Terminal, base station, communication method, and integrated circuit

Abstract

This terminal is provided with: a control circuit that determines control information pertaining to at least one of a calculation processing amount and power consumption, which are due to the use of an artificial intelligence model on a terminal side; and a transmission circuit that transmits the control information.

Description

Terminal, base station, communication method and integrated circuit

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

[0002] In recent years, the expansion and diversification of wireless services has led to the expectation of rapid development of the Internet of Things (IoT). Mobile communications are now being used in a wide range of applications, from smartphones and other information terminals to automobiles, homes, home appliances, and industrial equipment. To support this diversification, significant improvements in the performance and functionality of mobile communication systems are required, addressing various requirements, such as increased system capacity, an increased number of connected devices, and low latency. Fifth-generation mobile communication systems (5G) boast high-capacity and ultra-high-speed data transfer (eMBB: enhanced Mobile Broadband), massive machine-type communication (mMTC: massive Machine-Type Communication), and ultra-reliable and low-latency communication (URLLC), providing flexible wireless communications to meet diverse needs.

[0003] The 3rd Generation Partnership Project (3GPP), an international standardization organization, is working on the specification of New Radio (NR) as one of the 5G wireless interfaces.

[0004] RP-213599, “New SI: Study on Artificial Intelligence (AI) / Machine Learning (ML) for NR Air Interface,” Qualcomm (Moderator), December 2021.

[0005] However, there is room for improvement in the efficiency of wireless communications.

[0006] 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 communication.

[0007] A terminal according to one embodiment of the present disclosure comprises a control circuit that determines control information relating to at least one of computational processing volume and power consumption by using an artificial intelligence model on the terminal side, and a transmitting circuit that transmits the control information.

[0008] These comprehensive or specific aspects may be realized as a system, an apparatus, a method, an integrated circuit, a computer program, or a recording medium, or may be realized as any combination of a system, an apparatus, a method, an integrated circuit, a computer program, and a recording medium.

[0009] According to an embodiment of the present disclosure, it is possible to improve the efficiency of wireless communication.

[0010] Further advantages and benefits of one embodiment of the present disclosure will become apparent from the specification and drawings. Such advantages and / or benefits may be provided by some embodiments and features described in the specification and drawings, respectively, but not necessarily all of them may be provided to obtain one or more identical features.

[0011] 3GPP NR system architecture Schematic diagram showing functional separation between NG-RAN and 5GC Sequence diagram of RRC (Radio Resource Control) connection setup / reconfiguration procedures Schematic diagram showing usage scenarios for enhanced Mobile BroadBand (eMBB), massive Machine Type Communications (mMTC), and Ultra Reliable and Low Latency Communications (URLLC) Block diagram showing an exemplary 5G system architecture for a non-roaming scenario

[0012] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings.

[0013] For example, NR employs an access scheme based on Orthogonal Frequency Division Multiplexing (OFDM) for downlink transmission. Furthermore, Multiple-Input Multiple Output (MIMO) is adopted to increase data rates. To effectively utilize MIMO performance, closed-loop control is introduced, in which channel state information (CSI) is fed back from the terminal (e.g., user equipment (UE)) to the base station (e.g., gNB). CSI feedback is expected to achieve high performance with a smaller amount of control information.

[0014] 3GPP Release 18 (also referred to as Rel. 18, for example) discusses the use of artificial intelligence (AI) (hereinafter also referred to as "AI / ML") technologies such as machine learning (ML) in the NR radio interface (see, for example, Non-Patent Document 1). Use cases that utilize AI / ML technologies include, for example, CSI feedback, beam control, and position estimation.

[0015] For example, methods for compressing CSI information in the spatial and frequency domains using AI / ML technologies (hereinafter also referred to as CSI compression) are being considered for CSI feedback. For example, machine learning can automatically extract features by training an AI / ML model (e.g., an artificial intelligence model) such as a neural network using a huge amount of training data. CSI compression can utilize an autoencoder, an algorithm primarily used for dimensional compression and reconstruction of image data. For example, as shown in Figure 1, a CSI matrix represented by two-dimensional spatial and frequency domains is considered an image. The encoding unit (encoder) used for compression by the autoencoder is used for CSI compression processing on the terminal side, and the decoding unit (decoder) used for reconstruction is used for CSI reconstruction processing on the base station or network (e.g., also referred to as "base station / network").

[0016] Generally, high-performance AI / ML models have large model sizes, which can lead to large computational complexity or power consumption during processing on a device. Therefore, in practical application of AI / ML models, it is necessary to strike a balance between the improvement in performance (e.g., communication performance) achieved by using AI / ML and the computational complexity and power consumption on the device.

[0017] In use cases where AI / ML technology is utilized in the wireless interface and an AI / ML model (or part of the AI / ML model) is deployed on the terminal side, it is being considered to operate the AI / ML model using Model-ID-based Life Cycle Management (LCM). In the case of Model-ID-based LCM, the decision on whether to use an AI / ML model can be made by the base station / network side, regardless of whether the AI / ML model is trained on the terminal side or the network side.

[0018] For example, one of the LCM functions under consideration is to include a function for monitoring the performance of AI / ML models (referred to as performance monitoring). Performance monitoring of AI / ML models includes, for example, performance monitoring to check the behavior of new, unverified models or parameters (referred to as Category 1 monitoring), and performance monitoring to check whether the AI / ML models or parameters currently in use are suitable for the current environment (referred to as Category 2 monitoring).

[0019] For example, in performance monitoring to check the behavior of new, unverified models or parameters, the action after monitoring the performance of the AI / ML model(s) may be that the AI / ML model(s) do not need to be changed (or updated), the parameters of the AI / ML model(s) may be updated, the AI / ML model(s) themselves may be updated, or the settings of the preconfigured AI / ML model(s) may be deactivated or stopped.

[0020] Furthermore, for example, in performance monitoring to check whether the AI / ML model or parameters currently in use are suitable for the current environment, as an action after monitoring the performance of the AI / ML model, the AI / ML model currently in operation (activated) may not be changed (or updated), the parameters of the AI / ML model currently in operation (activated) may be updated, the AI / ML model currently in operation (activated) may be changed (switched) to another AI / ML model from a group of preconfigured AI / ML models, or the operation of the AI / ML model currently in operation (activated) may be deactivated, and a fallback may be made to a CSI reporting method that does not use AI / ML.

[0021] In Model-ID-based LCM, where the decision to use an AI / ML model is made by the base station / network, the base station / network is expected to identify information on the terminal's computational processing volume and power consumption in order to determine whether there is a balance between the communication performance resulting from the use of the AI / ML model and the terminal's computational processing volume and power consumption when using the AI / ML model.

[0022] Here, for example, a method is possible in which a terminal-side AI / ML model is implemented in a base station / network and the use of the AI / ML model in the terminal is determined based on the computational processing volume or power consumption of the AI / ML model. However, the actual implementation of the AI / ML model may differ depending on the terminal's hardware or software. For this reason, it is difficult for a base station / network to determine the actual computational processing volume or power consumption of the terminal-side AI / ML model without information from the terminal.

[0023] For example, there is room for consideration regarding a method for a terminal to report information regarding at least one of the amount of computational processing and power consumption (hereinafter also referred to as "amount of computational processing / power consumption") to a base station / network.

[0024] In one non-limiting embodiment of the present disclosure, a method for a terminal to report computational processing volume / power consumption to a base station / network is described. For example, in a use case in which an AI / ML model (or part of an AI / ML model) is deployed on the terminal side, when the base station / network decides whether to use the AI / ML model, the base station / network appropriately determines whether to use the terminal-side AI / ML model by taking into consideration the balance between communication performance resulting from the use of the AI / ML model and the computational processing volume / power consumption of the terminal. For example, the terminal reports information regarding the computational processing volume / power consumption resulting from the use of the terminal-side AI / ML model to the base station / network, thereby assisting the base station / network in determining whether to use the AI / ML model.

[0025] Non-limiting embodiments of the present disclosure will be described below.

[0026] [Overview of Communication System] A communication system according to an aspect of the present disclosure includes, for example, at least one base station and at least one terminal.

[0027] FIG. 2 is a block diagram showing a configuration example of a portion of a base station 100 according to an embodiment of the present disclosure, and FIG. 3 is a block diagram showing a configuration example of a portion of a terminal 200 according to an embodiment of the present disclosure.

[0028] 2, a receiver (e.g., corresponding to a receiver circuit) receives control information related to at least one of the computational load and power consumption due to the use of an AI / ML model on the terminal side. A controller (e.g., corresponding to a control circuit) controls the use of the AI ​​model on the terminal side based on the control information.

[0029] 3, a control unit (e.g., corresponding to a control circuit) determines control information related to at least one of the amount of computational processing and power consumption due to the use of an artificial intelligence model (AI / ML model) on the terminal side. A transmission unit (e.g., corresponding to a transmission circuit) transmits the control information.

[0030] (Embodiment 1) In this embodiment, terminal 200 reports information regarding the amount of computational processing / power consumption in terminal 200 to base station 100 (for example, base station / network). Based on the information reported from terminal 200, the base station / network determines whether to use an AI / ML model in terminal 200.

[0031] In this way, information regarding the amount of computational processing / power consumption due to the use of an AI / ML model on the terminal side in the terminal 200 assists the base station / network in determining whether to use an AI / ML model.

[0032] In addition, the information reported from terminal 200 may include either information regarding the computational processing amount of terminal 200 or information regarding the power consumption of terminal 200, or may include both information regarding the computational processing amount of terminal 200 and information regarding the power consumption of terminal 200.

[0033] For example, information regarding the amount of computational processing required by the use of an AI / ML model on the terminal 200 may include information regarding at least one of the FLOPs (Floating-point Operations) and memory size of the inference processing by the AI / ML model.

[0034] Furthermore, for example, information regarding power consumption due to the use of an AI / ML model on the terminal side in terminal 200 may include information regarding the absolute value of the amount of power consumption of terminal 200 when using the AI / ML model.

[0035] According to this embodiment, the base station / network can identify (or grasp) the computational processing amount / power consumption of terminal 200 due to the use of the terminal-side AI / ML model based on information regarding the computational processing amount / power consumption reported from terminal 200, and can therefore appropriately determine whether to use the terminal-side AI / ML model.

[0036] In addition, the report of information regarding the computational processing amount / power consumption from terminal 200 may be transmitted, for example, via U-plane, higher layer signaling (e.g., Radio Resource Control (RRC)), Medium Access Control-Control Element (MAC-CE), or uplink control information (UCI).

[0037] (Embodiment 2) For example, the power consumption of a device (e.g., terminal 200) may depend on various factors such as an operating state such as temperature, or other operations performed in parallel with an operation using an AI / ML model in terminal 200. Therefore, for example, when an absolute value of the amount of power consumption when using an AI / ML model in terminal 200 is reported, even if the amount of power consumption of terminal 200 when using an AI / ML model at a specific time is reported, the base station / network may not be able to appropriately determine whether to use the AI / ML model, and may not be able to appropriately determine the next operation after determining to use the AI / ML model.

[0038] Furthermore, for example, a process that does not involve frequent memory access but has a large amount of computational processing and a process that involves frequent memory access but has a small amount of computational processing may consume the same amount of power in one terminal implementation, but may consume different amounts of power in another terminal implementation. Thus, the power consumption of a terminal may vary depending on the implementation of the terminal. Therefore, for example, when the computational processing amount or memory size of terminal 200 is reported, even if the computational processing amount or memory size when an AI / ML model is used is reported, the relationship between the power consumption of terminal 200 and the computational processing amount or memory size is unclear, and the base station / network may not be able to appropriately determine whether to use the AI / ML model and may not be able to appropriately determine the next operation after determining to use the AI / ML model.

[0039] Therefore, in this embodiment, the terminal 200 reports to the base station / network, for example, information regarding a comparison between the computational processing amount / power consumption when an AI / ML model is used and the computational processing amount / power consumption when an AI / ML model is not used.

[0040] In this way, in this embodiment, the base station / network is assisted in making a decision on whether to use an AI / ML model by using information on a comparison between the amount of computational processing / power consumption when an AI / ML model is used in terminal 200 and the amount of computational processing / power consumption when an AI / ML model is not used.

[0041] In addition, the control information (or the object of comparison) reported from terminal 200 to the base station / network may be information relating to either the amount of computational processing or the amount of power consumption, or may be information relating to both the amount of computational processing and the amount of power consumption.

[0042] For example, the terminal 200 may report to the base station / network information regarding whether the computational processing amount / power consumption when using an AI / ML model is larger (or smaller) than the computational processing amount / power consumption when not using an AI / ML model.

[0043] In addition, in order for the base station / network to correctly understand the information regarding the comparison of computational processing amount / power consumption, it is expected that there is a shared understanding between the terminal 200 and the base station / network regarding the calculation method of the computational processing amount / power consumption used by the terminal 200 for the comparison (e.g., the method of determining the control information).

[0044] Sharing of understanding regarding the calculation method of the computational amount / power consumption used for comparison may be specified in advance by, for example, a standard. Furthermore, for example, when reporting information regarding the capability of terminal 200 or the applicability of an AI / ML model, terminal 200 may report to a base station / network the calculation method of the computational amount / power consumption used for comparison. Furthermore, the base station / network may configure, for terminal 200, the calculation method of the computational amount / power consumption used for comparison by a semi-static configuration such as RRC. Furthermore, when reporting information regarding the comparison of computational amount / power consumption to a base station / network, terminal 200 may also report the calculation method of the computational amount / power consumption used for comparison.

[0045] For example, the calculation method of the computational complexity / power consumption may be based on at least one of the average computational complexity / power consumption in a specified interval, the peak (maximum) computational complexity / power consumption occurring in a specified interval, or the total computational complexity / power consumption in a specified interval. For example, the specified interval may be a CSI computation time defined in a standard, or may be set by terminal 200 or a base station / network.

[0046] Furthermore, for example, the comparison of computational processing amounts may be performed based on at least one of the FLOPs and memory size of the inference processing by the AI / ML model.

[0047] According to this embodiment, the base station / network can identify (or grasp) the magnitude relationship between the computational processing amount / power consumption between the operation resulting from the use of the terminal-side AI / ML model and the existing operation (for example, the operation when the terminal-side AI / ML model is not used) under the same conditions of the terminal 200. In other words, the base station / network can accurately identify the state of the computational processing amount / power consumption when the terminal-side AI / ML model is used by comparing it with the computational processing amount / power consumption when the terminal-side AI / ML model is not used, regardless of the implementation of the terminal 200.

[0048] This allows the base station / network to appropriately determine whether to use the terminal-side AI / ML model, and also allows the base station / network to appropriately select the next action to take after determining whether to use the terminal-side AI / ML model (for example, whether to continue using the AI / ML model, or whether to stop the operation of the currently operating AI / ML model and fall back to a method that does not use AI / ML).

[0049] The terminal 200 may report information regarding the comparison of computational processing amount / power consumption via any of the U-plane, RRC, MAC-CE, or UCI.

[0050] (Embodiment 3) In this embodiment, terminal 200 reports, for example, information on the difference between the amount of computational processing / power consumption when an AI / ML model is used and the amount of computational processing / power consumption when an AI / ML model is not used, to the base station / network.

[0051] In this way, in this embodiment, the base station / network is assisted in making a decision on whether to use an AI / ML model by using information on the difference between the amount of computational processing / power consumption when an AI / ML model is used in terminal 200 and the amount of computational processing / power consumption when an AI / ML model is not used.

[0052] In addition, the control information (or the object for calculating the difference value) reported from terminal 200 to the base station / network may be information regarding either the amount of computational processing or the amount of power consumption, or may be information regarding both the amount of computational processing and the amount of power consumption.

[0053] For example, the terminal 200 may report to the base station / network, as information regarding the difference between the computational processing amount / power consumption when using an AI / ML model and the computational processing amount / power consumption when not using an AI / ML model, a value representing the absolute value of the difference, or a value representing one of multiple ranges obtained by dividing (or quantizing) the level of the difference value.

[0054] In addition, in order for the base station / network to correctly understand information regarding the difference in computational processing volume / power consumption, it is expected that there is a shared understanding between the terminal 200 and the base station / network regarding the method for calculating the difference in computational processing volume / power consumption in the terminal 200 (e.g., the method for determining control information).

[0055] Sharing of understanding regarding the calculation method of the difference value of the computational amount / power consumption may be specified in advance by, for example, a standard. Furthermore, for example, when reporting information regarding the capability of terminal 200 or the applicability of an AI / ML model, terminal 200 may report the calculation method of the difference value of the computational amount / power consumption to a base station / network. Furthermore, the base station / network may configure the calculation method of the difference value of the computational amount / power consumption for terminal 200 by a semi-static configuration such as RRC. Furthermore, when reporting information regarding the difference value of the computational amount / power consumption to a base station / network, terminal 200 may also report the calculation method of the difference value of the computational amount / power consumption.

[0056] For example, the calculation method for the difference value of the computational load / power consumption may be based on at least one of the average computational load / power consumption in a specified interval, the peak (maximum) computational load / power consumption occurring in a specified interval, or the total computational load / power consumption in a specified interval. For example, the specified interval may be a CSI computation time defined in a standard, or may be set by terminal 200 or a base station / network.

[0057] Furthermore, for example, the difference in the amount of computational processing may be calculated based on at least one of the FLOPs and memory size of the inference processing by the AI / ML model.

[0058] According to this embodiment, the base station / network can identify (or grasp) the relative value of the computational processing amount / power consumption of the terminal 200 between the operation resulting from the use of the terminal-side AI / ML model and the existing operation (for example, the operation when the terminal-side AI / ML model is not used). That is, the base station / network can accurately identify the computational processing amount / power consumption of the terminal-side AI / ML processing by comparing it with the computational processing amount / power consumption when the terminal-side AI / ML model is not used, regardless of the implementation of the terminal 200.

[0059] This allows the base station / network to appropriately determine whether to use the terminal-side AI / ML model, and also allows the base station / network to appropriately select the next action to take after determining whether to use the terminal-side AI / ML model (for example, whether to continue using the AI / ML model, or whether to stop the operation of the currently operating AI / ML model and fall back to a method that does not use AI / ML).

[0060] The report of information relating to the difference in computational processing amount / power consumption from terminal 200 may be transmitted by any of U-plane, RRC, MAC-CE, or UCI.

[0061] (Embodiment 4) In this embodiment, similarly to Embodiments 2 and 3, for example, terminal 200 may report to a base station / network information relating to a comparison or difference between the amount of computational processing / power consumption when an AI / ML model is used and the amount of computational processing / power consumption when an AI / ML model is not used.

[0062] Furthermore, in this embodiment, terminal 200 may report to a base station / network, for example, information regarding a comparison or difference between the amount of computational processing / power consumption when a certain AI / ML model (e.g., a first AI / ML model) is used and the amount of computational processing / power consumption when another AI / ML model (e.g., a second AI / ML model) is used. Furthermore, in this embodiment, terminal 200 may report to a base station / network, for example, information regarding a comparison or difference between the amount of computational processing / power consumption when a certain parameter (e.g., a first parameter) of an AI / ML model is used and the amount of computational processing / power consumption when a different parameter (e.g., a second parameter) of the AI / ML model is used.

[0063] In this way, in this embodiment, the base station / network is assisted in deciding whether to use an AI / ML model by comparing or using information on the difference between the amount of computational processing / power consumption when a certain AI / ML model is used in terminal 200 and the amount of computational processing / power consumption when a different AI / ML model or the same AI / ML model but with different parameters is used.

[0064] In addition, the control information reported from terminal 200 to the base station / network (or the object of comparison, or the object for calculating the difference value) may be information regarding either the amount of computational processing or the amount of power consumption, or information regarding both the amount of computational processing and the amount of power consumption.

[0065] For example, terminal 200 may report to the base station / network information regarding whether the computational processing amount / power consumption when using the AI / ML model currently pre-set in terminal 200 or the AI / ML model currently being used by terminal 200 is larger (or smaller) than the computational processing amount / power consumption when not using the AI / ML model, or information regarding the difference in computational processing amount / power consumption.

[0066] Furthermore, terminal 200 may report to the base station / network information regarding whether the computational processing amount / power consumption when using the AI / ML model currently preset in terminal 200 or the AI / ML model currently being used by terminal 200 is larger (or smaller) than the computational processing amount / power consumption when using an AI / ML model that may be preset in terminal 200 in the future, an AI / ML model that is not currently being used by terminal 200, or different parameters (unused parameters) of the AI / ML model currently preset in terminal 200 or the AI / ML model currently being used by terminal 200, or information regarding the difference in computational processing amount / power consumption.

[0067] In order for the base station / network to correctly understand the information regarding the comparison or difference in computational processing amount / power consumption, it is expected that there is a shared understanding between the terminal 200 and the base station / network regarding the method of calculating the computational processing amount / power consumption used for the comparison or the method of calculating the difference in computational processing amount / power consumption in the terminal 200 (e.g., the method of determining control information).

[0068] Sharing of knowledge regarding a calculation method for the computational amount / power consumption used for comparison or a calculation method for a difference value of the computational amount / power consumption in terminal 200 (hereinafter referred to as a "calculation method for the comparison or difference value") may be specified in advance by, for example, a standard. Furthermore, for example, when reporting information regarding the capability of terminal 200 or the applicability of an AI / ML model, terminal 200 may report the calculation method for the comparison or difference value to a base station / network. Furthermore, the base station / network may configure the calculation method for the comparison or difference value in terminal 200 by a semi-static configuration such as RRC. Furthermore, when reporting information regarding the comparison or difference value of the computational amount / power consumption to a base station / network, terminal 200 may also report the calculation method for the comparison or difference value.

[0069] For example, the calculation method for the comparison or difference value may be based on at least one of the average computational load / power consumption in a specified interval, the peak (maximum) computational load / power consumption occurring in a specified interval, or the total computational load / power consumption in a specified interval. For example, the specified interval may be a CSI computation time defined in a standard or may be set by terminal 200 or a base station / network.

[0070] Furthermore, for example, the difference in the amount of computational processing may be calculated based on at least one of the FLOPs and memory size of the inference processing by the AI / ML model.

[0071] According to this embodiment, the base station / network can identify (or grasp) the magnitude relationship or relative value of the computational workload / power consumption of terminal 200 between operation resulting from use of a terminal-side AI / ML model and existing operation (for example, operation when a terminal-side AI / ML model is not used), or operation resulting from a change to the AI / ML model or parameters. That is, similar to the second and third embodiments, the base station / network can accurately identify the computational workload / power consumption in terminal-side AI / ML processing that is independent of the implementation of terminal 200, and can also identify the magnitude relationship or relative value of the computational workload / power consumption between AI / ML models or parameters of AI / ML models that can be set in terminal 200.

[0072] This allows the base station / network to appropriately determine whether to use the terminal-side AI / ML model, and also allows the base station / network to appropriately select the next action to take after determining whether to use the terminal-side AI / ML model (for example, an action to continue using the AI / ML model, an action to switch the pre-configured or used AI / ML model or parameters, or an action to stop using the AI / ML model or the operation of the currently operating AI / ML model and fall back to a method that does not use AI / ML).

[0073] The report from terminal 200 of information regarding the comparison of computational processing amount / power consumption or the difference value may be transmitted by any of U-plane, RRC, MAC-CE, or UCI.

[0074] Furthermore, in this embodiment, the comparison of the computational processing amount / power consumption in the control information (e.g., information related to the comparison result or the difference value) reported from terminal 200 to the base station / network is not limited to a comparison of the computational processing amount / power consumption when an AI / ML model is used with the computational processing amount / power consumption when an AI / ML model is not used, and a comparison of the computational processing amount / power consumption when a certain AI / ML model is used with the computational processing amount / power consumption when a different AI / ML model or different parameters of the same AI / ML model are used.

[0075] For example, a comparison result or a difference value between the computational processing amount / power consumption when a certain AI / ML model is not used and the computational processing amount / power consumption when another AI / ML model or different parameters of the certain AI / ML model are used may be reported. Alternatively, a comparison result or a difference value between the computational processing amount / power consumption among multiple AI / ML models or parameters that are not currently used by terminal 200 may be reported.

[0076] Furthermore, in this embodiment, a reference model used for comparing computational complexity / power consumption or calculating a difference value in control information (e.g., information related to a comparison result or a difference value) reported from terminal 200 to a base station / network may be defined, and the computational complexity / power consumption when a certain AI / ML model is used may be compared with the computational complexity / power consumption of the reference model, the computational complexity / power consumption when a different AI / ML model or different parameters of the same AI / ML model may be compared with the computational complexity / power consumption of the reference model, or the computational complexity / power consumption when no AI / ML model is used may be compared with the computational complexity / power consumption of the reference model, and these comparison results or difference values ​​may be reported. The structure or parameters of the reference model for comparing computational complexity / power consumption may be notified or set to terminal 200 by the base station / network, or may be determined by terminal 200 and notified to the base station / network. Furthermore, the structure or parameters of the reference model may be defined by a standard.

[0077] Fifth Embodiment In this embodiment, the terminal 200 reports, for example, information on the usage rate or occupancy rate of the amount of computational processing / power consumption to the base station / network. In this way, in this embodiment, the information on the usage rate or occupancy rate of the amount of computational processing / power consumption in the terminal 200 helps the base station / network determine whether to use an AI / ML model.

[0078] The information reported from the terminal 200 may be information on the usage rate or occupancy rate of either the computational processing amount or the power consumption, or may be information on the usage rate or occupancy rate of both the computational processing amount and the power consumption.

[0079] For example, the information on the utilization rate or occupancy rate of the computational processing amount may be information on the utilization rate of a CPU (Central Processing Unit) when the AI / ML processing is executed by a general-purpose processor. For example, the information on the utilization rate or occupancy rate of the computational processing amount may be at least one of the current overall CPU utilization rate of the terminal 200, the CPU utilization rate occupied by computational processing when an AI / ML model is used, the CPU utilization rate occupied by computational processing when an AI / ML model is not used, and the CPU utilization rate occupied by computational processing when a different AI / ML model or the same AI / ML model with different parameters is used.

[0080] Furthermore, for example, when an accelerator dedicated to AI / ML processing is used, the information regarding the utilization rate or occupancy rate of the computational processing volume may be the utilization rate or resource occupancy rate of the computational processing when an AI / ML model is used in the accelerator.

[0081] The utilization rate or occupancy rate of the computational processing amount and the utilization rate or occupancy rate of the power consumption may be calculated and reported separately, or an integrated utilization rate or integrated occupancy rate calculated based on both the utilization rate or occupancy rate of the computational processing amount and the utilization rate or occupancy rate of the power consumption may be defined, and information on the integrated utilization rate or integrated occupancy rate may be reported.

[0082] In addition, in order for the base station / network to correctly understand information regarding the utilization rate or occupancy rate of the computational processing amount / power consumption, it is expected that there is a shared understanding between the terminal 200 and the base station / network regarding the method for calculating the utilization rate or occupancy rate of the computational processing amount / power consumption in the terminal 200 (e.g., the method for determining the control information).

[0083] Sharing of understanding regarding the calculation method for the utilization rate or occupancy rate of the computational amount / power consumption may be specified in advance by, for example, a standard. Furthermore, for example, when reporting information regarding the capability of terminal 200 or the applicability of an AI / ML model, terminal 200 may report the calculation method for the utilization rate or occupancy rate of the computational amount / power consumption to a base station / network. Furthermore, the base station / network may configure the calculation method for the utilization rate or occupancy rate of the computational amount / power consumption to terminal 200 by a quasi-static configuration such as RRC. Furthermore, when reporting information regarding the utilization rate or occupancy rate of the computational amount / power consumption to a base station / network, terminal 200 may also report the calculation method for the utilization rate or occupancy rate of the computational amount / power consumption.

[0084] For example, the calculation method of the usage rate or occupancy rate of the computational load / power consumption may be based on at least one of the average computational load / power consumption in a specified interval, the peak (maximum) computational load / power consumption occurring in a specified interval, or the total computational load / power consumption in a specified interval. For example, the specified interval may be a CSI computation time defined in a standard, or may be set by terminal 200 or a base station / network.

[0085] According to this embodiment, the base station / network can appropriately determine whether to use an AI / ML model, taking into consideration the usage rate or occupancy rate of the computational processing amount / power consumption of the terminal 200. Furthermore, the base station / network can appropriately select the next operation after determining whether to use an AI / ML model on the terminal side (including, for example, an operation to continue using the AI / ML model, an operation to switch the AI / ML model or parameters that are pre-set or used, or an operation to stop using the AI / ML model or the operation of the AI / ML model currently in operation and fall back to a method that does not use AI / ML, etc.).

[0086] For example, the base station / network may determine the use of an AI / ML model so that the usage rate or occupancy rate of the computational processing amount / power consumption of the terminal 200 does not exceed 100%.

[0087] In addition, the base station / network can be set so that the utilization rate or occupancy rate of the computational processing amount / power consumption on the terminal side is 100% or more. In this case, it is possible to realize processing on a best-effort basis.

[0088] The report of information on the usage rate or occupancy rate of the computational processing amount / power consumption from terminal 200 may be transmitted by any of U-plane, RRC, MAC-CE, or UCI.

[0089] The embodiments of the present disclosure have been described above.

[0090] (Variation 1) When a performance monitoring process is applied to terminal 200 (for example, when terminal 200 performs performance monitoring of an AI / ML model), the above-mentioned information on the computational processing amount / power consumption of terminal 200 may be reported from terminal 200 to a base station / network together with information on the performance monitoring process (for example, information on KPIs of the AI / ML model, information on failures of the AI / ML model).

[0091] Furthermore, the information regarding the computational processing amount / power consumption of terminal 200 described above may be reported from terminal 200 to the base station / network together with a signal for notifying the AI / ML capability of terminal 200.

[0092] Furthermore, the information regarding the amount of computational processing / power consumption of the terminal 200 may be reported from the terminal 200 to the base station / network when a trigger is notified from the base station / network, for example.

[0093] Furthermore, the above-described information on the computational processing amount / power consumption of terminal 200 may be reported from terminal 200 to a base station / network, for example, when a set timer expires or when the information on the computational processing amount / power consumption exceeds a threshold. Note that the timer time setting and the threshold for the information on the computational processing amount / power consumption may be specified in advance by a standard, may be set by terminal 200 and reported to the base station / network, or may be set in terminal 200 by semi-static configuration such as RRC by the base station / network.

[0094] (Variation 2) In the above embodiment, a case has been described in which terminal 200 reports to a base station / network information relating to a comparison result or a difference value of the computational processing amount / power consumption between two processes that differ in at least one of whether or not an AI / ML model is used, the AI / ML model, and parameters. However, this is not limited to this. For example, terminal 200 may report to a base station / network a comparison result (or a difference value) of the computational processing amount / power consumption between three or more processes that differ in at least one of whether or not an AI / ML model is used, the AI / ML model, and parameters. For example, terminal 200 may report the largest (or smallest) value of the computational processing amount / power consumption of the three or more processes.

[0095] (Variation 3) The computational processing amount / power consumption when an AI / ML model is used, reported by terminal 200, may be the computational processing amount / power consumption required for inference processing using the AI / ML model, the computational processing amount / power consumption required for performance monitoring, or the computational processing amount / power consumption for other purposes. Furthermore, the total of multiple processes (e.g., the computational processing amount / power consumption required for inference processing and the computational processing amount / power consumption required for performance monitoring) may be reported, or these may be reported separately. Furthermore, the timing of reporting the computational processing amount / power consumption for each process may be the same or different.

[0096] [Performance monitoring and examples of operations after performance monitoring] For example, performance monitoring of AI / ML models includes performance monitoring that checks the behavior of new, unverified models or parameters (e.g., referred to as Category 1 monitoring), and performance monitoring that checks whether AI / ML models or parameters currently in use are suitable for the current environment (e.g., referred to as Category 2 monitoring).

[0097] For example, Category 1 monitoring may be applied to performance monitoring of a group of AI / ML models preconfigured in the terminal 200. Also, Category 2 monitoring may be applied to performance monitoring of an AI / ML model that is currently activated. Also, different embodiments and modifications may be applied among the embodiments and modifications according to the performance monitoring category.

[0098] In Category 1 monitoring, the actions taken after monitoring the performance of the AI / ML model(s) may include not changing (or updating) the AI / ML model(s), updating the parameters of the AI / ML model(s), updating the AI / ML model(s) themselves, or disabling or terminating the settings of preconfigured AI / ML model(s).

[0099] In Category 2 monitoring, the actions taken after monitoring the performance of an AI / ML model may include: not changing (or updating) the currently activated AI / ML model; updating the parameters of the currently activated AI / ML model; changing the currently activated AI / ML model to another AI / ML model from a group of preconfigured AI / ML models; or deactivating the currently activated AI / ML model and falling back to a CSI reporting method that does not use AI / ML.

[0100] In functionality-based Life Cycle Management (LCM), which operation to perform after performance monitoring may be determined by terminal 200 depending on the implementation. On the other hand, in Model ID-based LCM, which operation to perform after performance monitoring may be determined by the base station / network.

[0101] [Learning method for AI / ML model] In CSI compression using a two-sided model in which an encoder for compression is used on the terminal side and a decoder for reconstruction is used on the base station / network side, any of the following methods may be applied as a learning method for the encoding unit (CSI compression unit) on the terminal side and the decoding unit (CSI reconstruction unit) on the base station / network side.

[0102] <Learning Method 1> Learning method 1 is joint learning of two-sided models on one side (terminal side or network side). In learning method 1, for example, both the CSI compressor and the CSI reconstructor are jointly trained on the network side. Information about the trained CSI compressor is delivered / transferred to the terminal. Alternatively, both the CSI compressor and the CSI reconstructor may be jointly trained on the terminal side.

[0103] <Learning Method 2> Learning Method 2 is joint learning of two-sided models on the network side and the terminal side. In Learning Method 2, for example, information such as forward propagation (FP) or backward propagation (BP) is exchanged between the terminal and the network, whereby both the terminal side and the network side jointly learn the CSI compression unit and the CSI reconstruction unit, respectively. Learning Method 2 also includes a method in which, after jointly learning two-sided models on one side (terminal side or network side) using Learning Method 1, the model on one side is fixed and information such as FP or BP is exchanged between the terminal and the network to learn the model on the other side.

[0104] <Learning Method 3> Learning method 3 is a method in which the terminal-side CSI compressor and the network-side CSI reconstructor are trained separately on the terminal side and the network side, respectively. In learning method 3, for example, the network side jointly trains both the CSI compressor and the CSI reconstructor, and after the network-side training is completed, the network side shares the data set used for training (e.g., input data and its output data used for training) with terminal 200. Terminal 200 trains the terminal-side CSI compressor using the data set shared from the network side.

[0105] For example, network-side performance monitoring based on target CSI reported by terminal 200 is useful when the entire model training is performed on the network side and for Category 2 monitoring with Model ID-based LCM, whereas terminal-side performance monitoring based on CSI reconstruction output notified by the network is useful when terminal 200 trains the network-side CSI reconstruction unit.

[0106] [Use Cases] The use cases to which the above-described embodiments or modifications are applied are not limited to CSI compression (CSI feedback enhancement), and can be applied to any use case in which an AI / ML model is deployed on the terminal side. For example, examples of use cases in which an AI / ML model is deployed on the terminal side include CSI prediction by a terminal-side AI / ML model, beam prediction in the spatial or time domain by a terminal-side AI / ML model, positioning accuracy enhancement by a terminal-side AI / ML model, and optimization of parameters and configurations of a wireless interface by AI / ML.

[0107] Furthermore, if the processor or AI / ML accelerator is general-purpose for AI / ML processing, AI / ML processing for different use cases may be executed simultaneously, or multiple AI / ML processing for different use cases may be selected.

[0108] [AI / ML Model] In non-limiting examples of the present disclosure, an "AI / ML model (or model)" may include a physical model, a binary model, an executable model, a converted model, a source code model, a non-executable model, a raw model, a logical model, or other model formats.

[0109] [Notification Regarding AI / ML Processing Capability of Terminal 200] Information indicating whether terminal 200 supports the functions, operations, or processes described in each of the above-described embodiments or modifications may be transmitted (or notified) from terminal 200 to base station 100, for example, as capability information or capability parameters of the terminal.

[0110] As described above, in CSI compression using a two-sided model in which an encoder for compression is used on the terminal side and a decoder for reconstruction is used on the base station / network side, it is desirable to monitor the performance of the AI / ML model to ensure the availability of the AI / ML model in the CSI compression unit on the terminal side and the AI / ML model in the CSI reconstruction unit on the base station / network side. For example, a case in which the availability of an AI / ML model deteriorates occurs when an AI / ML model trained in a certain environment (e.g., a specific cell, hereinafter, cell #A) is used in the same environment (e.g., cell #A) as the training environment, but the environment changes (e.g., cell #B) due to the mobility of terminal 200, etc., causing the AI / ML model currently being used to no longer match the training environment (e.g., cell #A), resulting in a deterioration in performance of the AI / ML model. In this case, terminal 200 is expected to take appropriate action, such as switching to an AI / ML model suitable for the changed environment (e.g., cell #B) (e.g., an AI / ML model trained in the cell #B environment) or stopping the CSI compression process using AI / ML.

[0111] Thus, in order to operate AI / ML models appropriately, it is expected that the performance of AI / ML models currently in operation, or AI / ML models that are not currently in operation but may be replaced, will need to be monitored.

[0112] In general, a high-performance AI / ML model has a large model size, and high computational capabilities are expected when the high-performance AI / ML model is implemented in the terminal 200. Therefore, the terminal 200 may have capabilities or limitations regarding the operations and processes using AI / ML that the terminal 200 can support (or simultaneously execute). In addition, it is desirable for the network side to grasp the AI / ML processing capabilities of the terminal 200.

[0113] On the other hand, the more AI / ML models there are, or the more processing processes associated with each AI / ML model (e.g., CSI reports based on inference processing (e.g., compressed CSI reports), performance monitoring, etc.), the larger and more complex the amount of information reporting the AI / ML processing capabilities of terminal 200 can become.

[0114] In the following, as an example, a case will be described in which terminal 200 is capable of implementing CSI reporting / monitoring according to "Model A," CSI reporting / monitoring according to "Model B," and existing CSI reporting / monitoring that does not apply AI / ML.

[0115] In this case, in addition to capability information indicating that each model and existing CSI reporting / monitoring can be performed, terminal 200 may also obtain information on whether performance monitoring of a CSI report according to model B can be performed simultaneously when performing a CSI report according to model A, whether performance monitoring of an existing CSI report to which AI / ML is not applied can be performed simultaneously when performing a CSI report according to model A, whether performance monitoring of a CSI report according to model A can be performed simultaneously when performing a CSI report according to model B, whether performance monitoring of an existing CSI report to which AI / ML is not applied can be performed simultaneously when performing a CSI report according to model B, whether performance monitoring of an existing CSI report to which AI / ML is not applied can be performed simultaneously when performing a CSI report according to model A, whether performance monitoring of a CSI report by Model B can be performed simultaneously when executing an existing CSI report that does not apply AI / ML; whether performance monitoring of a CSI report by Model B and performance monitoring of an existing CSI report that does not apply AI / ML can be performed simultaneously when executing a CSI report by Model A; whether performance monitoring of a CSI report by Model A and performance monitoring of an existing CSI report that does not apply AI / ML can be performed simultaneously when executing a CSI report by Model B; whether performance monitoring of a CSI report by Model A and performance monitoring of an existing CSI report that does not apply AI / ML can be performed simultaneously when executing a CSI report by Model B; and whether performance monitoring of a CSI report by Model A and performance monitoring of a CSI report by Model B can be performed simultaneously when executing an existing CSI report that does not apply AI / ML.

[0116] For example, in order to simplify notification regarding the AI / ML processing capabilities of terminal 200, notification of the AI / ML processing capabilities using an "AI / ML process set" and a "process set combination" may be applied.

[0117] For example, a processing set related to AI / ML processing may be defined for each AI / ML model. For example, in an AI / ML processing set, as shown in FIG. 4, each entry may correspond to a type of AI / ML processing, such as a CSI report (e.g., a CSI report based on inference processing (represented as "Inference") or an existing CSI report (represented as "Reporting")) and performance monitoring (e.g., represented as "Monitoring"). Furthermore, for example, notification of a processing set combination may be realized by notifying the processing set combination. For example, as shown in FIG. 5, when notifying the combination (A+B) of (processing set A, processing set B), it indicates that a CSI report based on inference processing by model A, which is the first entry of processing set A, and performance monitoring of a CSI report by model B, which is the second entry of processing set B, ("Model B monitoring"), can be executed simultaneously. Note that notification of a combination of three or more processing sets can be realized by adding more entries to each processing set.

[0118] Furthermore, as shown in FIG. 6 , for example, processing sets related to AI / ML processing may be defined for each AI / ML model, and each processing set may not have an entry but may include AI / ML processing executable by each AI / ML model. Furthermore, for example, processing set combinations may be defined by the model mapped to each entry. For example, the AI / ML processing (e.g., CSI reporting or performance monitoring) indicated may differ depending on the position of an entry in a processing set combination. For example, as shown in FIG. 7 , the first entry in a processing set combination may be defined as a CSI report, and the second entry may be defined as performance monitoring. In this case, when a combination (A+B) of (processing set A, processing set B) is notified, it indicates that a CSI report based on inference processing by model A and performance monitoring of the CSI report by model B can be performed simultaneously. Note that notification of three or more combinations can be achieved by increasing the number of entries in the processing set combination.

[0119] Furthermore, the AI / ML processing notified by an AI / ML processing set and a combination of AI / ML processing sets is not limited to a CSI report based on inference processing, an existing CSI report that does not apply AI / ML, and performance monitoring, and may also include processing necessary for other AI / ML operations (e.g., data collection). Furthermore, an AI / ML processing set and a combination of AI / ML processing sets may include capabilities related to multiple performance monitoring (e.g., multiple entries). For example, capabilities related to performance monitoring using a CSI reconstruction unit on the terminal side and performance monitoring without using a CSI reconstruction unit on the terminal side may be indicated separately.

[0120] Furthermore, the AI / ML processing capabilities of the terminal 200 are not limited to notifications simply indicating whether a certain capability is supported, but may also include, for example, capability priority information. Furthermore, when priorities are assigned to the AI / ML processing capabilities, each AI / ML process may be executed based on the set priority.

[0121] [Example of Operation of Base Station 100 and Terminal 200] FIG. 8 is a flowchart showing an example of operation of the base station 100 (base station / network) and the terminal 200. In FIG.

[0122] In FIG. 8, the terminal 200 calculates information relating to the amount of calculation processing / power consumption in the terminal 200 (S101).

[0123] The terminal 200 determines whether or not a condition (reporting condition) for reporting information on the calculated computational load / power consumption to the base station / network is satisfied (S102). If the reporting condition is not satisfied (S102: No), the terminal 200 may end the process shown in FIG. 8.

[0124] If the reporting conditions are met (S102: Yes), the terminal 200 reports information regarding the amount of computational processing / power consumption to the base station / network, for example, according to the method described in the above-mentioned embodiment or variant (S103).

[0125] The base station / network determines whether to use the terminal-side AI / ML model based on the information on the amount of computational processing / power consumption reported from the terminal 200 (S104).Then, the base station / network notifies the terminal 200 of the information on the determination of whether to use the terminal-side AI / ML model (S105).

[0126] [Regarding Architecture] Fig. 9 shows an example of an architecture including a function for processing AI / ML. Note that part of the network architecture may have the configuration shown in Fig. 9. Furthermore, the function for processing AI / ML may be included in an Access and Mobility Management Function (AMF) or a RAN (gNB).

[0127] [Configuration of Base Station] Fig. 10 is a block diagram showing an example configuration of base station 100. In Fig. 10, base station 100 has 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.

[0128] At least one of the control unit 101, the signal generation unit 102, the extraction unit 105, the demodulation unit 106, and the decoding unit 107 shown in Fig. 10 may be included in the control unit shown in Fig. 2. Also, the receiving unit 104 shown in Fig. 10 may be included in the receiving unit shown in Fig. 2.

[0129] The control unit 101, for example, determines control information related to the AI / ML-based operation of the terminal 200 and outputs the determined information to the signal generation unit 102. The control information related to the AI / ML-based operation of the terminal 200 may include, for example, at least one of the following: information related to the reporting of information related to the amount of computational processing / power consumption described above; information related to the terminal's decision on the use of an AI / ML model; and information related to performance monitoring of the AI / ML model. Furthermore, when the control unit 101 receives control information related to the AI / ML model of the terminal 200 from a function that processes AI / ML (for example, when input from the decoding unit 107), the control unit 101 may output the control information related to the AI / ML model to the signal generation unit 102.

[0130] Furthermore, the control unit 101 may output, for example, information regarding the computational processing amount / power consumption of the terminal 200 or a performance monitoring report from the decoding unit 107 to the function that processes AI / ML.

[0131] Furthermore, the control unit 101 determines, for example, information for the terminal 200 to receive a downlink signal, and outputs the determined information to the signal generation unit 102. The information for the terminal 200 to receive a downlink signal may include, for example, information on 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 on a coding and modulation scheme (e.g., MCS: Modulation and Coding Scheme).

[0132] Furthermore, the control unit 101 determines, for example, information used by the terminal 200 to transmit an uplink signal (e.g., a CSI report), 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 on resource allocation of an uplink data channel (e.g., a Physical Uplink Shared Channel (PUSCH)) or an uplink control channel (e.g., a Physical Uplink Control Channel (PUCCH)), information on a coding / modulation scheme (e.g., MCS), and information on a CSI report.

[0133] The signal generation unit 102 generates a data signal or a control signal bit sequence using, for example, information input from the control unit 101, and applies encoding as necessary. The signal generation unit 102 also modulates the encoded bit sequence to generate a modulated signal (for example, a symbol sequence), and maps the modulated signal to the radio resource specified by the control unit 101. The signal generation unit 102 outputs the mapped signal to the transmission unit 103.

[0134] The transmitting unit 103 performs, for example, OFDM or other transmission waveform generation processing on the signal input from the signal generating unit 102. Furthermore, in the case of OFDM transmission using a cyclic prefix (CP), for example, the transmitting unit 103 performs an inverse fast Fourier transform (IFFT) processing on the signal and adds a CP to the signal after the IFFT. Furthermore, the transmitting unit 103 performs, for example, RF processing such as D / A conversion or up-conversion on the signal and transmits the radio signal to the terminal 200 via an antenna.

[0135] The receiving unit 104 performs RF processing such as downconvert or A / D conversion on an uplink signal received from the terminal 200 via an antenna. In addition, in the case of OFDM transmission, the receiving unit 104 performs Fast Fourier Transform (FFT) processing on the received signal, and outputs the resulting frequency domain signal to the extracting unit 105.

[0136] The extraction unit 105 extracts, for example, based on information input from the control unit 101, a radio resource portion from which an uplink signal (for example, a PUSCH or a PUCCH) is transmitted, from the received signal input from the receiving unit 104, and outputs the extracted radio resource portion to the demodulation unit 106.

[0137] The demodulation unit 106 demodulates the uplink signal (for example, PUSCH or PUCCH) input from the extraction unit 105, for example, based on information input from the control unit 101. The demodulation unit 106 outputs the demodulation result to the decoding unit 107, for example.

[0138] The decoding unit 107 performs error correction decoding on the uplink signal (for example, PUSCH or PUCCH) based on, for example, information input from the control unit 101 and the demodulation result input from the demodulation unit 106, and obtains a decoded received bit sequence. For example, if the decoded received bit sequence includes information from the terminal 200 regarding the amount of calculation processing / power consumption, the decoding unit 107 outputs the information to the control unit 101.

[0139] 11 is a block diagram showing an example configuration of a terminal 200 according to an embodiment of the present disclosure. For example, in FIG. 11 , the terminal 200 includes a receiving unit 201, an extracting unit 202, a demodulating unit 203, a decoding unit 204, a control unit 205, a signal generating unit 206, and a transmitting unit 207.

[0140] At least one of the extraction unit 202, demodulation unit 203, decoding unit 204, control unit 205, and signal generation unit 206 shown in Fig. 11 may be included in the control unit shown in Fig. 3. Furthermore, the transmission unit 207 shown in Fig. 11 may be included in the transmission unit shown in Fig. 3.

[0141] The receiving unit 201 receives, for example, a downlink signal (e.g., a downlink data signal or a downlink control signal) from the base station 100 via an antenna, and performs RF processing such as downconvert or A / D conversion on the radio received signal to obtain a received signal (baseband signal). Furthermore, when receiving an OFDM signal, the receiving unit 201 performs FFT processing on the received signal to convert it into the frequency domain. The receiving unit 201 outputs the received signal to the extracting unit 202.

[0142] For example, based on information relating to the radio resource of the downlink control signal input from the control unit 205, the extraction unit 202 extracts a radio resource portion that may include a downlink control signal from the received signal input from the receiving unit 201, and outputs the extracted radio resource portion to the demodulation unit 203. Furthermore, based on information relating to the radio resource of the data signal input from the control unit 205, the extraction unit 202 extracts a radio resource portion that includes a downlink data signal, and outputs the extracted radio resource portion to the demodulation unit 203.

[0143] 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 .

[0144] The decoding unit 204 performs error correction decoding of the PDCCH or PDSCH using, for example, information input from the control unit 205 and 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.

[0145] The control unit 205 identifies information related to downlink transmission based on, for example, information obtained from the control signal input from the decoding unit 204, and outputs the information to the extraction unit 202, the demodulation unit 203, and the decoding unit 204. The control unit 205 also identifies information related to uplink transmission based on, for example, information obtained from the control signal input from the decoding unit 204, and outputs the information to the signal generation unit 206. The control unit 205 also generates information related to the amount of calculation processing / power consumption using the method described above, and outputs the information to the signal generation unit 206.

[0146] The signal generation unit 206 generates an uplink data signal or an uplink control signal based on the information on the computational processing amount / power consumption or the information on uplink transmission input from the control unit 205, encodes and modulates the bit string of the generated signal, and maps it to radio resources. The signal generation unit 206 outputs the uplink signal onto which the signal has been mapped to the transmission unit 207, for example.

[0147] The transmitter 207 generates a transmission signal waveform, such as OFDM, for the signal input from the signal generator 206. Furthermore, in the case of OFDM transmission or DFT-s-OFDM transmission using a CP, for example, the transmitter 207 performs IFFT processing on the signal and adds a CP to the signal after IFFT. Alternatively, when the transmitter 207 generates a single-carrier waveform such as a DFT-s-OFDM waveform, a DFT unit (not shown) may be added before the signal generator 206. Furthermore, the transmitter 207 performs RF processing, such as D / A conversion and up-conversion, on the transmission signal, and transmits the radio signal to the base station 100 via an antenna.

[0148] The above describes the embodiments according to non-limiting examples of the present disclosure.

[0149] In addition, different AI / ML models may be used depending on parameters such as the cell, site, Transmission and Reception Point (TRP), beam, location, terminal movement speed, wireless channel such as multipath, cell congestion status, and transmitted traffic.

[0150] Furthermore, "Mobility" used in the above embodiments may include handover in RRC CONNETED mode, reselection in RRC IDLE or RRC INACTIVE mode, and movement between TRPs, beam switching, and location movement.

[0151] Furthermore, in the present disclosure, the signal / message / signaling used for notification may be a control plane message (e.g., UCI or MAC-CE), an RRC signal, or a notification in DCI, which is physical layer signaling.

[0152] (Supplementary Note) Information indicating whether the terminal 200 supports the functions, operations, or processes described in each of the above-described embodiments and each supplementary note may be transmitted (or notified) from the terminal 200 to the base station 100, for example, as capability information or capability parameters of the terminal 200.

[0153] The capability information may include an information element (IE) that individually indicates whether or not the terminal 200 supports at least one of the functions, operations, or processes described in the above-described embodiments, modifications, and supplements. Alternatively, the capability information may include an information element that indicates whether or not the terminal 200 supports a combination of any two or more of the functions, operations, or processes described in the above-described embodiments, modifications, and supplements.

[0154] For example, the base station 100 may determine (or decide or assume) the functions, operations, or processes that the terminal 200 that transmitted the capability information supports (or does not support) based on the capability information received from the terminal 200. The base station 100 may perform operations, processes, or control according to the determination result based on the capability information. For example, the base station 100 may control processing related to an AI / ML model based on the capability information received from the terminal 200.

[0155] Note that the fact that terminal 200 does not support some of the functions, operations, or processes described in the above-described embodiments, modifications, and supplementary notes may be interpreted as meaning that such some of the functions, operations, or processes are restricted in terminal 200. For example, information or a request regarding such restrictions may be notified to base station 100.

[0156] Information regarding the capabilities or limitations of terminal 200 may, for example, be defined in a standard, or may be implicitly notified to base station 100 in association with information known at base station 100 or information transmitted to base station 100.

[0157] The above has described the embodiments, modifications, and supplementary notes according to a non-limiting example of the present disclosure.

[0158] (Control Signal) In the present disclosure, a downlink control signal (information) related to the present disclosure may be a signal (information) transmitted by a PDCCH of a physical layer, or may be a signal (information) transmitted by a MAC CE (Control Element) or RRC of a higher layer. Also, the downlink control signal may be a signal (information) that is specified in advance.

[0159] The uplink control signal (information) related to the present disclosure may be a signal (information) transmitted on a PUCCH in the physical layer, or a signal (information) transmitted on a MAC CE or RRC in a higher layer. The uplink control signal may also be a predefined signal (information). The uplink control signal may also be replaced with uplink control information (UCI), 1st stage sidelink control information (SCI), or 2nd stage SCI.

[0160] (Base Station) In the present disclosure, a base station may be a TRP (Transmission Reception Point), a cluster head, an access point, an RRH (Remote Radio Head), an eNodeB (eNB), a gNodeB (gNB), a BS (Base Station), a BTS (Base Transceiver Station), a parent device, a gateway, or the like. In sidelink communication, the base station may be replaced with a terminal. The base station may be a relay device that relays communication between an upper node and a terminal. The base station may be a roadside unit.

[0161] (Uplink / Downlink / Sidelink) The present disclosure may be applied to any of the uplink, downlink, and sidelink. For example, the present disclosure may be applied to the uplink PUSCH, PUCCH, and PRACH, the downlink PDSCH, PDCCH, and PBCH, and the sidelink PSSCH (Physical Sidelink Shared Channel), PSCCH (Physical Sidelink Control Channel), and PSBCH (Physical Sidelink Broadcast Channel).

[0162] The PDCCH, PDSCH, PUSCH, and PUCCH are examples of a downlink control channel, a downlink data channel, an uplink data channel, and an uplink control channel. The PSCCH and PSSCH are examples of a sidelink control channel and a sidelink data channel. The PBCH and PSBCH are examples of broadcast channels, and the PRACH is an example of a random access channel.

[0163] (Data Channel / Control Channel) The present disclosure may be applied to either a data channel or a control channel. For example, the channels of the present disclosure may be replaced with data channels such as PDSCH, PUSCH, and PSSCH, and control channels such as PDCCH, PUCCH, PBCH, PSCCH, and PSBCH.

[0164] (Reference Signal) In the present disclosure, a reference signal is a signal known by both a base station and a terminal, and may also be referred to as an RS (Reference Signal) or a pilot signal. The reference signal may be any of DMRS, CSI-RS (Channel State Information - Reference Signal), TRS (Tracking Reference Signal), PTRS (Phase Tracking Reference Signal), CRS (Cell-specific Reference Signal), and SRS (Sounding Reference Signal).

[0165] (Time Interval) In the present disclosure, the unit of time resource is not limited to one or a combination of slots and symbols, but may be, for example, a time resource unit such as a frame, a superframe, a subframe, a slot, a time slot, a subslot, a minislot, a symbol, an OFDM (Orthogonal Frequency Division Multiplexing Access) symbol, or an SC-FDMA (Single Carrier-Frequency Division Multiple Access) symbol, or another time resource unit. Furthermore, the number of symbols included in one slot is not limited to the number of symbols exemplified in the above-mentioned embodiment, and may be another number of symbols.

[0166] (Frequency Band) The present disclosure may be applied to either a licensed band or an unlicensed band.

[0167] (Communication) The present disclosure may be applied to communication between a base station and a terminal (Uu link communication), communication between terminals (Sidelink communication), and V2X (Vehicle to Everything) communication. For example, the channels of the present disclosure may be replaced with PSCCH, PSSCH, PSFCH (Physical Sidelink Feedback Channel), PSBCH, PDCCH, PUCCH, PDSCH, PUSCH, or PBCH.

[0168] The present disclosure may be applied to a terrestrial network, a non-terrestrial network (NTN) using satellites or highly advanced pseudo satellites (HAPS), or a terrestrial network in which the transmission delay is large compared to the symbol length or slot length, such as a network with a large cell size or an ultra-wideband transmission network.

[0169] (Antenna Port) An antenna port refers to a logical antenna (antenna group) consisting of one or more physical antennas. In other words, an antenna port does not necessarily refer to a single physical antenna, but may refer to an array antenna consisting of multiple antennas, etc. For example, the number of physical antennas an antenna port is composed of is not specified, and it is specified as the smallest unit by which a terminal can transmit a reference signal. In addition, an antenna port may also be specified as the smallest unit by which a weighting of a precoding vector is multiplied.

[0170] 5G NR System Architecture and Protocol Stack 3GPP continues work on the next release of fifth-generation cellular technology (also referred to simply as "5G"), which includes 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 at the end of 2017, allowing for the prototyping and commercial deployment of 5G NR-compliant devices (e.g., smartphones).

[0171] For example, the system architecture generally assumes a Next Generation - Radio Access Network (NG-RAN) comprising gNBs. The gNBs provide UE-side termination of the NG radio access user plane (SDAP / PDCP / RLC / MAC / PHY) and control plane (RRC) protocols. The gNBs are connected to each other via an Xn interface. The gNBs are also connected to a Next Generation Core (NGC) via a Next Generation (NG) interface, more specifically to an Access and Mobility Management Function (AMF) (e.g., a specific core entity performing AMF) via an NG-C interface, and to a User Plane Function (UPF) (e.g., a specific core entity performing UPF) via an NG-U interface. The NG-RAN architecture is shown in Figure 12 (see, for example, 3GPP TS 38.300 v15.6.0, section 4).

[0172] The NR user plane protocol stack (see, for example, 3GPP TS 38.300, section 4.4.1) includes a Packet Data Convergence Protocol (PDCP) sublayer (see Section 6.4 of TS 38.300), a Radio Link Control (RLC) sublayer (see Section 6.3 of TS 38.300), and a Medium Access Control (MAC) sublayer (see Section 6.2 of TS 38.300), which are terminated on the network side in the gNB. A new Access Stratum (AS) sublayer (Service Data Adaptation Protocol (SDAP)) has also been introduced above PDCP (see, for example, Section 6.5 of 3GPP TS 38.300). A control plane protocol stack has also been defined for the NR (see, for example, Section 4.4.2 of TS 38.300). An overview of Layer 2 functions is provided in Section 6 of TS 38.300. The functions of the PDCP sublayer, RLC sublayer, and MAC sublayer are listed in clauses 6.4, 6.3, and 6.2 of TS 38.300, respectively. The functions of the RRC layer are listed in clause 7 of TS 38.300.

[0173] For example, the Medium-Access-Control layer handles logical channel multiplexing and scheduling and scheduling-related functions, including handling various numerologies.

[0174] For example, the physical layer (PHY) is responsible for coding, PHY HARQ processing, modulation, multi-antenna processing, and mapping of signals to appropriate physical time-frequency resources. The physical layer also handles 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 for transmitting a specific transport channel, and each transport channel is mapped to a corresponding physical channel. For example, physical channels include the Physical Random Access Channel (PRACH), the Physical Uplink Shared Channel (PUSCH), and the Physical Uplink Control Channel (PUCCH) as uplink physical channels, and the Physical Downlink Shared Channel (PDSCH), the Physical Downlink Control Channel (PDCCH), and the Physical Broadcast Channel (PBCH) as downlink physical channels.

[0175] NR use cases / deployment scenarios may include enhanced mobile broadband (eMBB), ultra-reliable low-latency communications (URLLC), and massive machine-type communication (mMTC), which have diverse requirements in terms of data rate, latency, and coverage. For example, eMBB is expected to support peak data rates (20 Gbps in the downlink and 10 Gbps in the uplink) and effective (user-experienced) data rates approximately three times higher than those offered by IMT-Advanced. On the other hand, URLLC imposes stricter requirements on ultra-low latency (0.5 ms for user plane latency in both UL and DL) and high reliability (1-10-5 within 1 ms). Finally, mMTC preferably requires high connection density (1,000,000 devices / km in urban environments).２ ), wide coverage in adverse environments, and extremely long battery life (15 years) for a low-cost device may be desired.

[0176] Therefore, OFDM numerology (e.g., subcarrier spacing, OFDM symbol length, cyclic prefix (CP) length, number of symbols per scheduling interval) suitable for one use case may not be valid for another use case. For example, low-latency services may preferably require a shorter symbol length (and therefore a larger subcarrier spacing) and / or fewer symbols per scheduling interval (also referred to as TTI) than mMTC services. Furthermore, deployment scenarios with large channel delay spreads may preferably require a longer CP length than scenarios with short delay spreads. The subcarrier spacing may be optimized accordingly to maintain similar CP overhead. NR may support one or more subcarrier spacing values. Correspondingly, subcarrier spacings of 15 kHz, 30 kHz, 60 kHz, etc. are currently considered. The symbol length Tu and subcarrier spacing Δf are directly related by the formula Δf = 1 / Tu. Similar to LTE systems, the term "resource element" can be used to mean the smallest resource unit consisting of one subcarrier for the length of one OFDM / SC-FDMA symbol.

[0177] In the new radio system 5G-NR, for each numerology and each carrier, a resource grid of subcarriers and OFDM symbols is defined for each uplink and downlink. Each element of the resource grid is called a resource element and is specified based on a frequency index in the frequency domain and a symbol position in the time domain (see 3GPP TS 38.211 v15.6.0).

[0178] <Functional Separation Between NG-RAN and 5GC in 5G NR> Figure 13 shows the functional separation between NG-RAN and 5GC. The logical node of NG-RAN is gNB or ng-eNB. 5GC has logical nodes AMF, UPF, and SMF.

[0179] For example, gNB and ng-eNB host the following main functions: - Radio Resource Management functions such as Radio Bearer Control, Radio Admission Control, Connection Mobility Control, dynamic allocation (scheduling) of resources to UEs in both uplink and downlink; - IP header compression, ciphering and integrity protection of data; - AMF selection at UE attach time if routing to the AMF cannot be determined from the information provided by the UE; - Routing of user plane data towards the UPF; - Routing of control plane information towards the AMF; - Connection setup and release; - Scheduling and transmission of paging messages; - Scheduling and transmission of system broadcast information (sourced from the AMF or Operation, Admission, Maintenance (OAM)); - Configuration of measurements and measurement reports for mobility and scheduling; - Transport level packet marking in the uplink; - Session management; Support for network slicing; - QoS flow management and mapping to data radio bearers; - Support for UEs in RRC_INACTIVE state; - NAS message delivery function; - Radio access network sharing; - Dual connectivity; - Close coordination between NR and E-UTRA.

[0180] The Access and Mobility Management Function (AMF) hosts the following main functions: - Termination of Non-Access Stratum (NAS) signaling; - Security of NAS signaling; - Security control of Access Stratum (AS); - Signaling between Core Network (CN) nodes for mobility between 3GPP access networks; - Reachability to idle mode UEs (including control and execution of paging retransmissions); - Registration area management; - Support for intra-system and inter-system mobility; - Access authentication; - Access authorization including checking of roaming rights; - Mobility management control (subscription and policy); - Support for network slicing; - Selection of Session Management Function (SMF).

[0181] Furthermore, the User Plane Function (UPF) hosts the following main functions: - anchor point for intra-RAT / inter-RAT mobility (if applicable); - external PDU (Protocol Data Unit) session point for interconnection with data networks; - packet routing and forwarding; - packet inspection and policy rule enforcement for the user plane part; - traffic usage reporting; - uplink classifier to support routing of traffic flows to the data network; - branching point to support multi-homed PDU sessions; - QoS processing for the user plane (e.g. packet filtering, gating, UL / DL rate enforcement); - uplink traffic validation (mapping of SDF to QoS flows); - downlink packet buffering and triggering of downlink data notifications.

[0182] Finally, the Session Management Function (SMF) hosts the following main functions: session management; allocation and management of IP addresses for UEs; selection and control of UPF; configuration of traffic steering in the User Plane Function (UPF) to route traffic to the appropriate destination; policy enforcement and QoS of the control part; downlink data notification.

[0183] <RRC connection setup and reconfiguration procedure> Figure 14 shows some of the interactions between the UE, gNB, and AMF (5GC entities) when the UE transitions from RRC_IDLE to RRC_CONNECTED in the NAS part (see TS 38.300 v15.6.0).

[0184] RRC is a higher layer signaling (protocol) used to configure the UE and gNB. With 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 an INITIAL CONTEXT SETUP REQUEST. The gNB then activates AS security together with the UE. This is done by the gNB sending a SecurityModeCommand message to the UE, and the UE responding with a SecurityModeComplete message to the gNB. The gNB then sends an RRCReconfiguration message to the UE, and upon receiving an RRCReconfigurationComplete from the UE, the gNB performs reconfiguration to set up Signaling Radio Bearer 2 (SRB2) and Data Radio Bearer (DRB). For signaling-only connections, the steps related to RRCReconfiguration are omitted because SRB2 and DRB are not set up. Finally, the gNB notifies the AMF that the setup procedure is complete with an INITIAL CONTEXT SETUP RESPONSE.

[0185] Therefore, the present disclosure provides a 5th Generation Core (5GC) entity (e.g., AMF, SMF, etc.) that includes: a control circuit that, upon operation, establishes a Next Generation (NG) connection with a gNodeB; and a transmitter that, upon operation, transmits an initial context setup message to the gNodeB via the NG connection so that a signaling radio bearer between the gNodeB and a user equipment (UE) is set up. Specifically, the gNodeB transmits Radio Resource Control (RRC) signaling, including a resource allocation configuration information element (IE), to the UE via the signaling radio bearer. The UE then transmits in uplink or receives in downlink based on the resource allocation configuration.

[0186] <IMT Usage Scenarios Beyond 2020> Figure 15 shows some use cases for 5G NR. The 3rd Generation Partnership Project New Radio (3GPP NR) is considering three use cases envisioned by IMT-2020 to support a wide variety of services and applications. The first phase of specifications for enhanced mobile broadband (eMBB) has been completed. Current and future work includes standardization for ultra-reliable and low-latency communications (URLLC) and massive machine-type communications (mMTC), in addition to expanding support for eMBB. Figure 15 shows some examples of envisioned usage scenarios for IMT beyond 2020 (see, for example, ITU-R M.2083 Figure 2).

[0187] The URLLC use case has stringent performance requirements, such as throughput, latency, and availability. It is envisioned as one of the enabling technologies for future applications, such as wireless control of industrial production or manufacturing processes, remote medical surgery, automated power transmission and distribution in smart grids, and road safety. URLLC's ultra-high reliability is supported by identifying technologies that meet the requirements set by TR 38.913. Key requirements for NR URLLC in Release 15 include a target user plane latency of 0.5 ms on the uplink (UL) and 0.5 ms on the downlink (DL). The overall URLLC requirement for a single packet transmission is a block error rate (BLER) of 1E-5 for a 32-byte packet size with a user plane latency of 1 ms.

[0188] From a physical layer perspective, reliability can be improved in many possible ways. Current room for reliability improvement includes defining a separate CQI table for URLLC, more compact DCI formats, PDCCH repetition, etc. However, this room can be expanded to achieve ultra-high reliability as NR (with respect to the key requirements of NR URLLC) becomes more stable and developed. Specific use cases for NR URLLC in Release 15 include Augmented Reality / Virtual Reality (AR / VR), e-health, e-safety, and mission-critical applications.

[0189] Additionally, the technology enhancements targeted by NR URLLC aim to improve latency and reliability. Technology enhancements for latency improvement include configurable numerology, non-slot-based scheduling with flexible mapping, grant-free (configured grant) uplink, slot-level repetition in the data channel, and preemption in the downlink. Preemption means that a transmission with already allocated resources is stopped and the already allocated resources are used for another transmission with a later requested lower latency / higher priority. Therefore, a previously allowed transmission is preempted by a later transmission. Preemption is applicable regardless of the specific service type. For example, a transmission of service type A (URLLC) may be preempted by a transmission of service type B (eMBB, etc.). Technology enhancements for reliability improvement include a dedicated CQI / MCS table for a target BLER of 1E-5.

[0190] The use case of mMTC (massive machine type communication) is characterized by a very large number of connected devices that typically transmit relatively small amounts of data that are not sensitive to delays. These devices are required to be low-cost and have very long battery life. From the NR perspective, utilizing very narrow bandwidth portions is one solution that saves power and extends battery life from the UE's perspective.

[0191] As mentioned above, the scope of reliability improvement in NR is expected to be broader. One of the key requirements for all cases, for example, URLLC and mMTC, is high or ultra-high reliability. Several mechanisms can improve reliability from a radio perspective and a network perspective. Generally, there are two to three key areas that can help improve reliability. These areas include compact control channel information, repetition of data channels / control channels, and diversity in the frequency, time, and / or spatial domains. These areas are generally applicable to reliability improvement regardless of the specific communication scenario.

[0192] For NR URLLC, further use cases with more stringent requirements are envisaged, such as factory automation, transportation, and power distribution, such as high reliability (up to 10-6 level), high availability, packet size up to 256 bytes, time synchronization up to a few μs (depending on the use case, the value can be 1 μs or a few μs depending on the frequency range and low latency in the order of 0.5 ms to 1 ms (e.g., 0.5 ms latency on the targeted user plane)).

[0193] Furthermore, for NR URLLC, there may be several technical enhancements from the physical layer perspective. These technical enhancements include PDCCH (Physical Downlink Control Channel) enhancements for compact DCI, PDCCH repetition, and increased PDCCH monitoring. Also, UCI (Uplink Control Information) enhancements relate to enhanced Hybrid Automatic Repeat Request (HARQ) and CSI feedback enhancements. There may also be PUSCH enhancements related to minislot-level hopping and retransmission / repetition enhancements. The term "minislot" refers to a Transmission Time Interval (TTI) that contains fewer symbols than a slot (a slot comprises 14 symbols).

[0194] <QoS Control> The 5G Quality of Service (QoS) model is based on QoS flows and supports both QoS flows that require a guaranteed flow bit rate (Guaranteed Bit Rate QoS flows (GBR)) and QoS flows that do not require a guaranteed flow bit rate (non-GBR QoS flows). Thus, at the NAS level, a QoS flow is the finest granularity of QoS classification in a PDU session. A QoS flow is identified within a PDU session by a QoS Flow ID (QFI) carried in an encapsulation header over the NG-U interface.

[0195] For each UE, 5GC establishes one or more PDU sessions. For each UE, the NG-RAN establishes at least one Data Radio Bearer (DRB) for each PDU session, e.g., as shown above with reference to Figure 14. Additional DRBs for the QoS flows of that PDU session can be configured later (when this is up to the NG-RAN). The NG-RAN maps packets belonging to different PDU sessions to different DRBs. NAS-level packet filters in the UE and 5GC associate UL and DL packets with QoS flows, while AS-level mapping rules in the UE and NG-RAN associate UL and DL QoS flows with DRBs.

[0196] Figure 16 shows the non-roaming reference architecture for 5G NR (see TS 23.501 v16.1.0, section 4.23). An Application Function (AF) (e.g., an external application server hosting 5G services, as illustrated in Figure 15) interacts with the 3GPP core network to provide services. For example, it may access a Network Exposure Function (NEF) to support applications that affect traffic routing, or interact with a policy framework for policy control (e.g., QoS control) (see Policy Control Function (PCF)). Based on operator deployment, Application Functions considered trusted by the operator can interact directly with the associated Network Functions. Application Functions not permitted by the operator to directly access Network Functions interact with the associated Network Functions using an external exposure framework via the NEF.

[0197] Figure 16 further illustrates further functional units of the 5G architecture, namely, Network Slice Selection Function (NSSF), Network Repository Function (NRF), Unified Data Management (UDM), Authentication Server Function (AUSF), Access and Mobility Management Function (AMF), Session Management Function (SMF), and Data Network (DN, e.g., operator-provided services, Internet access, or third-party services). All or part of the core network functions and application services may be deployed and run in a cloud computing environment.

[0198] Therefore, the present disclosure provides an application server (e.g., an AF in a 5G architecture) comprising: a transmitter that, in operation, sends a request including QoS requirements for at least one of a URLLC service, an eMMB service, and an mMTC service to at least one of 5GC functions (e.g., an NEF, an AMF, an SMF, a PCF, an UPF, etc.) to establish a PDU session including a radio bearer between a gNodeB and a UE according to the QoS requirements; and a control circuit that, in operation, performs a service using the established PDU session.

[0199] The term "portion" used in this disclosure may be interchangeably used with other terms such as "circuitry," "device," "unit," or "module."

[0200] The present disclosure can be realized by software, hardware, or software in conjunction with hardware. Each functional block used in the description of the above embodiments may be partially or entirely realized as an LSI, which is an integrated circuit, and each process described in the above embodiments may be partially or entirely controlled by a single LSI or a combination of LSIs. The LSI may be composed of individual chips, or may be composed of a single chip that includes some or all of the functional blocks. The LSI may have data input and output. Depending on the degree of integration, the LSI may also be called an IC, system LSI, super LSI, or ultra LSI.

[0201] The integrated circuit method is not limited to LSI, and may be realized by a dedicated circuit, a general-purpose processor, or a dedicated processor. Also, a field programmable gate array (FPGA) that can be programmed after LSI manufacturing, or a reconfigurable processor that can reconfigure the connections and settings of circuit cells within the LSI, may be used. The present disclosure may be realized as digital processing or analog processing.

[0202] Furthermore, if an integrated circuit technology that can replace LSI emerges due to advances in semiconductor technology or other derivative technologies, it is natural that such technology may be used to integrate functional blocks. The application of biotechnology, etc. is also a possibility.

[0203] The present disclosure can be implemented in any type of apparatus, device, or system (collectively referred to as a communication apparatus) having a communication function. The communication apparatus may include a radio transceiver and processing / control circuitry. The radio transceiver may include a receiver and a transmitter, or both as functions. The radio transceiver (transmitter and receiver) may include a radio frequency (RF) module and one or more antennas. The RF module may include an amplifier, an RF modulator / demodulator, or the like. Non-limiting examples of communication devices include telephones (e.g., cell phones, smartphones), tablets, personal computers (PCs) (e.g., laptops, desktops, notebooks), cameras (e.g., digital still / video cameras), digital players (e.g., digital audio / video players), wearable devices (e.g., wearable cameras, smartwatches, tracking devices), game consoles, digital book readers, telehealth / telemedicine devices, communication-enabled vehicles or mobile transportation (e.g., cars, airplanes, ships), and combinations of the above devices.

[0204] The communication devices are not limited to portable or mobile devices, but also include any kind of non-portable or fixed equipment, devices, and systems, such as smart home devices (such as home appliances, lighting equipment, smart meters or measuring devices, control panels, etc.), vending machines, and any other "things" that may exist on an IoT (Internet of Things) network.

[0205] Communications include data communications via cellular systems, wireless LAN systems, communication satellite systems, and the like, as well as data communications via combinations of these.

[0206] A communications apparatus also includes devices such as controllers and sensors connected or coupled to a communications device that performs the communications functions described in this disclosure, such as controllers and sensors that generate control and data signals used by the communications device to perform the communications functions of the communications apparatus.

[0207] The communication apparatus also includes infrastructure facilities, such as base stations, access points, and any other apparatus, device, or system that communicates with or controls the various apparatuses listed above, but are not limited to these.

[0208] A terminal according to one embodiment of the present disclosure comprises a control circuit that determines control information relating to at least one of computational processing volume and power consumption by using an artificial intelligence model on the terminal side, and a transmitting circuit that transmits the control information.

[0209] In one embodiment of the present disclosure, the control information includes information regarding a comparison between at least one of the computational processing amount and power consumption when the artificial intelligence model is used and at least one of the computational processing amount and power consumption when the artificial intelligence model is not used.

[0210] In one embodiment of the present disclosure, the control information includes information regarding the difference between at least one of the computational processing amount and power consumption when the artificial intelligence model is used and at least one of the computational processing amount and power consumption when the artificial intelligence model is not used.

[0211] In one embodiment of the present disclosure, the control information includes information regarding a comparison or difference between at least one of the computational processing amount and power consumption when using a first artificial intelligence model and at least one of the computational processing amount and power consumption when using a second artificial intelligence model.

[0212] In one embodiment of the present disclosure, the control information includes information regarding a comparison or difference value between at least one of the computational processing amount and power consumption when using a first parameter of the artificial intelligence model and at least one of the computational processing amount and power consumption when using a second parameter of the artificial intelligence model.

[0213] In one embodiment of the present disclosure, the control information includes information regarding the usage rate or occupancy rate of at least one of the computational processing amount and power consumption in the terminal.

[0214] In one embodiment of the present disclosure, knowledge regarding how to determine the control information is shared between the network and the terminal.

[0215] In one embodiment of the present disclosure, information regarding the method is pre-specified in a standard, is reported from the terminal to the network when reporting the terminal's capability or capability information regarding the application of the artificial intelligence model, is configured in the terminal from the network, or is reported from the terminal to the network when reporting the control information.

[0216] In one embodiment of the present disclosure, when the terminal performs a performance monitoring process of the artificial intelligence model, the control information is reported from the terminal to the network together with information about the performance monitoring process.

[0217] In one embodiment of the present disclosure, the control information is reported from the terminal to the network together with the terminal's capability regarding the artificial intelligence model.

[0218] In one embodiment of the present disclosure, the control information is reported from the terminal to the network when a trigger is notified from the network, when a timer expires, and when information regarding at least one of the computational processing amount and power consumption exceeds a threshold.

[0219] A base station according to one embodiment of the present disclosure comprises a receiving circuit that receives control information regarding at least one of the computational processing volume and power consumption due to the use of an artificial intelligence model on a terminal side, and a control circuit that controls the use of the artificial intelligence model on the terminal side based on the control information.

[0220] In a communication method according to an embodiment of the present disclosure, a terminal determines control information related to at least one of a computational load and power consumption by using an artificial intelligence model on the terminal side, and transmits the control information.

[0221] In a communication method according to one embodiment of the present disclosure, a base station receives control information regarding at least one of the computational processing volume and power consumption due to the use of an artificial intelligence model on the terminal side, and controls the use of the artificial intelligence model on the terminal side based on the control information.

[0222] The disclosures of the specification, drawings and abstract contained in Japanese Patent Application No. 2023-129364, filed on August 8, 2023, are incorporated herein by reference in their entirety.

[0223] One embodiment of the present disclosure is useful in wireless communication systems.

[0224] 100 Base station 101, 205 Control unit 102, 206 Signal generation unit 103, 207 Transmission unit 104, 201 Reception unit 105, 202 Extraction unit 106, 203 Demodulation unit 107, 204 Decoding unit 200 Terminal 251 CSI generation unit 252 CSI reconstruction unit 253 Performance monitoring unit 254 CQI generation unit

Claims

1. A terminal comprising: a control circuit for determining control information relating to at least one of a computational processing amount and a power consumption by using an artificial intelligence model on the terminal side; and a transmission circuit for transmitting the control information.

2. The terminal described in claim 1, wherein the control information includes information regarding a comparison between at least one of the computational processing amount and power consumption when the artificial intelligence model is used and at least one of the computational processing amount and power consumption when the artificial intelligence model is not used.

3. The terminal described in claim 1, wherein the control information includes information regarding a difference between at least one of the computational processing amount and power consumption when the artificial intelligence model is used and at least one of the computational processing amount and power consumption when the artificial intelligence model is not used.

4. The terminal described in claim 1, wherein the control information includes information regarding a comparison or difference between at least one of the computational processing amount and power consumption when using a first artificial intelligence model and at least one of the computational processing amount and power consumption when using a second artificial intelligence model.

5. The terminal described in claim 1, wherein the control information includes information regarding a comparison or difference value between at least one of the computational processing amount and power consumption when a first parameter of the artificial intelligence model is used and at least one of the computational processing amount and power consumption when a second parameter of the artificial intelligence model is used.

6. The terminal according to claim 1, wherein the control information includes information regarding a usage rate or occupancy rate of at least one of the amount of calculation processing and power consumption in the terminal.

7. The terminal of claim 1, wherein knowledge of how to determine the control information is shared between a network and the terminal.

8. The terminal described in claim 7, wherein information regarding the method is predefined in a standard, is reported from the terminal to the network when reporting the terminal's capability or capability information regarding the application of the artificial intelligence model, is set in the terminal from the network, or is reported from the terminal to the network when reporting the control information.

9. The terminal according to claim 1, wherein, when the terminal performs a performance monitoring process of the artificial intelligence model, the control information is reported from the terminal to a network together with information related to the performance monitoring process.

10. The terminal according to claim 1, wherein the control information is reported from the terminal to a network together with the capability of the terminal with respect to the artificial intelligence model.

11. The terminal according to claim 1, wherein the control information is reported from the terminal to the network when a trigger is notified from the network, when a timer expires, and when information relating to at least one of the computational processing amount and power consumption exceeds a threshold.

12. A base station comprising: a receiving circuit that receives control information regarding at least one of the computational processing volume and power consumption due to the use of an artificial intelligence model on a terminal side; and a control circuit that controls the use of the artificial intelligence model on the terminal side based on the control information.

13. A communication method, comprising: a terminal determining control information relating to at least one of a computational processing amount and a power consumption by using an artificial intelligence model on the terminal side; and transmitting the control information.

14. A communication method in which a base station receives control information regarding at least one of the amount of computational processing and power consumption due to the use of an artificial intelligence model on the terminal side, and controls the use of the artificial intelligence model on the terminal side based on the control information.

15. An integrated circuit that controls the processing of a terminal, the processing including: determining control information relating to at least one of the amount of computational processing and power consumption by using an artificial intelligence model on the terminal side; and transmitting the control information.

16. An integrated circuit that controls the processing of a base station, the processing including: receiving control information regarding at least one of the amount of computational processing and power consumption due to the use of an artificial intelligence model on a terminal side; and controlling the use of the artificial intelligence model on the terminal side based on the control information.