely controlled by one LSI or a combination of LSIs. The LSI may be composed of individual chips, or may be composed of one chip in a manner that includes part or all of the functional blocks. The LSI may also include data input and output. Depending on the degree of integration, the LSI may also be referred to as an "IC (Integrated Circuit)", "System LSI", "Super LSI", or "Ultra LSI".

[0293] Integrated circuit implementation is not limited to LSIs; it can also be implemented using dedicated circuits, general-purpose processors, or dedicated processors. Alternatively, FPGAs (Field Programmable Gate Arrays), which can be programmed after LSI fabrication, or reconfigurable processors (RPs), which allow reconfiguration of the connections and settings of circuit blocks within the LSI, can be utilized. The present disclosure can also be implemented as digital or analog processing.

[0294] Furthermore, if semiconductor technology advances or other technologies evolve and a technology for integrated circuits emerges that replaces LSIs, it would be possible to use this technology to integrate functional blocks. There is also the possibility of applying biotechnology, etc.

[0295] The present disclosure can be implemented in all types of devices, equipment, and systems (collectively referred to as "communication devices") that have communication capabilities. A communication device may also include a wireless transceiver and processing / control circuitry. The wireless transceiver may also include a receiving unit and a transmitting unit, or perform the functions of these units. The wireless transceiver (transmitting unit, receiving unit) may also include an RF (Radio Frequency) module and one or more antennas. The RF module may also include an amplifier, an RF modulator / demodulator, or similar devices. Non-limiting examples of communication devices include: telephones (mobile phones, smartphones, etc.), tablet computers, personal computers (PCs) (laptops, desktops, notebook computers, etc.), cameras (digital cameras, digital video cameras, etc.), digital players (digital audio / video players, etc.), wearable devices (wearable cameras, smart watches, tracking devices, etc.), game consoles, e-book readers, telehealth / telemedicine (telehealth / medical prescription) devices, vehicles or transportation vehicles with communication capabilities (cars, airplanes, ships, etc.), and combinations of the above devices.

[0296] Communication devices are not limited to portable or mobile devices. They also include all types of devices, equipment, and systems that cannot be portable or fixed. Examples include smart home devices (such as appliances, lighting, smart meters or meters, control panels, etc.), vending machines, and all other "things" that can exist on an IoT (Internet of Things) network.

[0297] Communication includes data communication via a cellular system, a wireless LAN (Local Area Network) system, a communication satellite system, etc., as well as data communication via a combination of these systems.

[0298] Furthermore, the communication device also includes devices such as controllers and sensors that are connected or coupled to the communication equipment performing the communication functions described in the present invention. For example, the communication device may include a controller or sensor that generates control signals or data signals used by the communication equipment performing the communication functions of the communication device.

[0299] In addition, the communication device includes infrastructure equipment that communicates with or controls the above-mentioned non-limiting various devices, such as base stations, access points, and all other devices, equipment, and systems.

[0300] A base station of one embodiment of the present disclosure comprises: a control circuit that determines whether to notify the terminal of settings related to an artificial intelligence model on the terminal side of a wireless interface based on whether there is a request from the terminal; and a sending circuit that sends a signal related to the settings when it is decided to notify the settings.

[0301] In one embodiment of the present disclosure, the transmitting circuit transmits information related to a mapping between a wireless environment and an artificial intelligence model corresponding to the wireless environment.

[0302] In one embodiment of the present disclosure, the wireless environment is related to at least one of a cell, a site, a transceiver point (TRP), a beam, and a location.

[0303] In one embodiment of the present disclosure, the transmitting circuit transmits information related to a mapping between a wireless environment, an artificial intelligence model corresponding to the wireless environment, and a version of the artificial intelligence model.

[0304] In one embodiment of the present disclosure, the wireless environment is related to at least one of a cell, a site, a transceiver point (TRP), a beam, and a location.

[0305] A base station of one embodiment of the present disclosure comprises: a control circuit, which determines whether to notify the terminal of settings related to the artificial intelligence model on the terminal side based on information related to the artificial intelligence model for the wireless interface stored in the terminal; and a sending circuit, which sends a signal related to the setting when deciding to notify the setting.

[0306] A terminal according to an embodiment of the present disclosure comprises: a control circuit for determining whether to request notification of settings related to the artificial intelligence model on the terminal side from the network based on information related to the stored artificial intelligence model for the wireless interface; and a receiving circuit for receiving a signal related to the settings when notification of the settings is requested.

[0307] A terminal according to an embodiment of the present disclosure comprises: a transmitting circuit for transmitting information related to a stored artificial intelligence model for a wireless interface; and a receiving circuit for receiving a signal related to the setting when a notification of a setting related to the artificial intelligence model on the terminal side is decided based on the information.

[0308] In a communication method of one embodiment of the present disclosure, a base station performs the following steps: deciding whether to notify the terminal of settings related to an artificial intelligence model on the terminal side of a wireless interface based on whether there is a request from the terminal; and sending a signal related to the settings when deciding to notify the settings.

[0309] In a communication method of one embodiment of the present disclosure, the base station performs the following steps: deciding whether to notify the terminal of settings related to the artificial intelligence model on the terminal side based on information related to the artificial intelligence model for the wireless interface stored in the terminal; and sending a signal related to the setting when deciding to notify the setting.

[0310] In a communication method of one embodiment of the present disclosure, the terminal performs the following steps: deciding whether to request notification of settings related to the artificial intelligence model on the terminal side from the network based on information related to the stored artificial intelligence model for the wireless interface; and receiving a signal related to the settings when notification of the settings is requested.

[0311] In a communication method of one embodiment of the present disclosure, the terminal performs the following steps: sending information related to a stored artificial intelligence model for a wireless interface; and when a notification of a setting related to the artificial intelligence model on the terminal side is decided based on the information, receiving a signal related to the setting.

[0312] The disclosures of the specification, drawings, and abstract contained in Japanese patent application No. 2023-026228 filed on February 22, 2023 are incorporated herein by reference in their entirety.

[0313] Industrial Applicability

[0314] One embodiment of the present disclosure is useful for a wireless communication system.

[0315] Description of Reference Numerals

[0316] 100 base stations

[0317] 101, 205 Control Department

[0318] 102, 206 signal generating unit

[0319] 103, 207 Sending Department

[0320] 104, 201 Receiving Department

[0321] 105, 202 Extraction Department

[0322] 106, 203 Demodulation Unit

[0323] 107, 204 Decoding Unit

[0324] 200 Terminal

Claims

1. A base station comprising: a control circuit that determines whether to notify the terminal of settings related to the terminal-side artificial intelligence model for the wireless interface based on whether there is a request from the terminal; and The transmitting circuit transmits a signal related to the setting when deciding to notify the setting.

2. The base station according to claim 1, wherein The transmitting circuit transmits information related to a mapping between a wireless environment and an artificial intelligence model corresponding to the wireless environment.

3. The base station according to claim 2, wherein: The wireless environment is related to at least one of a cell, a site, a transceiver point (TRP), a beam, and a location.

4. The base station according to claim 1, wherein The transmitting circuit transmits information related to a mapping between a wireless environment, an artificial intelligence model corresponding to the wireless environment, and a version of the artificial intelligence model.

5. The base station according to claim 4, wherein: The wireless environment is related to at least one of a cell, a site, a transceiver point (TRP), a beam, and a location.

6. A terminal comprising: a control circuit that determines, based on the stored information related to the artificial intelligence model for the wireless interface, whether to request notification of settings related to the artificial intelligence model on the terminal side from the network; and The receiving circuit receives a signal related to the setting when notification of the setting is requested.

7. A communication method, wherein: The base station performs the following steps: determining whether to notify the terminal of settings related to the terminal-side artificial intelligence model for the wireless interface based on whether there is a request from the terminal; as well as When it is decided to notify the setting, a signal related to the setting is transmitted.

8. A communication method, wherein: The terminal performs the following steps: Determining whether to request notification of settings related to the terminal-side artificial intelligence model from the network based on the stored information related to the artificial intelligence model for the wireless interface; as well as When notification of the setting is requested, a signal related to the setting is received.

Images