Apparatus and method for reporting CSI in wireless communication system

Abstract

The present disclosure relates to a 5G or 6G communication system for supporting higher data rates beyond 4G communication systems such as LTE. The present disclosure provides a method for AI-based CSI reporting in a 5G or 6G communication system. The method performed by a base station in a wireless communication system includes: receiving capability information indicating whether a UE supports generation of CSI based on AI; receiving, based on the capability information, information on a change in an inference time related to AI-based CSI generation and information on a change in a channel state; identifying information on CSI reporting timing based on information on a change in reasoning time and a change in channel state; transmitting RRC configuration information including CSI report timing; the CSI-RS is transmitted; and receiving a CSI report based on the CSI-RS and the information on the CSI report timing.

Description

Technical Field

[0001] The present disclosure relates generally to a wireless communication system, and more particularly to a method and apparatus for reporting channel state information (CSI) in a wireless communication network. Background Art

[0002] Considering the development of wireless communication from generation to generation, technologies have been developed mainly for services for humans, such as voice calls, multimedia services, and data services. After the commercialization of the 5th generation (5G) communication system, the number of connected devices is expected to grow exponentially. These will be increasingly connected to the communication network. Examples of networked things can include vehicles, robots, drones, household appliances, displays, smart sensors connected to various infrastructures, construction machinery, and factory equipment. Mobile devices are expected to evolve in various form factors, such as augmented reality glasses, virtual reality headsets, and hologram devices. In order to provide various services by connecting hundreds of billions of devices and things in the 6G era, efforts have been made to develop improved 6G communication systems. For these reasons, 6G communication systems are called super 5G systems.

[0003] The 6G communication system, which is expected to be commercialized around 2030, will have a peak data rate of terabit (1,000 gigabits) per second (bps) and a radio latency of less than 100 μsec, and will therefore be 50 times faster than the 5G communication system and have 1 / 10 of its radio latency.

[0004] In order to achieve such high data rates and ultra-low latency, 6G communication systems have been considered to be implemented in the terahertz (THz) band (e.g., 95 gigahertz (GHz) to 3 THz band). It is expected that since the path loss and atmospheric absorption in the terahertz band are more serious than in the millimeter wave band introduced in 5G, the technology to ensure the signal transmission distance (that is, coverage) will become more critical. As the main technology to ensure coverage, it is necessary to develop radio frequency (RF) elements, antennas, new waveforms with better coverage than orthogonal frequency division multiplexing (OFDM) schemes, beamforming and massive multiple input multiple output (MIMO), full-dimensional MIMO (FD-MIMO), array antennas, and multi-antenna transmission technologies such as massive antennas. In addition, new technologies for improving the coverage of terahertz band signals have been discussed, such as metamaterial-based lenses and antennas, orbital angular momentum (OAM), and reconfigurable smart surfaces (RIS).

[0005] In addition, in order to improve spectrum efficiency and overall network performance, the following technologies have been developed for 6G communication systems: full-duplex technology for achieving simultaneous use of the same frequency resources at the same time for uplink transmission and downlink transmission; network technology for utilizing satellites, high altitude platform stations (HAPS), etc. in an integrated manner; improved network structure for supporting mobile base stations, etc. and achieving network operation optimization and automation, etc.; dynamic spectrum sharing technology through conflict avoidance based on prediction of spectrum use; use of artificial intelligence (AI) in wireless communication to improve overall network operation by utilizing AI from the design stage of developing 6G and internalizing end-to-end AI support functions; next-generation distributed computing technology for overcoming the limits of UE computing capabilities through achievable ultra-high performance communications and computing resources on the network (such as mobile edge computing (MEC), cloud, etc.). In addition, attempts continue to be made to strengthen connectivity between devices, optimize networks, promote softwareization of network entities, and increase the openness of wireless communications by designing new protocols to be used in 6G communication networks, developing mechanisms for achieving hardware-based security environments and secure use of data, and developing technologies for maintaining privacy.

[0006] It is expected that research and development of hyper-connected 6G communication systems, including human-to-machine (P2M) and machine-to-machine (M2M), will realize the next hyper-connected experience. In particular, it is expected that services such as truly immersive extended reality (XR), high-fidelity mobile holograms, and digital replicas can be provided through 6G communication systems. In addition, services such as remote surgery with enhanced safety and reliability, industrial automation, and emergency response will be provided through 6G communication systems, making it possible to apply technology in various fields such as industry, medical care, automobiles, and household appliances.

[0007] Channel state information (CSI) can be used to measure the state of a channel between a user equipment (UE) and a base station in a wireless communication system. In this case, the CSI reporting timing can be adjusted to report the CSI more efficiently. Therefore, a method for effectively managing CSI reporting configuration is considered. Summary of the invention

[0008] Technical Solution

[0009] Various embodiments of the present disclosure provide an apparatus and method capable of effectively providing a service in a wireless communication system.

[0010] According to various embodiments of the present disclosure, a method performed by a base station in a wireless communication system may include: receiving capability information from a user equipment (UE) indicating whether the UE supports generation of channel state information (CSI) based on artificial intelligence (AI); receiving information about a change in reasoning time related to AI-based CSI generation and information about a change in channel state from the UE based on the capability information; identifying information about CSI report timing based on the information about the change in reasoning time and the change in channel state; transmitting radio resource control (RRC) configuration information including the identified CSI report timing to the UE; transmitting a CSI reference signal (RS) to the UE; and receiving a CSI report based on the CSI-RS and the information about the CSI report timing from the UE.

[0011] According to various embodiments of the present disclosure, a method performed by a user equipment (UE) in a wireless communication system may include: transmitting capability information indicating whether the UE supports AI-based CSI generation to a base station; identifying status information of an AI chip that performs AI-based CSI generation; identifying a change in reasoning time related to AI-based CSI generation based on the identified status information of the AI ​​chip; transmitting information about the identified change in reasoning time related to CSI generation and information about a change in channel state to the base station; receiving RRC configuration information including CSI report timing from the base station, the CSI report timing being based on information about the change in reasoning time and the change in channel state; receiving a CSI-RS from the base station; and transmitting a CSI report to the base station based on the CSI-RS and the information about the CSI report timing.

[0012] According to various embodiments of the present disclosure, a base station in a wireless communication system may include at least one transceiver and at least one processor operably connected to the at least one transceiver. The at least one processor may be configured to: receive capability information indicating whether the UE supports the generation of AI-based CSI from a UE; receive information about a change in reasoning time related to AI-based CSI generation and information about a change in channel state from the UE based on the capability information; identify information about CSI report timing based on information about the change in reasoning time and the change in channel state; transmit RRC configuration information including the identified CSI report timing to the UE; transmit a CSI-RS to the UE; and receive a CSI report based on the CSI-RS and information about CSI report timing from the UE.

[0013] According to various embodiments of the present disclosure, a user equipment (UE) in a wireless communication system may include at least one transceiver and at least one processor operably connected to the at least one transceiver. The at least one processor may be configured to: transmit capability information indicating whether the UE supports the generation of AI-based CSI to a base station; identify the state information of an AI chip that performs AI-based CSI generation; identify the change of reasoning time related to AI-based CSI generation based on the identified state information of the AI ​​chip; transmit information about the identified change of reasoning time related to CSI generation and information about the change of channel state to the base station; receive RRC configuration information including CSI report timing from the base station, the CSI report timing is based on the information about the change of reasoning time and the change of channel state; receive CSI-RS from the base station; and transmit CSI report to the base station based on the CSI-RS and the information about CSI report timing.

[0014] Various embodiments of the present disclosure provide an apparatus and method capable of effectively providing a service in a wireless communication system.

[0015] In addition, the advantages obtained in the present disclosure are not limited to the aforementioned advantages, and other advantages not mentioned herein can be clearly understood by those skilled in the art to which the present disclosure belongs from the following description.

[0016] Before proceeding to the following detailed description, it may be helpful to set forth definitions of certain words and phrases used throughout this patent document: the term "including" and its derivatives mean including but not limited to; the term "or" is inclusive, meaning and / or; the phrases "associated with" and "associated with" and their derivatives may mean including, included within, interconnected with, containing, contained within, connected to or connected with, coupled to or coupled with, capable of communicating with, cooperating with, interleaved, juxtaposed, proximate to, bound to or bound with, having, having the property of, etc.; and the term "controller" means any device, system, or part thereof that controls at least one operation, such device may be implemented in hardware, firmware, or software, or some combination of at least two thereof. It should be noted that the functionality associated with any particular controller may be centralized or distributed, whether local or remote.

[0017] In addition, the various functions described below may be implemented or supported by one or more computer programs, each of which is formed by a computer-readable program code and embodied in a computer-readable medium. The terms "application" and "program" refer to one or more computer programs, software components, instruction sets, processes, functions, objects, classes, instances, related data or a portion thereof suitable for implementation with suitable computer-readable program code. The phrase "computer-readable program code" includes any type of computer code, including source code, object code and executable code. The phrase "computer-readable medium" includes any type of medium that can be accessed by a computer, such as a read-only memory (ROM), a random access memory (RAM), a hard drive, a compact disk (CD), a digital video disk (DVD) or any other type of memory. "Non-transitory" computer-readable media excludes wired, wireless, optical or other communication links that transmit transient electrical signals or other signals. Non-transitory computer-readable media include media that can store data permanently, as well as media that can store data and rewrite data later, such as rewritable optical disks or erasable memory devices.

[0018] Definitions for certain words and phrases are provided throughout this patent document, those of ordinary skill in the art should understand that in many, if not most instances, such definitions apply to prior, as well as future uses of such defined words and phrases. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] For a more complete understanding of the present disclosure and its advantages, reference is now made to the following description taken in conjunction with the accompanying drawings, in which like reference numerals represent like parts: Figure 1 A wireless communication system according to an embodiment of the present disclosure is shown; Figure 2 An example of the structure of a base station in a wireless communication system according to an embodiment of the present disclosure is shown; Figure 3 The structure of a terminal in a wireless communication system according to an embodiment of the present disclosure is shown; Figure 4 The present invention shows a process of using an AI model to report CSI in a wireless communication system according to an embodiment of the present disclosure; Figure 5 An example is shown in which the computational reasoning time of a GPU chip varies according to an AI algorithm in a wireless communication system according to an embodiment of the present disclosure; Figure 6 The signal flow of CSI reporting of a terminal and a base station in a wireless communication system according to an embodiment of the present disclosure is shown; Figure 7 A flowchart showing an operation of a base station configuring a CSI report in a wireless communication system according to an embodiment of the present disclosure; and Figure 8 A flowchart of an operation for configuring a CSI report of a terminal in a wireless communication system according to an embodiment of the present disclosure is shown.

[0020] With respect to the description of the drawings, the same or similar reference numbers may be used to refer to the same or similar elements. DETAILED DESCRIPTION

[0021] Discussed below Figures 1 to 8 The various embodiments used to describe the principles of the present disclosure in this patent document are only for illustration and should not be interpreted in any way as limiting the scope of the present disclosure. Those skilled in the art will understand that the principles of the present disclosure can be implemented in any appropriately arranged system or device.

[0022] The terms used in this disclosure are only used for the purpose of describing specific embodiments, and are not intended to limit other embodiments. Unless the context is clearly different, singular expressions may include plural expressions. All terms (including technical and scientific terms) used herein may have the same meaning as those generally understood by those of ordinary skill in the field disclosed in this disclosure. It will also be understood that terms (such as those defined in commonly used dictionaries) should be interpreted as having meanings consistent with their meanings in the context of the relevant field, and will not be interpreted in an idealized or overly formal sense, unless explicitly defined in this article. Optionally, the terms defined in this disclosure should not be interpreted as excluding embodiments of the present disclosure.

[0023] In various embodiments of the present disclosure described below, for example, a hardware-based method is described. However, since various embodiments of the present disclosure include techniques using both hardware and software, a software-based method is not excluded in the embodiments of the present disclosure.

[0024] For convenience of explanation, the terms used hereinafter to refer to components of a device (e.g., controller, processor, artificial intelligence (AI) model, encoder, decoder, autoencoder (AE), neural network (NN) model, etc.) and the terms referring to data (e.g., signal, feedback, report (or reported), information, parameter, value, bit, codeword, etc.) are exemplified. Therefore, the present disclosure is not limited to the terms described below, and therefore other terms having the same technical meaning may also be used.

[0025] In addition, although the present disclosure describes various embodiments by using terms used in some communication standards (eg, the 3rd Generation Partnership Project (3GPP)), this is only for exemplary purposes. Various embodiments of the present disclosure can be easily modified and applied to other communication systems.

[0026] Figure 1A wireless communication system according to various embodiments of the present disclosure is shown. As part of a node using a radio channel, Figure 1 The base station 110, the terminal 120 and the terminal 130 are exemplified in the wireless communication system of FIG. Figure 1 Only one base station is shown in the figure, but other base stations that are the same as or similar to the base station 110 may also be included.

[0027] The base station 110 is a network infrastructure that provides radio access to the terminals 120 and 130. The base station 110 has a coverage area that is limited to a specific geographical area based on the distance at which a signal can be transmitted. In addition to the term "base station", the base station 110 may also be referred to as an "access point (AP)", "eNodeB (eNB)", "gNodeB (gNB)", "fifth generation (5G) node", "sixth generation (6G) node", "radio point", "transmit / receive point (TRP)" or other terms with equivalent technical meanings.

[0028] As a device used by a user, each of the terminal 120 and the terminal 130 communicates with the base station 110 through a radio channel. Optionally, at least one of the terminals 120 and 130 may operate without user participation. That is, as a device for performing machine type communication (MTC), at least one of the terminals 120 to 130 may not be carried by a user. In addition to the term "terminal", each of the terminals 120 and 130 may be referred to as a "user equipment (UE)", "mobile station", "subscriber station", "customer premises equipment (CPE)", "remote terminal", "wireless terminal", "electronic device", "user device", or other terms with equivalent technical meanings.

[0029] In addition, the base station 110, the terminal 120, and the terminal 130 may transmit and receive radio signals in a millimeter wave (mmWave) frequency band (e.g., 28 GHz, 30 GHz, 38 GHz, 60 GHz, above 60 GHz, etc.). In this case, in order to improve the channel gain, the base station 110, the terminal 120, and the terminal 130 may perform beamforming. In this article, beamforming may include transmit beamforming and receive beamforming. That is, the base station 110, the terminal 120, and the terminal 130 may assign directivity to the transmit signal and / or receive signal. To this end, the base station 110 and the terminals 120 and 130 may select the service beams 112, 113, 121, and 131 through a beam search or beam management process. After the service beams 112, 113, 121, and 131 are selected, subsequent communications may be performed through resources having a quasi-co-location (QCL) relationship with the resources used to transmit the service beams 112, 113, 121, and 131.

[0030] Figure 2 An example of a structure of a base station in a wireless communication system according to an embodiment of the present disclosure is shown. According to various embodiments of the present disclosure, for convenience, the base station 110 may be referred to as a network. Figure 2 The exemplary structure of can be understood as the structure of the base station 110. Hereinafter, the terms "... unit", "... means" and the like mean a unit that processes at least one function or operation and can be implemented in hardware or software or a combination of hardware and software.

[0031] refer to Figure 2 , the base station 110 may include a wireless communication circuit 210 , a backhaul communication circuit 220 , a storage device 230 , and a controller 240 .

[0032] The wireless communication circuit 210 performs the function of transmitting and receiving signals through a wireless channel. For example, the wireless communication circuit 210 performs the conversion function between the baseband signal and the bit stream according to the physical layer standard of the system. For example, in data transmission, the wireless communication circuit 210 generates complex symbols by encoding and modulating the transmission bit stream. In addition, in receiving data, the wireless communication circuit 210 restores the received bit stream by demodulating and decoding the baseband signal. In addition, the wireless communication circuit 210 up-converts the baseband signal into a radio frequency (RF) band signal and then transmits the RF signal through the antenna, and down-converts the RF signal received through the antenna into a baseband signal.

[0033] To this end, the wireless communication circuit 210 may include a transmit filter, a receive filter, an amplifier, a mixer, an oscillator, a digital-to-analog converter (DAC), an analog-to-digital converter (ADC), etc. In addition, the wireless communication circuit 210 may include multiple transmit / receive paths. In addition, the wireless communication circuit 210 may include at least one antenna array, which is composed of multiple antenna elements. From a hardware aspect, the wireless communication circuit 210 may be composed of a digital unit and an analog unit, and the analog unit may be composed of multiple sub-units according to operating power, operating frequency, etc.

[0034] The wireless communication circuit 210 may transmit and receive signals. To this end, the communication circuit 210 may include at least one transceiver. For example, the wireless communication circuit 210 may transmit synchronization signals, reference signals, system information, messages, control messages, streams, control information, data, etc. In addition, the wireless communication circuit 210 may perform beamforming.

[0035] As described above, the wireless communication circuit 210 transmits and receives signals. Therefore, the wireless communication circuit 210 may be referred to as a transmitter, a receiver, or a transceiver. In addition, in the following description, transmission and reception performed through a wireless channel are used to imply that the above-mentioned processing is performed by the wireless communication circuit 210.

[0036] The backhaul communication circuit 220 provides an interface for performing communication with different nodes in the network. That is, the backhaul communication circuit 220 converts a bit stream transmitted from a base station to different nodes (such as different access nodes, different base stations, upper nodes, core networks, etc.) into a physical signal, and converts a physical signal received from a different node into a bit stream.

[0037] The storage device 230 stores data such as basic programs, applications, configuration information, etc. for the operation of the base station 110. The storage device 230 may include a memory. The storage device 230 may be composed of a volatile memory, a non-volatile memory, or a combination of a volatile memory and a non-volatile memory. In addition, the storage device 230 provides the stored data according to the request of the controller 240. According to an embodiment, the storage device 230 may store learning data for AI-based CSI reporting, and the stored learning data may be applied to the neural network structure of the AI-based CSI reporting.

[0038] The controller 240 controls the overall operation of the base station 110. For example, the controller 240 may transmit and receive signals via the wireless communication circuit 210 or the backhaul communication circuit 220. In addition, the controller 240 writes data into the storage device 230 and reads the data. In addition, the controller 240 may perform the functions of the protocol stack required in the communication standard. To this end, the controller 240 may include at least one processor.

[0039] Figure 2 The structure of the base station 110 is merely an example of a base station, and examples of base stations that implement various embodiments of the present disclosure are not limited to Figure 2 That is, the structure may be added, deleted or partially changed according to various embodiments.

[0040] Despite Figure 2 The base station is described as one entity in the specification, but the present disclosure is not limited thereto. The base station according to various embodiments of the present disclosure may be implemented to constitute an access network having not only an integrated deployment but also a distributed deployment. According to an embodiment, the base station may be divided into a central unit (CU) and a digital unit (DU). The CU may be implemented to perform the functions of an upper layer (e.g., a packet data convergence protocol (RRC)), and the DU may be implemented to perform the functions of a lower layer (e.g., a medium access control (MAC) and a physical (PHY)). The DU of the base station may constitute beam coverage on a radio channel.

[0041] Figure 3 The structure of a terminal in a wireless communication system according to an embodiment of the present disclosure is shown. Figure 3The exemplary structure of can be understood as the structure of terminals 120 and 130. Hereinafter, the terms "... unit", "... means" and the like mean a unit that processes at least one function or operation and can be implemented in hardware or software or a combination of hardware and software.

[0042] refer to Figure 3 , the terminal 120 or 130 may include a communication circuit 310 , a storage device 320 , and a controller 330 .

[0043] The communication circuit 310 performs the function of transmitting and receiving signals through a wireless channel. For example, the communication circuit 310 performs the conversion function between the baseband signal and the bit stream according to the physical layer standard of the system. For example, in data transmission, the communication circuit 310 generates complex symbols by encoding and modulating the transmission bit stream. In addition, in receiving data, the communication circuit 310 recovers the received bit stream by demodulating and decoding the baseband signal. In addition, the communication circuit 310 up-converts the baseband signal into a radio frequency (RF) band signal and then transmits the RF signal through the antenna, and down-converts the RF signal received through the antenna into a baseband signal. For example, the communication circuit 310 may include a transmit filter, a receive filter, an amplifier, a mixer, an oscillator, a DAC, an ADC, etc.

[0044] In addition, the communication circuit 301 may include multiple transmit / receive paths. In addition, the communication circuit 301 may include an antenna unit. The communication circuit 301 may include at least one antenna array, which is composed of multiple antenna elements. From a hardware perspective, the communication circuit 301 may be composed of digital circuits and analog circuits (e.g., radio frequency integrated circuits (RFICs)). In this article, the digital circuit and the analog circuit may be implemented as one package. In addition, the communication circuit 301 may include multiple RF chains. In addition, the communication circuit 301 may perform beamforming. In order to assign directivity to the signal to be transmitted / received according to the setting of the controller 305, the communication circuit 301 may apply beamforming weights to the signal. According to an embodiment, the communication circuit 301 may include a radio frequency (RF) block (or RF unit). The RF block may include a first RF circuit associated with an antenna and a second RF circuit associated with baseband processing. The first RF circuit may be referred to as an RF antenna (RF-A). The second RF circuit may be referred to as an RF baseband (RF-B).

[0045] In addition, the communication circuit 301 can transmit / receive signals. To this end, the communication circuit 301 may include at least one transceiver. The communication circuit 301 may receive a downlink signal. The downlink signal may include a synchronization signal (SS), a reference signal (RS) (e.g., a cell-specific reference signal (CRS), a demodulation (DM)-RS), system information (e.g., MIB, SIB, remaining system information (RMSI), other system information (OSI)), a configuration message, control information or downlink data, etc. In addition, the communication circuit 301 may transmit an uplink signal. The uplink signal may include a random access-related signal (e.g., a random access preamble (RAP) (message 1 (Msg1), message 3 (Msg3))), a reference signal (e.g., a sounding reference signal (SRS), DM-RS) or a power headroom report (PHR), etc.

[0046] In addition, the communication circuit 301 may include different communication modules to process signals of different frequency bands. In addition, the communication circuit 301 may include multiple communication modules to support multiple different radio access technologies. For example, different radio access technologies may include Bluetooth Low Energy (BLE), Wireless Fidelity (Wi-Fi), WiFi Gigabyte (WiGig), cellular networks (e.g., Long Term Evolution (LTE), New Radio (NR)), etc. In addition, different frequency bands may include Super High Frequency (SHF) (e.g., 2.5 GHz, 5 GHz) bands and millimeter wave (e.g., 38 GHz, 60 GHz, etc.) bands. In addition, the communication circuit 301 may use the same type of wireless access technology on different frequency bands (e.g., unlicensed bands for Licensed Assisted Access (LAA), Citizen Broadband Radio Service (CBRS) (e.g., 3.5 GHz)).

[0047] The communication circuit 301 transmits and receives signals as described above. Therefore, the communication circuit 301 may be referred to as a transmitter, a receiver, or a transceiver. In addition, in the following description, transmission and reception performed through a radio channel are used to imply that the above-mentioned processing is performed by the communication circuit 301.

[0048] The storage device 320 stores data such as basic programs, applications, configuration information, etc. for the operation of the terminal 120. The storage device 320 may be composed of a volatile memory, a non-volatile memory, or a combination of a volatile memory and a non-volatile memory. In addition, the storage device 320 provides the stored data according to a request of the controller 330. According to an embodiment, the storage device 320 may store learning data for AI-based CSI reporting based on a CSI reporting format configured by the base station.

[0049] The controller 330 controls the overall operation of the terminals 120 and 130. For example, the controller 330 may transmit and receive signals via the communication circuit 310. In addition, the controller 330 writes data into the storage device 320 and reads the data. In addition, the controller 330 may perform the functions of the protocol stack required in the communication standard. To this end, the controller 330 may include at least one processor. The controller 330 may include at least one processor or microprocessor, or may be a part of a processor. In addition, the communication circuit 310 and a part of the controller 330 may be referred to as a communication processor (CP). The controller 330 may include various modules for performing communication. According to various embodiments, the controller 330 may control the terminal to perform operations according to the various embodiments described below.

[0050] Although Figure 3 Not shown, but according to various embodiments of the present disclosure, the terminals 120 and 130 may further include a measurement unit for the AI ​​chip. According to an embodiment, the measurement unit included in the terminal may detect the state of the AI ​​chip (for example, including various physical parameters of the AI ​​chip), including the temperature of the AI ​​chip. According to an embodiment, the state of the AI ​​chip detected by the measurement unit may affect the calculation performed by the AI ​​chip. According to an embodiment, the controller 330 may monitor the state information of the AI ​​chip in real time based on the state of the AI ​​chip detected by the measurement unit or the scheduling state configured in the terminal from the base station. The controller 330 may measure the change in the calculation reasoning time based on the state information of the AI ​​chip monitored in real time.

[0051] In one embodiment, the AI ​​model based on neural network learning can be operated by the controller 330 and the storage device 320. In this case, the controller 330 may be composed of one or more processors. One or more processors may include general-purpose processors such as a central processing unit (CPU), an application processor (AP), a digital signal processor (DSP), etc., dedicated graphics processors such as a graphics processing unit (GP) and a visual processing unit (VPU), or functions of an AI-specific processor such as a neural processing unit (NPU). One or more processors may provide control to process input data according to a predefined operation rule or AI model stored in the storage device 320. Alternatively, when one or more processors are AI-specific processors, the AI-specific processor may be designed with a hardware structure specifically for processing a specific AI model. The AI-specific processor may not be included in the controller 330, but may be included as a separate component.

[0052] According to an embodiment, the predefined operation rules or AI models are characterized in that they are created by learning. In this article, creation by learning means that the basic AI model is learned by using multiple learning data according to the learning algorithm, thereby creating a predefined operation rule or AI model to perform the required features (or purposes). Learning can be implemented by the device itself in which the AI ​​according to the present disclosure is executed, or it can be implemented by a separate server and / or system. Examples of learning algorithms include supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning, but are not limited to the foregoing examples. The controller 330 can learn events that are about to occur, decisions made, and information collected or input through a learning algorithm. The controller 330 can store such learning results in a storage device 320 (e.g., a memory).

[0053] The AI ​​model may be composed of multiple neural network layers. Each of the multiple neural network layers has multiple weight values, and the neural network calculation is performed by calculating between the calculation results of the previous layer and the multiple weight values. The multiple weight values ​​of the multiple neural network layers can be optimized by the learning results of the AI ​​model. For example, during the learning process, multiple weight values ​​can be updated to reduce or optimize the loss value or cost value obtained from the AI ​​model. The artificial neural network may include a deep neural network (DNN), such as a convolutional neural network (CNN), a deep neural network (DNN), a recursive neural network (RNN), a restricted Boltzmann machine (RBM), a deep belief network (DBN), a bidirectional recursive deep neural network (BRDNN), a deep Q network, etc., but is not limited to the aforementioned examples.

[0054] In one embodiment, the controller 330 may execute an algorithm for performing operations related to AI-based channel state information (CSI) reporting or feedback. In one embodiment, the AI ​​model learned to perform AI-based CSI feedback may be constructed as hardware in the controller 330, included as software, or constructed by a combination of hardware and software. In other words, the controller 330 may include an AI-based CSI feedback controller. The AI-based CSI feedback controller may determine whether to monitor the AI ​​model that performs AI-based feedback, determine whether to report the monitoring results, determine whether the AI ​​model that performs AI-based CSI feedback fails, or determine whether to use AI-based CSI feedback. In addition, according to various embodiments, the controller 330 may include an update unit.

[0055] The update unit can obtain updated data (e.g., data related to CSI feedback between the terminal and the base station) through a learning process between the terminal and the base station, and can reconstruct parameter values ​​constituting the neural network based on the data (e.g., neural network structure, information of each node layer, weight value information between nodes). The AI-based CSI feedback controller and update unit can be a command / code residing in the controller 330 or a storage space for storing commands / codes as a command set or code stored in the storage device 320, or can be a part of the circuit constituting the controller 330. According to various embodiments, the controller 330 can control the terminals 120 and 130 to perform operations according to various embodiments.

[0056] Figure 3 The structures of the terminals 120 and 130 are merely examples of terminals, and examples of terminals for performing various embodiments of the present disclosure are not limited to Figure 3 That is, the structure may be added, deleted or partially changed according to various embodiments.

[0057] For ease of explanation, the following description is based on the AI ​​model included in the terminals 120 and 130. That is, an AI model including a specific neural network structure and learned by a specific algorithm may be included in the terminals 120 and 130. However, the present disclosure is not limited thereto and may also be applied to the AI ​​model included in the base station 110.

[0058] Technologies related to AI-based CSI feedback may include performing learning and configuration based on a specific algorithm to apply CSI feedback as a specific AI model, collecting learning data required in the process of learning the specific AI model, and verifying the performance of the learned specific AI model. Specifically, with respect to verifying the performance of the learned specific AI model, the present disclosure may also include an operation in which the terminal monitors the AI ​​model and reports the monitoring results to perform CSI feedback through the optimal AI model.

[0059] When describing the AI ​​model-based CSI reporting method of the present disclosure, for the sake of convenience of explanation, an autoencoder (AE) is used as an AI model, for example. However, the present disclosure is not limited to this, and is obviously applicable to all AI models that can perform CSI compression when performing CSI reporting. In this article, AE may refer to a structure with the same input and output (including an AI model with a bottleneck structure). AE can compress the CSI measured by the terminal in the form of a low-dimensional vector. In other words, the terminal can generate compressed CSI through the encoder of AE by using the measured complete CSI, and the compressed CSI can be transmitted to the base station. Therefore, the base station can receive explicit CSI feedback instead of implicit CSI feedback. AE can have advantages in reporting through a CSI compression scheme.

[0060] For example, since the true value of the original data is known, the AE can accurately evaluate the AI ​​model even when the AE is inferring the performance of the AI ​​model. That is, since the input value of the AE is known, the performance of the AE can be measured by comparing the output value of the AE with the input value. In the case where the AE knows the true value of the original data, it can accurately predict which value can be output based on the value input to the AE. In addition, since the AE is highly dependent on data, the AE can also be used for anomaly detection of unlearned data. However, since the AE (or AI model) is an AI model based on data, the performance stability may be degraded. Therefore, the terminal can request the base station to execute a procedure for monitoring and reporting the accuracy (or reliability) of the AE.

[0061] Figure 4 The process of reporting CSI using an AI model in a wireless communication system according to an embodiment of the present disclosure is shown. Although the autoencoder is described as an AI model for CSI reporting (or CSI feedback), the present disclosure is not limited thereto. In this article, it is assumed that the autoencoder for CSI reporting is an AI model that has performed learning for reporting CSI based on a specific learning algorithm.

[0062] refer to Figure 4 , the autoencoder 400 may be an AI model that learns CSI reporting or feedback between a user equipment (UE) and a base station (BS). The UE may generate CSI by preprocessing channel information estimated based on a measurement result of a signal received from the BS. For example, the preprocessing may include eigenvalue decomposition (EVD) or singular value decomposition (SVD). In this article, the generated CSI may refer to the complete CSI. The complete CSI may be input to the encoder 410 of the UE, which is the input of the autoencoder 400, and thus compressed CSI may be generated. The UE may transmit the compressed CSI to the BS, and the BS may restore the compressed CSI through the decoder 420. In this case, the decoder 420 of the BS may be the output of the autoencoder 400. According to the foregoing description, the autoencoder 400 may learn CSI compression, which may be used for CSI feedback between the UE and the BS. The autoencoder 400 may perform explicit CSI feedback, rather than implicit CSI feedback, through feedback based on the learned CSI compression method. The CSI compression method may require the autoencoder 400 to have stable performance and high accuracy.

[0063] Therefore, according to an embodiment, the UE needs to manage the process of the auto encoder 400 by continuously and periodically reporting to the BS the evaluation result obtained by monitoring the performance of the auto encoder 400. Based on the foregoing description, an operation for more efficiently performing CSI reporting based on the AI ​​mode will be described according to various embodiments of the present disclosure. Hereinafter, the CSI described for this purpose may refer to at least one of the full CSI and the compressed CSI.

[0064] According to various embodiments of the present disclosure, the AI ​​chip included in the UE or BS may perform various functions including the aforementioned CSI compression. The various functions performed by the AI ​​chip may include AI-related physical layer functions, modem functions, application processor (AP) functions, etc., and are not limited to autoencoders for CSI compression. According to an embodiment, various AI algorithms may be executed according to the various functions to be performed. However, the various AI-related functions may have different characteristics, such as the amount of calculation to be performed in each AI chip, the scheduling of AI algorithms, etc. When performing calculations, when various functions of AI chips with their own characteristics are included in the same AI chip, mutual calculations may be affected.

[0065] Accordingly, even if the same function is executed in an AI chip, the timing of the completion of the calculation based on the AI ​​algorithm may be different each time. In order to solve the aforementioned problem, the AI-related functions to be executed can be controlled by being included in a separate chip, which may cause a great cost burden and physical limitations. Therefore, for hardware and cost efficiency, the present disclosure discloses various embodiments on the premise of executing various AI-related functions in a shared AI chip.

[0066] In the case where various AI-related functions are performed in a shared AI chip, as described above, when the CSI compression and reporting operation of the autoencoder is performed as one of the AI-related functions, the timing of the completion of the calculation may not be accurately predicted. For example, when the timing of the completion of the calculation of the UE's CSI compression cannot be accurately predicted, the AI ​​calculation may not be completed at the timing set for the UE, which is set to generate and compress CSI (hereinafter, interchangeably referred to as "CSI calculation") and report it. In this case, there may be a situation where the UE requests the BS to provide resources for CSI reporting again or abandons the CSI report of the corresponding channel, which may lead to resource waste and delay problems.

[0067] In order to solve the aforementioned problems, the present disclosure provides a method and apparatus capable of adaptively determining and configuring CSI compression reporting timing when the CSI calculation timing varies according to computing resources of an AI chip shared by other AI functions.

[0068] Figure 5An example is shown in which the computational reasoning time of a graphics processing unit (GPU) chip varies according to an AI algorithm in a wireless communication system according to an embodiment of the present disclosure. More specifically, Figure 5 Figure 500 shows the change in computational inference time that occurs when each GPU chip is shared by various AI algorithms. According to various embodiments of the present disclosure, the "change in computational inference time" described below may be referred to as various terms having the same or similar meanings within the scope of the present disclosure, such as "change in inference time", "time change information", "time information", "computation information", etc. According to an embodiment, "change in inference time" may refer to the uncertainty in the timing of computational completion, depending on the addition (or deletion, change) of AI functions relative to non-AI-based functions (or AI-based functions). According to an embodiment, "change in inference time" refers to the uncertainty in the timing of computational completion according to the AI ​​function, and may include a single value or a range of values ​​including time units. Alternatively, but not limited to this, it may also include information containing numerical values.

[0069] refer to Figure 5 , the vertical axis of the graph 500 represents the computational inference time (ms) of each AI algorithm, and the horizontal axis represents different GPU chips (GPU1 510, GPU2 520, GPU3 530). According to an embodiment, each AI algorithm may be not only an algorithm applied to the same AI function (e.g., CSI compression, etc.), but also an algorithm applied to different AI functions.

[0070] According to an embodiment, the AI ​​chip of the present disclosure may include Figure 5 GPU chip. Unlike a central processing unit (CPU) that includes multiple processing cores and is composed of millions of transistors, a GPU may be a processor composed of smaller and more specialized cores than a CPU. The CPU and GPU work together as important computing engines for processing data, but a GPU chip may be more suitable for deep learning and AI training of large-capacity specific data sets, such as multiple neural network layers or two-dimensional (2D) images. Therefore, in order to perform a wide range of workloads such as deep learning or AI, a GPU may be more suitable as a shared chip for executing various AI algorithms. However, according to various embodiments of the present disclosure, this is merely an example, and a shared AI chip is not limited to a GPU chip, and obviously, a shared AI chip may include various processor chips capable of executing AI algorithms similar or equivalent thereto.

[0071] refer to Figure 5, when AI algorithms 501, 502, 503, 504 are implemented in GPU1 chip 510, in GPU2 chip 520, or in GPU3 chip 530, each AI algorithm may have a different computational inference time. In addition, each GPU chip may have a different computational inference time required for each AI algorithm. The computational inference time may be different due to GPU contention, queuing delay, etc. Figure 5 As shown, even if the algorithm performs the same AI function, the computational reasoning time may vary for each shared AI chip, or the computational reasoning time may vary according to the AI ​​algorithm of the shared AI chip. As described above, the variation in the computational reasoning time of the AI ​​function may cause problems in the case of fixed CSI report timing. To solve this problem, various embodiments of the present disclosure disclose a method and apparatus for adaptively determining and configuring CSI report timing based on channel status and variation in computational reasoning time.

[0072] Figure 6 1 shows a signal flow for CSI reporting of a UE and a BS in a wireless communication system according to an embodiment of the present disclosure. More specifically, referring to Figure 6 , flowchart 610 shows the signal flow of periodic CSI reporting of UE and BS, and flowchart 620 shows the signal flow of aperiodic CSI reporting of UE and BS. In addition, flowchart 630 shows the signal flow of semi-persistent CSI reporting of UE and BS. However, but not limited to this, it is obvious that as long as a person of ordinary skill in the art can clearly understand, the various steps of CSI reporting according to various embodiments of the present disclosure are also applicable to various CSI reports between UE and BS. In addition, according to various embodiments of the present disclosure, at least one of all, some or some combinations of the steps described below may be included, and it is obvious that within the achievable range, some combinations of steps of periodic, aperiodic or semi-persistent CSI reporting are possible. Hereinafter, the operation of CSI reporting is described by including all signal flows of periodic, aperiodic and semi-persistent CSI reporting.

[0073] although Figure 6 Although not shown in the figure, according to an embodiment, the UE may transmit capability information including information on whether the AI ​​function is supported to the BS. The capability information transmitted by the UE may include information on whether the UE is capable of generating and reporting AI-based CSI. The capability information transmitted by the UE may be information reported to the BS periodically, or may be information reported to the BS based on a request of the BS. According to an embodiment, the capability information reported by the UE may be transmitted via high-layer signaling (e.g., RRC signaling) or MAC CE.

[0074] According to an embodiment, a BS that has received capability information from a UE may identify whether the UE supports AI functions including AI-based CSI operations (CSI generation or reporting). When the UE does not support the AI ​​function, the BS may transmit configuration information related to non-AI-based CSI reporting to the UE. When the UE supports the AI ​​function, the BS may also transmit configuration information related to AI-based CSI reporting or configuration information related to reporting of calculation of inference time changes (hereinafter interchangeably referred to as inference time change reports). According to an embodiment, the configuration information transmitted by the BS may be transmitted via at least one of higher layer signaling (e.g., RRC signaling), MAC CE, and downlink control information (DCI).

[0075] According to an embodiment, the BS may configure the UE to periodically report the reasoning time change according to a determined timing based on the configuration information related to the reporting of the reasoning time change. According to an embodiment, the BS may configure the UE to detect the state of the AI ​​chip based on the configuration information related to the reporting of the reasoning time change, and report the reasoning time change when the trigger condition is met. According to various embodiments of the present disclosure, without being limited to the aforementioned configuration information, it is obvious that the UE is able to report the reasoning time change (for example, when it is pre-configured in the UE) regardless of the configuration information of the BS.

[0076] According to various embodiments of the present disclosure, the aforementioned steps for transmitting capability information of a UE or configuration information for a BS to transmit a report on reasoning time variation may be applied to all or part of flowcharts 610 , 620 , and 630 , or may be omitted.

[0077] refer to Figure 6 In step 611, 621 or 631, the UE may transmit information about the change in reasoning time to the BS. According to an embodiment, in addition to the information about the change in reasoning time, the UE may also transmit information about the channel state to the BS. According to an embodiment, the change in reasoning time of the present disclosure may include: when an AI function including AI-based CSI compression must be performed, a CSI report timing that is changed compared to the timing of reporting CSI based on general CSI configuration information.

[0078] According to an embodiment, the change in inference time may include a single value, or may include multiple values (e.g., from N ms to M ms, where N and M are positive integers and N < M), depending on the AI algorithm applied to the AI function (e.g., AI-based CSI compression, etc.). When the change in inference time includes a single value, the single value may be a value representing multiple values depending on the AI algorithm, and may be at least one of, for example, the sum, weighted sum, average, and maximum value of the multiple values. The information about the change in inference time transmitted by the UE to the BS may include the change in inference time for an AI-based CSI report or the change in inference time for various AI functions executed in the UE's AI chip. The information about the change in inference time transmitted by the UE may include the computation time required to generate an AI-based CSI.

[0079] In addition, the information about the change in inference time transmitted by the UE may include the difference (or offset) in the computation time required to generate an AI-based CSI (or the time required to generate a non-AI-based CSI) compared to before the report, or may include the absolute value of the computation time to generate an AI-based CSI. Alternatively, the information about the change in inference time transmitted by the UE may include information about a quantization value that divides the difference in the computation time required to generate an AI-based CSI (or the time required to generate a non-AI-based CSI) compared to before the report into different values for each section.

[0080] According to an embodiment, the UE may measure the computation time to generate CSI based on the state of the AI chip that executes the AI function. As Figure 5 shown, even in the case of the same AI function (e.g., CSI compression, etc.), the computation time may be determined differently according to the algorithm to be applied. The UE may identify the state of the AI chip to measure the computation time for compressing and generating CSI. According to an embodiment, the UE may detect the temperature of the AI chip or a change in temperature. Alternatively, the UE may identify the various AI functions being executed in the AI chip or the scheduling of the algorithms to be applied to the various AI functions. The state of the AI chip identified by the UE may include the physical attribute information of the aforementioned AI chip (e.g., temperature or a change in temperature, etc.) or information about the algorithms of the AI functions. The UE may identify the state of the AI chip in real time and identify the computation time to generate an AI-based CSI based on the identified state of the AI chip.

[0081] According to an embodiment, the UE may have information about a preset reasoning time. The reasoning time information preset to the UE may include a combination of each pre-measured AI function and the reasoning time information of the AI ​​function. Alternatively, the information about the preset reasoning time may include a mapping relationship between each pre-measured AI function and the reasoning time information of each AI function. According to an embodiment, the UE may identify the calculation time for generating AI-based CSI based on the information about the preset reasoning time and the AI ​​function implemented in the AI ​​chip.

[0082] According to an embodiment, the UE may identify information about the change in reasoning time based on the calculation time for generating the identified AI-based CSI. The UE may transmit information about the change in reasoning time to the BS. According to an embodiment, the UE may transmit information about the change in reasoning time to the BS based on configuration information related to the reasoning time change report received from the BS. According to an embodiment, based on the configuration information related to the reasoning time change report received from the BS, when the trigger condition is met, the UE may transmit information about the change in reasoning time to the BS. The trigger condition may be any reference value related to the state of the AI ​​chip detected by the UE. For example, the trigger condition configured by the BS for the UE may include a case where the temperature of the AI ​​chip exceeds a specific value, a case where the number of AI functions scheduled in the AI ​​chip exceeds a specific reference, and a case where the calculation time for generating the identified AI-based CSI exceeds a specific reference. According to an embodiment, the information about the change in reasoning time transmitted by the UE may be transmitted by at least one of high-level signaling (e.g., RRC signaling), MAC CE, and DCI.

[0083] In steps 612, 622, or 632, the BS may recognize and transmit configuration information for CSI reporting to the UE. Figure 6 It is shown that the configuration information is transmitted through high-layer signaling (e.g., RRC signaling), but the present disclosure is not limited thereto, and according to an embodiment, the BS may obviously transmit the configuration information through MAC or DCI, and may transmit the configuration information not only through a downlink control channel but also through a data channel. According to an embodiment, in addition to the configuration information related to the general CSI report, the BS may also transmit additional configuration information identified based on the information about the inference time change to the UE.

[0084] According to an embodiment, in case of aperiodic CSI reporting (see 620) or semi-persistent CSI reporting (see 630), configuration information for CSI reporting transmitted by the BS to the UE may include information indicating receiving a CSI reference signal (RS) and reporting CSI based on a trigger signal (e.g., MAC CE or DCI) received by the UE from the BS. According to various embodiments of the present disclosure, in aperiodic CSI reporting (see 620) or semi-persistent CSI reporting (see 630), configuration information for CSI reporting transmitted by the BS to the UE may include an indicator (e.g., an offset indicator), such as n time slots after the trigger signal related to the CSI-related transmission / reception operation timing of the UE.

[0085] According to an embodiment, the BS may compare information about the inference time change received from the UE or the change in the channel state with a threshold. The BS may identify the timing for CSI reporting based on the comparison result. According to an embodiment, a case where the change in the channel state is not significant may include a case where the channel coherence time is long. According to an embodiment, the change in the channel state may be determined based on whether the channel is stable or the mobility of the UE. According to an embodiment, the change in the channel state may be identified based on an uplink signal (e.g., a sounding reference signal (SRS)) transmitted by the UE to the BS. According to an embodiment, all thresholds described below are not limited to the same threshold and may include corresponding different values. The case of "exceeding" may include "greater than or equal to".

[0086] According to an embodiment, when the change in the inference time for CSI generation (e.g., compression and generation) is less than a threshold, the BS may perform the CSI reporting process without configuring an additional CSI reporting timing. That is, the BS may transmit configuration information (e.g., configuration information for non-AI-based CSI reporting) to the UE without configuring an additional CSI reporting timing. For example, the BS may determine that the inference time for AI-based CSI generation (e.g., compression and generation) has no significant effect on the CSI reporting timing that the BS is initially to configure, and may not perform configuration of the additional CSI reporting timing in addition to configuring the general CSI reporting timing.

[0087] According to an embodiment, when a change in the inference time for CSI generation (e.g., compression and generation) is greater than a threshold but a change in the channel state is less than a specific value (e.g., when the channel coherence time is greater than a specific value), the BS may change the CSI report timing based on information about the inference time change. According to an embodiment, when a change in the inference time for CSI generation (e.g., compression and generation) is greater than a change in the channel state (e.g., channel coherence time), the BS may identify an additional CSI report offset (in addition to an existing CSI report offset) based on information about the inference time change. According to an embodiment, in the case of periodic CSI reporting, the BS may change the CSI report timing by increasing the period for CSI-RS transmission and CSI reporting.

[0088] Alternatively, the BS may change the CSI reporting timing by postponing the offset of the CSI reporting. According to an embodiment, in the case of non-periodic CSI reporting (see 620) or semi-persistent CSI reporting (see 630), the BS may change the CSI reporting timing by indicating that CSI reporting is performed in n time slots after the trigger signal based on the received information about the reasoning time change. According to an embodiment, the BS may identify the value with the largest change in each AI algorithm as the change value of the reasoning time based on the received information about the reasoning time change, and may identify the identified change value of the reasoning time as the changed CSI reporting timing in addition to the existing CSI reporting timing.

[0089] Since the demand for performing CSI reporting immediately is relatively low in an environment where the channel state does not change rapidly, the BS can identify the timing obtained by adding the change value of the inference time to the existing CSI reporting timing and allocate the timing to the UE instead of configuring an additional CSI report, thereby effectively performing the process for resource allocation. According to an embodiment, the BS can transmit the configuration information to the UE by including information indicating that the UE can perform CSI reporting in the identified changed CSI reporting timing. The configuration information can be transmitted to the UE by being included in the high-level configuration process of the aforementioned steps 612, 622, and 632.

[0090] According to an embodiment, when the change in the inference time for CSI generation (e.g., compression and generation) is greater than a threshold and the change in the channel state is greater than a specific value (e.g., when the channel coherence time is less than a specific value), the BS can identify (in addition to the existing CSI report offset) an additional CSI report offset based on information about the change in inference time. In the case where the channel state changes significantly, it may be difficult to reallocate the CSI report when the CSI report timing is postponed or missed. Therefore, the BS can identify the offset for the additional CSI report. According to an embodiment, with respect to AI-based CSI generation (e.g., compression and generation), the UE may not be able to complete the CSI calculation (e.g., compression and generation) at the timing of the first CSI report configured by the BS. In this case, the UE may not be able to perform CSI reporting at the allocated timing of the first CSI report.

[0091] However, when an offset for additional CSI reporting (e.g., timing for a second CSI report) is configured, the UE may perform CSI reporting at the timing for the second CSI report. For example, when the inference time variation is 2 ms to 5 ms, the BS may also recognize additional CSI reporting timing offsets corresponding to 2 ms and 5 ms in addition to the CSI reporting timing. According to an embodiment, when the UE is able to perform CSI reporting based on the timing for the first CSI report, the BS may reallocate resources allocated for the timing of the second CSI report to the UE. According to an embodiment, in order to indicate the timing of the second CSI report, the BS may transmit information related to the additional CSI reporting timing. For example, the information related to the additional CSI reporting timing transmitted by the BS may include an indicator indicating n time slots after the first CSI reporting timing.

[0092] Furthermore, the number of additional CSI reporting timings is not limited to one, and the information related to CSI reporting timings transmitted by the BS may include multiple additional CSI reporting timings and indicators thereof. The configuration information may be transmitted to the UE through a high-level configuration process included in the aforementioned steps 612, 622, and 632.

[0093] According to an embodiment, the BS may transmit information about the additional CSI reporting timing to the UE in advance (e.g., RRC signaling) before receiving information about the change in reasoning time from the UE. According to an embodiment, the BS may configure the pre-configured information about the additional CSI reporting timing to the UE before performing (e.g., generating and compressing) AI-based CSI, or may configure the information through a reconfiguration process when performing AI-based CSI. When the information about the change in reasoning time identified by the UE is reported to the BS, the UE that has pre-configured the information about the additional CSI reporting timing may identify the additional CSI reporting timing based on the information about the identified change in reasoning time and the information configured from the BS.

[0094] For example, the information about the additional CSI reporting timing preconfigured by the BS for the UE may include an additional reporting timing indicator. When the indicator is set to 0, the UE may identify the lowest value (e.g., 2 ms) of the inference time variation (e.g., 2 ms to 5 ms) as the first reporting timing, and the highest value (e.g., 5 ms) as the second reporting timing, and may report the CSI based on the identified reporting timing. When the indicator is set to 1, the UE may identify the average value (e.g., 3.5 ms) of the inference time variation (e.g., 2 ms to 5 ms) as the first reporting timing, and the highest value (e.g., 5 ms) as the second reporting timing, and may report the CSI based on the identified reporting timing. However, this is merely an example, and it is apparent that the information about the CSI reporting timing preconfigured to the UE and the indicator contained in the information may be implemented with a variety of parameters or contents.

[0095] According to an embodiment, the BS may transmit configuration information including information on the identified additional CSI reporting timing to the UE in addition to the CSI reporting offset. According to an embodiment, the BS may transmit a control indication related to the information on the identified additional CSI reporting timing to the UE.

[0096] According to an embodiment, when the change in reasoning time for CSI generation (e.g., compression and generation) is greater than a threshold value and the change in channel state is greater than a specific value (e.g., when the channel coherence time is less than a specific value), the BS may deactivate AI-based CSI feedback (e.g., AI-based CSI compression and generation). According to an embodiment, the threshold value of the change in reasoning time may include a threshold value prepared for a change in reasoning time that is too large to identify a reporting timing for AI-based CSI reporting. When the change in reasoning time includes a significantly large value, the BS may recognize that it cannot support the AI-based CSI compression operation of the UE, and may therefore instruct the UE to deactivate the AI-based CSI function.

[0097] According to an embodiment, after deactivating the AI-based CSI function, the BS may receive information about the change in reasoning time from the UE again, and may compare it with the threshold value. The BS may identify whether to activate the AI-based CSI function based on the comparison result between the change in reasoning time or the change in channel state and each threshold value. When a specific condition is met, the BS may instruct the UE to activate the AI-based CSI function, and may perform the AI-based CSI function again. Figure 6 steps.

[0098] According to an embodiment, the BS may identify a CSI report category based on information about a change in reasoning time received from the UE. Obviously, the BS may identify the CSI report category without considering the information about the change in reasoning time. The CSI report category identified by the BS may include information about UEs classified based on specific criteria. The CSI report category may include information about a single report timing. For example, category A may include information indicating that there is an additional CSI report timing offset after n time slots, or category B may include information indicating that the CSI report timing is changed after n time slots. However, this is merely an example, and the information about CSI reporting indicated by the BS and contained in one category (for CSI reporting operations) may be unified, or may include specific separate information.

[0099] According to an embodiment, the BS may identify a CSI reporting category corresponding to the UE in the CSI reporting category based on the capability information received from the UE, and may transmit information about the identified CSI reporting category to the UE. According to an embodiment, the BS may identify a CSI reporting category corresponding to the UE in the CSI reporting category based on the information about the change in reasoning time received from the UE, and may transmit information about the identified CSI reporting category to the UE.

[0100] In step 613, 624 or 634, the BS may transmit a reference signal related to the CSI to the UE. The reference signal transmitted by the BS may be a CSI-RS, but is not limited to this embodiment. For example, it is obvious that a synchronization signal block (SSB) may be used instead of a reference signal. Figure 6 The operation of the CSI-RS shown and described is also included in the embodiments of the present disclosure. According to an embodiment, the BS may periodically transmit the CSI-RS to the UE. In the case of aperiodic CSI reporting (see 620) or semi-persistent reporting (see 630), in step 623 or 633, the BS may transmit a trigger signal for transmitting the CSI-RS in advance. According to an embodiment, the trigger signal transmitted by the BS may include a MAC CE or a DCI.

[0101] According to an embodiment, in the case of aperiodic CSI reporting (see 620), the CSI report of the UE may be triggered by DCI. According to an embodiment, in the case of aperiodic CSI reporting (see 620), the CSI report of the UE may be triggered by a subselection indicator of a MAC CE and a DCI. According to an embodiment, in the case of semi-persistent reporting (see 630), the CSI report of the UE may be performed by an activation command based on a MAC CE. According to an embodiment, in the case of semi-persistent CSI reporting (see 630), the CSI report of the UE related to the data channel may be triggered by a DCI.

[0102] According to an embodiment, the trigger signal transmitted by the BS may be a signal used as a reference offset to indicate at least one of transmission / reception timing of a CSI-RS, reporting timing of CSI, and additional reporting timing of CSI.

[0103] In step 614, 625 or 635, the UE may transmit a CSI report to the BS. According to an embodiment, the CSI report transmitted by the UE to the BS may be CSI compressed by AI-based CSI generation. According to an embodiment, the UE may perform CSI reporting based on configuration information of CSI reporting received from the BS. According to an embodiment, the UE may perform CSI reporting based on the configuration information, including at least one operation of CSI reporting identified by the BS in step 622 or 632. According to an embodiment, the UE may transmit a CSI report according to a general CSI reporting procedure without additional configuration information for AI-based CSI reporting. According to an embodiment, the UE may transmit a CSI report at a CSI reporting timing changed based on the configuration information for CSI reporting.

[0104] In step 615, 626 or 636, the UE may transmit an additional CSI report to the BS. According to an embodiment, the UE may identify an additional CSI report offset (in addition to the existing CSI report offset) based on the configuration information about the CSI report received from the BS. The UE may transmit an additional CSI report to the BS based on the additional CSI report offset. Obviously, under the condition based on the configuration information received from the BS, the additional CSI report may be performed to replace the general CSI reporting process, and may be performed as an additional process regardless of whether the CSI reporting process based on the general CSI reporting process is performed.

[0105] In step 616, the BS may transmit a periodic CSI-RS to the UE.

[0106] According to the periodic CSI reporting process, the CSI reporting operations performed in steps 618, 614, and 615 may be similarly performed. According to various embodiments of the present disclosure, it is apparent that the aforementioned steps may be repeatedly performed.

[0107] In step 619, 627 or 637, the UE may transmit information about the change in the new calculation reasoning time to the BS. Specifically, the UE may identify the change in the activated AI function, such as activation or deactivation of another AI function included in the UE, while performing repeated CSI transmission / reception processes with the BS. By activating (or deactivating) another AI function in the AI ​​chip included in the UE, the calculation time of the algorithm used for AI-based CSI generation may be changed. Therefore, the UE may identify the change in the new calculation reasoning time and may transmit its information to the BS. According to an embodiment, the UE may transmit information about the change in reasoning time and information about a reconfiguration request (e.g., an RRC reconstruction message) for requesting new configuration information caused thereby to the BS. In the case of non-periodic CSI reporting (see 620) or semi-persistent CSI reporting (see 630), after step 628 or 638, the CSI reporting operation performed after step 623 or 633 may be similarly performed. According to various embodiments of the present disclosure, the aforementioned steps may be repeatedly performed. According to various embodiments of the present disclosure, the steps may be performed again. Figure 6 each step.

[0108] although Figure 6 Not shown, but according to various embodiments of the present disclosure, the BS or UE may deactivate another AI function during the execution of the aforementioned steps or in at least one step before or after the execution of these steps. Specifically, according to an embodiment, the UE may report information about the state of the AI ​​chip to the BS. According to an embodiment, the UE may transmit information about the state of the AI ​​chip to the BS periodically or when a trigger condition is met. The trigger condition may be any reference value related to the state of the AI ​​chip detected by the UE. For example, the trigger condition configured by the BS for the UE may include a case where the temperature of the AI ​​chip exceeds a specific value, a case where the number of AI functions scheduled in the AI ​​chip exceeds a specific reference, and a case where the calculation time for generating the identified AI-based CSI exceeds a specific reference.

[0109] According to an embodiment, the BS may identify the AI ​​function with the lowest priority among the AI ​​functions included in the AI ​​chip of the UE based on the information about the state of the AI ​​chip received from the UE. The BS may instruct the UE to deactivate the identified AI function with the lowest priority. Since the BS instructs to deactivate the AI ​​function with the lowest priority, the calculation reasoning time of the UE for performing AI-based CSI can be reduced.

[0110] According to an embodiment, the BS may instruct the UE to activate a deactivated AI function included in the AI ​​chip of the UE based on the information on the state of the AI ​​chip received from the UE.

[0111] The operation of reporting the aforementioned information about the state of the AI ​​chip and its specific description may be similar to the operation of reporting the information about the reasoning time change of the UE disclosed in steps 611, 621, or 631.

[0112] According to various embodiments of the present disclosure, a BS and a UE may include Figure 6 At least one of all or some steps disclosed in the related description and some combinations thereof, for improved CSI reporting operation. For example, according to the embodiment, a CSI-RS transmission / reception operation and a CSI reporting operation may be included, but an operation of reporting information on UE reasoning time variation may not be included; or, only a CSI reporting operation may be included, but not a CSI-RS transmission / reception operation or a UE capability information transmission / reception operation.

[0113] Figure 7 The flow chart of the operation of configuring CSI reporting by a BS in a wireless communication system according to an embodiment of the present disclosure is shown. Figure 7 Obviously, in addition to the periodic, aperiodic or semi-persistent CSI reporting process, the operation for CSI reporting is also applicable to various CSI reports between the UE and the BS, as long as it is clearly understood by a person of ordinary skill in the art. In addition, according to various embodiments of the present disclosure, at least one of the combinations of all, some or some steps described below may be included, and obviously, within the achievable range, some combinations of the steps of periodic, aperiodic or semi-persistent CSI reporting are possible. Hereinafter, the operation of CSI reporting is described by including all signal flows of periodic, aperiodic and semi-persistent CSI reporting.

[0114] In step 710, the BS may receive capability information of the UE from the UE. According to an embodiment, the BS may receive capability information including information whether the UE supports an AI function from the UE. The capability information received by the BS may include information whether the UE is capable of generating and reporting AI-based CSI. The capability information received by the BS may be information reported to the BS periodically, or may be information reported to the BS based on a trigger of the BS. According to an embodiment, the capability information received by the BS may be transmitted through high-layer signaling (e.g., radio resource control (RRC) signaling) or a medium access control (MAC) control element (CE).

[0115] According to an embodiment, a BS that has received capability information from a UE may identify whether the UE supports AI functions including AI-based CSI operations (CSI generation or reporting). When the UE does not support the AI ​​function, the BS may transmit configuration information related to non-AI-based CSI reporting to the UE. When the UE supports the AI ​​function, the BS may also transmit configuration information related to AI-based CSI reporting or configuration information related to reporting of calculation of inference time changes (hereinafter interchangeably referred to as inference time change reports). According to an embodiment, the configuration information transmitted by the BS may be transmitted via at least one of higher layer signaling (e.g., RRC signaling), MAC CE, and DCI.

[0116] According to an embodiment, the BS may configure the UE to periodically report the reasoning time change according to a determined timing based on the configuration information related to the reporting of the reasoning time change. According to an embodiment, the BS may configure the UE to detect the state of the AI ​​chip based on the configuration information related to the reporting of the reasoning time change, and report the reasoning time change when the trigger condition is met. According to various embodiments of the present disclosure, without being limited to the aforementioned configuration information, it is obvious that the UE is able to report the reasoning time change regardless of the configuration information of the BS (for example, when it is pre-configured in the UE).

[0117] According to various embodiments of the present disclosure, the aforementioned steps including the UE transmitting capability information or the BS transmitting configuration information about the report on the reasoning time change may be included in whole or in part, or may be omitted.

[0118] In step 720, the BS may receive information about the reasoning time from the UE. According to an embodiment, in addition to the information about the change in the reasoning time, the BS may also receive information about the channel state from the UE. According to an embodiment, the change in the reasoning time of the present disclosure may include a CSI report timing that changes when an AI function including AI-based CSI compression must be performed, compared to the timing of reporting CSI based on general CSI configuration information. According to an embodiment, the change in reasoning time may include a single value, or may include multiple values ​​(e.g., 2 ms to 5 ms), depending on the AI ​​algorithm applied to the AI ​​function (e.g., AI-based CSI compression, etc.). The information about the change in reasoning time received by the BS from the UE may include a change in the reasoning time of the AI-based CSI report or a change in the reasoning time for various AI functions performed by the AI ​​chip of the UE. The information about the change in reasoning time received by the BS may include the computation time required to generate the AI-based CSI.

[0119] According to an embodiment, the UE may measure the calculation time for generating CSI based on the state of the AI ​​chip performing the AI ​​function. Figure 5As shown, even in the case of the same AI function (such as CSI compression, etc.), the calculation time may be determined differently according to the algorithm to be applied. The UE can identify the state of the AI ​​chip to measure the calculation time for compressing and generating CSI. According to an embodiment, the UE may detect the temperature of the AI ​​chip or a change in temperature. Alternatively, the UE may identify the various AI functions being executed in the AI ​​chip or the scheduling of algorithms to be applied to various AI functions. The state of the AI ​​chip identified by the UE may include physical property information of the aforementioned AI chip (such as temperature or a change in temperature, etc.) or information about the algorithm of the AI ​​function. The UE can identify the state of the AI ​​chip in real time, and identify the calculation time for generating AI-based CSI based on the identified state of the AI ​​chip.

[0120] According to an embodiment, the UE may have information about a preset reasoning time. The reasoning time information preset to the UE may include a combination of each pre-measured AI function and the reasoning time information of the AI ​​function. Alternatively, the information about the preset reasoning time may include a mapping relationship between each pre-measured AI function and the reasoning time information of each AI function. According to an embodiment, the UE may identify the calculation time for generating AI-based CSI based on the information about the preset reasoning time and the AI ​​function implemented in the AI ​​chip.

[0121] According to an embodiment, the UE may identify information about the change in reasoning time based on the calculation time for generating the identified AI-based CSI. The UE may transmit information about the change in reasoning time to the BS. According to an embodiment, the UE may transmit information about the change in reasoning time to the BS based on configuration information related to the reasoning time change report received from the BS. According to an embodiment, based on the configuration information related to the reasoning time change report received from the BS, when the trigger condition is met, the UE may transmit information about the change in reasoning time to the BS. The trigger condition may be any reference value related to the state of the AI ​​chip detected by the UE. For example, the trigger condition configured by the BS for the UE may include a case where the temperature of the AI ​​chip exceeds a specific value, a case where the number of AI functions scheduled in the AI ​​chip exceeds a specific reference, and a case where the calculation time for generating the identified AI-based CSI exceeds a specific reference. According to an embodiment, the information about the change in reasoning time transmitted by the UE may be transmitted by at least one of high-level signaling (e.g., RRC signaling), MAC CE, and DCI.

[0122] According to various embodiments of the present disclosure, the aforementioned steps including receiving, by the BS, information about the change in reasoning time or the BS transmitting report configuration information of the change in reasoning time may be included in whole or in part, or may be omitted.

[0123] In step 730, the BS may identify configuration information about CSI reporting and transmit the identification result to the UE. The configuration information transmission of the BS is not limited to the configuration information transmission through high-layer signaling (e.g., RRC signaling). According to an embodiment, the BS may obviously transmit configuration information through MAC or DCI, and may transmit information not only through a downlink control channel but also through a data channel. According to an embodiment, in addition to the configuration information related to the general CSI report, the BS may also transmit additional configuration information identified based on information about the inference time variation to the UE.

[0124] According to an embodiment, in case of aperiodic CSI reporting or semi-persistent CSI reporting, configuration information for CSI reporting transmitted by the BS to the UE may include information indicating receiving CSI-RS and reporting CSI based on a trigger signal (e.g., MAC CE or DCI) received by the UE from the BS. According to various embodiments of the present disclosure, in aperiodic CSI reporting or semi-persistent CSI reporting, configuration information for CSI reporting transmitted by the BS to the UE may include an indicator (e.g., an offset indicator), such as n time slots after the trigger signal related to the CSI-related transmission / reception operation timing of the UE.

[0125] According to an embodiment, the BS may compare the information received from the UE about the inference time change or the change in the channel state with the threshold. The BS may identify the timing for CSI reporting based on the comparison result. According to an embodiment, a situation where the change in the channel state is not significant may include a situation where the channel coherence time is long. According to an embodiment, the change in the channel state may be determined based on whether the channel is stable or the mobility of the UE. According to an embodiment, the change in the channel state may be identified based on an uplink signal (e.g., a sounding reference signal (SRS)) transmitted by the UE to the BS. According to an embodiment, all thresholds described below are not limited to the same threshold and may include corresponding different values. The situation of "exceeding" may include "greater than or equal to".

[0126] According to an embodiment, when the change in the inference time for CSI generation (e.g., compression and generation) is less than a threshold, the BS may perform the CSI reporting process without configuring an additional CSI reporting timing. That is, the BS may transmit configuration information (e.g., configuration information for non-AI-based CSI reporting) to the UE without configuring an additional CSI reporting timing. For example, the BS may determine that the inference time for AI-based CSI generation (e.g., compression and generation) has no significant effect on the CSI reporting timing that the BS is initially to configure, and may not perform configuration of the additional CSI reporting timing in addition to configuring the general CSI reporting timing.

[0127] According to an embodiment, when the change in reasoning time for CSI generation (e.g., compression and generation) is greater than a threshold value but the change in channel state is less than a specific value (e.g., when the channel coherence time is greater than a specific value), the BS may change the CSI reporting timing based on information about the change in reasoning time. According to an embodiment, when the change in reasoning time for CSI generation (e.g., compression and generation) is greater than the change in channel state (e.g., channel coherence time), the BS may identify an additional CSI reporting offset (in addition to the existing CSI reporting offset) based on information about the change in reasoning time. According to an embodiment, in the case of periodic CSI reporting, the BS may change the CSI reporting timing by increasing the period for CSI-RS transmission and CSI reporting. Alternatively, the BS may change the CSI reporting timing by postponing the offset of the CSI reporting.

[0128] According to an embodiment, in the case of non-periodic CSI reporting or semi-persistent CSI reporting, the BS may change the CSI reporting timing by indicating that CSI reporting is performed in n time slots after the trigger signal based on the received information about the change in reasoning time. According to an embodiment, the BS may identify the value with the largest change in each AI algorithm as the change value of the reasoning time based on the received information about the change in reasoning time, and may identify the change value of the identified reasoning time as the changed CSI reporting timing in addition to the existing CSI reporting timing. Since the need to perform CSI reporting immediately is relatively low in an environment where the channel state does not change rapidly, the BS may identify the timing obtained by adding the change value of the reasoning time to the existing CSI reporting timing, and allocate the timing to the UE instead of configuring an additional CSI report, thereby effectively performing the process for resource allocation. According to an embodiment, the BS may transmit configuration information to the UE by including information indicating that the UE is capable of performing CSI reporting in the identified changed CSI reporting timing.

[0129] According to an embodiment, when the change in the inference time for CSI generation (e.g., compression and generation) is greater than a threshold and the change in the channel state is greater than a specific value (e.g., when the channel coherence time is less than a specific value), the BS can identify (in addition to the existing CSI report offset) an additional CSI report offset based on information about the change in inference time. In the case where the channel state changes significantly, it may be difficult to reallocate the CSI report when the CSI report timing is postponed or missed. Therefore, the BS can identify the offset for the additional CSI report. According to an embodiment, with respect to AI-based CSI generation (e.g., compression and generation), the UE may not be able to complete the CSI calculation (e.g., compression and generation) at the timing of the first CSI report configured by the BS.

[0130] In this case, the UE may not be able to perform CSI reporting at the allocated timing of the first CSI report. However, when an offset for additional CSI reporting (e.g., timing for the second CSI report) is configured, the UE may perform CSI reporting at the timing for the second CSI report. For example, when the inference time variation is 2 ms to 5 ms, the BS may also recognize additional CSI reporting timing offsets corresponding to 2 ms and 5 ms in addition to the CSI reporting timing. According to an embodiment, when the UE is able to perform CSI reporting based on the timing for the first CSI report, the BS may reallocate resources allocated for the timing of the second CSI report to the UE. According to an embodiment, in order to indicate the timing of the second CSI report, the BS may transmit information related to the additional CSI reporting timing. For example, the information related to the additional CSI reporting timing transmitted by the BS may include an indicator indicating n time slots after the first CSI reporting timing. In addition, the number of additional CSI reporting timings is not limited to one, and the information related to the CSI reporting timing transmitted by the BS may include multiple additional CSI reporting timings and indicators thereof.

[0131] According to an embodiment, the BS may transmit information about the additional CSI reporting timing to the UE in advance (e.g., RRC signaling) before receiving information about the change in reasoning time from the UE. According to an embodiment, the BS may configure the information about the pre-configured additional CSI reporting timing to the UE before performing (e.g., generating and compressing) AI-based CSI, or may configure the information through a reconfiguration process when performing AI-based CSI. When the information about the change in reasoning time identified by the UE is reported to the BS, the UE that has pre-configured the information about the additional CSI reporting timing may identify the additional CSI reporting timing based on the identified information about the change in reasoning time and the information configured from the BS.

[0132] For example, the information about the additional CSI reporting timing preconfigured by the BS for the UE may include an additional reporting timing indicator. When the indicator is set to 0, the UE may recognize the lowest value (e.g., 2 ms) of the inference time variation (e.g., 2 ms to 5 ms) as the first reporting timing, and the highest value (e.g., 5 ms) as the second reporting timing, and may report the CSI based on the identified reporting timing. When the indicator is set to 1, the UE may recognize the average value (e.g., 3.5 ms) of the inference time variation (e.g., 2 ms to 5 ms) as the first reporting timing, and the highest value (e.g., 5 ms) as the second reporting timing, and may report the CSI based on the identified reporting timing. However, this is merely an example, and it is apparent that the information about the CSI reporting timing preconfigured to the UE and the indicator contained in the information may be implemented with a variety of parameters or contents.

[0133] According to an embodiment, the BS may transmit configuration information including information on the identified additional CSI reporting timing to the UE in addition to the CSI reporting offset. According to an embodiment, the BS may transmit a control indication related to the information on the identified additional CSI reporting timing to the UE.

[0134] According to an embodiment, when a change in inference time for CSI generation (e.g., compression and generation) is greater than a threshold value and a change in channel state is greater than a specific value (e.g., when the channel coherence time is less than a specific value), the BS may deactivate AI-based CSI feedback (e.g., AI-based CSI compression and generation). According to an embodiment, the threshold value of the change in inference time may include a threshold value prepared for a change in inference time that is too large to identify a report timing for AI-based CSI reporting.

[0135] When the change in reasoning time includes a significantly large value, the BS may recognize that it cannot support the AI-based CSI compression operation of the UE, and may therefore instruct the UE to deactivate the AI-based CSI function. According to an embodiment, after deactivating the AI-based CSI function, the BS may receive information about the change in reasoning time from the UE again, and may compare it with the threshold. The BS may identify whether to activate the AI-based CSI function based on the comparison result between the change in reasoning time or the change in channel state and each threshold. When specific conditions are met, the BS may instruct the UE to activate the AI-based CSI function, and may perform the above operation again. Figure 7 steps.

[0136] According to an embodiment, the BS may identify the CSI report category based on information about the change in reasoning time received from the UE. Obviously, the BS may identify the CSI report category without considering the information about the change in reasoning time. The CSI report category identified by the BS may include information about UEs classified based on specific criteria. The CSI report category may include information about a single report timing. For example, category A may include information indicating that there is an additional CSI report timing offset after n time slots, or category B may include information indicating that the CSI report timing is changed after n time slots.

[0137] However, this is only an example, and information about CSI reports indicated by the BS and included in one category (for operation of CSI reporting) may be unified or may include specific separate information. According to an embodiment, the BS may identify a CSI reporting category corresponding to the UE in the CSI reporting category based on capability information received from the UE, and may transmit information about the identified CSI reporting category to the UE. According to an embodiment, the BS may identify a CSI reporting category corresponding to the UE in the CSI reporting category based on information about a change in reasoning time received from the UE, and may transmit information about the identified CSI reporting category to the UE.

[0138] According to various embodiments of the present disclosure, the aforementioned steps including the BS identifying configuration information about CSI reporting or the BS transmitting report configuration information of inference time variation may be included in whole or in part, or may be omitted.

[0139] In step 740, the BS may transmit a CSI-RS to the UE and receive a CSI report from the UE. According to an embodiment, the BS may transmit a reference signal related to the CSI to the UE. The reference signal transmitted by the BS may be a CSI-RS, but is not limited to this embodiment. For example, it is obvious that a synchronization signal block (SSB) may be used instead of a reference signal. Figure 7 The operation of the CSI-RS shown and described is also included in the embodiments of the present disclosure. According to an embodiment, the BS may periodically transmit the CSI-RS to the UE. In the case of aperiodic CSI reporting or semi-persistent reporting, the BS may transmit a trigger signal for transmitting the CSI-RS in advance. According to an embodiment, the trigger signal transmitted by the BS may include a MAC CE or a DCI. According to an embodiment, the trigger signal transmitted by the BS may be a signal used as a reference offset to indicate at least one of the transmission / reception timing of the CSI-RS, the reporting timing of the CSI, and the additional reporting timing of the CSI.

[0140] According to an embodiment, the BS may receive a CSI report from the UE. According to an embodiment, the CSI report received by the BS from the UE may be CSI compressed by AI-based CSI generation. According to an embodiment, the UE may perform CSI reporting based on configuration information of the CSI report received from the BS. According to an embodiment, the UE may perform CSI reporting based on configuration information including at least one operation for CSI reporting identified by the BS. According to an embodiment, the UE may transmit a CSI report according to a general CSI reporting process without additional configuration information for AI-based CSI reporting. According to an embodiment, the UE may transmit a CSI report at a CSI reporting timing changed based on the configuration information for CSI reporting. According to an embodiment, the BS may receive an additional CSI report from the UE. According to an embodiment, the UE may identify an additional CSI reporting offset (in addition to an existing CSI reporting offset) based on the configuration information about the CSI report received from the BS. The UE may transmit an additional CSI report to the BS based on the additional CSI reporting offset.

[0141] According to the embodiment, although Figure 7 Not shown, but the BS can receive information about the change of the new calculation reasoning time from the UE. Specifically, the UE can identify the change of the activated AI function, such as the activation or deactivation of another AI function included in the UE, while performing repeated CSI transmission / reception processes with the BS. By activating (or deactivating) another AI function in the AI ​​chip included in the UE, the calculation time of the algorithm for AI-based CSI generation can be changed. Therefore, the UE can identify the change of the new calculation reasoning time and can transmit its information to the BS. According to an embodiment, the UE can transmit information about the change of the reasoning time and information about the reconfiguration request (e.g., RRC reconstruction message) caused thereby for requesting new configuration information to the BS.

[0142] According to various embodiments of the present disclosure, although Figure 7 Not shown, the BS or UE may deactivate another AI function during the execution of the aforementioned steps or in at least one step before or after the execution of these steps. Specifically, according to an embodiment, the UE may report information about the state of the AI ​​chip to the BS. According to an embodiment, the UE may transmit information about the state of the AI ​​chip to the BS periodically or when a trigger condition is met. The trigger condition may be any reference value related to the state of the AI ​​chip detected by the UE. For example, the trigger condition configured by the BS for the UE may include a case where the temperature of the AI ​​chip exceeds a specific value, a case where the number of AI functions scheduled in the AI ​​chip exceeds a specific reference, and a case where the calculation time for generating the identified AI-based CSI exceeds a specific reference.

[0143] According to an embodiment, the BS may identify the AI ​​function with the lowest priority among the AI ​​functions included in the AI ​​chip of the UE based on the information about the status of the AI ​​chip received from the UE. The BS may instruct the UE to deactivate the identified AI function with the lowest priority. Since the BS instructs the deactivation of the AI ​​function with the lowest priority, the computational reasoning time of the UE for performing AI-based CSI may be reduced. According to an embodiment, the BS may instruct the UE to activate the deactivated AI function included in the AI ​​chip of the UE based on the information about the status of the AI ​​chip received from the UE. The operation of reporting the aforementioned information about the status of the AI ​​chip and its specific description may be similar to that in Figure 7 The operation of reporting information about the UE's inference time variation is disclosed.

[0144] According to various embodiments of the present disclosure, it is obvious that the aforementioned steps can be performed repeatedly. According to various embodiments of the present disclosure, Figure 7 Some of the various steps in the may be omitted or performed again. According to various embodiments of the present disclosure, the aforementioned steps including the BS receiving information about the change in reasoning time or the BS transmitting report configuration information about the change in reasoning time may be included in whole or in part, or may be omitted.

[0145] Figure 8 The flow chart of the operation for configuring the CSI report of the UE in the wireless communication system according to the embodiment of the present disclosure is shown. Figure 8 Obviously, in addition to the periodic, aperiodic or semi-persistent CSI reporting process, the operation for CSI reporting is also applicable to various CSI reports between the UE and the BS, as long as it is clearly understood by a person of ordinary skill in the art. In addition, according to various embodiments of the present disclosure, at least one of the combinations of all, some or some steps described below may be included, and obviously, within the achievable range, some combinations of the steps of periodic, aperiodic or semi-persistent CSI reporting are possible. Hereinafter, the operation of CSI reporting is described by including all signal flows of periodic, aperiodic and semi-persistent CSI reporting.

[0146] In step 810, the UE may transmit capability information of the UE to the BS. According to an embodiment, the UE may transmit capability information including information on whether the UE supports an AI function to the BS. The capability information transmitted by the UE may include information on whether the UE is capable of generating and reporting CSI based on AI. The capability information transmitted by the UE may be information reported to the BS periodically, or may be information reported to the BS based on a trigger of the BS. According to an embodiment, the capability information transmitted by the UE may be transmitted via high-layer signaling (e.g., RRC signaling) or MAC CE.

[0147] According to an embodiment, the BS that has received the capability information from the UE can identify whether the UE supports an AI function including AI-based CSI operation (CSI generation or reporting). When the UE does not support the AI function, the BS can transmit configuration information related to non-AI-based CSI reporting to the UE. When the UE supports the AI function, the BS can also transmit configuration information related to AI-based CSI reporting or configuration information about the report on the change in the calculation inference time (hereinafter interchangeably referred to as the inference time change report). According to an embodiment, the configuration information transmitted by the BS can be transmitted by at least one of high-layer signaling (e.g., RRC signaling), MAC CE, and DCI.

[0148] According to an embodiment, the BS can configure the UE to periodically report the inference time change according to a determined timing based on the configuration information related to the report on the inference time change. According to an embodiment, the BS can configure the UE to detect the state of the AI chip based on the configuration information related to the report on the inference time change and report the inference time change when a trigger condition is met. According to various embodiments of the present disclosure, and not limited to the foregoing configuration information, it is obvious that the UE can report the inference time change independently of the configuration information of the BS (e.g., when it is pre-configured in the UE).

[0149] According to various embodiments of the present disclosure, the foregoing steps including the UE transmitting the capability information or the BS transmitting the configuration information about the inference time change may be included in whole or in part, or may be omitted.

[0150] In step 820, the UE can measure the inference time and can transmit information about the inference time to the BS. According to an embodiment, in addition to the information about the change in the inference time, the UE can also transmit information about the channel state to the BS. According to an embodiment, the change in the inference time of the present disclosure can include the CSI reporting timing that changes compared to the timing of reporting CSI based on general CSI configuration information when an AI function including AI-based CSI compression must be performed. According to an embodiment, the change in the inference time can include a single value or can include multiple values (e.g., from N ms to M ms, where N and M are positive integers and N < M)), depending on the AI algorithm applied to the AI function (e.g., AI-based CSI compression, etc.).

[0151] When the change in reasoning time includes a single value, the single value may be a value representing multiple values ​​depending on the AI ​​algorithm, and may be, for example, at least one of the sum, weighted sum, average value, and highest value of multiple values. The information about the change in reasoning time transmitted by the UE to the BS may include the change in reasoning time of the AI-based CSI report or the change in reasoning time of various AI functions executed in the AI ​​chip of the UE. The information about the change in reasoning time transmitted by the UE may include the calculation time required to generate the AI-based CSI. In addition, the information about the change in reasoning time transmitted by the UE may include the difference (or offset) of the calculation time required to generate the AI-based CSI (or the time for generating the non-AI-based CSI) before the report, or may include the absolute value of the calculation time for generating the AI-based CSI. Alternatively, the information about the change in reasoning time transmitted by the UE may include information about a quantized value that divides the difference in the calculation time required to generate the AI-based CSI (or the time for generating the non-AI-based CSI) before the report into different values ​​for each segment.

[0152] According to an embodiment, the UE may measure the calculation time for generating CSI based on the state of the AI ​​chip performing the AI ​​function. Figure 5 As shown, even in the case of the same AI function (such as CSI compression, etc.), the calculation time may be determined differently according to the algorithm to be applied. The UE can identify the state of the AI ​​chip to measure the calculation time for compressing and generating CSI. According to an embodiment, the UE may detect the temperature of the AI ​​chip or a change in temperature. Alternatively, the UE may identify the various AI functions being executed in the AI ​​chip or the scheduling of algorithms to be applied to various AI functions. The state of the AI ​​chip identified by the UE may include physical property information of the aforementioned AI chip (such as temperature or a change in temperature, etc.) or information about the algorithm of the AI ​​function. The UE can identify the state of the AI ​​chip in real time, and identify the calculation time for generating AI-based CSI based on the identified state of the AI ​​chip.

[0153] According to an embodiment, the UE may have information about a preset reasoning time. The reasoning time information preset to the UE may include a combination of each pre-measured AI function and the reasoning time information of the AI ​​function. Alternatively, the information about the preset reasoning time may include a mapping relationship between each pre-measured AI function and the reasoning time information of each AI function. According to an embodiment, the UE may identify the calculation time for generating AI-based CSI based on the information about the preset reasoning time and the AI ​​function implemented in the AI ​​chip.

[0154] According to an embodiment, the UE may identify information about the reasoning time change based on the calculation time used to generate the identified AI-based CSI. The UE may transmit information about the identified reasoning time change to the BS. According to an embodiment, the UE may transmit information about the reasoning time change to the BS based on configuration information related to the reasoning time change report received from the BS. According to an embodiment, based on the configuration information related to the reasoning time change report received from the BS, when the trigger condition is met, the UE may transmit information about the reasoning time change to the BS. The trigger condition may be any reference value related to the state of the AI ​​chip detected by the UE.

[0155] For example, the trigger condition configured by the BS for the UE may include a case where the temperature of the AI ​​chip exceeds a specific value, a case where the number of AI functions scheduled in the AI ​​chip exceeds a specific reference, and a case where the calculation time for generating the identified AI-based CSI exceeds a specific reference. According to an embodiment, the information about the change in reasoning time transmitted by the UE may be transmitted through at least one of high-layer signaling (e.g., RRC signaling), MAC CE, and DCI.

[0156] According to various embodiments of the present disclosure, the aforementioned steps including the UE measuring the change of the reasoning time or the BS receiving the information about the change of the reasoning time may be included in whole or in part, or may be omitted.

[0157] In step 830, the UE may receive configuration information about CSI reporting from the BS. The UE's reception of configuration information is not limited to configuration information transmission through high-layer signaling (e.g., RRC signaling). According to an embodiment, the UE may obviously receive configuration information through MAC or DCI, and may receive information not only through a downlink control channel but also through a data channel. According to an embodiment, in addition to configuration information related to general CSI reporting, the UE may also receive additional configuration information identified based on information about inference time variation from the BS.

[0158] According to an embodiment, in case of aperiodic CSI reporting or semi-persistent CSI reporting, the configuration information for CSI reporting received by the UE from the BS may include information indicating receiving CSI-RS and reporting CSI based on a trigger signal (e.g., MAC CE or DCI) received by the UE from the BS. According to various embodiments of the present disclosure, in aperiodic CSI reporting or semi-persistent CSI reporting, the configuration information for CSI reporting received by the UE from the BS may include an indicator, such as n time slots after the trigger signal related to the CSI-related transmission / reception operation timing of the UE.

[0159] According to an embodiment, the BS may compare the information received from the UE about the inference time change or the change in the channel state with the threshold. The BS may identify the timing for CSI reporting based on the comparison result. According to an embodiment, a situation where the change in the channel state is not significant may include a situation where the channel coherence time is long. According to an embodiment, the change in the channel state may be determined based on whether the channel is stable or the mobility of the UE. According to an embodiment, the change in the channel state may be identified based on an uplink signal (e.g., a sounding reference signal (SRS)) transmitted by the UE to the BS. According to an embodiment, all thresholds described below are not limited to the same threshold and may include corresponding different values. The situation of "exceeding" may include "greater than or equal to".

[0160] According to an embodiment, when the change in the inference time for CSI generation (e.g., compression and generation) is less than a threshold, the BS may perform the CSI reporting process without configuring an additional CSI reporting timing. That is, the BS may transmit configuration information (e.g., configuration information for non-AI-based CSI reporting) to the UE without configuring an additional CSI reporting timing. For example, the BS may determine that the inference time for AI-based CSI generation (e.g., compression and generation) has no significant effect on the CSI reporting timing that the BS is initially to configure, and may not perform configuration of the additional CSI reporting timing in addition to configuring the general CSI reporting timing.

[0161] According to an embodiment, when the change in reasoning time for CSI generation (e.g., compression and generation) is greater than a threshold value but the change in channel state is less than a specific value (e.g., when the channel coherence time is greater than a specific value), the BS may change the CSI reporting timing based on information about the change in reasoning time. According to an embodiment, when the change in reasoning time for CSI generation (e.g., compression and generation) is greater than the change in channel state (e.g., channel coherence time), the BS may identify an additional CSI reporting offset (in addition to the existing CSI reporting offset) based on information about the change in reasoning time. According to an embodiment, in the case of periodic CSI reporting, the BS may change the CSI reporting timing by increasing the period for CSI-RS transmission and CSI reporting. Alternatively, the BS may change the CSI reporting timing by postponing the offset of the CSI reporting.

[0162] According to an embodiment, in the case of non-periodic CSI reporting or semi-persistent CSI reporting, the BS may change the CSI reporting timing by indicating that CSI reporting is performed in n time slots after the trigger signal based on the received information about the change in reasoning time. According to an embodiment, the BS may identify the value with the largest change in each AI algorithm as the change value of the reasoning time based on the received information about the change in reasoning time, and may identify the change value of the identified reasoning time as the changed CSI reporting timing in addition to the existing CSI reporting timing. Since the need to perform CSI reporting immediately is relatively low in an environment where the channel state does not change rapidly, the BS may identify the timing obtained by adding the change value of the reasoning time to the existing CSI reporting timing, and allocate the timing to the UE instead of configuring an additional CSI report, thereby effectively performing the process for resource allocation. According to an embodiment, the BS may transmit configuration information to the UE by including information indicating that the UE is capable of performing CSI reporting in the identified changed CSI reporting timing.

[0163] According to an embodiment, when the change in the inference time for CSI generation (e.g., compression and generation) is greater than a threshold and the change in the channel state is greater than a specific value (e.g., when the channel coherence time is less than a specific value), the BS can identify (in addition to the existing CSI report offset) an additional CSI report offset based on information about the change in inference time. In the case where the channel state changes significantly, it may be difficult to reallocate the CSI report when the CSI report timing is postponed or missed. Therefore, the BS can identify the offset for the additional CSI report. According to an embodiment, with respect to AI-based CSI generation (e.g., compression and generation), the UE may not be able to complete the CSI calculation (e.g., compression and generation) at the timing of the first CSI report configured by the BS.

[0164] In this case, the UE may not be able to perform CSI reporting at the allocated timing of the first CSI report. However, when an offset for additional CSI reporting (e.g., timing for the second CSI report) is configured, the UE may perform CSI reporting at the timing for the second CSI report. For example, when the inference time variation is 2 ms to 5 ms, the BS may also recognize additional CSI reporting timing offsets corresponding to 2 ms and 5 ms in addition to the CSI reporting timing. According to an embodiment, when the UE is able to perform CSI reporting based on the timing for the first CSI report, the BS may reallocate resources allocated for the timing of the second CSI report to the UE. According to an embodiment, in order to indicate the timing of the second CSI report, the BS may transmit information related to the additional CSI reporting timing. For example, the information related to the additional CSI reporting timing transmitted by the BS may include an indicator indicating n time slots after the first CSI reporting timing. In addition, the number of additional CSI reporting timings is not limited to one, and the information related to the CSI reporting timing transmitted by the BS may include multiple additional CSI reporting timings and indicators thereof.

[0165] According to an embodiment, the BS may transmit information about the additional CSI reporting timing to the UE in advance (e.g., RRC signaling) before receiving information about the change in reasoning time from the UE. According to an embodiment, the BS may configure the pre-configured information about the additional CSI reporting timing to the UE before performing (e.g., generating and compressing) AI-based CSI, or may configure the information through a reconfiguration process when performing AI-based CSI. When the information about the change in reasoning time identified by the UE is reported to the BS, the UE that has pre-configured the information about the additional CSI reporting timing may identify the additional CSI reporting timing based on the information about the identified change in reasoning time and the information configured from the BS.

[0166] For example, the information about the additional CSI reporting timing preconfigured by the BS for the UE may include an additional reporting timing indicator. When the indicator is set to 0, the UE may recognize the lowest value (e.g., 2 ms) of the inference time variation (e.g., 2 ms to 5 ms) as the first reporting timing, and the highest value (e.g., 5 ms) as the second reporting timing, and may report the CSI based on the identified reporting timing. When the indicator is set to 1, the UE may recognize the average value (e.g., 3.5 ms) of the inference time variation (e.g., 2 ms to 5 ms) as the first reporting timing, and the highest value (e.g., 5 ms) as the second reporting timing, and may report the CSI based on the identified reporting timing. However, this is merely an example, and it is apparent that the information about the CSI reporting timing preconfigured to the UE and the indicator contained in the information may be implemented with a variety of parameters or contents.

[0167] According to an embodiment, the BS may transmit configuration information including information on the identified additional CSI reporting timing to the UE in addition to the CSI reporting offset. According to an embodiment, the BS may transmit a control indication related to the information on the identified additional CSI reporting timing to the UE.

[0168] According to an embodiment, when the change in reasoning time for CSI generation (e.g., compression and generation) is greater than a threshold and the change in channel state is greater than a specific value (e.g., when the channel coherence time is less than a specific value), the BS may deactivate AI-based CSI feedback (e.g., AI-based CSI compression and generation). According to an embodiment, the threshold of the change in reasoning time may include a threshold prepared for when the change in reasoning time is too large to identify the reporting timing for AI-based CSI reporting. When the change in reasoning time includes a significantly large value, the BS may recognize that it cannot support the AI-based CSI compression operation of the UE, and may therefore instruct the UE to deactivate the AI-based CSI function. According to an embodiment, after deactivating the AI-based CSI function, the BS may receive information about the change in reasoning time from the UE again, and may compare it with the threshold. The BS may identify whether to activate the AI-based CSI function based on the comparison result between the change in reasoning time or the change in channel state and each threshold. When a specific condition is met, the BS may instruct the UE to activate the AI-based CSI function, and may perform again Figure 6 steps.

[0169] According to an embodiment, the BS may identify the CSI report category based on information about the change in reasoning time received from the UE. Obviously, the BS may identify the CSI report category without considering the information about the change in reasoning time. The CSI report category identified by the BS may include information about UEs classified based on specific criteria. The CSI report category may include information about a single report timing. For example, category A may include information indicating that there is an additional CSI report timing offset after n time slots, or category B may include information indicating that the CSI report timing is changed after n time slots.

[0170] However, this is only an example, and information about CSI reports indicated by the BS and included in one category (for operation of CSI reporting) may be unified or may include specific separate information. According to an embodiment, the BS may identify a CSI reporting category corresponding to the UE in the CSI reporting category based on capability information received from the UE, and may transmit information about the identified CSI reporting category to the UE. According to an embodiment, the BS may identify a CSI reporting category corresponding to the UE in the CSI reporting category based on information about a change in reasoning time received from the UE, and may transmit information about the identified CSI reporting category to the UE.

[0171] According to various embodiments of the present disclosure, the aforementioned steps including the BS identifying the configuration information about the CSI report or the UE receiving the configuration information about the reasoning time variation may be included in whole or in part, or may be omitted.

[0172] In step 840, the UE may receive a CSI-RS from the BS and may transmit a CSI report to the BS. According to an embodiment, the BS may receive a reference signal related to the CSI from the BS. The reference signal received by the UE may be a CSI-RS. According to an embodiment, the UE may periodically receive a CSI-RS from the BS. In the case of aperiodic CSI reporting or semi-persistent reporting, the UE may receive a trigger signal for transmitting the CSI-RS in advance. According to an embodiment, the trigger signal received by the UE may include a MAC CE or a DCI. According to an embodiment, the trigger signal received by the UE may be a signal used as a reference offset to indicate at least one of the transmission / reception timing of the CSI-RS, the reporting timing of the CSI, and the additional reporting timing of the CSI.

[0173] According to an embodiment, the UE may transmit a CSI report to the BS. According to an embodiment, the CSI report transmitted by the UE to the BS may be a CSI compressed by the AI-based CSI generation. According to an embodiment, the UE may perform CSI reporting based on the configuration information of the CSI report received from the BS. According to an embodiment, the UE may perform CSI reporting based on the configuration information, the configuration information including at least one operation for CSI reporting identified by the BS. According to an embodiment, the UE may transmit a CSI report according to a general CSI reporting process without additional configuration information for AI-based CSI reporting. According to an embodiment, the UE may transmit a CSI report at a CSI report timing changed based on the configuration information for CSI reporting. According to an embodiment, the UE may transmit an additional CSI report to the BS.

[0174] According to an embodiment, the UE may identify an additional CSI reporting offset (in addition to the existing CSI reporting offset) based on the configuration information about the CSI report received from the BS. The UE may transmit an additional CSI report to the BS based on the additional CSI reporting offset. Obviously, under the condition based on the configuration information received from the BS, the additional CSI report may be performed to replace the general CSI reporting process, and may be performed as an additional process regardless of whether the CSI reporting process based on the general CSI reporting process is performed.

[0175] According to the embodiment, although Figure 8 Not shown, but the BS can receive information about the change of the new calculation reasoning time from the UE. Specifically, the UE can identify the change of the activated AI function, such as the activation or deactivation of another AI function included in the UE, while performing repeated CSI transmission / reception processes with the BS. By activating (or deactivating) another AI function in the AI ​​chip included in the UE, the calculation time of the algorithm for AI-based CSI generation can be changed. Therefore, the UE can identify the change of the new calculation reasoning time and can transmit its information to the BS. According to an embodiment, the UE can transmit information about the change of the reasoning time and information about the reconfiguration request (e.g., RRC reconstruction message) caused thereby for requesting new configuration information to the BS.

[0176] According to various embodiments of the present disclosure, although Figure 8 Not shown, the BS or UE may deactivate another AI function during the execution of the aforementioned steps or in at least one step before or after the execution of these steps. Specifically, according to an embodiment, the UE may report information about the state of the AI ​​chip to the BS. According to an embodiment, the UE may transmit information about the state of the AI ​​chip to the BS periodically or when a trigger condition is met. The trigger condition may be any reference value related to the state of the AI ​​chip detected by the UE. For example, the trigger condition configured by the BS for the UE may include a case where the temperature of the AI ​​chip exceeds a specific value, a case where the number of AI functions scheduled in the AI ​​chip exceeds a specific reference, and a case where the calculation time for generating the identified AI-based CSI exceeds a specific reference.

[0177] According to an embodiment, the BS may identify the AI ​​function with the lowest priority among the AI ​​functions included in the AI ​​chip of the UE based on the information about the status of the AI ​​chip received from the UE. The BS may instruct the UE to deactivate the identified AI function with the lowest priority. Since the BS instructs the deactivation of the AI ​​function with the lowest priority, the computational reasoning time of the UE for performing AI-based CSI may be reduced. According to an embodiment, the BS may instruct the UE to activate the deactivated AI function included in the AI ​​chip of the UE based on the information about the status of the AI ​​chip received from the UE. The operation of reporting the aforementioned information about the status of the AI ​​chip and its specific description may be similar to that in Figure 8 The operation of reporting information about the UE's inference time variation is disclosed.

[0178] According to various embodiments of the present disclosure, it is obvious that the aforementioned steps can be performed repeatedly. According to various embodiments of the present disclosure, Figure 8 Some of the various steps in the may be omitted or performed again. According to various embodiments of the present disclosure, the aforementioned steps including the BS receiving information about the change in reasoning time or the BS transmitting report configuration information about the change in reasoning time may be included in whole or in part, or may be omitted.

[0179] According to various embodiments of the present disclosure, a method performed by a base station in a wireless communication system may include: receiving capability information indicating whether the terminal supports generation of channel state information (CSI) based on artificial intelligence (AI) from a terminal; receiving information about a change in reasoning time related to AI-based CSI generation and information about a change in channel state from the terminal based on the capability information; identifying information about CSI report timing based on the information about the change in reasoning time and the change in channel state; transmitting radio resource control (RRC) configuration information including the identified CSI report timing to the terminal; transmitting a CSI reference signal (RS) to the terminal; and receiving a CSI report from the terminal, the CSI report being based on the CSI-RS and the information about the CSI report timing.

[0180] According to an embodiment, identifying information about CSI reporting timing may include: when the change in reasoning time exceeds a first threshold and the change in channel state is less than a second threshold, identifying an offset for CSI reporting as a value greater than the change in reasoning time, or when the change in reasoning time exceeds a first threshold and the change in channel state is greater than a second threshold, identifying multiple offsets for CSI reporting.

[0181] According to an embodiment, the method may further include: when the change in inference time exceeds a third threshold and the change in channel state is greater than a fourth threshold, identifying whether to deactivate AI-based CSI generation of the terminal, and transmitting information indicating deactivation of AI-based CSI generation of the terminal to the terminal based on the identification result.

[0182] According to an embodiment, the method may further include: identifying whether to deactivate at least one of the different AI functions included in the terminal based on the change in the inference time; and transmitting information indicating deactivation of at least one of the different functions included in the terminal to the terminal based on the identification result.

[0183] According to an embodiment, the method may further include: transmitting a trigger signal related to the aperiodic CSI report to the terminal. The trigger signal may be one of a medium access control (MAC) control element (CE) and downlink control information (DCI).

[0184] According to various embodiments of the present disclosure, a method performed by a terminal in a wireless communication system may include: transmitting capability information indicating whether the terminal supports AI-based CSI generation to a base station; identifying status information of an AI chip that performs AI-based CSI generation; identifying changes in reasoning time related to AI-based CSI generation based on the identified status information of the AI ​​chip; transmitting information about the identified changes in reasoning time related to CSI generation and information about changes in channel states to the base station; receiving RRC configuration information including CSI report timing from the base station, the CSI report timing being based on information about changes in reasoning time and changes in channel states; receiving CSI-RS from the base station; and transmitting a CSI report to the base station based on the CSI-RS and the information about the CSI report timing.

[0185] According to an embodiment, when the change in reasoning time exceeds a first threshold and the change in channel state is less than a second threshold, the information about CSI report timing received from the base station may include an offset of the CSI report having a value greater than the change in reasoning change; or when the change in reasoning time exceeds a first threshold and the change in channel state is greater than a second threshold, the information about CSI report timing received from the base station may include multiple offsets for CSI reporting.

[0186] According to an embodiment, the method may further include: receiving, from the base station, information indicating deactivation of AI-based CSI generation of the terminal based on a change in inference time and a change in a channel state.

[0187] According to an embodiment, the method may further include: receiving, from the base station, information indicating deactivation of at least one of the different AI functions included in the terminal based on a change in the inference time.

[0188] According to an embodiment, the method may further include receiving a trigger signal related to aperiodic CSI reporting from a base station. The trigger signal may be one of MAC CE and DCI.

[0189] According to various embodiments of the present disclosure, a base station in a wireless communication system may include at least one transceiver and at least one processor operably connected to the at least one transceiver. The at least one processor may be configured to: receive capability information indicating whether the terminal supports the generation of AI-based CSI from a terminal; receive information about changes in reasoning time related to AI-based CSI generation and information about changes in channel status from the terminal based on the capability information; identify information about CSI report timing based on information about changes in reasoning time and changes in channel status; transmit RRC configuration information including the identified CSI report timing to the terminal; transmit CSI-RS to the terminal; and receive a CSI report from the terminal, the CSI report being based on the CSI-RS and information about CSI report timing.

[0190] According to an embodiment, in order to identify information about CSI report timing, at least one processor may be configured to: identify an offset for CSI reporting as a value greater than the change in reasoning time when the change in reasoning time exceeds a first threshold and the change in channel state is less than a second threshold, or identify multiple offsets for CSI reporting when the change in reasoning time exceeds a first threshold and the change in channel state is greater than a second threshold.

[0191] According to an embodiment, at least one processor may be configured to: identify whether to deactivate AI-based CSI generation of the terminal when a change in inference time exceeds a third threshold and a change in channel state is greater than a fourth threshold, and transmit information indicating deactivation of AI-based CSI generation of the terminal to the terminal based on the identification result.

[0192] According to an embodiment, at least one processor may be configured to: identify whether to deactivate at least one of the different AI functions included in the terminal based on a change in inference time; and transmit information indicating deactivation of at least one of the different functions included in the terminal to the terminal based on the identification result.

[0193] According to an embodiment, at least one processor may be configured to transmit a trigger signal related to aperiodic CSI reporting to the terminal. The trigger signal may be one of MAC CE and DCI.

[0194] According to various embodiments of the present disclosure, a terminal in a wireless communication system may include at least one transceiver and at least one processor operably connected to the at least one transceiver. The at least one processor may be configured to: transmit capability information indicating whether the terminal supports AI-based CSI generation to a base station; identify state information of an AI chip that performs AI-based CSI generation; identify changes in reasoning time associated with AI-based CSI generation based on the identified state information of the AI ​​chip; transmit information about the identified changes in reasoning time associated with CSI generation and information about changes in channel state to the base station; receive RRC configuration information including CSI report timing from the base station, the CSI report timing being based on information about changes in reasoning time and changes in channel state; receive CSI-RS from the base station; and transmit a CSI report to the base station based on the CSI-RS and information about CSI report timing.

[0195] According to an embodiment, when the change in reasoning time exceeds a first threshold and the change in channel state is less than a second threshold, the information about CSI report timing received from the base station may include an offset of the CSI report having a value greater than the change in reasoning change; or when the change in reasoning time exceeds a first threshold and the change in channel state is greater than a second threshold, the information about CSI report timing received from the base station may include multiple offsets for CSI reporting.

[0196] According to an embodiment, at least one processor may be configured to receive, from a base station, information indicating deactivation of AI-based CSI generation of a terminal based on a change in inference time and a change in a channel state.

[0197] According to an embodiment, at least one processor may be configured to receive, from a base station, information indicating deactivation of at least one of different AI functions included in the terminal based on a change in reasoning time.

[0198] According to an embodiment, at least one processor may be further configured to receive a trigger signal related to aperiodic CSI reporting from a base station. The trigger signal may be one of a MAC CE and a DCI.

[0199] The methods according to the embodiments disclosed in the claims and / or the specification of the present disclosure may be implemented in hardware, software or a combination of both.

[0200] When implemented in software, a computer-readable recording medium for storing one or more programs (i.e., software modules) may be provided. The one or more programs stored in the computer-readable recording medium are configured to be executed by one or more processors in an electronic device. The one or more programs include instructions for allowing the electronic device to execute the method according to the embodiments disclosed in the claims and / or specification of the present disclosure.

[0201] The program (i.e., software module or software) may be stored in random access memory, non-volatile memory, including flash memory, read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), magnetic disk storage, compact disk-ROM (CD-ROM), digital versatile disk (DVD) or other forms of optical storage, and magnetic tape cassettes. Alternatively, the program may be stored in a memory configured with a combination of some or all of these storage media. In addition, the number of configured memories may be plural.

[0202] In addition, the program may be stored in an attachable storage device that can access the electronic device through a communication network such as the Internet, an intranet, a local area network (LAN), a wide area LAN (WLAN), or a storage area network (SAN), or a communication network configured by a combination of networks. The storage device may access the device for executing the embodiments of the present disclosure via an external port. In addition, an additional storage device on the communication network may access the device for executing the embodiments of the present disclosure.

[0203] In the aforementioned specific embodiments of the present disclosure, the components included in the present disclosure are expressed in singular or plural form according to the specific embodiments provided herein. However, singular or plural expression is appropriately selected for the situation provided for convenience of explanation, and therefore the present disclosure is not limited to single or multiple components. Therefore, components expressed in plural form can also be expressed in singular form, and vice versa.

[0204] Although the present disclosure has been shown and described with reference to certain preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the present disclosure as defined by the appended claims. Therefore, the scope of the present disclosure is not limited by its detailed description but by the appended claims, and all differences within the equivalents of the said scope will be construed as being included in the present disclosure.

[0205] Although the present disclosure has been described with various embodiments, various changes and modifications may be suggested to one skilled in the art. The present disclosure is intended to encompass such changes and modifications as fall within the scope of the appended claims.

Claims

1. A method performed by a base station in a wireless communication system, the method comprising: Receiving capability information from a user equipment UE indicating whether the UE is capable of performing channel state information CSI generation based on artificial intelligence AI; receiving, from the UE based on the capability information, information related to a change in reasoning time and information related to a change in a channel state, the reasoning time being related to the AI-based CSI generation; identifying information related to CSI reporting timing based on the information related to the change in the inference time and the change in the channel state; transmitting radio resource control (RRC) configuration information including the identified CSI reporting timing to the UE; Transmitting a CSI reference signal RS to the UE; as well as A CSI report based on the CSI-RS and information related to the CSI reporting timing is received from the UE.

2. The method of claim 1, wherein: Identifying information related to the CSI reporting timing includes: When the change in the inference time exceeds a first threshold and the change in the channel state is less than a second threshold, identifying the offset for the CSI report as a value greater than the change in the inference time, or When the change in the inference time exceeds the first threshold and the change in the channel state is greater than the second threshold, a plurality of offsets for the CSI reporting are identified.

3. The method of claim 1, further comprising: When the change in the inference time exceeds a third threshold and the change in the channel state is greater than a fourth threshold, determining whether to disable the AI-based CSI generation of the UE; as well as Information indicating disabling of the AI-based CSI generation of the UE is transmitted to the UE based on the identified result.

4. The method of claim 1, further comprising: determining whether to deactivate at least one of different AI functions included in the UE based on the change in the inference time; as well as Information indicating disabling of at least one of the different functions included in the UE is transmitted to the UE based on the identified result.

5. The method of claim 1, further comprising: A trigger signal related to aperiodic CSI reporting is transmitted to the UE, wherein the trigger signal is one of a medium access control MAC control element CE or downlink control information DCI.

6. A method performed by a user equipment UE in a wireless communication system, the method comprising: Transmitting capability information indicating whether the UE is capable of performing channel state information CSI generation based on artificial intelligence AI to the base station; Identify status information of an AI chip that performs the AI-based CSI generation; identifying a change in inference time associated with the AI-based CSI generation based on the identified AI chip's state information; transmitting information related to the identified change in inference time associated with the AI-based CSI generation and information related to the change in channel state to the base station; receiving, from the base station, radio resource control (RRC) configuration information including CSI reporting timing based on information related to a change in the inference time and information related to a change in the channel state; receiving a CSI reference signal RS from the base station; as well as The CSI report is transmitted to the base station based on the CSI-RS and information related to the CSI report timing.

7. The method of claim 6, wherein: When the change in the inference time exceeds a first threshold and the change in the channel state is less than a second threshold, the information related to the CSI report timing received from the base station includes an offset for the CSI report, the offset including a value greater than the change in the inference change, or When the change in the inference time exceeds the first threshold and the change in the channel state is greater than the second threshold, the information related to the CSI report timing received from the base station includes a plurality of offsets for the CSI report.

8. The method of claim 6, further comprising: Information indicating deactivation of the AI-based CSI generation of the UE based on the change in the inference time and the change in the channel state is received from the base station.

9. The method of claim 6, further comprising: Information indicating deactivation of at least one of different AI functions included in the UE based on the change in the reasoning time is received from the base station.

10. The method of claim 6, further comprising: A trigger signal related to aperiodic CSI reporting is received from the base station, wherein the trigger signal is one of a medium access control element MAC CE and downlink control information DCI.

11. A base station in a wireless communication system, the base station comprising: Transceiver; as well as a controller, the controller being coupled to the transceiver, Wherein, the controller is configured as: Receiving capability information from a user equipment UE indicating whether the UE is capable of performing channel state information CSI generation based on artificial intelligence AI; receiving, from the UE based on the capability information, information related to a change in inference time and information related to a change in a channel state, the inference time being related to the AI-based CSI generation; identifying information related to CSI reporting timing based on the information related to the change in the inference time and the change in the channel state; transmitting radio resource control (RRC) configuration information including the identified CSI reporting timing to the UE; transmitting a CSI reference signal RS to the UE; and A CSI report based on the CSI-RS and information related to the CSI reporting timing is received from the UE.

12. The base station according to claim 11, wherein: The controller processor is further configured to: When the change in the inference time exceeds a first threshold and the change in the channel state is less than a second threshold, identifying the offset for the CSI report as a value greater than the change in the inference time, or When the change in the inference time exceeds the first threshold and the change in the channel state is greater than the second threshold, a plurality of offsets for the CSI reporting are identified.

13. The base station according to claim 11, wherein: The controller processor is further configured to: When the change in the inference time exceeds a third threshold and the change in the channel state is greater than a fourth threshold, determining whether to disable the AI-based CSI generation of the UE; as well as Information indicating disabling of the AI-based CSI generation of the UE is transmitted to the UE based on the identified result.

14. The base station according to claim 11, wherein: The controller is also configured to: determining whether to deactivate at least one of different AI functions included in the UE based on the change in the inference time; as well as Information indicating disabling of at least one of the different functions included in the UE is transmitted to the UE based on the identified result.

15. The base station according to claim 11, wherein the controller is further configured to: transmit a trigger signal related to aperiodic CSI reporting to the UE, wherein: The trigger signal is one of a medium access control element MAC CE and downlink control information DCI.

Images